THE BIOLOGY AND MANAGEMENTOF THE

RIVER DEE

INSTITUTE of TERRESTRIAL ECOLOGYNATURAL ENVIRONMENT RESEARCH COUNCIL

á

Natural Environment Research Council

INSTITUTE OF TERRESTRIAL ECOLOGY

The biology and managementof the

River Dee

Edited byDAVID JENKINS

Banchory Research StationHill of Brathens, Glassel

BANCHORYKincardineshire

2

Printed in Great Britain byThe Lavenham Press Ltd, Lavenham, Suffolk

NERC Copyright 1985

Published in 1985 byInstitute of Terrestrial EcologyAdministrative HeadquartersMonks Wood Experimental StationAbbots RiptonHUNTINGDON PE17 2LS

BRITISH LIBRARY CATALOGUING-IN-PUBLICATION DATAThe biology and management of the River Dee.—(ITE symposium,

ISSN 0263-8614; no. 14)1. Stream ecology—Scotland—Dee River 2. Dee, River (Grampian)I. Jenkins, D. (David), 1926– II. Institute of Terrestrial EcologyIll. Series574.526323'094124 OH141ISBN 0 904282 88 0

COVER ILLUSTRATIONRiver Dee west from Invercauld, with the high corries and plateau of1196 m (3924 ft) Beinn a'Bhuird in the background marking thewatershed boundary (Photograph N Picozzi)The centre pages illustrate part of Grampian Region showing thewater shed of the River Dee.

AcknowledgementsAll the papers were typed by Mrs L M Burnett and Mrs E J PAllen, ITE Banchory. Considerable help during the symposiumwas received from Dr N G Bayfield, Mr J W H Conroy andMr A D Littlejohn. Mrs L M Burnett and Mrs J Jenkins helpedwith the organization of the symposium. Mrs J King checkedall the references and Mrs P A Ward helped with the finalediting and proof reading. The photographs were selected byMr N Picozzi. The symposium was planned by a steeringcommittee composed of Dr D Jenkins (ITE), Dr P S Maitland(ITE), Mr W M Shearer (DAES) and Mr J A Forster (NCC).

The Institute of Terrestrial Ecology (ITE) was established in 1973, fromthe former Nature Conservancy's research stations and staff, joinedlater by the Institute of Tree Biology and the Culture Centre of Algae andProtozoa. ITE contributes to, and draws upon, the collective knowledgeof the 13 sister institutes which make up the Natural EnvironmentResearch Council, spanning all the environmental sciences.

The Institute studies the factors determining the structure, compositionand processes of land and freshwater systems, and of individual plantand animal species. It is developing a sounder scientific basis forpredicting and modelling environmental trends arising from natural orman-made change. The results of this research are available to thoseresponsible for the protection, management and wise use of our naturalresources.

One quarter of ITE's work is research commissioned by customers,such as the Department of Environment, the European EconomicCommunity, the Nature Conservancy Council and the OverseasDevelopment Administration. The remainder is fundamental researchsupported by NERC.

ITE's expertise is widely used by international organizations in overseasprojects and programmes of research.

Dr D JenkinsInstitute of Terrestrial EcologyBanchory Research StationHill of Brathens, GlasselBANCHORYKincardineshireAB3 4BY033 02 (Banchory) 3434

Contents

Preface

The physical background of the River Dee(Judith K Maize Is, Dept of Geography, University of Aberdeen)

Hydrology of the River Dee and its tributaries(J S Warren, North East River Purification Board, Aberdeen)

5

7

23

Land use within the catchment of the River Dee 29(J S Smith, Dept of Geography, University of Aberdeen)

The chemistry of the river system 34(K B Pugh, North East River Purification Board, Aberdeen)

Vegetation of the River Dee 42(N T H Holmes, Nature Conservancy Council, Peterborough)

Vegetation of the valley floor of the River Dee 56(J A Forster and Jacqui Green, Nature Conservancy Council, Aberdeen)

Invertebrates of the River Dee 64(M B Davidson*, R P Owen* and M R Youngt,

*North East River Purification Board, Aberdeen,tDept of Zoology, University of Aberdeen)

Vertebrates, except salmon and trout, associated with the River Dee 83(D Jenkins* and M V Be Ilt,

*Institute of Terrestrial Ecology, Banchory,tlnstitute of Marine Biochemistry, Aberdeen)

Agricultural land use beside the River Dee 94(H L Black, North of Scotland College of Agriculture, Aberdeen)

Forestry beside the River Dee 100(H G Miller* and G Tuleyt,

*Dept of Forestry, University of Aberdeen,tForestry Commission, Banchory)

River engineering and the control of erosion(B B Willetts, Dept of Engineering, University of Aberdeen)

Water abstraction from the River Dee(I D Brown, Dept of Water Services, Grampian Regional Council)

Fishery management on the River Dee(J Oswald, Glen Tanar Estates, Aboyne)

Amenity and recreation on the River Dee(R D Watson, North East Mountain Trust, Aberdeen)

Salmon catch statistics for the River Dee, 1952-83(W M Shearer, Dept of Agriculture & Fisheries for Scotland, Montrose)

The status of the River Dee in a national and international context(P S Maitland, Institute of Terrestrial Ecology, Edinburgh)

Concluding comments(C H Gimingham, Dept of Plant Science, University of Aberdeen)

105

111

117

121

127

142

149

3

4

Appendix I: List of participants

Appendix ll: List of plant and animal species referred to in the text

152

154

Preface

The aims of this symposium were to describe differentinterests in the River Dee, to assess existing knowl-edge, to identify gaps in knowledge and to determinewhether there are conflicts between the differentinterests and, if so, ways of reconciling them.

Many people are interested in the River Dee. It iswidely known as one of the more beautiful parts ofeastern Scotland attracting many tourists each year.The strong associations of the valley with Royalty havegiven it the name 'Royal Deeside'. Recently, manynew people have come to make their homes inDeeside, not only because of its accessibility toAberdeen, but because it is a lovely place in which tolive. Scenic beauty adds therefore to the local econ-omy, with the river as the focus of the main valley.

The river is also important as a source of drinkingwater, as it is very largely free of pollution; it suppliesthe city of Aberdeen, as well as villages in the valley.At least until recently, the Dee has been one of themost important salmon rivers in northern Scotlandand, through salmon fishing, it makes an essentialcontribution to the local economy. It has been carefullymanaged as a salmon fishery for over 100 years, andrecent decreases in numbers of fish caught by rod andline, together with the threats of increased abstractionof water from the river, are a cause of concern topeople interested in the fishery.

However, tourism and the salmon fishery are not theonly important contributors to the local economy.Agriculture, forestry, and other sport are also impor-tant interests affecting the river. Farmers graze theirstock adjacent to the river and are responsible for thefarming practices which determine the chemistry ofthe water running off their land. Under the influence ofprice differences and subsidy, agriculture is changing,with fewer cattle, more grain, new crops, and ever

earlier ploughing after the harvest. In the mid-reaches,a great extent is forested by both private estates andthe Forestry Commission, and the upper valley is oneof the most densely wooded parts of Britain.

The river is of value to geographers, at least partlybecause the course and the banks have been littledisturbed. The purity of its water typifies highlandstreams running over granite, and, to the ecologist, theDee is an excellent example of a nutrient-poor riverwith little ecological change from source to mouth.These features make the river and its associatedtributaries and lochs of considerable interest to naturalscientists, and several wildlife species found on and bythe waterways are important to conservationists. Thisscientific interest is not merely academic, but is sharedby a large and ever-growing number of amateurnaturalists who not only watch birds, but also studythe flowers, butterflies, moths and the interesting wildmammals of the area.

The expansion in the human population at the Aber-deen end of the valley and increased wealth havegiven rise to new water-based recreational activities,including water ski-ing, canoeing, wind-surfing and,most recently, skin-diving. These activities, togetherwith walking and picnicking, and camping where this ispermitted by the landowners, put pressure on thebanks of the river. All these interests and activitiestestify to the importance of the river system; and thesymposium reflected a need to discuss whether futuremanagement of the area needs greater liaison andco-ordination to ensure that conflict is avoided anddevelopments agreed, and that the river as a whole iswisely managed.

David JenkinsHill of BrathensJuly 1985

5

á

The physical background of the River DeeJUDITH K MAIZELSDepartment of Geography, University of Aberdeen

1 IntroductionThis paper examines the physical background to theDee and its catchment from 4 approaches: (i) themajor geomorphic characteristics of the catchment,including its solid geology and topography, the de-velopment of the drainage network, its glacial historyand associated glacial and fluvioglacial deposits; (ii) thenature of the bank and bed materials, and thelongitudinal profile of the river; (iii) the development ofmeanders, bars and islands in the river; (iv) someevidence for channel changes since the mid-19thcentury, relating these to concepts of dynamic equilib-rium. The discussion suggests further lines of researchnecessary for understanding more fully the re-lationships between fluvial processes and associatedlandforms of the Dee.

2 Catchment geomorphology2.1 Solid geology and topography

The Dee catchment covers an area of about 2100 km2,dominated by Precambrian metamorphic and igneousrocks (Figure 1). These rocks comprise the olderMoine schists and granulites in the west, and theyounger south-west/north-east striking Da !radianschists and gneisses, together with outcrops oflimestones, black schists, quartzites, quartz schistsand quartz-porphyry dykes in the centre and east of thecatchment (Bremner 1912, 1922; Fraser 1963; John-stone 1966). The oldest igneous rocks are associatedwith the period of Caledonian mountain uplift andoverfolding, which occurred during Lower Palaeozoictimes. Older basic intrusions were subsequently meta-morphosed to produce hornblende schist and epidior-ites, now forming such mountain areas as Morven(871 m above Ordnance Datum (OD)), while the acidic'newer' granitic rocks now form the most extensivebedrock type in the catchment, representing some52% of the area. These granites also form the highestmassifs in the catchment, including Lochnagar(1154 m above OD) and Mt Keen (938 m) in the south,Hill of Fare (471 m) in the north-east, and the Cairn-gorm plateau to the north, with a number of summitsrising above 1000 m along the Dee watershed. TheDee rises on the Braeriach plateau at about 1220 m,while the maximum elevation of the catchment isreached on Ben Macdui, at 1309 m.

These upland massifs are deeply dissected by the Deeand its tributaries, forming both deep narrow valleysand broad enclosed rock basins or 'howes' (eg Howeof Cromar). The basin areas create 'open' reaches onthe Dee (eg Bremner 1912) and are believed to resultfrom a combination of deep chemical weathering ofthe granites (eg Smith 1981) and selective deep glacialerosion (Bremner 1931).

7

2.2 Development of the drainage network

During the Mesozoic era, the Grampians are believedto have been mantled by a series of sedimentary coverrocks (Bremner 1942), and to have been upliftedseveral times during the Tertiary period. Each period ofuplift generated a new 'cycle of erosion' such that theGrampians were subjected to extensive sub-aerialweathering to produce a succession of peneplanederosion surfaces, with residual granite mountainridges, or summits forming the monadnock level of theCairngorm plateau (Johnstone 1966; Sissons 1967).

The uplifted Cairngorm surface may have exhibited anoverall west-east slope (but see Sissons 1967, p24), sothat the 'proto-Dee' drainage system that developedon that surface drained towards the east, and graduallybecame superimposed on to the underlying crystallinebase; the resulting dendritic river system (Brown1985, Figure 1) was largely discordant with theunderlying geology (Bremner 1912; Smith 1981)(Figure 1).

This model of drainage development is, however,complicated by the existence of a number of apparentanomalies. Bremner (1912), for example, claimed thatevidence of misfit streams and dry valleys, anomalousvalley size, valley orientation, stream length andsub-catchment size in several tributary basins wasindicative of river capture (eg the upper Gairn mayhave been lost to the Feshie catchment and the upperTarf to the Tilt). In addition, geological controls ondrainage are clearly important in the many reacheswhere streams are locally adjusting to softer rocktypes, structural weaknesses, strike directions andfault lines. The Dee itself appears to follow a west-to-east orientated tectonic depression between the maingranitic intrusions (eg Bremner 1912, 1922; Johnstone1966).

2.3 Glacial history and associated deposits

The whole of the Dee catchment was mantled duringthe last glaciation, the Late Devensian, by the Scottishice sheet, which flowed generally from west to east,and reached a maximum thickness of over 1000 m(Clapperton & Sugden 1977; Lowe & Walker 1977).The ice sheet probably retreated and wasted awayduring the period between 16 000 and 13 500 years BP(Vasari 1977). A later, relatively short-lived period ofglacier advance, the Loch Lomond Advance, occurredin the Dee catchment between c 10 650 and 10 145years BP (Vasari 1977), but was confined to the corriesof the Cairngorms and southern Grampians (egSugden & Clapperton 1975).

The de-glaciation of the Late Devensian Scottish icesheet was associated with the formation of an

'New

er'

gran

iteC

ambr

o-O

rdov

icia

nH

ombl

ende

sc

hist

Lim

esto

ne

Dar

k sc

hist

D

a !ra

dian

Mic

a sc

hist

and

gne

iss

Sch

ist

and

gran

ulite

M

oini

an

Oth

er l

ithol

ogie

s

/ -•

I %./.

• I +

Br

I +

+

...

......

..,•

+/

.1.

+

7 '+

+,

' ,

I ÷

'1

\ \

++

+4-

'

- -.

./3

3•=

C7

+

3+

+

3 3

3

+3

34-

.1-

+A

byn

+

++

D

•:::,

B

an

++

•••

4.+

.1.

4.

I''

Bal

l +

41

+

4 4-

/

+

+

•+

++

I+

4

+

•4

+

/

+

+

++

++

.---

- +

4+

..-

--+

+ (

1\1

+ i

.-

-....

.---

2s '--

---.

0 10

20

.++

. 1 /

/

„/.

+

N

\ +

km

Aby

n —

Abo

yne

Bal

l —

Bal

late

r

Bal

m

— B

alm

oral

Ban

—

Ban

chor

y

Br

— B

raem

ar

D

— D

inne

t

AB

ER

DE

EN

Nor

thS

ea

Fig

ure

1. G

ener

aliz

ed s

olid

geo

logy

of

the

Dee

cat

chm

ent.

Bas

ed o

n G

eolo

gica

l Sur

vey

of S

cotla

nd,

1:25

3440

, S

heet

9 (

1913

) &

She

et

12 (

1910

); a

nd J

ohns

tone

(196

6)

Co

extensive and varied series of glacial and fluvioglaciallandforms and sediments in the Dee catchment (egBremner 1912, 1931; Clapperton & Sugden 1972;Sugden & Clapperton 1975; Smith 1981). Many valleyswere widened and over-deepened by glacial erosion,and valley-side slopes were scarred by meltwaterchannels and mantled with till and hummockymoraines, lateral and terminal moraines, kame ter-races, kames and eskers.

The deep rock basins (and gorges?) (eg at Braemar,Ballater, Dinnet, and some upper glens) have beeninfilled with coarse outwash sands and gravels de-posited by meltwaters from the decaying ice masses.Kettle hole lakes, such as Lochs Davan and Kinord,formed around large, isolated, wasting ice masses.

The sand and gravel outwash deposits have sincebeen dissected into a series of terraces. Up to 4 or 5such terraces or terrace fragments have been identi-fied in the wider parts of the Dee valley (eg betweenCambus O'May and Aboyne). These terraces exhibitdistinctive braided palaeochannel patterns represent-ing the glacial Dee meltwater streams (Bremner 1931)(Figures 2i, ii) which extended across the formerlybroad valley floor. The terrace sediments comprisestratified sands and gravels (Institute of GeologicalSciences (IGS) 1977), as well as coarse cobble andboulder beds representing deposition in streams witha much higher discharge and tractive velocity than inthe present-day Dee.

The terrace sequence probably dates from the finalstages of de-glaciation, with downcutting associatedwith a major change in the hydrological regime of theDee as ice disappeared from the catchment. Thisdowncutting may well have culminated, between9700 and 8000 years BP (cf Forth Valley; Sissons1967), in the formation of the buried channel thatappears to underlie the lower Dee, where an IGSborehole extended to 15 m below the present riverlevel (at Bridge of Dee; IGS 1977). Since that time, theDee has infilled this channel and appears to have beenonly slowly modifying its narrow floodplain from astate of long-term dynamic equilibrium (see later,section 5), despite land use and settlement changes inthe catchment over the past 5000 years (Edwards1975).

3 Bank and bed materials, and the longitudinal profile,of the Dee3.1 Bank materials

The Dee and its tributaries have access to a widerange of source materials and flow through a variety ofsubstrates and bank materials, reflecting the numer-ous different bedrock types and extensive glacial,fluvioglacial, periglacial and slope deposits in thecatchment.

3.1.7 Bedrock

Bedrock forms the channel boundary where moreresistant rock types, especially quartz-porphyry dykes

9

and certain granites, are present, acting to constrictthe channel and often creating rapids or waterfallsbounded by near-vertical walls. Where erosion hascaused block collapse into the stream, the bedrockbanks act as a major contributor of large boulderymaterial to the stream bed.

3.1.2 Till

In many parts of the Dee valley, bedrock is mantled bytill (glacial drift), ranging from only a few metres thick(eg Bremner 1925) to tens of metres thick. The tillbanks form relatively soft, readily saturated banks,which provide a wide range of sediment sizes and rocktypes to the river sediment load.

3.1.3 Slope deposits

In the upland valleys, streams flow through bankswhich comprise alluvial fan, solifluction, landslip andScree deposits, of a variety of ages, and whichcontribute heterogeneous, poorly sorted, coarse,angular, non-cohesive materials to the water.

3.1.4 Fluvioglacial outwash deposits

The most extensive river bank material, particularly inthe middle and lower reaches of the Dee and itstributaries, is represented by the late Devensian sandand gravel deposits forming the terrace sequence.These deposits are readily erodible, non-cohesive,coarse, poorly sorted deposits, providing an abundantsupply of heterogeneous material to the rivers (IGS1977). Bank collapse is through grain-by-grain removal,and, while the smaller sizes may be moved duringmoderate and high river flows, the larger cobbles andboulders may rarely be moved and accumulate at thefoot of the bank.

3.1.5 'Recent' alluvium

'Recent' alluvium represents bar and overbank de-posits that have accumulated to form the present riverfloodplain. These sediments generally comprise up to30 cm of fine-grained sands and silts (IGS 1977), butoften include lenses or sheets of pebbles and cobblesderived from flood events which deposited overbankgravel bars and sheets. The alluvium is found only onthe present floodplain and lowest terrace, and isabsent on the higher fluvioglacial terraces. The alluv-ium forms a relatively thin superficial cover, andappears to be accumulating only slowly. It forms low,unstable banks that are readily undercut and sloughinto the river.

3.1.6 Artificial embankments

Artificial embankments have been constructed at anumber of reaches on the Dee as a means of floodprotection, bank stabilization and channelization (Wil-letts 1985). These embankments include revetments,comprising boulders held in wire baskets, concretepiles and fences, and stone and earth levees.

3.1.7 Vegetation

The majority of the river banks, especially in the middleand lower reaches of the Dee, are vegetated, thereby

i.A

boyn

e G

lidin

g C

lub

terr

aces

I 1

'•".

""4.

!‘•

,/ •

,t

-,••

1,1

I

- "

• •

4

•

1‘

56

Dru

mna

gesk

te

rrac

es

k i

,

e

1

•t

Ii.

0 ,

Airs

triP

-.

.--

- -

, ...

,.s

-i

41,_

.. ,

...1

,..,

-4

• \

.- •,

',s-

-7

...- ..

...;-

17:1

(71

"C

.)";

,7••

••••

%-r

il-;

:-:-

-.:-

----

-.

. •

i,

--

) (-

--',

0C

a 14

/1-_

_••

. 7...

....,,

...

.,7

; --

..1

t.

/1

'''-

i "

-.

i A

,

-".

\

.'-:

' •.

: -

s'

' S

1 ...

.. ..)

1.

1 ...

. \

1 '

:".

.. -.

'S

- • .

=.

1

s \

.:...

.. 1

• %

a?

ti

I )

/I

1 i

\.".

%

‘ "

. . . .

' '

.-

048

:s '

,,

, - ,

...

-,

ss-\

\,:.

‘ '1

%f

'"

11

I •

...

l ...

...

.. 7

, A

•

• '

1

I li

t 4,

7.1:

:' 1.

/.....

;,.)

7,'

--' ,

s

lis

.„(

-,''

'' -

- -

el;

''',

"/ "

r e'

'

' •n

• i,.

.• -

-',/

,..47

,j''

'//"

. ...

.•

''..,

.•--

',

, ..‘

":".

-.

. Z

.-

- '//

/---

111

V

:pe-

e2

_.,

-

-'`I-

----

--:.

:: --

,.^

"P

.''.

''' '-i

-'7

7;:?

:;::.7

.-..,

--,,

.-;-

-„

,- ....-

, „ -.,;

„ ,.....

...__

_ .._

-_,,,

,,,,

- --

---.

)--

- --

-t.'

---.

...,

::-..4

,-

,1...

...

....:.

.....-

.-2.

...,

I

giV

er D

ee...

. ...

111

500

49

98

0

_

-r-

r11

%

I I

i A

(1"

I I"

vV

I

1

I"

Ip

•1

I •

,•

•a

.-•

1/1

• •

76;

Fig

ure

2.

Pal

aeoc

hann

el

syst

ems

on

sele

cted

D

eete

rrac

es.

Bas

ed

on

inte

rpre

tati

on

of

1:11

000

aeri

alph

otog

raph

s (G

ram

pian

Reg

iona

l C

ounc

il

1981

)

--A

—.

Maj

or t

erra

ce

bluf

f

Min

or

terr

ace

bluf

f

Pal

aeoc

hann

el

cour

see •

Shi

ngle

500

50

-y-

dam

•••

98 •• •

••

acting to increase resistance to bank erosion andoverall bank stability. Other gravel-bank rivers in theUK with tree-lined channels tend to be about 30%narrower than streams with grassy banks (Charlton etal. 1978), as a result of increased bank resistance.

3.1.8 Discussion

Many banks along the Dee exhibit a composite verticaland downstream stratigraphy, with different sectionsof the bank containing different materials, each exhibit-ing a different texture, permeability, cohesion, tensilestrength, saturation limit, and susceptibility to bothsub-aerial and fluvial erosion (eg Thorne & Lewin 1979;Simons & Li 1982). Bank materials often differ onopposite banks, and vary rapidly downstream (egCharlton 1982). The extent of the different bankmaterials in the Dee river system would prove aworthwhile project for future investigation. Spatialpatterns of bank erosion are highly variable, andexercise a major control on river channel geometry,and on rates of sediment supply, transport anddeposition in the river.

3.2 Bed sediments

The Dee is characterized by its pebbly bed sedimentsand may be classified as a 'gravel-bed river'. Althoughit also includes boulder bed and bedrock reaches, theDee flows mostly over a stony shingle bed, with clasts(pebbles) derived predominantly from local inputs andreworking of till and outwash bank sediments. The sizeof material depends both on local sediment suppliesand on the tractive force of flow. Clast sizes exhibit anoverall decrease in size downstream (eg preliminarydata on Figure 3), but show local peaks in size wherecoarse-grained banks are exposed or where localresidual deposits of larger cobbles or boulders arepresent.

The bed sediments comprise a wide range of rocktypes, but are dominated by granites (eg 53% granitesin the Dinnet-Potarch reach), together with a largevariety of schists and gneisses (totalling 36%). Most ofthese clasts are rounded, suggesting a relatively longdistance of transport, either in the present river or in aformer meltwater river. The sediments contain littlefine matrix material (sand and silt), which tends to beflushed out during high flow events and to beredeposited farther downstream as bar-top and over-bank deposits on the floodplain (Willetts 1985). Thecoarser sediments are dominated by a strong imbri-cated fabric, with a stable, densely packed orientationparticularly resistant to erosion, forming an 'armoured'bed. As a result, the bed sediments tend to be farmore resistant to stream scour and removal than theadjacent bank sediments, and hence lateral bankretreat is a more likely response to flood erosion thanis bed degradation.

3.3 Longitudinal profile

The longitudinal profile of a river represents thelong-term response of fluvial processes (especially

11

downstream changes in stream discharge), operatingin a catchment of varied rock structure, with its ownhistory of tectonic and climatic change, vegetation, soiland land use patterns. The Dee exhibits a distinctiveconcave-upward longitudinal profile over its streamlength of about 140 km, decreasing in gradient down-stream. The long profile does not follow a smoothdownstream curve in gradient, but in detail comprisesa series of irregular steep sections alternating withsections of anomalously low channel gradient (Brem-ner 1912) (Figure 3). Although the mean gradient fromsource to mouth is about 8.9 m km-1, the upperreaches of Glen Dee are considerably oversteepened,with a drop of over 600 m in just 3.8 km, producing anaverage local gradient of 160 m km-1. The NaturalEnvironment Research Council (NERC) (1975) meanslope measure of 'S1085' (mean gradient betweenpoints at 10% and 85% along the channel length) isonly 3.4 m km-1 and is more representative of gra-dients over most of the Dee. Nevertheless, steeperreaches occur, especially over bedrock (eg7.3 m km-1 in the Cambus O'May reach; 4.1 m km-1in the Crathie reach; and 4.3 m km-1 in the Potarchreach), while lower gradients occur where the valleyfloor widens out and the channel is able to meanderthrough more erodible gravel and alluvial sediments(eg 2.2 m km-1 in the Braemar reach; 2.8 to0.7 m km-1 in the Dinnet-Aboyne reach; and 1.2 to0.8 m km-1 in the lower Dee east of Park). In most ofthese wider, more open sections of the valley,sediment infilling has taken place behind downstreamrock bars or gorge constrictions which have acted aslocal 'base levels' for upstream aggradation.

The longitudinal profile of the Dee apears to haveresulted from a combination of glacial and fluvioglacialerosion of the (weathered?) bedrock to form a seriesof steep gorge or rock-walled reaches and deep rockbasins, followed by infilling of these basins withfluvioglacial and fluvial sediments. The longitudinalprofile of the present river probably differs significantlyfrom the profiles of the river during Late-glacial andearly Post-glacial times. The gradients of the differentterrace surfaces, however, have not yet been deter-mined, while the depth of sand and gravel infill abovethe bedrock surface also remains unknown.

4 Fluvial processes: meander and bar development4.1 Introduction

The Dee exhibits a range of channel patterns reflectingfluvial processes operating in a variety of bed and bankmaterials, valley floor and sediment supply conditions.The traditional classification of channel patterns is intostraight, meandering and braided patterns (Leopold & •Wolman 1957). This approach, however, while provid-ing a convenient framework for channel patterndescription on the Dee, is a somewhat artificialapproach to the study of fluvial processes becausesimilar processes of scour and deposition operate ineach channel environment. Bars and islands, forexample, are common to both straight and meander-

12

Elevation(m)

'Gorge' reaches

Channelsinuosity

Meanderspacing

(m)

Maximummean particle

diameter(cm)

900

800

700

600

500

400

300

200

100

0

1.8

1.6

1.5

1.2

1.02000

1500

1000

500

100

80Bankfull 60chennelwidth 40

(m) 20

40

30

20

A

tG GL1r

Tributaries, starting upstream:Al It a' Gharbh, Geldie, Lui, Ey, Quoich, Clunie, Feardar, Gelder,Girnock, Gairn, Muick, Tullich, Tanar, Canny, Feugh, LeucharSettlements, starting upstream:Mar, Braemar, Crathie, Balmoral, Cambus O'May, Dinnet,Aboyne, Potarch, Banchory, Bridge of Dee

M Br C Ba CM D T A P.• • • • • •

140 130 120 110 100 90 80 70 60 50 40 30 20 10 0

Distance upstream (km)

Figure 3. Longitudinal profile and downstream changes in channel characteristics of the River Dee

BD

ing reaches. In addition, this classification of channelpattern types is normally limited to alluvial channels, iechannels that are free to alter the geometry of theirchannel boundaries as a means of adjusting to anychanges in flow hydraulics. This condition is notalways present in the Dee, particularly where bedsediment is too coarse or too well imbricated to bereadily moved, or where relatively resistant banksformed by rock, till, outwash or concrete all act topreclude free bank retreat and unconfined meanderdevelopment.

4.2 Characteristics of straight reaches on the Dee

The most important flow characteristic of straightgravel-bed reaches is that of flow deflection, by which,even in a straight channel, the main flow path and thelocation of the dominant flow cell meander alternatelyfrom one bank to the other (eg Leopold 1982). Thispattern of flow, in turn, tends to result in alternatingzones of bed scour to form 'pools', and in zones ofsediment concentration to form 'riffles' and lateralbars.

A riffle is a zone of shallow flow over a coarse-grainedbed, presenting high resistance to flow and littlesediment storage. A bar is a large depositional unit(longer than the associated channel width) of sandsand gravels acting as a major storage zone for bedloadsediment (Church & Jones 1982). Bars and rifflesdevelop and evolve during high flow events when bedsediment becomes mobile; at low flows, most barsand riffles remain inactive or emerge as shingle bars orislands.

A straight reach is defined as extending for a distanceof 10-14 times the channel width, as within thisdistance one would otherwise expect a meander tohave developed (see below). On the Dee, 30 suchstraight reaches were identified, averaging 811 ±322 m in length, with a mean downstream spacing of4.19 ± 3.31 km, representing about 16% of the totalmain stream length. These straight reaches includedgravel, bedrock and artificially embanked reaches.

In 15 of the straight reaches examined in detail fromaerial photographs, riffle spacing was found to average5.5 times channel width, but with a correlationcoefficient of only +0.49 (ie only 24% statisticalexplanation). Hence, mean riffle spacing conforms tothe geometric pattern recognized on most rivers,where pool-riffle spacing is generally 5-7 times channelwidth. However, the large variability in spacingappears to reflect the numerous local controls onpatterns of flow and sedimentation (eg sources ofsediment input, the presence of coarse bed materialand constrictions of the channel caused by bedrockoutcrops and relatively resistant high banks of till andoutwash).

Bars and riffles are also seen to be extremely variablein terms of their location within the channel and their

13

size, shape and orientation across the channel (Figure4). Some riffles extend for only 30 m, while othersextend for over 200 m downstream; in some reaches,sequences of 'gravel tongues' extend as indistinctshallows for hundreds of metres, paralleled by longscour pools at the foot of one or both banks. Riffleforms include narrow, linear gravel accumulationsclose to one bank, small patches of coarse sedimentattached to one bank, and broad shingle zonesextending partly or wholly across the channel bed.

4.3 Characteristics of meandering reaches on the Dee

With increased discharge, stream power and sedimenttransport, the strength and asymmetry of individualflow cells increase sufficiently to allow banks to bemore readily eroded to form a meander bend (egLeopold 1982).

On the Dee, 221 meanders have been identified, witha mean spacing of 609 ± 400 m, ie an averagewavelength of 1200 m, and with both spacing andwavelengths generally increasing downstream withincreases in discharge and channel capacity (especiallychannel width) (Figure 3). Mean wavelength (X) over5 km reaches has been found to be related to bothmean annual flood (Qma) (only for reaches at Polhollick,Woodend and Park) and to channel width (W), esti-mated from Ordnance Survey (OS) maps, such that:

The wavelength-discharge relationship conformsclosely to that established for other gravel-bed rivers(eg Dury 1977; Charlton 1982; Hey & Thorne 1984),but the wavelength-width relationship is defined by amuch higher coefficient (37.8) than is normally (ie10-14) found. Hence, it appears that wavelengths inthe Dee tend to be much greater than would beexpected for equivalent width (and discharge?) con-ditions in fine-grained, lowland floodplain channels.These large wavelengths are found on a particularlynarrow floodplain, rarely exceeding 850 m, and gener-ally averaging less than 400 m. As a result, overallchannel sinuosity (defined as the ratio betweenchannel length and meander axis length) is relativelylow (Figure 3), averaging only 1.27 ± 0.18. Only 35 km reaches exhibit sinuosities exceeding 1.5 (ie'meandering'), while over 60% of reaches exhibitvalues of less than 1.25. The predicted amplitudes(Lewin & Brindle 1977) of the Dee range betweenc 620 m and 3.2 km, thereby far exceeding the actualwidth of much of the active floodplain.

The distinctive meander pattern of the Dee, compris-ing long wavelength, low curvature meanders, inter-spersed with numerous straight reaches, may resultfrom several mechanisms. First, weak meander de-velopment may result from rapid rates of bank erosion,thereby preventing full development of secondary

1 4

yi

i. West Dinnet reach

Garlogie reach

iii. Kincardine O'Neil reach

v. East Potarch reach

Potarch Bridge

a0s:•

" fI

vi. West Banchory reach

Pool

Riffle

Migrating gravel bars

Dinnet Bridge

dill Ihilig-----6.111P 411)1

alb

111111%.

iv. Tillydrine reach

Banchory Bridge

040110111

Shingle bar and bedrock outcrop

Approx scale

0 500

Figure 4. Pool and riffle sequences in straight reaches of the Dee. Based on interpretation of 1:11000 aerialphotographs (GRC 1981)

flows and the stable development of meander forms(Newson 1981; Ferguson & Werritty 1983). Second, itis possible that the meander pattern may be partlyinherited from a period of different hydrological con-ditions in the past (eg wetter Atlantic period; Smith1981), since when lower river discharges may haveprecluded any major channel re-adjustment (Ferguson1981), largely because channel sediments are socoarse. Third, the presence of resistant terrace bluffs,together with rock-walls and artificial embankments,

i. Chest of Dee — Linn of Dee reach

ii. Mar Lodge — Victoria Bridge reach

Girnock — Ballater reach

Victoria Bridge

Gairn

..• Floodplain deposits• • •

• • Fluvioglacial terrace deposits. •

Valley-side slopes

Figure 5. Examples of the confined meander forms on the Dee

15

have acted as a major control on meander develop-ment by reducing the floodplain width to below thetheoretical meander amplitude. As a result, much ofthe Dee exhibits the distorted meander forms charac-teristic of 'second degree meander confinement'(Lewin & Brindle 1977), where meanders includebox-shaped reaches with one or more right-angledbends, ensconced or entrenched meanders, and longstraight reaches hugging the edge of the valley floor(Figure 5). An analysis of the distribution of free and

iv. Dinnet — Aboyne reach

v. Kincardine O'Neil reach

vi. Woodend reach

1

km

2 3

16

confined meanders in the Dee catchment would bevaluable in determining the role of these lateralconstraints on meander development.

4.4 Development of bars and islands

4.4.7 Origins of bars and islands in gravel-bed rivers

Bars (active bedforms) and islands (inactive and oftenvegetated) can form in a number of possible ways,both depositional and erosional, and at a variety ofdates during the river's history. Many bars and islandstend to be both polygenetic and of varied age at anyone site and over a downstream distance. The mainorigins of bars and islands may be classified as primary(ie depositional) and secondary (ie erosional), althoughmost forms are derived from a combination of theseprocesses.

The primary origins include:

i. deposition of a lateral or point bar;ii. deposition of a mid-channel (medial) bar, nor-

mally located near a significant local source ofsediment supply (eg just downstream of ameander bend);

iii. deposition of a bar at a tributary junction whereexcess sediment brought down by the tributaryis dumped as backwater forms behind the mainstream (Church & Jones 1982).

The secondary origins result from the dissection of anyprimary bar form by chute (flood) channels (Church &Jones 1982). During flood events, a bar becomespartly or wholly submerged, and a flood channel isfrequently scoured into the back (ie at the foot of thechannel bank) of the bar or down the bar front. Duringsuccessive floods, the chutes can migrate upchannel,progressively incising into the bar; the chute channelitself may eventually become the dominant channel,and, as its banks are progressively widened, it maydevelop its own system of associated channel bars.Often, a series of partly or fully formed chute channelsdevelops at different elevations on the bar duringdifferent flood events.

4.4.2 Bars and islands in the River Dee

In the Dee, 38 island reaches have been identified(from OS maps) averaging 250 ± 240 m in length, andwith a highly variable downstream spacing, averaging3.28 ± 4.18 km. Preliminary examination suggeststhat bars and islands of all 4 origins (see above) arepresent (see Figure 6). Lateral and point bars are beingactively formed at many sites, but the majority of bars,especially point bars, represent deposits that havebeen dissected and reworked at a variety of periodsduring the river's recent history.

Most islands in the Dee are now largely or completelyinactive. The highest islands often lie at the sameheight as adjacent outwash terraces, suggesting thatsome islands at least are thousands of years old. TheDee is characterized by long stretches free of islands,separated by zones of intense and active sedimenta-

tion which form major 'megaform' bar assemblages(Church & Jones 1982), as in the Dinnet and Drum-nagesk reaches. These complex systems of bars,locally varying in origin, age, rates of development andflow thresholds for development, need to be studied inmore detail, in order to understand the spatial patternsof fluvial processes in the Dee.

5 Channel changes5.1 Historical channel changes

Evidence has been presented earlier (section 2.3) oflong-term changes in the hydrological regime andchannel pattern of the Dee, associated with deglacia-tion and terrace formation at the end of the LateDevensian. Evidence of more recent channel changessince the mid-19th century is also available from OSmaps (at scales of 1:25 000, 1:10 560 and 1:10 000,dating from 1864/66, 1900, 1962-69) and from aerialphotographs (1:11 000, 1974, 1977, 1981; GrampianRegional Council). Most historical channel changesprobably occurred during short-term, high-magnitudeflow events (ie bankfull and, particularly, overbankfloods), as in many other gravel-bed and upland rivers(eg Anderson & Calver 1980; Werritty & Ferguson1980; Werritty 1984). Although some reaches may beunstable and vulnerable to change even at low flows(eg Church & Jones 1982), it appears that majorchange is very localized, and much of the channel mayremain unchanged for many decades (Ferguson 1981).

Major floods occurred in the Dee during the historicalperiod, and could account for significant channelchanges. For the period 1864/66 to 1900, major floodsoccurred in 1873 and 1895, and for 1900 to 1969, in1913, 1920, 1937 and 1951 (eg Warren 1985). Indeed,a number of historical channel changes associatedwith floods have been noted in the local literature. Inparticular, the effects of the 1829 'Muckle Spate'(McConnochie 1900; Bremner 1922; Wyness 1950),the 1920 floods (Bremner 1922) and the 1956 Cairn-gorm floods (Baird & Lewis 1957) have been widelyreported.

Initial examination of the available evidence suggeststhat significant changes in channel course, pattern andwavelength have occurred at numerous sites on theDee, in association with periods of local bank retreat,meander migration and abandonment, bar accretion,chute channel formation and island development(Figures 7 & 8). These changes have been concen-trated in zones of local bank instability on the widerparts of the floodplain that present some potential forlateral channel expansion or channel initiation, and forincreased sediment availability and reworking.

5.2 Estimates of rates of channel change and bank retreat

No field measurements of the rates of bank retreat areyet available for streams in the Dee catchment.However, measurements from historical map evi-dence provide estimates of overall net rates of bankrecession over periods of between 34 and 104 years

i. Crathes reach

Murtle reach

Tanar

•

Aboyne reach

AboyneBridge• . •

iv. Tullich reach

Tullich •

. • -

Fungle

v. Meikle Inverey reach

vi. Woodend reach

vii. Drumnagesk reach

• . Floodplain deposits

.• .

cz°

Fluvioglacialterrace deposits

Valley-side slopes

0 1

km

Figure 6. Examples of dissected lateral bar, point bar, tributary bar, and bar assemblage deposits in the Dee

Levee

17

18

WI

— •—• — ••- -"••

Inyerey

Footbridge

--Suspension bridge

N.

.—.....,

.--- .'1", ..pi • :

....\ V' 1 __ ..b.

..J 'b • \1.2° • \s'

/ 1 500

1866

1900

""-

1969

Figure 7. Example of channel change on the River Dee since the mid-19th century: Inverey reach, 1866-1969

brain

---------

0 500

MurtIe House

(Table 1 i). Measurements taken from 14 zones ofsignificant channel change on the Dee indicate that netrates of bank retreat have averaged 0.70 ± 0.67m yr-1, ranging from as little as 0.04 m yr-1 to asmuch as 2.3 m yr-1. The latter retreat occurred in theMurtle reach between the 1864/66 and 1962/63surveys and represented the development of a mean-der extending some 225 m across the floodplain.Rates of bank retreat have been highly variablebetween sites, while net annual rates of retreat havevaried by up to 6 times at individual sites through time.

The zones of bank retreat, meander development andchannel expansion appear, however, to be consist-ently matched by adjacent zones of sedimentation andbank accretion. Numerous examples of channel nar-rowing and bar abandonment to form floodplaindeposits during the historic period can be identified onthe Dee. Measurements of these changes at 22 sites(Table 1 ii) indicate that net rates of bank accretionhave averaged 0.72 ± 0.48 m yr-1, a rate of changealmost identical to the rates of change estimated forbank retreat.

Hence, where channel change has occurred, the netrates of change produced by erosion and depositionare of a similar order of magnitude. This similarity inrates of bank retreat and accretion may mean thatpatterns of erosion and deposition in the Dee since themid-19th century have developed in a state of dynamicequilibrium, or 'passive disequilibrium' (Ferguson1981), with no overall degradation or aggradation in the

1864-66

Ewe Haugh

River Dee

1962-63

_ Ardo

— — — — — — —South Deeside Road

House

19

Figure 8. Example of channel change on the River Dee since the mid-19th century: Murtle reach1864/66-1962/63

river. The absence of any major overall change in theDee may reflect the relative coarseness of the bed andbank material, the resistance of confining boundariesto the narrow floodplain, and the absence of anysignificant change in hydrological regime caused byany changes in catchment climate or land use.However, .this hypothesis of no long-term, overall,progressive change in the Dee needs further testing.In addition, recent and present-day channel changesalso need to be determined at selected vulnerablesites to establish whether or not rates of channelchange have remained consistent, despite possiblechanges in hydrological conditions since the 1950s(Warren 1985). A further important research topic liesin the development of a predictive model for channelchange in relation to the threshold discharges requiredfor geomorphic changes at a given site throughout thehydrological year (Harvey et al. 1979; Gregory &Madew 1982). Many reaches of the Dee remainunstable and liable to dramatic modification duringannual flood events; some means of predicting sitesusceptibility and likely geomorphic responses togiven flood conditions would prove invaluable to allusers and managers of the river and its floodplain.

6 ConclusionThe Dee catchment covers an area of 2100 km2 and isdominated by granites and schists forming the mainupland massifs of the catchment, rising to Ben Macduiat 1309 m above OD. The whole catchment wasice-covered during the last glaciation, providing alegacy of extensive glacial and fluvioglacial deposits,

20

Table 1. Estimated net rates of bank erosion and bank accretion on the River Dee, 1864-1969

including outwash sediments infilling the main valleys.These sediments have been dissected into a series ofterraces, comprising coarse sands, gravels and cob-bles which accumulated in high-discharge braidedmeltwater streams, and which formed during de-glaciation. The Dee and its tributaries thus flowthrough a variety of bedrock types, till, slope deposits,outwash, alluvial sands and gravels on the floodplain,and artificial embankments, each bank type respond-ing differently to fluvial erosion, and influencing rivermorphology. The Dee is a gravel-bed river, comprisingan armoured bed of coarse granite and schist pebbles,and exhibits an irregular long profile with steep, narrowsections alternating with low gradient sections wherethe valley floor opens out and bedrock basins havebeen infilled.

The Dee exhibits straight, meandering and braidedreaches, characterized by irregular patterns of pools,riffles and bars. Pool-riffle spacing in the Dee conformsto that of most other rivers, averaging 5.5 timeschannel width, although high variability in spacingreflects the dominance of local geomorphic controlson flow and sediment conditions. The Dee is a lowsinuosity river, averaging only 1.27, with a meanderspacing of 609 m. Meander wavelength (X) was foundto be related to mean annual flood (Oma) by thefunction X = 93.3 0m2.52, and to channel width (W) byk = 37.8 Wa91. Predicted meander amplitudes associ-ated with measured wavelength and width far exceedthe available floodplain width, resulting in the develop-ment of a range of confined meander forms. Bars andislands in the Dee are derived from the deposition of

lateral, point and confluence bars, but most bars andislands have been dissected and reworked by floodflows, and some have become abandoned and vege-tated. Significant changes in channel course haveoccurred at numerous sites since the mid-19th cen-tury, through bank retreat, meander evolution, baraccretion, chute channel formation and island develop-ment. Net rates of bank retreat in 14 reaches haveaveraged 0.7 m yr-', while net rates of bank accretionin 22 reaches have averaged 0.72 m yr-1, suggestingthat patterns of erosion and deposition in the Dee havebeen operating in a state of dynamic equilibrium, withno long-term progressive change.

This paper presents the results of a preliminary surveyof the geomorphology of the Dee and its catchment.Five important fields of study can be identified asrequiring further detailed investigation.

i. River terraces and valley fill deposits need to befully surveyed to determine in detail theirmorphological and sedimentological character-istics; boreholes through the valley fill are neces-sary, together with more precise dating of thedeposits and further palaeohydrological analysisof the terrace/infill sequence.

ii. Bank materials and rates of erosion need to bedetermined through a detailed survey of thephysical characteristics and spatial distribution ofthe different bank materials, together with de-tailed monitoring of bank retreat rates (using fieldand air surveys) at selected, vulnerable sites.

iii. Bed sediments and bedforms need to be studiedin detail to determine the size, shape, lithology,orientation of bedload sediments as indicators ofsource inputs, downstream variations, extent ofarmouring, rates of transport, and thresholdvelocities. The range of bar types, and rates ofbar and sediment movement also need to bedetermined.

iv. Hydraulic relations between the main flow para-meters (width, depth, slope, velocity, discharge,sediment discharge), geometric characteristicsof channel planform (sinuosity, wavelength,meander form), bedforms and patterns ofchange all need to be clearly established for theDee, mainly as a basis for developing a model ofchannel dynamics.

v. Channel changes over long-term, historical andshort-term timescales need to be analysed for alarge number of reaches on the Dee and itstributaries, together with analysis of the associ-ated geomorphic, hydraulic and human controlson these changes, as a basis for developing apredictive model of channel change.

7 ReferencesAnderson, M. G. & Calver, A. 1980. Channel plan changesfollowing large floods. In: Timescales in geomorphology, edited byR. A. Cullingford, D. A. Davidson & J. Lewin, 43-52. Chichester:John Wiley.

21

Baird, P. 0. & Lewis, W. V. 1957. The Caimgorms floods, 1956.Scott. geogr. Mag., 73, 91-100.

Bremner, A. 1912. The physical geology of the Dee Valley.Aberdeen: Aberdeen University Press.

Bremner A. 1922. Deeside as a field for the study of geology.Deeside Field, 1, 33-36.

Bremner, A. 1925. Limits of the valley glaciation in the basin of theDee. Trans. Edinb. GeoL Soc., 11, 61-68.

Bremner, A. 1931. The valley glaciation in the district round Dinnet,Cambus O'May and Ballater. Deeside Field, 5, 15-24.

Bremner, A. 1942. The origin of the Scottish river system. Scott.geogr. Mag., 58, 15-20, 54-59, 99-103.

Brown, I. D. 1985. Water abstraction from the River Dee. In: Thebiology and management of the River Dee, edited by D. Jenkins,111-116. (ITE symposium no. 14.) Abbots Ripton: Institute ofTerrestrial Ecology.

Charlton, F. G. 1982. River stabilization and training in gravel-bedrivers. In: Gravel-bed rivers, edited by R. D. Hey, J. C. Bathurst &C. R. Thorne, 635-657. Chichester: John Wiley.

Charlton, F. G., Brown, P. M. & Bensson, R. W. 1978. Thehydraulic geometry of some gravel rivers in Britain. (Report no. IT180.) Wallingford: Hydraulics Research Station.

Church, M. & Jones, D. 1982. Channel bars in gravel-bed rivers. In:Gravel-bed rivers, edited by R. D. Hey, J. C. Bathurst & C. R. Thorne,291-338. Chichester: John Wiley.

Clapperton, C. M. & Sugden, D. E. 1972. The Aberdeen and Dinnetglacial limits reconsidered. In: Northeast Scotland geographicalessays, edited by C. Clapperton, 5-11. Aberdeen: Aberdeen Uni-versity Press.

Clapperton, C. M. & Sugden, D. E. 1977. The Late Devensianglaciation of north-east Scotland. In: Studies in the ScottishLate-glacial environment, edited by J. M. Gray & J. J. Lowe, 1-14.Oxford: Pergamon.

Dury, G. H. 1977. Peak flows, low flows, and aspects of geomorphicdominance. In: River channel changes, edited by K. J. Gregory,61-74. Chichester: John Wiley.

Edwards, K. J. 1975. Aspects of the prehistoric archaeology of theHowe of Cromar. In: Quaternary studies in north-east Scotland,edited by A. M. D. Gemmell, 82-87. Aberdeen: Aberdeen UniversityPress.

Ferguson, R. I. 1981. Channel forms and channel changes. In:British rivers, edited by J. Lewin, 90-125. London: Allen & Unwin.

Ferguson, R. I. & Werritty, A. 1983. Bar development and channelchanges in the gravelly River Feshie, Scotland. In: Modern andancient fluvial systems, edited by J. D. Collinson & J. Lewin,181-193. Oxford: Blackwell Scientific.

Fraser, W. E. 1963. Geology and structure. In: The northeast ofScotland, 3-15. Aberdeen: British Association for the Advancementof Science.

Gregory, K. J. & Madew, J. R. 1982. Land use change, floodfrequency and channel adjustments. In: Gravel-bed rivers, edited byR. D. Hey, J. C. Bathurst & C. R. Thorne, 757-781. Chichester: JohnWiley.

Harvey, A. M., Hitchcock, D. H. & Hughes, D. J. 1979. Eventfrequency and morphological adjustment of fluvial systems in uplandBritain. In: Adjustments of the fluvial system, edited by D. D.Rhodes & G. P. Williams, 139-167. London: Allen & Unwin.

Hey, R. D. & Thorne, C. R. 1984. Flow processes and river channelmorphology. In: Catchment experiments in fluvial geomorphology,edited by T. P. Burt & D. E. Walling, 489-514. Norwich: Geo Books.

Institute of Geological Sciences. 1977. Sand and gravel resourcesof the Grampian region. (IGS report 77/2.1 London: HMSO.

Johnstone, G. S. 1966. The Grampian highlands. 3rd ed. (Britishregional geology.) London: HMSO.

Leopold, L. B. 1982. Water surface topography in river channels andimplications for meander development. In: Gravel-bed rivers, editedby R. D. Hey, J. C. Bathurst & C. R. Thorne, 359-388. Chichester:John Wiley.

22

Leopold, L.B. & Wolman, M. G. 1957. River channel patterns:braided, meandering and straight. (Professional paper no. 282-B.)U.S. Geological Survey.

Leopold, L. B., Wolman, M. G. & Miller, J. P. 1964. Fluvialprocesses in geomorphology. Oxford: Freeman.

Lewin, J. & Brindle, B. J. 1977. Confined meanders. In: Riverchannel changes, edited by K. J. Gregory, 221-233. Chichester:John Wiley.

Lowe, J. J. & Walker M. J. C. 1977. The reconstruction of theLate-glacial environment in the southern and eastern GrampianHighlands. In: Studies in the Scottish Late-glacial environment,edited by J. M. Gray & J. J. Lowe, 101-118. Oxford: Pergamon.

McConnochie, A. I. 1900. Deeside. Aberdeen: L. Smith.

Natural Environment Research Council. 1975. Flood studiesreport. 5v. London: NERC.

Newson, M. D. 1981. Mountain streams. In British rivers, edited byJ. Lewin, 59-89. London: Allen & Unwin.

Simons, D. B. & Li, R-M. 1982. Bank erosion on regulated rivers. In:Gravel-bed rivers, edited by R. D. Hey, J. C. Bathurst & C. R. Thorne,717-754. Chichester: John Wiley.

Sissons, J. B. 1967. The evolution of Scotland's scenery. Edin-burgh: Oliver & Boyd.

Smith, J. S. 1981. The scenery of the Grampian Region. DeesideField, 17, 11-16.

Sugden, D. E. & Clapperton, C. M. 1975. The deglaciation of UpperDeeside and the Cairngorm Mountains. In: Quaternary studies innorth-east Scotland, edited by A. M. D. Gemmel!, 30-38. Aberdeen:Aberdeen University Press.

Thorne, C. R. & Lewin, J. 1979. Bank processes, bed materialmovement, and planform development in a meandering river. In:Adjustments of the fluvial system, edited by D. D. Rhodes & G. P.Williams, 117-137. London: Allen & Unwin.

Vasari, Y. 1977. Radiocarbon dating of the Late-glacial and EarlyFlandrian vegetational succession in the Scottish Highlands and theIsle of Skye. In: Studies in the Scottish Late-glacial environment,edited by J. M. Gray & J. J. Lowe, 143-162. Oxford: Pergamon.

Warren, J. S. 1985. Hydrology of the River Dee and its tributaries.In: The biology and management of the River Dee, edited by D.Jenkins, 23-28. (ITE symposium no. 14.) Abbots Ripton: Institute ofTerrestrial Ecology.

Werritty, A. 1984. Stream response to flash floods in uplandScotland. In: Catchment experiments in fluvial geomorphology,edited by T. P. Burt & D. E. Walling, 537-560. Norwich: Geo Books.

Werritty, A. & Ferguson, R. I. 1980. Pattern changes in a Scottishbraided river over 1, 30 and 200 years. In: Timescales in geomor-phology, edited by R. A. Cullingford, D. A. Davidson & J. Lewin,53-68. Chichester: John Wiley.

Willetts, B. 1985. River engineering and the control of erosion. In:The biology and management of the River Dee, edited by D. Jenkins,105-110. (ITE symposium no. 14.) Abbots Ripton: Institute ofTerrestrial Ecology.

Wyness, F. 1950. Royal Deeside Aberdeen: Deeside Field Club.

Hydrology of the River Dee and its tributariesJ S WARRENNorth East River Purification Board, Aberdeen

1 IntroductionThe River Dee drains a catchment of 2100 km2 on theeastern side of the Grampian Mountains. In the first20 km, the land falls 974 m from the summit of BenMacdui to the river gauging station at Mar. In theremaining 90 km, the river falls a further 335 m beforedischarging to the sea at Aberdeen.

This paper first shows the dimension of the waterresource by illustrating the spatial distribution of theprecipitation within the catchment, and converting thisto monthly averages. It then draws attention to theparticular nature of the river regime by contrasting itwith the regime of the River Ugie (discharging 40 kmto the north) and comparing the 'effective rainfall' ofthe Dee catchment with its discharge characteristics.

The distribution of the data collection sites is alsoexamined, and may assist readers to assess theavailability of material for further study.

2 Rainfall, climatological and river records2.1 Rainfall

Rainfall records appear to have commenced in the Deevalley with the establishment of a rain gauge atBraemar in 1856. Since then, the records of a further70 rain gauges have been added to the archive. Not allof these gauges are extant. Girdleness discontinuedrecording in 1982 after 120 years, and Drum Castle in1975 after 109 years.

A number of rainfall stations also record climatologicaldata such as temperature, windspeed, and so on.From time to time, special research is initiated bygovernment- or university-sponsored bodies and addi-tional data are collected. These special projects oftenbecome the base for a scientific paper, but theirfindings rarely form part of any national archive. Forexample, both the Central Electricity Generating Boardand the Meteorological Office have automatic climat-ological stations on the hills in catchments adjacent tothe Dee. The CEGB was studying the relationshipbetween performance and power lines. The Meteor-ological Office is recording weather conditions on thehills and evaluating their sensors and instrumentation.Research is also being carried out in the Cairngorms bythe Institute of Hydrology on the physical processes ofsnowmelt.

River levels have been recorded in the Dee catchmentsince 1892, but the early records are sparse andsporadic. Some early river levels were recorded at theold Bridge of Dee and in Park Estate, and (for varyingperiods) at Balmoral, Dinnet and Cults. At present, thecatchment is served by 10 flow recording stations

(Table 1, Figure 1). Several sites are also gaugedintermittently in support of special projects, such asthe acid rain investigation and low flow yield studies.

Table 1. Flow recording stations in the River Dee catchment in 1984

*Height records not yet converted to flow**Under construction

23

2.2 Climate and precipitation

Precipitation within the Dee valley ranges in structurefrom haar on the coastal strip, which for this purposecan be ignored, to heavy convectional rain withintensities in excess of 25 mm 11-1. Snow accounts fora significant proportion of the precipitation which goesinto storage, although it is of lesser intensity than thatexpected of rain. Annual precipitation ranges from700 mm on the coast to over 2000 mm in themountains (approximately 1000 mm over the catch-ment as a whole).

The bulk of the catchment's precipitation is providedby Atlantic depressions. In winter, these pass up thewest coast of Scotland, their associated troughs andfrontal systems passing over the Grampian Mountains.Autumn precipitation is largely convectional and borneon the wind circulation in the rear of depressions to theeast of Scotland. Much of this precipitation is inter-cepted by the surrounding hills, and in winter isdeposited as snow. The snow may run off precipitatelywith a change of air mass, but more commonlyaccumulates on the hills until the gentle thaws ofspring.

Figure 2 shows the annual average isohyets in mmover the Dee catchment for the standard period1941-70. The main points are the light precipitation onthe coast compared with the surrounding hills and thearea of comparatively light precipitation betweenBanchory and Ballater. In 1941-70, the averagemonthly precipitation in the catchment above Wood-end did not drop below 70 mm or rise above 125 mm(Figure 3). As would be expected, the winter monthswere wetter, but the spring was rather drier thanmid-summer. With a difference of only 55 mm of

24

2

1400

Recording river flow stations A

Rain gauge sites

, •-..., -. 1\ t\....1

- •

BALLATER •.7 •

sf„

A--'-7----"BRAEMAR•f

1

Figure 1. Rainfall and river flow stations within the Dee catchment in use in 1984

rainfall equivalent, the Dee catchment cannot be saidto have a well-marked seasonal variation of actualprecipitation.

However, if the average evapotranspiration is takeninto consideration, the seasonal variation of 'effectiveprecipitation' becomes apparent. Effective precipit-ation is defined as 'the proportion of actual precipitationwhich remains after the losses to the atmosphere byevaporation and transpiration have been subtracted'

Figure 2. Annual average rainfall in the Dee catchment, 1941-70

vtwer Don

BANCHORY

ABERDEEN

(MacDonald & Partners 1975), ie the precipitationavailable to run immediately down the river channel orto be held in storage. A diagram of effective precipit-ation would look very similar to that of a 'simple' riverregime (see below, and Figure 3).

3 River regimeA 'simple' river regime exhibits one period of both highand low flows annually. In Britain, this would be due tothe effective rainfall in summer being lower than that

JF MA MJJA SOND

Figure 3. Average monthly rainfall and evapotrans-piration above Woodend, 1941-70

of winter, with no appreciable storage of snow in theupper catchment to augment spring river flows. Theregime of the Dee, however, is not 'simple', butmoderately complex, and it conforms to that regimeknown as 'alpine'. Figure 4 compares the regimes ofthe Dee and Ugie. The units used in this diagram arelitres per second per km2 (I s-' km-2) of catchmentand the difference between the 2 values is due to thedifference in rainfall in the 2 catchments. The meanmonthly flows of the Ugie conform in pattern to that ofa simple regime, whilst those of the Dee exhibit a

110

100

90

80

70

60

50

40

30

20

10

40

30

ID RainfallKi Evapotranspiration

River Dee--- River Ugie

20 I II __II Ii__I I

I IL-1 __I

I10 I I--I II I

JFMAMJ JASOND

'hump' or an increase of flows in the spring when thesnow stored in the upper catchment itself melts andruns off.

4 RunoffFigure 5 shows the mean monthly flows for the Dee atWoodend for 1941-70 plotted in the same units (mmof rainfall equivalent) and on the same graph as theeffective precipitation for the catchment. In August toFebruary, effective precipitation exceeded runoff. Inthese months, water was stored both as increasedgroundwater and as snow on the hills. In March andApril, flows increased whilst effective precipitationbecame less. This increase in flow was due tosnowmelt which continued for a further 3 months,although it was only rarely detectable on the Woodendhydrograph during July. At this time of the year, smallvariations of flow could be attributed to either snow-melt or diurnal evapotranspiration.

It is difficult to be specific about the amount of snowstored in the upper catchments, as it varied spatiallyboth vertically and horizontally. For relatively smallcatchments, an estimate may be made by eye. Therainfall equivalent of lying snow would be about one

120

110

100

90

80

E 70

60

50

40

30

20

10

— Precipitation--- Runoff

L_ L_

L__

JFMA.MJJASOND

25

Figure 4. Average monthly river flows in the Dee at Figure 5. Effective precipitation and runoff at Wood-Woodend, 1930-77, and Ugie at Inverugie, 1977-79 end, 1941-70

26

tenth of the snow depth, with older snow havinggreater density than fresh falls. It has been calculatedthat an experimental catchment at an elevation of950 m on the western side of the Cairngorm Moun-tains discharged 1250 mm of rainfall equivalent ofsnowmelt in 1984. At Woodend, the rainfall equivalentof snow for the entire catchment was in the order80 mm or 109.6 Mm3 of water.

Sections of the 1984 hydrograph for Woodend (Figure6) are included to show the effect of diurnal snowmelton river levels in April and the decrease of this effectby June. The areas between the peaks and troughs inthis hydrograph are due entirely to snowmelt. The areabelow the troughs is due to groundwater discharge.The hydrological day starts at 0900 hours and this timeis marked in Figure 6 by the vertical line on thehorizontal scale. Discharge is indicated by the verticalscale. The first peak shown, which would haveoccurred _about 0300 h on 26 April, represents adischarge during the 24-hour period of 2 Mm3 of waterdue entirely to snowmelt. By June, the contribution toflow due to snow was barely detectable and the peakoccurred at approximately 1400 h, thus taking about24 h to travel from the upper catchment to Woodend.

Daily mean flows in rivers are usually examined bymeans of a flow distribution diagram plotted on Log xProbability paper. This diagram gives the percentage of

120

100

90

80

70

60

20

15

10

April

26

8 9

27

June

Date

daily means that are equal to or greater than a certainfigure. Selected values from a flow distribution dia-gram for Woodend are listed in Table 2. Further valuescould be obtained by replotting the values given andinterpolation.

*% of all daily mean flows on which this value (cumecs) wasreached or exceeded

5 High flowsA most useful method of expressing flood statistics isby referring to the 'return period'. For example, a floodequal to or greater than 100 cumecs (m3 s-1) can beexpected to occur at Woodend every year (selecteddata in Table 3), and is therefore said to have aone-year return period. Statistically, a flood of1500 cumecs can be expected to occur once every

28 29 30

10

Figure 6. Hydrographs for the Dee at Woodend in April and June 1984, showing snowmelt

11 12

Table 3. Estimated high and low flow return periods for Woodend

100 years, and this would be said to have a 100-yearreturn period. Return periods for events at any locationare usually plotted on Log x Probability paper and it isbest to refer to these diagrams if a full range of returnperiods is to be examined.

Wyness (1968) records a spate on the Dee in 1738,and found that similar floods occurred on the Dee with'startling regularity every 30 years'. A flood with a30-year return period would give a flood flow ofapproximately 1000 cumecs at Woodend. Although a30-year return period is a statistical concept and doesnot mean that a flood of this magnitude will return at30-year intervals, it gives a datum by . which to judgehigher floods.

The 1829 flood is estimated to have had a maximumdischarge of 1900 cumecs at Woodend (NERC 1975),but this value is probably over-estimated. However, itwas certainly in excess of 1000 cumecs. Reliablerecords of floods (in terms of flow) do not existbetween 1829 and 1929, but the flood of 1920 wasalmost certainly over 1000 cumecs; 1920 is almost amultiple of 30 years from 1829. Two more floods over1000 cumecs have occurred since then, the first in1937, followed by a space of 17 years before thesecond flood in 1951, which is again nearly 30 yearssince 1920.

If one ascribes anything other than chance toWyness's observations, a flood of 1000 cumecs isimminent. Examination of the Woodend record for1930-1984 shows that there were 5 floods greaterthan 500 cumecs in the first 25 years, and that therehave been 8 greater than this value since, but noneover 1000 cumecs as previously indicated.

By using water levels alone, as opposed to flowfigures, we can extend the continuous record back to1920. December and January were the months inwhich the greatest number of occurrences took place(Table 4), although not necessarily the months which

Table 4. Annual maximum flood for Woodend, 1920-83

MonthEvents

MonthEvents

J FMAMJ J A SOND13 3 3 2 1 0 1 5 4 8 7 17

Table 5. Annual minimum daily mean at Woodend, 1930-84

J FMAMJ J A SOND0 0 1 0 2 5 16 14 10 5 0 1

27

produced the highest floods. In contrast, June did notproduce a single flood event. It can be concluded thatrarely does the annual maximum flood occur in themonths of seasonal snowmelt, although many areassociated with heavy rain and a change of air masswhich also melts snow. The flood of 1937 whichoccurred in January was probably due to this cause.However, the extremeflood of 1829 occurred inAugust and was associated with intense convectionalprecipitation, as was the flood of 1920.

An analysis of hydrographs within the catchment mayprovide mean relationships between various locations,which can often be invaluable in estimating the heightand time of arrival of a flood at a downstream site,given the height and time of its peak at some positionupstream. For example, a peak of 400 cumecs atWoodend can be expected to have risen to500 cumecs when arriving at Park one hour later.Peaks of lesser magnitude travel correspondinglymore slowly. This is a mean relationship and thecontribution made by the intervening catchment maydiffer slightly from flood to flood. Similar relationshipsexist between all stations within the Dee catchment,but this example is sufficient to show the variedinformation that may be gleaned from the hydro-graphs.

6 Low flowsJuly or August are usually the months in which thelowest flows are recorded in the Dee. However, inyears with low summer rainfall, the recession of thehydrograPh may extend into September or evenOctober. Low flows have also occurred in winter dueto potential runoff being locked in snow, frozengroundwater and river ice. Table 4 shows one occasionbetween 1930 and 1984 when the lowest annual flowoccurred in March. This was in 1963 during a particu-larly hard winter. The December occasion was in 1958.

The magnitude and occurrence of minimum dailymean flows may be examined with the aid of a lowflow frequency diagram (cf selected data in Table 3),which is very similar in structure to the flood frequencydiagram. These diagrams for the Dee, and indeed forall rivers in the north-east, are kept constantly underreview by the River Purification Board.

There has.been much discussion in recent years aboutthe possibility that summer flows are becomingprogressively lower. It has been suggested thatchanges of land use, increase of drainage, or forestryactivities are to blame. Figure 7 shows the annualminimum daily mean flows for Woodend from 1930 to1984. For the 25 years 1930-54, the minimum did notfall below 5 cumecs. In the succeeding 25 years,1955-79, it dropped below this figure on 7 occasions,which may suggest that summer flows are gettinglower. However, no significant statistical trend hasbeen identified.

28

13

12

11

10

9

6

5

4

31930 1940 1950

Figure 7. Annual minimum daily mean flows at Woodend, 1930-84

In 1975, Sir M MacDonald and Partners carried out aregional study of water resources on behalf of theNorth East of Scotland Water Board. They concludedthat the period prior to 1972 should be subject to acorrection applied on a sliding scale from 1934 until1972. These corrections effectively reduced the com-puted annual minimum daily flows by approximately2 cumecs. Figure 7 incorporates these correctedminimum flows. A student of low flows in the Deeshould be aware that there are 2 sets of records for theperiod 1934-72. The National Archive will hold theoriginal records, although these are now likely to havebeen converted from the imperial units in which theywere taken.

Referring to the early records, Sir M MacDonald andPartners state 'in view of the various uncertainties, asophisticated correction procedure is hardly justified'.However, subsequent records appear to indicate thatsome correction is justified, and I conclude that thelower summer flows since 1955 really were lowerthan originally assumed. Whether this phenomenonwas due to changes in catchment characteristics or tochanges in climate deserves further investigation.

7 SummaryThe Dee is the major drainage basin on the easternside of the Cairngorms. Average annual rainfall varies

1960 1970 1980

between 700 mm on the coast to 2000 mm on thehills where over 25% may fall as snow. Snowremaining in the upper catchment over winter andproviding excess runoff during the spring gives theriver an 'alpine' regime.

At Woodend, no flows below 3 cumecs or 2.2 I s-1km-2 of catchment have been recorded. Ninety percent of Woodend flows are above 10 cumecs and 5%are above 90 cumecs. The highest flood recorded wasin 1829 and has been estimated to have produced aflow of 1900 cumecs at Woodend, or 1387 I s-1 km-2.

Quantitative records of hydrological phenomena withinthe catchment exist from 1829, although early recordsare sparse. The present-day network of hydrologicalsensors is sufficient for overall resource management,but would require augmenting for special or specificstudies.

8 ReferencesMacDonald, Sir M. & Partners. 1975. Regional water resourcesstudy for the North East of Scotland Water Board. Vol. 2. Cam-bridge: Sir M. MacDonald & Partners Ltd.

Natural Environment Research Council. 1975. Flood studiesreport. Vol. IV. London: NERC.

Wyness, F. 1968. Royal valley, the Aberdeenshire Dee. Aberdeen:Alex P. Reid.

Land use within the catchment of the River Dee

J S SMITHDepartment of Geography, University of Aberdeen

1 IntroductionThe paper establishes the origins and physical charac-ter of the environment of the catchment of the Dee.The human exploitation patterns over the last 5000years are examined and demonstrated to have had, atthe least, minor impact throughout the catchment,notably on the characteristics of soil and plant com-munities. The major impacts have been concentratedon lower Deeside where agrarian efforts have beenconcentrated, and where the intensity of usage hasbeen greatest. In the upper parts of the catchment, thefrontier of permanent settlement has receded sincethe middle of the 18th century, and the intensity ofusage has subsequently diminished. The improvingmovement in agriculture and the Victorian develop-ment of sporting land uses are regarded as particularlyformative periods in the development of the contem-porary cultural landscape. It is argued that the historicalchanges in land use discussed, while clearly affectingthe lives of the contemporary inhabitants and substan-tially modifying the character of Deeside landscapes,were not dramatic or rapid enough to have substan-tially modified the Dee's discharge or water quality,although a number of natural habitats were eitherreduced in area or eliminated.

2 Physical environmentThe planimetric shape of the Dee catchment isrectangular, about 140 km long and averaging 20 kmbroad, but tapering markedly eastwards over the last25 km. Its physiographic character is an incised majorvalley set within a series of dissected plateau surfaces,valley benches and inland basins. The Dee valleywestwards of Kincardine O'Neil is narrow and en-trenched, but eastwards the middle and lower reachesopen out with the channel flowing through a series ofpre-Quaternary basins with lowland characteristics,notably Cromar and Feughside.

Less than 20% of the overall catchment lies below244 m above OD, arguably a significant altitude forsuccessful ripening of cereals at this latitude, and thismore agriculturally productive ground lies mainly in thelast 35 km of the catchment's long axis. The majorelements of catchment scenery are generally ofpre-Quaternary age (Smith 1981), a feature whichdistinguishes Deeside from most other parts of theScottish highlands which tend to exhibit a greaterdegree of glacial modification of their physical land-scape.

The headwaters of the Dee originate in the Cairn-gorms. These hills collectively form the largest area ofcontinuous high ground in Britain, thus acting as amajor originator of weather events whose effects are

29

transmitted downstream into middle and lowerDeeside in terms of flood peaks. The overall length ofthe catchment and its east-west orientation results inits upper parts, especially the valley floors, experienc-ing a relatively continental climate, with a combinationof wide annual ranges of temperature and compara-tively low rainfall. Exceptionally low temperatures arerecorded in valley floor situations in winter, and atBraemar air frosts may occur from late Augustonwards. Increasing altitude, on the other hand,dramatically shortens the growing season and both inthe past and at present markedly affects both thecharacteristics and the utilization of the bio-geographical resource.

Reconstruction of the natural environment of thecatchment prior to human interference is facilitated bythe general uniformity of the geological base (Maize Is1985), which consists of very extensive areas of acidicgranites with restricted outcrops of base-rich horn-blende schists and metamorphic limestones. Thelatter are located on the eastern flanks of the Strachanbasin, on the valley floor between Aboyne and Birse,and as localized outcrops on the Baddoch, the Clunieand the flanks of Morrone in the Braemar area.Although largely concealed by glacial drift, the meta-morphic limestones are mirrored by the distribution oflime kilns. Hornblende schists outcrop notably in theMorven mass, and in less extensive outcrops on thebenches and plateaux of the middle Muick. While thesummits of the highest hills carry skeletal soils,bedrock outcrops and peat, in the valleys the parentmaterials tend to be dominated by soils developed onfluvioglacial sands and gravels containing high propor-tions of granitic materials.

The surviving oakwoods at Dinnet and Craigendarroch,albeit semi-natural, recall the character of the originallowland climax woodland prior to human settlement.Historical accounts dealing with Culblean in Cromarconfirm the more widespread extent of oak woodlandeven in the historical past. The present natural climatictreeline of native conifers lies around 700 m aboveOD, although the tree stumps of pine around theflanks of the Cairngorms are evidence of a still highertreeline during the climatic optimum (Nethersole-Thompson & Watson 1974).

While the changing post-glacial climate must haveproduced altitudinal migration of the natural life zones,by far the most important factor affecting the presentcharacteristics of the natural biogeographical resourcein terms of surviving natural woodland and soilcharacteristics has been human exploitation patternsover the last 5000 years. These changes, which were

30

ultimately to affect, in varying degrees, almost thewhole Dee catchment, are likely to have had at leastminor repercussions on water and sediment budgets.However, the only hard evidence published to daterelates to the Cromar area (Edwards 1979), and dealswith the prehistoric farming and stock-rearing phaseand its effects on rates of soil erosion. Recentlypublished work by the Institute of Hydrology on theeffects of extensive conifer afforestation in certainwestern upland catchments in Britain suggests that,as the canopy develops, water release from suchcatchments may decrease by up to 20%. It would beunwise to extrapolate such figures to a Deesidecontext, and, in any case, established sporting in-terests and shortage of plantable ground would seemto preclude further extensive afforestation in theforeseeable future.

3 Prehistoric land usageAlthough the cumulative impact of human interferenceon the natural environment of the Dee valley de-creases westwards, in terms of woodland cover only afew isolated but important stands of native forestsurvive today, the remainder having been cleared forcultivation, grazing and timber production. Inroads intothe natural forests proceeded earliest on the easternvalley floor and terraces. Here, lighter fluvioglacial soilswith initial brown forest soil characteristics werecleared as early as 3000 BC by the first pioneerfarmers. The recent excavation of a substantialNeolithic timber hall nearly 23 m in length at Balbridie,near Crathes, is evidence not only of considerableagricultural activity in lower Deeside (Ralston 1984),but also of the availability of oak in quantity and size forconstruction activities. Like the Mesolithic hunter-gatherers they followed, these communities lived nearthe river, although seasonal mobility was an essentialpart of their modus vivendi. The work of Edwards(1979) at Braeroddach and Loch Davan on middleDeeside reveals that clearance and grazing acceleratedrates of soil erosion in Cromar.

By the Bronze Age, evidence is firm for repeatedcycles of clearance and regeneration, and the impress-ive density of field systems and tracks in Cromar(Ogston 1931) is supplemented by additional evidenceof agricultural activity surviving beyond the presentlimit of cultivation. According to Edwards' sites, thepollen evidence suggests 'extensive and sustained'usage of the ground for both pastoral and agriculturalactivities.

However, the total package of prehistoric field monu-ments on Deeside suggests that permanent humanendeavour was concentrated in those parts of thecatchment east of Bal later, although sites do occur asfar west as Crathie.

4 The medieval periodBy the medieval period, with increasing human popul-ation, colonization on Deeside extended westwards to

conflict with largely recreational activities ongoing inthe 'hunting forests' of Morven, Culblean, Strathdeeand Mar on upper Deeside. In the days of Robert theBruce, such hunting forests extended almost to theoutskirts of the burgh of Aberdeen. The land grants ofthe Irvines of Drum and the Burnards of Leys includedin their charters clauses involving a degree of con-servation of these earliest of 'reserves'. The mainthrust of management was the maintenance of 'vertand venison'.

The pressures on these medieval reserves, until the16th century largely the property of Scottish Kings heldvia their trustees, were not only through directclearance for agriculture, but for firewood, grazinganimals and for useful domestic products such asagricultural implements and candle-fir. Indirect effectswrought by goats and other grazing animals must alsohave been substantial (Mather 1969). They clearlyimpeded the 'natural shifting of the forest stance'observed by Pennant in Ballochbuie as late as the 18thcentury. Increasing inroads into the natural woodlandled to a decrease in species variety and resulted in theearly demise of the wolf and brown bear.

Medieval Deeside was locally governed by the Earl ofMar, whose major power centre lay at KindrochitCastle, Braemar, with a subordinate centre at Migvie,Cromar. The higher glens and mountains of upperDeeside were used seasonally and were activelyconserved as hunting land for local aristocracy androyal visitors. In middle and lower Deeside, as thedistribution of castles and tower houses shows, theland had been parcelled up amongst feudal barons,and the adjoining lands used agriculturally for subsist-ence purposes. The evident difficulties of obtainingprime timber for castle and tower house constructionreflected the demise of the original woodland, longsince broken up and cleared by increasing populationpressures. As the 1696 Poll Book reveals (Walton1950), population density closely mirrored the inherentproductive capacity of the land. In Deeside, the bestpotential arable land lay either in the lowland basins ofCromar and Feughside, or on the alluvial and terracedbottomlands of the Dee and its major tributaries.

As the Earldom declined, so new powerful familiesintruded into Mar, notably the Farquharsons in upperDeeside, and the Gordons in middle Deeside. The clangroups which developed in Farquharson territorygreatly increased agriculture and stock-rearing press-ures in upper Deeside. In the upper parts of the valleyand its major tributaries, the herds, notably blackcattle, were at least as significant as crops in the localeconomy, and the production of the essential grainproducts on the arable ground was only possiblethrough the often distant shealing of the stock duringthe summer. The Cromar population appears to havehad shealing rights in upper Strathdon, while similarsummer pasturing patterns on Deeside, notably inGlen Shee, Glen Cal later and Glen Baddoch, extendedthe ground utilized, at least on a temporary basis.

5 Improved agriculture and recent land useConservation measures were introduced in Mar by thelandowners at least as early as the 16th century, in anattempt to reduce grazing in the pinewoods andencourage regeneration. Although the populationgrowth throughout Deeside was frequently checkedby feuding and famine, gradually agricultural coloniz-ation pushed higher up the glens and hillsides, withshealings on occasion being taken into permanentagricultural occupation. By the 18th century, this landpressure, combined with a marked decline in thedemand for deer stalking in the old grand style, meantthat deer forests were confined to the most remoteareas, while in the middle and lower Deeside catch-ments the new agricultural techniques of drainage andenclosure were creating a new agrarian landscape.

This phase of agricultural improvement was tempor-arily checked in upper Deeside at least by the militaryinvolvement of the clans during the Jacobite period(1689-1746), culminating in the appearance of 300Invercauld Farquharsons at Culloden. Following thedemise of the Stuart cause, a combination of landforfeiture and changing aspirations amongst the land-owners introduced new attitudes to land use andestate management (Smith 1979). New lairds like theDuke of Fife acquired substantial lands in upperDeeside (Balmoral and Mar), and others (eg Farqu-harson of Monaltrie) subsequently regained theirforfeited -properties. Together, they embarked onambitious programmes of road, bridge and landimprovement, including the development of the miner-al waters at Pannanich (Smith 1983). Commercialism,that is the seeking of economic returns on investment,already apparent in the lowland pattern of individualfarms set within apportioned and compact holdings,was mirrored in the upland areas by the first seriouscommercial plantations of conifers, the creation ofestate arboreta, and the improvement of hill stock.

The fashion of deer stalking was revived in Deesidefrom the 1790s onwards, and by 1826 the first MarLodge, which was to be largely destroyed by themuckle spate in August 1829, was advertized forletting as the centre of 'the finest shooting district inScotland' at £1,800 per year. Attempts to stock thehigh glens like Lui as sheep walks were quicklyabandoned in favour of more lucrative shooting lets,and extensive areas within the upland catchmentpreviously used for extensive grazing or shealing werelet either as deer forests or grouse moors. Thisresurgence of sporting land uses caused a withdrawalof agriculture from the highest tidemark of the early18th century, and effectively ended the shealingsystem. The purchase of Balmoral in the 1850s by thePrince Consort provided a further burst of publicity tothe land use of sporting lets, and shooting lodgesstaffed by a seasonal population appeared in even theremotest glens.

A marked feature of the Victorian period on Deeside,particularly well exemplified in Glen Tanar, was the

effect a single laird's tastes could have on the culturallandscape (Mather 1969). Over the same period, treeplanting became increasingly popular, initially withScots pine and larch, and later with North Americanconifers. Since 1919, the Forestry Commission haveacquired and planted land (eg at Alltcailleach), whileseveral private estates have long regarded forestry asan integral part of estate land use, notably at GlenTanar and Invercauld. There has been substantialreplanting of those areas felled during the 2 WorldWars, and also of ground devastated by the great 1953gale.

6 Contemporary land use and population trendsThe altitudinal and environmental contrasts existingwithin the Dee catchment are reflected in the variednature of the agricultural economy and current landuse (Table 1). Middle and upper Deeside are character-ized by the dominance of upland and hill farms, whileeast of Kincardine O'Neil farming takes on the charac-ter of lowland agriculture. The key restrictive featuresare shortness of growing season, the lack of improvedground for cultivation and cropping, and a seasonalmoisture deficit. In the west, stock rearing is thetraditional farming economy, while eastwards crop-ping and occasionally dairying are dominant. Whilethese patterns reflect the environmental limitations foragriculture, recent agricultural trends also reflectnational and European agricultural policies, as a resultof which farmers perhaps determine their economiesthrough their pockets rather than with their hearts. Inthe upland areas where units are substantial and spana variety of terrain, scope remains for landowners toseek a measure of integration between the array ofland uses on a particular estate (eg between forestryand stock rearing), but only if the mechanism ofsubsidy support is devised with such aims in mind.

Table 1.Land uses in the catchment of the River Dee, Grampian, in1980 (sources: June 1980 Agricultural Returns, OrdnanceSurvey maps and unpublished information)

*Deer forest and rough grazings overlapped

31

Generally, maximum populations on Deeside wereattained in the early 18th century in the uplandparishes, and in the early 19th century in the hill-edgeand lowland parishes, reflecting the continued pace ofagricultural reclamation up to the hill edge. However,by the second half of the 19th century, with theexception of the small burghs and the newly created

32

railway villages like Lumphanan and Torphkns, many ofwhich also functioned as Victorian summer resorts,the present pattern of decline in landward ruralpopulations had become almost universal. Improvedefficiency and mechanization in the present century,and particularly since the Second World War, havegreatly reduced the number of people employed onthe land. Amalgamation of farm holdings also con-tributes to the process of rural depopulation.

An increasingly important dimension of establishedland use patterns, both on the riverside and in the hills,has been the demands of weekend visitors, notablyfrom Aberdeen and Dundee. Their usage and accessto the land resource increasingly enters into the landuse decisions of both farmers and landowners. Inevit-ably, the question arises whether such informal accessand utilization require a measure of formalization, orindeed may be tapped for estate revenue. Theestablishment of the Glen Tanar Charitable Trust,together with a Visitor Centre and Countryside Rangerin middle Deeside, may well be a pointer towardsmore formalized access in upper Deeside in the future.One result of increased scientific knowledge of theDee catchment is the identification and scheduling ofregionally and nationally important nature conservationsites. It is important to meld nature conservationinterest with both the land use practices of the peoplewho make their living by the land and the requirementsof visitors who seek their recreation there.

7 DiscussionWhile this conference concentrates on the riverchannel and its immediate environs, a total catchmentviewpoint is the classic unit of study of geomorpho-logists, watershed engineers and geographers. Workon modelled catchments, notably in North America,has demonstrated that major changes in land use suchas deforestation, afforestation and even agriculturaldrainage have downstream effects on channel andfloodplain form and water quality, via the proportion ofwater and sediment released from the appropriatecatchment. Much of this work, however, is derivedfrom studies of extremely dramatic manipulated landuse changes on very small modelled catchments,rather than on real-world situations like the Dee.

Changes in land use beside the Dee, while far reachingover the lengthy period of time outlined above, havebeen relatively slow, and the characteristics of thecatchment are large and contrasting. With the notableexception of water abstraction, particularly appropriatein 1984 when water flows reached the lowest figuresince records began in 1929, the Dee and its tributar-ies remain unregulated. It is a wild river, by comparisonwith many of its highland and European counterparts.In nature conservation terms, the natural instability ofwater channels creates many interesting albeitephemeral habitats, eg on point bars, backwaters andmid-channel bars, even within the floodplain. Totalregulation through engineering works precludes the

emergence of these new habitats, and therefore maybe regarded is deleterious to the total biologicalpackage, although a farmer concerned with problemsof bank erosion and periodic flooding would viewregulation in a rather different way.

At present, it seems unlikely that any major changes inland use within the catchment will occur which wouldproduce any significant effects on the Dee's naturalbiology and regime. On the other hand, rationalizationof Common Market agricultural production to matchEuropean demand would clearly sound the death knellto a large part of marginal Scottish agriculture. It wouldleave a void to be filled by either forestry or recreation,or a combination of both.

The likely effects of extensive conifer plantation inmiddle and upper Deeside on water yields as the treesreach maturity may be argued to be substantially lessthan the 20% decrease estimated for the wetter westof the country where ground-storey vegetation is grassrather than heather. Nonetheless, the Deeside catch-ment lacks the appropriate evapotranspirationmeasurements to enable meaningful predictions to becalculated. Although the proximity of both commercialplantations and native woodlands to the main publicroads in upper Deeside suggests a heavily forestedcatchment, measurements summed per km squareusing the most recent Land Ranger 1:50 000 Ord-nance Survey sheets (early 1980s) reveal that thecatchment west of Kincardine O'Neil (catchment area1365 km2) carries 147 km2 woodland (a cover of10.8%), while the remainder (catchment area735 km2) carries 117 km2 woodland (a cover of 16%).A key figure not available at present is the area ofground west of Kincardine O'Neil which remainsunplanted and below the economic treeline.

With increasing leisure time, demand for watersideaccess on Deeside seems likely to increase, and it isnotable that such access points on lower Deeside,where demand is potentially greatest, are very few (egPotarch). Informal waterside access is long estab-lished on the Feugh, the Quoich and the lower Tanar.

Writing in 1840, the minister of Drumoak parishcommented that 'the super-abundant waters of about900 square miles feed its current, but so clean are itsgravelly banks and pebbly channel that its waters rivalin purity the most limpid streams in Scotland . . . thisbeautiful river, although superior to its gentler neigh-bour, the Don, in the value of its salmon fishings, hasnever, like it, been rendered serviceable to themanufacturer; nor does it compensate this unprofit-ableness to the manufacturer by rivalling the Don inusefulness to the agriculturalists, for instead of fatten-ing his meadows with a rich alluvium, its inundationscarry off the best part of the soil and deposit in itsplace a bed of sand possessed of little vegetativepower; while its wintry torrent, unwilling to berestrained, contends powerfully with the embank-

ments which have been raised to protect the haughsfrom its destructive ravages, and occasions greattrouble and expense to the proprietors, for repairingthe breaches which it frequently makes' (New Statis-tical Account, Drumoak, 1841).

The papers in this symposium illuminate the extent towhich this characterization of the Dee in terms ofcleanness and irregularity of regime remains truetoday. The land use within the Dee catchment, whichis the logical response to the environmental setting ofvalley, hill edge and mountain, makes the catchmentscenically attractive and ecologically diverse. Althoughthis diversity of land use is traditionally the response offarmers to difficult agricultural environments, thepattern is ultimately dependent on the CommonAgricultural Policy.

While nature conservation and recreation will beincreasingly important land uses in Deeside in the 21stcentury, it is to be hoped that policies will beadministered with a view to retaining a workingagricultural population in the catchment, as there arelimitations to the attractiveness and usefulness of awholly depopulated countryside. The present situationseems to offer the best compromise between ameasure of resource productivity and a measure ofecological diversity. The pressures on the Dee and itschannel and channel-side habitats are likely to comefrom demands for water abstraction and watersiderecreation, rather than from agricultural land uses.

8 SummaryWhile the management strategies of the catchmentover the past 5000 years may be described in generalterms, a lack of basic data on these systems and onpast physical and biological conditions within the Deeand its tributaries makes it impossible to commentauthoritatively about the effects of history and land useon the channel. The major changes in the nature of theplant communities described spanned thousands ofyears, and therefore the published data on smallmodelled catchments in comparable latitudes whichdeal with rapid and large-scale changes in land use donot seem to be relevant. The attributes of lack ofregulation which make the Dee increasingly importantat a time when many major rivers in Europe have beendeveloped for reservoir construction and navigationhave been perceived to conflict with full agriculturalusage of the floodplain. The natural instability of themajor part of the channels of the Dee and its tributaries

33

and their irregularity of regime create interestingchannel and floodplain habitats. The value of the totalDee biological package is increased by the range ofbiogeographical zones through which it flows. Majorchanges in catchment land uses in the future, egextensive afforestation, depend on national and Com-mon Market strategies, rather than on the inclinationsor tastes of the individual farmer or landowner. Themost likely changes for the future, in the absence ofdeliberate policies formulated to achieve the contrary,are those of limited afforestation, further rural de-population of the agricultural areas, and the provisionof increased and probably formalized opportunities forrecreation, both in waterside and hill locations. Itwould be unfortunate if the present land use diversitywhich is scenically attractive and ecologically diverseshould dramatically change in favour of less diversesystems. Prime consideration should be given to themaintenance of a working agricultural populationwithin the catchment. The requirements of natureconservation and informal recreation might be moreacceptable to farmers and landowners if the reductionin land use opportunity they frequently bring with themcould be financially recompensed.

9 ReferencesEdwards, K. J. 1979. Palynological and temporal inference in thecontext of prehistory with special reference to the evidence fromlake and peat deposits. Journ. archaeol. Sci., 6, 255-270.

Maize Is, J. K. 1985. The physical background of the River Dee. In:The biology and management of the River Dee, edited by D. Jenkins,7-22. (ITE symposium no. 14.) Abbots Ripton: Institute of TerrestrialEcology.

Mather, A. 1969. The development and characteristics of presentland use. In: Royal Grampian country, edited by K. Walton.Aberdeen: Department of Geography, University of Aberdeen.

Nethersole-Thompson, D. & Watson, A. 1974. The Cairngorms.London: Collins.

Ogston, A. 1931. The prehistoric antiquities of the Howe of Cromar.Aberdeen: Third Spalding Club.

Ralston, I. B. M. 1982. A timber hall at Balbridie farm. AberdeenUniv. Rev., 168, 238-249.

Ralston, I. B. M. 1984. Notes on the archaeology of Kincardine andDeeside district. Deeside Field, 18, 73-83.

Smith, J. S. 1979. War and peace in the Braes of Mar. AberdeenUniv. Rev., 161, 36-43.

Smith, J. S. 1981. The scenery of the Grampian Region. DeesideField, 17, 11-16.

Smith, J. S. 1983. Lairds and landscape in middle Deeside.Aberdeen Univ. Rev., 169, 40-45.

Walton, K. 1950. Distribution of population in Aberdeenshire, 1696.Scott. geogr. Mag., 66, 17-26.

34

The chemistry of the river systemK B PUGHNorth East River Purification Board, Aberdeen

1 Introduction -This paper describes the chemical quality of the waterof the Dee, its tributaries and its related waters in1980-84.

Draining the south-east side of the Cairngorm range,falling 1310 m in 140 km from its source eastwards tothe sea at Aberdeen, collecting water from a totalcatchment area of 2100 km2, and having a long-termaverage flow of approximately 50 cumecs, the Dee issurpassed in size only by the Tweed, Tay and Spey,and is thus Scotland's fourth largest river. Rising at theWells of Dee (NH 938 989) and Pools of Dee (NH 973008) and receiving contributions from the Lui, Clunie,Gairn, Muick, Tanar, Dye and Feugh (Figure 1), theriver drains high-altitude mountain and moorland withbare rock or hill peat, and, successively, heather,conifer and deciduous woodlands and, finally, lightlystocked pasture interspersed with small pockets ofarable land. Generally, the soils, derived from graniticparent material, are sandy, podzolic and freely drained.Thus, with an absence of heavy industry, and sufferingonly the mildest of insults from agriculture, tourismand urbanization (the human population of Braemar,Ballater, Aboyne and Banchory together totals lessthan 10000), the Dee remains oligotrophic fromsource to estuary. It is one of the few British rivers thatcan still claim such status throughout its length.

2 Analytical methods and site selectionApart from the measurement of temperature and thechemical fixation of dissolved oxygen in the field, alldeterminations were conducted in the laboratory. Forroutine surveys, samples of river water were returnedfrom the field in clean (acid washed and then copiously

o Routine surveys

* Headwaters surveys

Spey

0 10

km

BRAEMAR

TAY RPB

BALLATER

rinsed with de-ionized water), sample-rinsed glassbottles without the addition of preservatives, and theiranalysis was commenced within 24 hours. Standardmethods of water analysis are employed (StandingCommittee of Analysts, various dates).

For studies of the acidification of headwaters and forsurveys of base and trace metal status of the rivers,samples were drawn into similarly cleaned polypropy-lene bottles. Bottles receiving samples for the metalsurveys were pre-dosed with Analar concentratednitric acid, and the bottles filled with sample to yield afinal acid concentration of approximately 0.5%. For theacidification studies, the analysis of all determinands,except the metals, was performed on untreatedsamples, and thereafter Analar nitric acid was added tothe remainder of the sample to a final concentration of0.5% before metal analysis. The metal analysis wasperformed on unfiltered samples.

Determination of metals was by atomic emission orabsorption spectrophotometry with atom excitation ina flame or electrothermal graphite furnace, as appropri-ate.

Sampling sites for routine river surveys were strateg-ically selected to provide an assessment of the wholelength of the Dee, as well as the tributaries in thecatchment. Account was taken of the position ofeffluent discharges, water abstraction and the conflu-ence of tributaries in this selection. There wasfrequently a correspondence between sites forchemical, biological and hydrological data collectionwithin the programmes of work of the River Board (cfWarren 1985; Davidson et al. 1985).

Don

ABOYNE

BANCHORY

Kincardine

Figure 1. Sites used for chemical sampling in streams in the Dee catchment in 1980-84 (cf Table 1)

- ABERDEEN

NORTHSEA

3 Water qualityFor general purposes, river water quality in thenorth-east of Scotland is described by reference to awater quality index (WOI) (Scottish DevelopmentDepartment 1976; Bolton et al. 1978). Chemical waterquality is thereby classified according to placement ona scale of whole numbers from poorest (0) to best(100) quality. With this system, cognisance is taken of9 major riverine characteristics (dissolved oxygen,BOD (biochemical oxygen demand), ammonium-N,pH, total oxidized nitrogen, orthophosphate-P, sus-pended solids, conductivity and temperature).

3.1 The River Dee

Figure 2 illustrates the mean water quality indices of

ClunieWater

RiverMuick

R.Gairn Dinnet B.

RIVER DEE

Figure 2. Water quality indices for the Dee and its tributaries

up to 56 samples drawn from each sampling site(Table 1) in 1980-84. The main watercourse was ofhigh quality (WQI 94) throughout its length, withonly one slight depression (WQI = 92) observed atSilverbank, immediately downstream of the dischargeof effluent from Banchory sewage treatment works.The effects of this discharge were quickly attenuatedby virtue of the high dilution generally available.

3.2 The tributaries

All the tributaries in the upper and middle reaches ofthe catchment (Clunie, Muick, Gairn, Dye and Feugh)were likewise of pristine character. There was a slightdepression of quality in the Dinnet which drains theLoch Kinord/Davan complex (see below).

Watersof Dye& Feugh

Banchory STW

Crynoch& ArdoeBurns

Table 1 Sites for sampling the quality of river water in the catchment of the Dee in 1980-84

*SD = Standard deviation tn = Number of observations

LeuchaBurn

ElrickBurn

Brodiach Burn

100

Culter Burn 80

60

40

WQI

20

35

36

The Elrick, a small stream draining the WesthillIndustrial Estate, received several polluting dischargesapart from surface water runoff. Consequently, itsquality was much reduced (mean WQ1 = 43). Itdrained into the larger Brodiach, where quality im-proved (mean WQ1 = 75) as dilution and self-purification exerted their ameliorating influences. Inturn, the Brodiach drained into the Leuchar, thedischarge stream from the Loch of Skene. TheLeuchar exhibited some of the characteristics of itseutrophic source, and hence a lower mean WQ167). There was a gradual improvement in quality, againbecause of self-purification and increased volumeflow, as this stream drained into the Culter (mean WQ1= 82), and thence to the Dee at Peterculter.

The Crynoch and Ardoe, though exhibiting depressedwater quality because of septic tank discharges, wereso small that they had little effect on the quality of themain watercourse.

Table 2. River water quality in the Dee catchment (mean values for 1980-84)

All results expressed in mg I-1 (except conductivity -SS Suspended solids Alk = Alkalinity (as mg CaCO

Table 2 lists the detailed water characteristics of thecatchment. The low nutrient status of the Dee and thetributaries of its upper and middle reaches is self-apparent. With BOD concentrations less than1 mg I-1, ammonium-N concentrations less than50 ig 1-1, total oxidized nitrogen (TON) concentrationsless than 1 mg 1-1, microgram quantities of nitrite-Nand orthophosphate-P, and conductivities less than100 j.tSiemens cm-1, the waters may be described asoligotrophic. With increasing distance down the catch-ment, the silicate concentration increased. This in-crease reflects progressively greater weathering (natu-ral and agriculturally induced) and drainage. The waterquality data returned for the Dye are consistent withthose reported by Reid et al. (1981) in their studies ofthe weathering rates in the soils of that catchment.

By extreme contrast, the Elrick was polluted; it hadhigh BOD, ammonium-N, phosphate and chlorideconcentrations, and conductivity. The remaining burns,except those draining Lochs Davan and Skene, fellbetween these extremes and exhibited the anticipated

characteristics. The Ardoe, as well as receiving nut-rient enrichment from septic tank discharges andsome agricultural activity (silage and fertilizer runoff),was impounded by a small dam behind which veget-able matter accumulated in large quantities and thendecomposed, thus generating further nutrient ions.

A few small streams of variable water quality drain theAberdeen urban area, and discharge into the harbourand the tidal reaches of the river below the Bridge ofDee. These streams also had a negligible effect on themain river, and surveys of the harbour area confirmedthat nutrient-depleted fresh water entered the harbourunder Victoria Bridge.

3.3 Lochs Kinord and Davan

The results of a survey in June 1982 (Table 3) confirmMarren's (1979) description of these lochs. Kinord(surface area 0.82 km2; mean depth 1.52 m), whosemain feeder is the Vat draining the Culblean Hill

liSiemens cm-1 - and pH)3 H) Cond = Conductivity

granite, was virtually oligotrophic. A minimal diatomfilm was observed through clear water on submergedstones and rocks, but aquatic vegetation was absent.By- contrast, Davan (surface area 0.42 km2; meandepth 1.22 m), whilst not exhibiting major differencesin nutrient chemistry, had both a higher alkalinity and ahigher total ion concentration (conductivity), and acarpet of aquatic vegetation, which elevated the pHduring the daylight hours. There are 3 small feederstreams to Loch Davan. The largest of these, Milton ofLogie, drains the agricultural Howe of Cromar carryingwith it not only minor quantities of fertilizer runoff buttreated effluent from the Logie Coldstone sewagetreatment works. The effect was minimal and, atworst, Loch Davan should be considered only slightlymesotrophic. (A high chloride concentration and con-ductivity observed in the Glendavan and near its mouthin Loch Davan was found to be due to leachate from anew stockpile of road salt/grit nearby.)

The Dinnet (-= Monandavan) which carries the effluentfrom both lochs to the Dee was therefore slightly

Table 3. Water quality in Lochs Davan and Kinord in June 1982

All results expressed in mg I-1 (except conductivity - ItSiemens crn-' - and pH)TOC = Total organic carbon TON = Total oxidized nitrogen Alk = Alkalinity (as mg CaCO3 Cond =Conductivity*One high value removed from the mean - cf textNames of feeder streams in italics

nutrient-enriched relative to the oligotrophic water ofthe main river and its upper tributaries (Table 2).

3.4 The Loch of Skene

Since 1970, and perhaps even earlier, the Loch ofSkene (surface area 1.2 km2; mean depth 1.43 m) hasexhibited a summer phytoplankton bloom of nuisanceproportions; the dominant organism of the bloom haslatterly been a cyanobacteriUm, Aphanizomenon flos-aquae. In a bid to understand this system, River Boardstaff surveyed the loch from April 1978 to October1979 (Owen 1983), and the situation observed thenremains essentially unchanged today. Figure 3 pre-sents a general summary of the variation of the lochchemistry with time.

Chlorophyll 'a' exhibited a pronounced maximumbetween June and September in each year, withassociated smaller increases in spring and autumn.Microsqopic examination of the phytoplankton indi-cated a succession; the smaller chlorophyll peakswere associated with the presence of a diatompopulation, with Stephanodiscus spp. dominating,whilst the major chlorophyll peaks were associatedwith Aphanizomenon. The latter became so prolificthat its small filaments bound loosely together inbundles which eventually coalesced into substantialrafts.

Soluble reactive silicate concentrations reflected theperiods of diatom ascendancy; for example, low meanvalues (160 jig 1-1) were recorded in May in bothyears. During the summer, Aphanizomenon domin-ated and, as the diatom population declined, thesilicate concentration increased rapidly to values be-tween 5000 and 6000 ug 1-1, before decreasing againin the autumn.

This loch had moderately soft water with an overallannual mean calcium concentration of 19.7 mg(range 17.4-23.9). The buffering capacity of the lochwas thus not particularly great.

Soluble orthophosphate and total oxidized nitrogenshowed interesting differences in the timing of theirhighest and lowest concentrations, relative to phyto-

37

plankton periodicity. Orthophosphate certainly de-clined slightly during periods of diatom growth, forexample from a winter concentration of 33 jig 1-1 to aspring value of 7 fig 1-1. However, dramatic andsustained increases in orthophosphate occurred dur-ing the peak Aphanizomenon growth periods, reachingas much as 255 jig 1-1 before declining rapidly as theautumn diatom blooms occurred.

Total oxidized nitrogen (TON) behaved in a completelydifferent manner. Periods of peak phytoplankton pro-duction coincided with low TON, and vice versa. Lowvalues of 50-100 jig 1-' occurred frequently during thesummer Aphanizomenon growth peaks, in contrast torecorded winter maxima in excess of 3000 jig 1-1.

Each decline in chlorophyll 'a' was followed by anincrease in ammonium nitrogen concentrations. Ineach year, these concentrations were highest after thedecline of Aphanizomenon growth peaks, reachingmean concentrations as high as 481 [ig 1-1.

Generally, each peak of chlorophyll 'a' coincided with arise in pH. These elevated values, between pH 9.5 and10.0, persisted for many weeks in summer, and valuesas high as pH 11.0 were recorded.

During the study, an attempt was made to assess thenutrient inputs to the loch. Some of these inputs arosefrom the small sewage treatment works serving thevillages of Dunecht (population 104) and Kirkton ofSkene (population 242), which discharge to the Kin-nernie and Kirktonbridge draining into the loch. Owen(1983) concluded that these 2 sewage works con-tributed 12% orthophosphate, 47% total inorganicnitrogen and 10.9% BOD to the total annual inputs tothe loch, and considered these of minor importancerelative to the whole loch catchment. The sum of themany small domestic and farm effluent dischargesand, perhaps, the agricultural use of fertilizers in thecatchment, was more important than the sewageworks in contributing nutrients to the Loch of Skene.

The Loch of Skene is thus the largest body ofeutrophic water in the whole of the Dee catchment,and has a wide annual fluctuation in chemistry and

38

a)

200

400

0

8

4

0

160

80

0

4

2

9

7

16

8

NH4 & TON

Silicate

Phosphate

pH

*lb

Ternperature

-

Chlorophyll 'a'

APHANIZOMENON

DIATOMS DOMINATE

TONNH4-N

0 AMJ J AS ONDJ FMAMJ J ASON

1978 1979Figure 3. Water chemistry in the Loch of Skene

biology. It finally discharges via the Leuchar (Table 2),which similarly had the most fluctuating chemicalquality of all the monitored watercourses. Happily, asreported above, the quality of the loch water, even atits worst, improved during its passage along theLeuchar and Culter so that the loch had no notableinfluence on the quality of the Dee itself.

4 Base and trace metals in the catchmentStarting in May 1984, surveys of the metal status ofriver water at selected sites in the Dee catchmenthave been made every 3 months. Whilst it is too earlyto comment on the seasonal variation of the data,Table 4 gives an indication of the order of magnitude ofthe concentrations found in the 2 surveys done so far.To aid assessment, the water quality standards (max-imum allowable concentration (MAC) or guideline (GL)

Table 4. Base and trace metals in waters in the Dee catchment (means of 2 surveys in 1984)

'Maximum allowable concentrations (except * = Guideline concentrations) in water for direct abstraction forpotable supplies (as defined by EEC Directive 80/778)

concentration) set by the EEC Directive (80/778) fordirect abstraction for potable supplies are also re-ported. Clearly, the results confirm the chemicallydepleted nature of the main watercourse and theMuick, and parallel the nutrient enrichment (by weath-ering, natural eutrophication and pollution) of thetributaries in the lower part of the catchment. TheMilltimber site on the Dee is one of 7 in the north-eastof Scotland routinely sampled under the auspices ofthe Department of the Environment Harmonised

Table 5. Base and trace metals in the rivers of north-east Scotland (means of 2 surveys in 19841

Monitoring Scheme. These sites are at the mostaccessible, lowest, non-tidally influenced point on themajor rivers, and serve as a summation of their waterquality. Table 5 compares the rivers of the area andillustrates the meagre composition of the water of theDee catchment with respect to the metal elements.Only the data for the Spey are comparable.

5 Surveys of headwatersThe growing concern about the 'acid rain' phe-nomenon prompted a long-term monthly baselinemonitoring exercise in order to assess changes withtime in water quality (several anions and cations apartfrom hydrogen are of interest) in those streams mostlikely to be at risk. Accepting that there are severalprocesses by which streamwater might becomeacidified, apart from wet and dry deposition from theatmosphere, it was reasoned that acidification would

be most pronounced in those poorly buffered watersdraining the thin soils on the acidic granite of theCairngorm range. Accordingly, 82 sites in the Spey andthe Dee catchments were sampled chemically andbiologically (invertebrate analysis) in August 1983.Nearly all the sites were in small headwater catch-ments and most were above the treeline in order thatthe effects of afforestation could be eliminated fromthe assessment. A minimum number of contrastingsites was included.

39

40

Alkalinity provides an indication of the ability of a riverwater to buffer changes in acidity. The accumulatedsurvey data were grouped on the basis of alkalinity andvery tentatively described as shown in Table 6. Of the15 sites in the 'At Risk?' category, 11 were in theupper Dee catchment. Only one site, the Dubh Lochabove Loch Muick, exhibited a value (pH 5.4) below pH5.6, the value returned when carbon dioxide is in freeexchange equilibrium with de-ionized water, and be-low which acidification is assumed. Only one other siteexhibited a value less than pH 6.0.

Table 6. Classification of results of a survey of headwaters in theCairngorms in August 1983

6 SummaryScotland's fourth largest river catchment, the Dee, is,for the most part, in pristine condition. In 1980-84, the

Table 7 Sites in the Dee catchment investigated in the 'Headwaters survey programme'

*See Table 7 for site definition

*Alkalinities determined during the August 1983 survey

Samples are now drawn monthly from 3 or 4 'AtRisk?', 3 or 4 'Medium Risk' and 3 'No Risk' sites ineach of the Dee (Table 7) and Spey catchments. Table8 contains the mean results obtained since May 1984for some of the more important ions for the sites in theDee catchment. Again, the data confirm the meagrechemistry and poorly buffered nature of the uppercatchment. The sites in the Muick sub-catchment (Site

Table 8. Quality of river water in the headwaters of the DeeMay-September 1984)

Site number*

Alk mg 1-1pHSO4 mg l'

Na mg HK mg HCa mg I-1Mg mg H

Al Ilg HZn 11g HMn Rg I-1Fe 1.tg H

89 = Loch Muick Outlet and Site 92 = Glas Allt, afeeder stream to Loch Muick) returned the lowest pHvalues, which were only marginally above the cut-offvalue of pH 5.6. Clearly, these sites need closescrutiny and the processes of their acidification (if,indeed, such a trend emerges) could form an interest-ing research project. Whilst the results for the remain-ing sites gave no cause for alarm, their chemistry andbiology will continue to be monitored. The biological(invertebrate) assessments revealed depleted faunas,but these have been ascribed to the extreme oligo-trophy rather than to the effects of acid rain (MDavidson, pers. comm.). This conclusion is supportedby the fact that several of the sites are below lochoutlets, and are thus buffered from the effects of acidrain.

main river and the upper tributaries (Clunie, Muick,Gairn, Dye, and Feugh) exhibited a high degree ofoligotrophy. Some of the tributaries of the middle andlower reaches (Dinnet, Leuchar, Ardoe and Crynoch)experienced nutrient enrichment by fertilizer runoffand/or effluent discharged from scattered septic tanksand a few small sewage works serving the localcommunities. However, the latter tributaries were so

catchment (mean values from monthly surveys

'At Risk?' sites 'Medium Risk' sites 'No Risk' sites89 92 79 101 102 97 103 93 91 80

3 3 6 7 16 12 13 27 24 315.71 5.75 6.49 6.60 7.03 6.75 6.97 7.19 7.10 7.283.5 3.0 2.8 2.8 3.4 5.3 3.6 4.4 2.9 4.52.1 2.3 2.7 2.4 2.0 4.7 3.5 5.5 4.0 2.60.3 0.2 0.3 0.3 0.5 0.5 0.4 0.6 0.8 0.91.2 0.7 1.5 1.8 4.8 3.1 2.8 5.7 4.3 10.70.5 0.4 0.4 0.4 1.0 0.8 0.9 3.0 2.4 2.3

96.7 64.6 55.2 64.0 33.4 63.9 33.4 131.0 48.9 27.24.9 1.1 1.1 1.0 3.0 2.3 1.3 4.3 1.3 0.9

21.1 3.5 1.0 2.3 6.6 30.6 1.9 64.6 14.9 2.964.7 5.0 10.1 25.0 69.3 138.4 20.1 304.8 127.9 35.1

small relative to the main watercourse of the catch-ment that they exerted a negligible effect on thequality of water in the Dee. Relative to the rest of thenorth-east of Scotland, pollution in the catchment wasminimal and only the newly developing WesthillIndustrial Estate (immediately to the west of Aber-deen) depleted water quality, but then in a minorstream where its effects were quickly attenuated.

In parallel with the oligotrophy, the waters of thecatchment were poorly buffered and exhibited lowconcentrations of base and trace metals, reflecting thepredominantly acidic, granitic geology of the area. Theheadwaters, particularly in the Muick catchment, mayhave been subject to acidification and at risk in thecontext of acid rain (but this hypothesis was notproven and requires continued monitoring).

Of the major pieces of standing water (Loch Muickforms 80% of the standing water in Grampian Region),only the Loch of Skene was eutrophic, with aconsequent annual proliferation of diatoms (Stephano-discus spp.) and then a blue-green alga (Aphanizo-menon spp.), the latter to massive proportions. Again,the Leuchar which drained the loch was relativelysmall, and dilution by the Brodiach and self-purificationduring its passage along the Culter ensured negligibleeffects in the main watercourse. As already implied,the Dubh Loch and Loch Muick were the most acidicbodies of water. Lochs Kinord and Davan were virtuallyoligotrophic, and at worst should be considered onlyslightly mesotrophic.

The size and chemical nature of the Dee set it apartfrom many other rivers, and suggest that care beexerted in its management to maintain its status andto preserve a fine example of a river which isoligotrophic from source to mouth.

41

7 AcknowledgementsThe author gratefully acknowledges the work of hiscolleagues and staff in the generation of the datareported in this paper, and thanks the River Inspectorand the North East River Purification Board forpermission to publish it.

The conclusions and views expressed are those of theauthor and not necessarily those of the Board.

8 ReferencesBolton, P. W., Currie, J. C., Tervet, D. J. & Welsh, W. T. 1978. Anindex to improve water quality classification. Wat. Pollut. Control,77, 271-284.

Davidson, M. B., Owen, R. P. & Young, M. R. 1985. Invertebratesof the River Dee. In: The biology and management of the River Dee,edited by D. Jenkins, 64-82. (ITE symposium no. 14.) Abbots Ripton:Institute of Terrestrial Ecology.

Marren, P. 1979. Muir of Dinnet. Portrait of a National NatureReserve. Aberdeen: Nature Conservancy Council.

Owen, R. P. 1983. Causes and effects of nuisance populations ofphytoplankton in the Loch of Skene, Aberdeenshire. Aberdeen:North East River Purification Board.

Reid, J. M., MacLeod, D. A. & Cresser, M. S. 1981. Factorsaffecting the chemistry of precipitation and river water in an uplandcatchment. J. Hydrol. (Amst.), 50, 129-145.

Scottish Development Department. 1976. Development of awater quality index. (Report ARD 3.) Edinburgh: Scottish Develop-ment Department, Engineering Division.

Standing Committee of Analysts. (Various dates.) Methods for theexamination of waters and associated materials. London: HMSO.

Warren, J. S. 1985. Hydrology of the River Dee and its tributaries.In: The biology and management of the River Dee, edited by D.Jenkins, 23-28. (ITE symposium no. 14.1 Abbots Ripton: Institute ofTerrestrial Ecology.

42

Vegetation of the River DeeN T H HOLMESNature Conservancy Council, Peterborough

I IntroductionBetween 1978 and 1982, the Nature ConservancyCouncil (NCC) commissioned surveys of British riversto provide not only the basis for a classification of plantcommunities in Britain (Holmes 1983), but also adetailed picture of the communities of individual rivers.This paper presents the data from the River Dee andassesses them in relation to 1041 other rivers sur-veyed throughout Scotland, Wales and England.

Although lakes, rivers and streams represent less than1% of the land surface of Britain, they are the mostubiquitous of our remaining natural, or semi-natural,habitats. Unlike many terrestrial habitats which aretotally destroyed by man's activities, rivers tend to bemodified rather than destroyed. Such modificationsoften change the character of the vegetation foundwithin a river, and p-ollutidn, drainage, river regulation,water abstraction, hydro-electric schemes and re-creational activities are often the causes of suchchanges.

The 2 factors which have greatest influence on thecommunities of plants in rivers are current velocity andgeology. These 2 factors interact to provide a wealth ofvariety both in terms of the substrates they create onthe bed of rivers for plants to root in, and of thechemistry of the water in which the plants are bathedand supported. Details of the relationships of plants inrivers to the physical and chemical environments inwhich they occur are to be found in Butcher (1933),Has lam (1978), Westlake (1973), and Whitton (1972).However, it is the current of rivers which separatesthem from other open water habitats such as lakesand ponds. Currents within a river may vary dramati-cally in velocity from site to site, and from time to time,yet they constantly flow in a unidirectional manner.This constancy of direction assures a steady supply ofnutrients and particulate matter down a river over avariety of physical features.

Wildlife habitats are frequently classified into typesaccording to the plants found within them. Rivers areno exception (Butcher 1933; Has lam 1978; Has lam &Wolseley 1982; Holmes 1983), but they are notori-ously difficult to classify because they frequentlychange dramatically from source to mouth. The floralassemblages of upland sections often have little incommon with those in the lower reaches. In general,the higher the altitude at which rivers rise, and thelonger their length, the greater the plant variety. Evengreater variety occurs if rivers traverse differentgeological strata. Because harder rocks are generallyfound at higher altitudes than soft rocks, and becausethe latter yield nutrients more readily, the trophic

status of rivers generally increases on passing down-stream.

The Nature Conservation Review (Ratcliffe 1977)describes over 700 key nature conservation sites inBritain, and 20 sites are included because runningwater represents a key interest therein. The Dee isone of the 20 and is described as 'the finest exampleof a large oligotrophic river in Britain'.

The Dee is one of 4 large rivers described in theReview; the others are the Spey, Tweed and Wye.Compared with the Dee, all are described as consider-ably more nutrient- and base-rich and they rise at muchlower altitudes. The Spey comes closest to the Dee incharacter, because it too flows over resistant rocks;these rocks are either deficient in nutrients or incap-able of releasing them readily. However, the nutrient-poor water of the Spey as it flows through Loch Insh inthe upper half of its catchment is richer than that of theDee in its lowermost reaches.

2 Vegetation of the River DeeThe plant communities of the Dee received scantattention in the past. In the pioneering work on Britishrivers by Butcher (1933), vegetation surveys were notdone in Scotland. In the late 1960s and early 1970s,the Dee was included as part of the national surveysdesigned to find and assess sites of nature conserva-tion importance. This programme culminated in theproduction of the Nature Conservation Review, withthe Dee included as a nationally important site. Sincethat time, Has lam (1982) has surveyed several hun-dred rivers and tributaries, yet the Dee escaped herattention, probably because it was included within theReview. The neighbouring Spey, Ythan and Don were,however, surveyed.

3 MethodThe survey sites are 1 km lengths of river, each sitebeing 5-10 km apart. Surveying is done only during lowflows when water is clear and visibility good. For suchsurveys, it is imperative to wade into the riverwherever it is shallow enough; this too can beachieved safely only during low flow periods. Plants inrivers are most obvious between May and Septemberand surveys are confined to these periods. Algalsamples are not taken but all species visible to thenaked eye, and (theoretically) identifiable in the field,are recorded. The survey at each site includes thewhole river and its immediate bank from one side tothe other. The upper limit of the bank for inclusionwithin the survey was the estimated height of the riverrunning at mean water level. Details of the method-ology are given by Holmes (1983).

3.1 Survey of the River Dee

The vegetation survey of the Dee was done inSeptember 1979 and 14 sites were investigated (Table1). The main channel of the river had its plantsrecorded separately from those found at the edge ofthe water on the banks. Within each site, records weremade from 2 consecutive 0.5 km lengths, and theabundance of each individual species was scored on adominance-rarity scale of 1-5. This score relates theabundance of one species against all the otherspresent within the section; it cannot be used forassessing the standing crop of individual species orthe whole community. A second abundance score wasalso employed which indicates the actual percentagearea covered by individual species.

During surveys of each 1 km length, records weremade of the stability of the river, the substrate, depth,width and velocity characteristics of the channel, theslope and structural make-up of the banks, and theland use immediately adjacent to the river.

4 ResultsTable 2 lists the species found in the 14 sitessurveyed. The relative abundance of each species is

Table 2. Species recorded from 14 1 km sites on the River Dee. Numbers refer to the relative abundance on a scale 1-5, 1 being rarest and 5being dominant. The Table simplifies data held by NCC on the relative abundance of the above species within the channel and onthe immediate banksides

Table 1. Sites of 1 km length surveyed on River Dee for macro-phytes in September 1979. The grid reference refers to the'exact point in km length': ie 0.5 = middle of 1 km length,0.0 = upstream limit of 1 km, 0.99 = lower limit of 1 km

indicated by the scale of 1-5. As few species werepresent in abundance, the actual abundance figures forindividual species have been omitted but an indicationof the total standing crop for the whole I_ km length isgiven at the bottom of the Table. The data for the 2

43

44

Table 2—continued

Sites down the River Dee

Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Anornobryum %forme 1Amblystegiurn fluviatile 1 2 1 2 1 2 3 3 4A. riparium 1 1Atrichum tenellum 1 1Aulacornnium palustre 1Barbula recurvirostra 1Blindia acuta 1 3 1 2 2 4 1 1 1Brachytheciutn mildeanum 1B. plumosurn 1 1 1 1Bryum alpinum 1 1 1 1B. argenteum 1B. pallens 1B. pseudotriquetrurn 1 2 1 3 2 4 4 1 3 4 1 2 1Calliergon cuspidatum 1 1 1 3 3 3 3 1C. giganteum 1 1 1Climacium dendroides 1Dichodontium pellucidurn 1 1Dicranella palustris 1 1 2 1 1D. rufescens 1 1Drepanocladus revolvens 1 1 1 1Fissidens viridulus 1Fontinalis ant0yretica 1 2 2 4 3 4 4 4 1 5 5 5 5 5F. squarnosa 1 1 1 4 3 4 2 3 4 2 1Gymnostomum aeruginosum 1 1Hygrohypnum luridum 1H. ochraceum 1 1 2 1 4 2 1 1 4 4 5 1H. smithii 1Hyocomium armoricum 1 2 1 2 1 1Mniutn hornurn 1 1Oligotrichurn hercynicurn 1 1Philonotis fontana 1 1 1Plagiomniurn rostratum 1 1Pogonatum nanum 1 1Polytrichum commune 1P. juniperinurn 1Racomitrium aciculare 2 3 2 5 5 5 4 2 1R. aquaticum 1Rhizomnium punctatum 1Rhynchostegium riparioides 1 1 1 2 2 5 1 2 2 2 1Schistidium agassizii 1 1 4 2 4 1 1S. alpicola 1 2 2 3 3 3 4 2 1 4 5 2 1 1Scorpidiurn scorpioides 1 1 1 1Sphagnum spp. 3 1 1

Achillea ptarmica 1 2 2 1 2 1 3 2 2 2 2 2Angelica sylvestris 1 1 1 1 1 1 1 1Callitriche hamulata 1 1 1 1 1 1C. stagnalis 1 1 1 1Caltha palustris 1 1 2 1 1 1 2 2 2 2 2 2Equisetum arvense 1 1 1 1 1 1 1 1E. fluviatile 1 1 1 1 1 1Filipendula ulmaria 1 1 1 1 1 1 2 3 1 1Galium boreale 1 2 2 1 1Littorella uniflora 1 1 1 1 1 1 3 1Lupinus nootkatensis 1 1 2 1 2 1 1Mentha aquatica 1 1 1 2 1 3 3 2 1 2 2 2Mirnulus guttatus x luteus 1 1 1Myosotis scorpioides 1 1 2 2 2 2 2 2 2Myriophyllurn alterniflorum 5 2 2 2 1 2 2 5 1 4 2Petasites hybridus 1Polygonum amphibiurn 1Potentilla erecta 1 1 1Ranunculus aquatilis 1R. flammula 1 2 2 2 2 2 2 2 3 3 3 2 3 2Ribes rubrurn 1 1Rorippa sylvestris 1 3 3 1 2Sagina procumbens 1 1 1Salix spp. 1 1 1 1 2 3 2 1 2 1 1 3 4

5 6 7 8 9 10

1 1 2 2 2 42 1 2

1 1 1 1 1 11

1

1 1 12 2 3 1 1

1 11

2 2 1 1 1 13 3 2 2 1

1 11 2 1 2 1

13 4 4 5 5 5

1 1 1 1 1 12 1 1 1 2

12 2 12 2 1 2 1

2 2

1

1

1 1 1 4 2 3

1 1 1 1 1 1

3 1 3 5 5 <1

5 1 5 1 3 10

0.5 km lengths within each site have been amalg-amated, as have the records for the riverside speciesand those on the bank.

The diversity of plants within the different stretches isshown in Figure 1. Each diagram shows the number ofspecies recorded and the proportion of the totalspecies assemblage represented by large algae andlichens, mosses and liverworts, vascular cryptogams,dicotyledons and monocotyledons. Figures 2 and 3show substrate, flow characteristics and the slope ofthe river at individual sites.

Sites down the River Dee

2 4 1 2 1 11 1 1 1

2 1 1 1 21 1 2 1 3 11 1 1 1 2 2

11 12 13 14

2 11 1

2 3 5 3

11 21 11 21 1

1

1

1

4

3

1

1

5

1355

1

1

20

1123

1 2 3 2

1 1 1 11

1

5 5 5 5

1 1 1

2

3

7

45

4.1 Details of the aquatic flora

4.1.1 Flora of river above Whitebridge

By Whitebridge, the Dee has already tumbled its wayalong 20 km of a precipitous and steep-sided course,which rises high in the Cairngorms at over 1200 mabove OD. The flora is sparse, either wiped clean byerosive floods in spring and summer or frozen andgouged by ice pushed down the river in winter. Onrocks within the channel which are afforded protectionby larger boulders upstream from them, small patchesof the delicate liverworts Hygrobiella laxifolia, Cephalo-zia bicuspidata and Anthelia julacea are found.

vL)

1

Linn

of

Dee

•

...

BR

AE

MA

R

•C

RA

TH

IE

5

Riv

er M

uick

.•••

••

AB

OY

NE

10

Fig

ure

1. P

lant

com

mun

ity

stru

ctur

e in

the

14

1km

str

etch

es

of t

he R

iver

Dee

. T

he la

rger

the

pie

, th

e m

ore

spec

ies

Riv

er D

on

Wat

er o

f F

eugh

•B

AN

CH

OR

Y

4,1,

4%

Mon

ocot

yled

ons

Mos

ses

& l

iver

wor

ts

AB

ER

DE

EN

Substrate

Substrate River Dee

Substrate River Spey

Silt, peat or clay

Pebbles, gravel or sand

Amongst these may also be found the stunted, wirygrowths of the moss Blindia acuta. Where pebbles andcobbles are not cemented between large boulders orsheet rock, they are rolled over by the force of thewater and are uncolonized by perennial species.

The largest rocks are occasionally colonized by largerliverworts and mosses, Nardia compressa being themost conspicuous as it occurs on permanently, as wellas occasionally, submerged rock surfaces. The inter-mittently submerged parts of rocks are colonized byScapania and Marsupella species, as well as by themosses Racomitrium aciculare and R. aquaticum. Thedrier surfaces of large boulders within the channel are

Slow or negligible

Moderate

Velocity

Velocity

Velocity

Bedrock, boulders or cobbles Rapid or fast

Figure 2. Substrate and velocity characteristics of 14 1 km sites of the River Dee compared to18 sites surveyed on the River Spey. Each column shows the proportion of each category asa percentage of the whole;

47

also colonized by a variety of lichens, liverworts andmosses. Barbula species and Andreaea rothii areconspicuous.

During the warmer summer months when the waterruns clear, a number of algae take advantage of thebare rock surfaces underwater. Early in the year, whenthe temperature of the water is barely above freezing,thin silvery-brown slimy films cover stones; these areusually chrysophytes mixed with an assortment ofsmall diatoms. As the temperature increases,filamentous green algae replace them. The greenalgae rarely have sufficient nutrients or time toproduce large standing crops. Two algae, in particular,

48

15

14

13

12

11

10

9

8

7

6

5

4

3

2

1

River Dee

1 2 3 4 5 6 7 8 9 10 11 12 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18Sites

Figure 3. Slope characteristics of the River Dee compared to the River Spey in sites surveyed for plants

are worthy of note. Tolypothrix is a drab blackish-brown cyanophyte (a blue-green alga). It grows insmall, interwoven, wool-like colonies on rock surfacesexposed to the force of the current. It has special cellswithin its filaments called heterocysts which can 'fix'nitrogen. These plants are ideally suited for life in theextreme upper reaches of streams above the line ofcultivation or afforestation. Haematococcus is a uni-cellular green alga with a bright red pigment whichovershadows its chlorophyll. It grows only in the smallwater-filled depressions in sheet rocks and boulderswhich are temporary rock pools above the high watermark of the river. These pools turn bright red as theywarm rapidly in the long summer days and theirtemperatures reach levels 3 times higher than that ofthe river.

Marginal vegetation in the upper reaches of the Dee isusually sparse as the edges are often strewn withboulders and disturbed by the rapid rises and falls inlevel. Where large boulders occur, bryophytes aresparse, but where gravels and sand are trappedbetween these boulders a mixture of bryophytes andhigher plants can survive. The richest plant commun-ities occur where the edge of the river laps againstheathland with boulders embedded in the thin soils.The rocks are often covered by the liverworts andmosses listed for Site 1 in Table 1. Species such asBryum pseudotriquetrum, Philonotis fontana, Brachy-thecium rivulare and Polytrichum commune also growon the occasional sandy gravels which are mixed withthe peat and the soil at the edge.

The higher plants which dominate the river's edge arepredominantly those which inhabit the surroundingland. Nardus, Anthoxanthum, Erica, Calluna and Tricho-phorum are common, whilst Viola palustris, Potentilla

River Spey

palustris, Pinguicula vulgaris and Narthecium ossi-fragum are commonest at the riverside.

The small tributaries and flushes which enter the mainriver in the upper 20 km show a full range ofcharacteristics, which span from precipitous drops of600 m in less than 1.5 km to small flushes draining thenarrow strip of boggy heathland on the valley floor. Theformer are either completely devoid of vegetation oronly sparsely colonized by bryophytes previouslymentioned, or occasionally by Racomitrium lanugin-osum and rare species such as Hygrohypnum molle,Andreaea nivalis and Marsupella condensata. Thelatter have richer floras which contrast with that of themain river because of the lack of erosive floods. Apartfrom Sphagna, the flushes are often dominated byhigher plants which can tolerate acidic or dystrophicconditions. Potamogeton polygonifolius is the onlytruly aquatic plant to thrive, but Potentilla erecta,Juncus bulbosus, Drosera rotundifolia, Carex nigra, C.rostrata, Eriophorum angustifolium and Ranunculusflammula may be found invading from the edges.

4.1.2 Whitebridge to Lui water confluence; sites 1 and 2

Between Whitebridge and Lui, the Dee flows througha narrow U-shaped glacial valley and begins itscharacteristic, shifting, meandering course (Figures 2& 3). Because sheets of bedrock are common, theriver has some stability. Elsewhere, boulders, cobblesand pebbles are strewn across the valley amongstgravel and coarse sand. The steep slope of the riverresults in high current velocities in association withlarge rock substrates. The valley is a mixture of grassymoorland, heathland and coniferous forests, and thisaffects the flora of the river.

Where the bed of the river is not continually over-turned by floods, a productive bryophyte community

thrives, with Marsupella, Nardia, Scapania andSolenostoma common (Table 2). On smaller boulders,Anthelia and Blindia dominate, but these are frequentlycovered in algal growths during the summer. Wherethe river shifts its course, piling great heaps ofboulders and gravels where it flowed only the daypreviously, transient filamentous green algal and slimydiatom growths are the only taxa to survive. Thesedifferences are illustrated in the different cover valuesfor sites 1 and 2, the former having a much moreconstrained, and therefore more stable, channel and amore productive, stable, perennial flora.

The flora of the banks also reflects the stability of theriver and whether it is contained within a permanentchannel. Where constant changes in course occur, theflora is very sparse and only a few quick-growing,long-rooted, annuals are present. In more stable areas,where gravel and rocks combine, a rich bryophyte andangiosperm assemblage develops. The bryophytesoccur on the rock surfaces, on the peat-stained sandygravels between the rocks, and on the steep, heavysoil banks themselves. The rocks are preferentiallycolonized by the same species as occur within theriver, but on the gravels Brachythecium fivulare,Philonotis fontana, Bryum pseudotriquetrum, Bryumalpinum, Anomobryum filiforme, Dichodontium pelluci-dum and Oligotrichum hercynicum are present. On theearth banks and where peat is exposed, the mossesCampylopus flexuosus, Polytrichum species and Di-cranella heteromalla occur alongside liverworts suchas Pe Ilia epiphylla, Preissia quadrata and Blasia pusilla.The last mentioned is perfectly adapted to thesenutrient-poor habitats because it harbours its owncolonies of the cyanophyte Nostoc which can fixatmospheric nitrogen.

The flowering plants on the banks of the river arepredominantly those which also occur in the habitatsadjacent to the river. The most typical are summarizedin Table 2. In addition to those listed are heathlandspecies such as Erica tetraliX and Calluna vulgaris,trees such as Betula pendula and alpine Salix species,together with Pinguicula vulgaris on moss cushionsand Pedicularis palustris on margins with a perman-ently retained high water level due to bouldersimpeding flow downstream. Although Nardus stricta isdominant, a wide variety of sedges is present andseveral occur preferentially at the river's edge.

4.1.3 Inverey to Dinnet; sites 3-7

Between Inverey and Dinnet, the Dee drops from350 m to 150 m above OD, making an averagedescent of under 3 m km-1. This descent is remark-ably gradual and even (Figure 3), but within the sectionsome stretches fall less than 1 m km-1. Anotherfeature of this long stretch of the Dee is that it flowswithin a narrow glacial valley with steep sides risingfrom the valley floor for about 50 km. Around Braemar,the river meanders extravagantly along the widest partof the valley floor which approaches 1 km wide.

49

However, 10 km downstream, it is confined to apermanent channel by steep wooded slopes whichrise directly from the river banks.

Land use within this section of the Dee is predomi-nantly rough grassland, but scrub and arable cultivationare interspersed among the grasslands. Almost with-out exception, the lower slopes of the hills which risefrom the valley floor are covered with trees. In places,these trees are purely deciduous, in others they aremixed assemblages, and in others they are pureconifer plantations. In this section, the adjacent habit-ats begin to have less effect on the flora of the river.

Figure 2 shows that the river bed is composed ofcoarse substrates and that velocity is rapid or fast.Only occasionally are there sluggish sections wherewater is impounded by natural rock obstacles. Theexcessive meandering over the wide valley floor nearBraemar manifests itself in the substrate features; site3 has fewer boulders and large rocks than any othersection of the river, and instead is dominated bypebbles and gravels.

The flora of the river in this section maintains many ofthe alpine features prevalent upstream. Many of themosses and liverworts are still present, but thesegradually become less important on passing down-stream (Figure 1). Filamentous algae still dominate theunstable substrates because they can begin growthafter the worst of the winter and spring floods.Perennial bryophytes are very sparse. Although inlocalized and sheltered stretches they may cover up to20% of the rock surfaces, in general the mean cover ofperennial plants for the whole section is only around2%.

As the common liverworts of sites 1 and 2 disappear,their place downstream in the river is generally takenby mosses. Blindia acuta still grows in flattened, wirycushions on stones in torrents, but it is now associatedwith other mosses instead of liverworts. Both Fontin-alis squamosa and F. antipyretica are widely distri-buted, but the former shows a distinct preference fortorrential currents and deep water. The latter tends togrow in totally submerged habitats as well as in thosesubject to periodic submergence. In this respect,it shares the alternately wet and dry rock surfaceswith Rhynchostegium riparioides, Hygrohypnumochraceum, Racomitrium aciculare, Schistidium alpico-la and S. agassizii. The last mentioned is a commonspecies of Scandinavian meltwater rivers (Nyholm1956), but Until 1967 had not been recorded fromBritain (Birks & Birks 1967). At that time, it was knownonly from waterlogged hollows on Ben Lawers. In1975, however, it was found to be common in torrentstretches of the River Tees (Holmes 1976) in UpperTeesdale. Since that time, the author has found it inover 1.0 large rivers in the highlands of England, Walesand Scotland.

50

Truly aquatic plants are rare in the river, Callitricheharnulata and Myriophyllurn alterniflorurn being theonly widespread species. Both occur only wheregravels are compacted between large stones andwhere the bed of the river does not constantly move inresponse to floods. At Braemar, they are the dominantriver plants but they still fail to cover more than 1% ofthe river. In isolated sheltered stretches of the river,other submerged aquatic flowering plants occur also.These include Littorella uniflora, Scirpus fluitans, Jun-cus bulbosus and Callitriche stagnalis. All but the firstmentioned prefer peat interspersed amongst thesubstrate, as well as slack water velocity. They arethus characteristic of backwaters where the channelhas changed course and where gravel and cobble spitsprotect small pockets of standing water retained in theold channel.

The bank flora in this stretch of the Dee is much moredistinct from that of the adjacent land than it was in theupper reaches. Shingle and rocks constitute 75% ofthe banks, whilst earth accounts for 15% and coarsesands account for the remaining 10%. Where rockspredominate, it is the bryophytes which dominate theflora, but where finer deposits are cemented betweenlarger rocks flowering plants prevail. In these stableareas, perennials such as Juncus acutiflorus, Carexnigra and C. rostrata are common. These uprightstands create bank stability and retain organic matterto create mini-niches suitable for such plants as Calthapalustris, Galiurn boreale and Ranunculus flarnrnula.Whilst these species rarely occur except where thewater laps at their roots, another contrasting commun-ity exists on the fast-draining shingles dominated bycobbles and pebbles. Only annuals or tenaciouslyrooted plants can survive in such environmentsbecause they are as frequently destroyed as they arecreated. Three species are characteristic of the area:Viola tricolor, Tussilago farfara, Lupinus nootkatensis.The latter is a rare naturalized plant which thrives onthe shingles of a few Scottish rivers.

4.1.4 Dinnet to Banchory; sites 8-10

From Dinnet to about 10 km upstream of Banchory,the Dee continues to meander across a gravel,flat-bottomed, U-shaped river valley. Above Banchory,the river becomes confined to a much narrower valleywith wooded slopes rising directly from the banks. Theslope of the river channel increases marginally, butsubstrate and water velocity characteristics remainsimilar to upstream (Figures 2 & 3). This section of theriver also sees an intensification of agricultural use inthe valley floor and a decrease in conifer afforestationon its slopes.

Because of the gradual downstream changes inphysical characteristics and in land use, the flora of theriver also changes slightly. Gone are the speciescharacteristic of the upper 2 sites, and those thatcharacterized the reaches between Braemar andDinnet are also less frequent. Although filamentous

green algae still dominate the river in summer,encrusting and foliose lichens are also present. Abovethe permanent water mark, Derrnatocarpon fluviatile isseen, whilst several species of encrusting Verrucariaoccur permanently submerged.

This section of the river is not noteworthy for theabundance of any particular species, but is a transitionarea where the alpine elements are all but replaced bythose characteristic of lowlands. Juncus acutiflorusremains dominant on the banks, but, instead of beingassociated with Anthoxanthurn, Nardusand Molinia, itgradually becomes associated with Phalaris arundin-acea and Agrostis stolonifera. Similarly, Fontinalis isthe dominant bryophyte, but, instead of being associ-ated with liverworts Racomitriurn aciculare and Schist-idiurn agassizii, it is found more commonly withSchistidiurn alpicola and Arnblystegiurn fluviatile. Thelatter is associated with more basic rocks and isindicative of the influence exerted by the meta-morphosed limestone which outcrops in the Dinnetarea.

4.1.5 Banchory to Aberdeen; sites 11-14

From Banchory to its mouth, the Dee flows within awider valley, with more gently inclined slopes. Theagricultural land use intensifies but never encroachesdirectly on to the bank vegetation. The velocity of theriver gradually decreases in parallel with the decreasein slope (Figures 2 & 3), yet the substrate of the rivershows less dramatic changes.

The increased richness of the surrounding land, whichmakes the lower Dee suitable for more intenseagricultural use, also has an effect on the flora of theriver. Below Banchory, there is a genuine increase introphic status of the community. Several new plantspecies are present which do not occur upstream.However, the most notable feature is that very littlechange occurs in the flora of the river itself, but all thechanges are evident at the water/bank interface. Thespecies of the banksides are therefore very different.

Because the substrate characteristics of the river itselfhave changed little (Figure 2), the dominant taxa in theDee right to its discharge into tidal waters remaingreen algae, lichens and bryophytes. The banks,however, become more stable, and the vegetationincorporates detritus and other fine particles washeddown by the river in spate. The result is a richer soilcapable of supporting 3 types of reed, Sparganiurnerecturn, Phalaris arundinacea and Glyceria maxima.All these 3 species occur where stable, richer banksare at least close to the low water level of the river sothat their roots are constantly in contact with thewater. Several other flowering plants are also confinedto, or only common on, the banks of the lower reachesof the Dee. This situation contrasts markedly with thatfor the plants within the channel itself.

5 Tributaries of the River Dee and standing watersNo detailed surveys of the tributaries have been doneby either the author or by Has lam. However, aboveBraemar, the character and flora of tributaries such asLui, Ey, Quoich and Clunie are similar in many respectsto that of the Dee itself. Bryophytes, particularlyliverworts, dominate. Below Braemar, the characteris-tics of the large tributaries frequently differ from thoseof the main river, and further investigations arerequired. Of particular concern are the Muick, whichdischarges from Loch Muick and therefore shouldhave a relatively stable flow regime; the Feughsystem, which drains a less steeply sloping valley witha wide and fertile river corridor; and the Leuchar,which drains flat land to the north of the Dee anddischarges from Loch of Skene.

The lack of knowledge concerning these potentiallyvery different watercourses requires remedy, as theyprobably represent a totally different riverine wildliferesource. They are doubly important because theirinfluence on the main river flora is almost nil and verylocal. This situation contrasts with the influence oftributaries on other rivers such as the Tweed (Holmes& Whitton 1975), and they may be of unique scientificinterest. In the case of the Tweed, the Teviotcompletely changes the floral assemblage of the mainriver below its inflow.

Within the Dee system are a number of lakes andother areas of standing water which have floras inmarked contrast to that of the Dee itself. Indeed, someare far richer in aquatic plants. However, the Deeneither flows through any of them nor is its florainfluenced in any way by them. The standing waters alldischarge into small tributaries and, as these enter theDee, their influence is totally eclipsed by the over-riding size and power of the main river. In this respect,the Dee has a most unusual flora because neither thestanding waters within the catchment nor its tributar-ies exert any influence on the floral assemblages ofthe main river.

Although the botanical importance of mountain andmoorland lochs and pools, as well as the more lowlandlochs such as Kinord, Davan and Skene, cannot beoverlooked, they may be treated as entities in theirown right and they are not considered in this paper.

6 DiscussionThe Nature Conservation Review (Ratcliffe 1977)described the Dee as the finest example of anoligotrophic river in Britain, at a time when fewcomparative surveys had been undertaken. Sincethen, national vegetation surveys have been carriedout by NCC in an attempt to classify rivers according totheir plant communities (Holmes 1983). This classific-ation allows the flora of the Dee to be put into anational perspective and its significance for natureconservation assessed.

51

Over 1000 sites were investigated on over 100 riversthroughout Britain, with data collected in a standardmanner using a checklist of species. Analyses of thedata used 223 taxa, and sites were classified by astandard program called TWINSPAN (Hill 1979). Theanalyses split the 1055 sites initially into 2 unequalgroups of 656 and 399, the former lowland and morechemically rich and the latter more upland and lessrich. The second split then divided these 2 groups into4 major groups of significance. Group A had sitescharacteristic of lowland areas and a rich geology andchemistry; group B had sites associated with sand-stone or hard limestones and intermediate chemistry;group C had sites associated with resistant rock andnutrient-poor water; group D had sites associated withhighlands and acid, nutrient-poor water. The pro-gramme then produced 56 meaningful communitytypes. Details on the distribution throughout Britain ofthese communities, and their plant assemblages, aregiven in Holmes (1983). The relationship of thesecommunities to water chemistry and the physicalenvironment of rivers is given in Holmes and Newbold(1984) and Holmes (in prep.).

Although Butcher (1933) had pioneered work onclassifying British rivers according to their flora, by1977 there existed only a descriptive account whichdid not include rivers in Wales and Scotland. TheNature Conservation Review developed the classific-ation further and the publications of Has lam (1978,1982) added immeasurably to the knowledge of theflora in the rivers of Britain. The classification of plantcommunities in rivers by Holmes (1983) is possibly thefirst rigorous analytical study which provides anobjective classification enabling the significance ofplant communities to be evaluated.

The Dee was one of 19 Scottish rivers which flow intothe North Sea to be surveyed. A meaningful evaluationof the floral assemblages of the Dee is possible only ifthey are compared to relatively large rivers of a similarsize in Britain.

Table 3 lists the plant community types found in theDee alongside 5 other large rivers which flow in agenerally west to east direction before discharging intothe North Sea. Table 3 shows that 4 of the 56 nationalcommunity types were evident in the Dee and 20 inthe combined 88 sites surveyed on the 6 rivers. The 4community types to which the sites on the Dee wereassigned by the TWINSPAN analyses were:

Sites 1 and 2 (Whitebridge to Lui confluence),assigned to community type D3ii, a national rivervegetation type described by Holmes (1983) as 'moor-land rivers with boulders and adjacent peat';Sites 3-7 (Inverey to Dinnet) assigned to communitytype D1 i, described as 'mountain rivers with gravelmargins' ;Sites 8-10 (Dinnet to Banchory) assigned to commun-ity C4iii, described as 'highland rivers with gravel andpeat';

52

Table 3. Six large rivers surveyed in Scotland which flow into the North Sea. From left to right are listed the vegetation types of the riversfrom source to mouth as given in Holmes (1983). Emboldened sites are oligotrophic, single underlined sites are meso-eutrophiccommunities and double underlined sites are eutrophic

FindhornSpeyDonDeeTeith

Tweed

D2i D21 D2i Dli Dli C4iii C4iii D1 i Cliii Cliii Cliii C4iiD4i D3ii D4i C4iu D41 C4iii C3i C4iii C3i C3i C3i C3i C4i C2v C2v C2v Cli C4i C4iDli Dli C4ii C4iii C4iii C4iii C4i Bli Bli Bli Bli Bli Bli BliD3ii D3ii Dli Dli Dli Dli Dli C4iii C4iii C4iii C4i C4i C4i C4iD3i Dlii C4iii Dlii C2iv Cliii C2iii A2ii

Dli C4iii C4iii Bliii Bliu B1 iii B1 iii Blui Bliii B2i B2i B2i B2i B2i B2i B2i B2i B2i B2i

Sites 11-14 (Banchory to tidal influence) assigned tocommunity C4i, described as 'large oligotrophic Scot-tish rivers'.

The Dee is shown by the analyses to have a floralassemblage which in the upper half of the river isdefinitely oligotrophic but in the lower reachesbecomes no more than oligo-mesotrophic (Holmes &Newbold 1984). In this respect, it had features incommon with the Findhorn and the Spey, but theindividual community types were frequently different.All 3 rivers illustrate that the accepted normal pro-gression for large rivers rising in highlands is notalways met, and some do not go from oligotrophicuplands to mesotrophic middle reaches and then tomeso-eutrophic or eutrophic communities in the lowerreaches. The more generally accepted pattern ofdownstream succession of communities is seen,however, in the Don and Tweed.

Although there are 3 examples in north-east Scotlandof oligotrophic large rivers, the national surveys ind-icate this type of river to be rare; the only other largeriver in Britain found to have such a low trophicexpression in its flora was the Conway in north Wales(Holmes 1983). However, it is not uncommon for smallrivers in England, Wales or Scotland to contain only anacidic, ultra-oligotrophic flora (ie New Forest streamsin Hampshire or precipitous rivers in western Scot-land).

National surveys have thus confirmed that the Dee ismost unusual, but more detailed comparisons arerequired to see how its flora relates to other largerivers in eastern Scotland.

Table 4 lists the community types found in the 6 riverslisted in Table 3; by cross-reference it is possible tosee how the floral assemblage of the Dee relates toothers in the area.

The Don is the nearest large river to the Dee. Theupper reaches of the Don are less oligotrophic and itslower reaches approach eutrophic status. The Donlacks the alpine elements of the Dee flora, and theupper reaches of the Don have a flora akin to thatfound between Lui and Dinnet on the Dee. This is thecommunity of oligotrophic, gravel-edged, rivers andnot one of boulders. Whereas the Dee retains anoligotrophic (D) plant community until about 150 m

B2i Al iii

above OD, the Don loses its oligotrophes at about350 m. Below this level, the Don has 3 different typesof mesotrophic (C) communities but all are very closelyrelated and similar to those of the lower reaches of theDee. The 7 sites surveyed on the lower half of the Donwere all classified in the meso-eutrophic communityBli. Reference to Table 4 shows the general lack ofbryophytes, particularly alpine ones, and the abundantoccurrence of flowering plants with requirements forgreater physical stability of their environment andhigher nutrient levels. Indicative species include Sol-anum dulcamara, Elodea canadensis, Polygonumamphibium, Lemna minor, Potamogeton crispus andGlyceria maxima;

The Spey rises in mountainous country with anoligotrophic expression in its flora which changes to amesotrophic flora as the river approaches the flatvalley of Loch lnsh Marshes. This mesotrophy isretained in the plant community expression through-out the remainder of the river. The most interestingfeature of the flora of the Spey is that, although thetrophic status of the river is only marginally differentfrom that of the Dee, its physical structure results insubtle differences in plant communities. In the upperreaches, sites 1, 3 and 5 all have floral assemblagesclassified as D4i, described in Holmes (1983) as'slow-flowing upland rivers', whilst site 2 is classifiedas D3ii, similar to that of the upper Dee. Reference toFigure 3 shows that the slope of the Spey valley is verydifferent and the resultant differences in current andsubstrate characteristics (Figure 2) are responsible forthe differences in floral composition. The most impor-tant differences between the Dee and Spey are seenin the middle reaches of the Spey around Loch Inshwhere its valley is broad and shallowly sloping (seeFigure 3). Sites 7-11 on the Spey were surveyed in thisarea and 4 out of the 5 were classified as C3i. Thiscommunity is described as characteristic of 'uplandrivers of fen and bog'. This community type isexceedingly rare in Britain and the characteristicspecies are given in Table 4. The most significantfeature of the flora is the rich emergent fringingvegetation, which invades from the stable banks intothe more slowly flowing water of the river itself. In thecase of the Spey, Carex aquatilis is diagnostic.

The Findhorn is a large river which also does notprogress beyond a mesotrophic flora in its lowerreaches. Like the Dee, 50% has an oligotrophic flora

and 50% mesotrophic flora (Table 3). The differencesin flora between the 2 are only subtle. The differentcommunity structures are explained by the upper sitesof the Findhorn being surveyed where the river wasnot constrained to a definite channel as in the upperDee. From sites 4-8, the Findhorn has a flora similar tothat of the Dee between Balmoral and Banchory.However, below this, the community of the Findhornis dominated by Cl iii (described as 'upland, rapid riverswith acid water but rich substrates') instead of C4iwhich occurs in the lower Dee. These communitydifferences indicate that the Findhorn flows oversandstone in the lower reaches, whilst the Deecontinues its course flowing over less rich rocks.

Further south, the Teith also has an oligotrophic-mesotrophic expression but its assemblage differscompletely from that of the Dee. Out of the 7oligotrophic and mesotrophic communities recorded

for the Teith, only one is of a type found on the Dee.The oligotrophic communities of the Teith are richer inflowering plants because of the stability of the riverwhich results from its passage through 3 lochs. Themesotrophic communities in the lower reaches alsodiffer from those of the Dee in being tree-lined, onricher geological strata and more stable. In theextreme lower reaches, it has a eutrophic plantcommunity where the river flows over rich old marinedeposits at the head of the Firth of Forth.

Table 4. Species which characterize the 20 community types found in 6 rivers in Scotland which flow into the North Sea (see Table 3). Fordetails see Holmes (1983). Letters indicate the frequency of individual species expected to occur in sites assigned to eachcommunity type. D = >75% recorded occurrence; A = 50-74% recorded occurrence; F = 20-49% recorded occurrence; blank =absent or up to 19% recorded occurrences. Emboldened communities are the 4 community types present in the River Dee

British river community types

D2i D3i D3ii D4i Cli Cliii C2iii C2iv C2v C3i C4i C4ii C4iii Bli Bliii B2i Aliii A2ii

The Tweed is also a river of prime nature conservationimportance and was awarded grade 1 status in theNature Conservation Review (Ratcliffe 1977). TheReviewdescribes the river as eutrophic throughout itslength because of its passage over rich geologicalstrata, in contrast to that of the Dee. The communitiesshown in Table 3, and the species assemblagecharacteristic of those species shown in Table 4

F F AF

F F ADDDA A F F

FD DF FAA DF FA A F AF F FADDDDA A FF F

A ADF F FA ADADAD

A F FFFFDF F F

F F F A

D DF A FFFFFF F A

AA A F F AA F AD D F F

F F F A F DADAFDADDD

D F AF F F DDA A

AD F F DD DFDDFDDADDDA A A F

F A AF ADFF F F ADDD

ADF F F ADD A F A

D ADAD D AD DSDDFDD

ADDDF A FD DAFD D DADDA A F FAF F F F F F

F F

F F FF AA FD D A

F FF A FD DFD DDA A FF DD

DFF F

53

54

Table 4.—continued

British river community types

Species Dli D1 ii D2i D3i D3ii D4i Cli Cliii C2iii C2iv C2v C3i C41 C4ii C4iii Bli Bliii B2i Al iii A2ii

Senecio aquaticus F F F F F A F F FDeschampsia cespitosa ADA A F A AA A F ADF A AF F FFontinalis antipyretica DADF F F DD A ADDDD DDDRhynchostegium riparioides DDA A F FDDDDDFDDDDDDF FFilamentous algae DDDDDDDDDDDADDDDDDF FSalix spp. DDADF ADDDD ADDDDDDD ADTrees A AF DA F ADDDD A A ADAD A AAAgrostis stolonifera DDFDF ADDDDDDDDDDDDDDJuncus acutiflorus DDADDDDDDDDDDDDDDDF AJ. effusus DDADDDFDADDDADDDDF FDGlyceria fluitans DD A F DF A A ADDDDDD A A F FEquisetum arvense A F F F A A F FD ADADDD FMentha aquatica A F F F DADF DDDDDDDD A AMyosotis scorpioides A A F F AA A A F DDDDDDDDDDFilipendula ulmaria DAF F A ADD ADDDDDDD A F AVeronica beccabunga F A A F F F AFDF ADDCallitriche stagnalis FFFFF F A F FDFFDAF F FDermatocarpon fluviatile A F F F F ADF FF F FEquisetum palustre F F A F F F A FOenanthe crocata F F A D F F F FThamnobryum alopecurum F F AA A F F F F FAngelica sylvestris F A F F F A A F ADDAAF A F FEleocharis palustris A F F A ADA AF DDIris pseudacorus F F F F A FStachys palustris F F F F A FEquisetum fluviatile A D D ACarex remota A A F F FC. aquatilis ADF F FPhalaris arundinacea A F A ADDDDDDADDDDDAlopecurus geniculatus A F F A A D F F ASparganium erectum F A A ADFDDDDDS. emersum F F D A FPotamogeton natans F DF A AF F FStellaria alsine F F F F A F FCinclidotus fontinaloides DF F F F F DDMimulus guttatus OF F F AA FDM. guttatus x luteus hybrids F F F A F FHildenbrandia rivularis A F F FDDCollema fluviatile A F F FCladophora glomerata D F ADDDDVaucheria sessilis F F F F F F A ADDSolanum dulcamara A A A F DANasturtium officinale F F F A A F FDDRonppa sylvestris A AD FPetasites hybridus A AEphydatia fluviatilis F F ACladophora aegagropila DDDLunularia cruciata AImpatiens glandulifera ADAmblystegium riparium E A F ABrachythecium rutabulum F A F FMyriophyllum spicatum A ARanunculus fluitans F DFElodea canadensis F F ADD A APolygonum amphibium ADDFDLemna minor A F F AAPotamogeton crispus F DD A AGlyceria maxima A F AEnteromorpha spp. ADARanunculus calcareus FDF FMyosoton aquaticum F F AEpilobium hirsutum F ADDScirpus sylvaticus A A APotamogeton perfoliatus F AP. pectinatus ADZannichellia palustris A FRanunculus sceleratus F ACarex riparia F AVeronica anagallis-aquatica F AJuncus inflexus F AApium nodiflorum DAScrophularia auriculata A FRoriPpa amphibia AR. palustris AScirpus lacustris A

confirm the differences, although it is difficult tounderstand how the uppermost reaches of the Tweedcould be regarded as eutrophic. What is clear, how-ever, is that the uppermost reaches of the Tweed havefloral assemblages akin to those of the lower Dee, yetdownstream the Tweed gradually increases in trophicstatus until it reaches eutrophy. The differences inflora between the 2 rivers clearly illustrate the com-bined effect of underlying rocks and of current, whichresult from the altitudinal source of the river and itsdescent from that height.

6.1 Conservation significance of the Dee

The selection of key sites for nature conservation isbased upon a number of criteria, the principal 2 beingnaturalness and typicalness, together with criteriasuch as fragility, diversity of features and of flora andfauna, and rarity. If a representative series of rivers isto be included within the national framework ofhabitats, it is imperative to choose large and smallexamples which illustrate the variations of the re-source on a national scale. In Ward's (1981) list of the20 largest rivers in Britain, the Dee is ranked as no. 14.Its length of 116 km makes it the tenth longest, itsarea of 1370 km2 makes it the twenty-first largest, andits mean discharge of 35.7 cumecs ranks it aseighteenth. Four other large rivers are considered tohave high nature conservation significance. They arethe Tweed, Spey, Wye and Avon, all of which havetheir interests described in the Nature ConservationReview.

It has already been shown that the flora of the Deeretains a uniquely low trophic status along its entirelength, which is lower than in the Conway, Spey andFindhorn, rivers most closely resembling it. In addition,the plant communities of the Dee show a uniquesuccession on passing downstream. The Dee cantherefore be regarded as the river which exemplifiesthe oligotrophic, large, British river.

One of the 2 main site selection criteria thus puts theDee at the very top of the list. The other, naturalness,may be the reason why. In Ward's list, the Dee rankswith the Spey as the most natural. For one reason oranother, many of Britain's longest rivers have beenexploited by man and their physical character changed.In so doing, their floral characteristics have changedalso. The second of the 2 key nature conservationselection criteria again places the Dee at the top of thelist.

Because the floral assemblage is such an importantelement in the selection of sites of high natureconservation importance, there can be no doubt thatthe Dee is a nationally important river.

55

7 SummaryNational botanical surveys of British rivers have ind-icated that the River Dee has a unique assemblage ofplants. Unlike most large rivers which rise in theuplands, the Dee increases its trophic status onlymarginally before discharging into the sea. For thisreason, the flora begins as oligotrophic and thengradually increases to oligo-mesotrophic at its mouth.In this respect, it resembles only the Spey, Findhornand Conway. In both naturalness and typicalness, theDee is nationally important as a site for natureconservation.

8 ReferencesBirks, H. H. & Birks, H. J. B. 1967. Grimmia agassizii (Su II. & Lesq.)Jaeg. in Britain. Trans. Br. bryol. Soc., 5, 215-217.

Butcher, R. W. 1933. Studies on the ecology of rivers. I. On thedistribution of macrophytic vegetation in the rivers of Britain. J.Ecol., 21, 58-91.

Has lam, S. M. 1978. River plants: macrophytic vegetation of watercourses. Cambridge: Cambridge University Press.

Has lam, S. M. 1982. Vegetation in British rivers. Vol. 1: Text. Vol. II:River maps. (CST note 32.) Banbury: Nature Conservancy Council.

Has lam, S. M. & Wolseley, P. A. 1982. River vegetation: itsidentification, assessment and management. Cambridge: Cam-bridge University Press.

Hill, M. 0. 1979. TWINSPAN - a FORTRAN program for arrangingmultivariate data in an ordered two-way table by classification of theindividuals and attributes. Ithaca, N.Y.: Section of Ecology andSystematics, Cornell University.

Holmes, N. T. H. 1976. The distribution and ecology of Grimmiaagassizii (Su II. & Lesq.) Jaeg. in Teesdale. J. Bryol., 9, 275-278.

Holmes, N. T. H. 1980. Preliminary results from river macrophytesurvey and implication for conservation. (CST note 24.1 Banbury:Nature Conservancy Council.

Holmes, N. T. H. 1983. Typing British rivers according to their flora.(Focus on nature conservation no. 4.1 Banbury: Nature ConservancyCouncil.

Holmes, N. T. H. & Newbold, C. 1984. River plant communities -reflectors of water and substrate chemistry. (Focus on natureconservation no. 9.1 Peterborough: Nature Conservancy Council.

Holmes, N. T. H. & Whitton, R. A. 1975. Macrophytes of the RiverTweed. Trans. Proc. bot. Soc. Edinb., 42, 383-395.

Nyholm, E. 1956. Illustrated moss flora of Fennoscandia. II. Musci,fasc. 2. Lund: Gleerup.

Ratcliffe, D. A., ed. 1977. A nature conservation review: theselection of biological sites of national importance to natureconservation in Britain. 2v. Cambridge: Cambridge University Press.

Ward, R. C. 1981. River systems and river regimes. In: British rivers,edited by J. Lewin, 1-33. London: Allen & Unwin.

Westlake, D. 1973. Aquatic macrophytes in rivers: a review. PolskieArchwm Hydrobiol., 20, 31-40.

Whitton, B. A. 1972. Environmental limits of plants in flowingwaters. In: Conservation and productivity of natural waters, editedby R. W. Edwards & D. J. Garrod, 3-19. (Symp. Zoological Society ofLondon no. 29.1 London: Academic Press.

56

Vegetation of the valley floor of the River DeeJ A FORSTER* and JACQUI GREENNature Conservancy Council, 17 Rubislaw Terrace,Aberdeen(*now University Office, University of Aberdeen)

1 IntroductionThis paper describes the vegetation present in the Deevalley in April-November 1979 and makes recom-mendations for its conservation. A survey was com-missioned by the Nature Conservancy Council (NCC)because scarce types of semi-natural vegetation wereknown to occur in the lower reaches of the river,together with rare plants and invertebrates, eg Mack-ay's horsetail and Kentish glory moth, and it wasdesirable to record the location, extent and compos-ition of these communities. Also, in considering riverconservation and management, knowledge wasneeded about all the vegetation growing nearby.

The study area was defined so as to include onlyvegetation which was either directly influenced byriver processes, such as erosion, deposition andflooding, or subject to management directly related tothe river, eg salmon fishing, bank reclamation andwater abstraction. In the upper reaches, the riversidevegetation differed little from that in the surroundingcatchment area, and consequently the study areaextended no higher up the Dee than Whitebridge atthe foot of the mountain slopes.

Figure 1 gives a theoretical cross-section through themain valley showing typical features. The floodplain, iethe area subject to periodic inundation, where soils arethe product of river-borne deposition, and where watertables are generally high, provided the obvious latit-udinal delimitation to ihe study area, but this area wasextended to include a strip up to 50 m wide on eachside of the floodplain where semi-natural vegetationoccupied adjacent slopes. The whole is described asthe 'valley floor', and the total area studied wasc 93 km2 (Table 1).

Steepbluff

IAbandonedchannel

I 50 m I14•• •1

Riverbank

1.1 General characteristics of the valley floor

Along the main river and its tributaries, landform andvegetation are closely related, not only because of thedirect influence of the physical environment on thetype of species present, but also because man's useof the land is a response to topography. The nature ofthese landforms and their relationship to the plantcommunities are briefly outlined as a prelude to themore detailed description of vegetation.

Deposits of shingle, to a greater or lesser extentstable, are common; Maize Is (1985) describes theconditions under which they form. Land beyond theriver banks, wherever the gradient is low, is liable tocontinuing deposition of sediment and detritus infloods. Most alluvial flats have relatively good soils,but, because of the vulnerability to flooding and, inplaces, the difficulty of surface drainage, cultivationcan be difficult and some areas are consequently leftas permanent grassland. Within the plain, swampvegetation occurs in abandoned river channels andother wet hollows. Where only a narrow strip offloodplain lies between the river and a neighbouringbluff, difficulty of access has sometimes led to theretention of natural or semi-natural woodland or to theabandonment of agriculture and the establishment ofsecondary natural vegetation.

The edge of the floodplain is often bounded by steepslopes marking the limit of erosive activity of the mainriver. These slopes are often too steep or narrow to beworth cultivating, and are sometimes difficult to grazewith domestic stock. Consequently, woodland hastended to develop. In the lower sections of the valley,these bluffs have often been planted, particularly withbroadleaved trees (eg ash, wych elm or beech). In

Shingleisland

FloodplainStudy area

Figure 1. Definition of study area with typical features of the Dee valley

I 50 m I

places, they contain some of the few examples oflow-altitude natural woodland, for example the oak,birch and hazel wood at Quithel, between Aboyne andKincardine O'Neil.

1.2 Data

The survey providing the main source of data for thispaper was a qualitative description of the semi-naturalvegetation of the valley floor. Because linearity in thedisposition of riverside vegetation and great variabilitybetween sections of the valley were expected, thewhole of the defined study area was examined. Allareas of semi-natural vegetation were outlined from airphotographs and the dominant plants then recordedon field visits. Plant communities in some represent-ative 2 m x 2 m quadrats were assessed in order torelate the vegetation to earlier classifications (McVean& Ratcliffe 1962; Birse & Robertson 1976). Additionalinformation was used for the valley of the Feugh andother tributaries. For this paper, the classification ofthe vegetation is simplified and English names aregiven to most units. Individual species are discussedonly briefly. The plant community named by thedominant species is a convenient shorthand for adescription of the site as a whole. The distribution andextent of coverage of identified types of vegetationwere estimated within the valley floor of the Dee, andthe valley was divided into sections according to themain villages.

Table 1. Area (ha) of different sections of the valley floor of the River Dee occupied by different types of plant cover (<0.1% =

Some of the vegetation types occurred in such smallareas that they could often have been missed, andthese types are likely to be under-recorded. There wasalso a considerable gradation between different types,and the separation of types where no sample quadratswere described was to some extent arbitrary. Thepaper is descriptive rather than analytical, as little workhas been done to relate vegetation and environment inthe study area. The full account of the study isavailable in the library at NCC, Aberdeen.

2 Results of the survey2.1 Woods

2.1.1 Scots pine

Most of the pinewoods in the Dee catchment wereplanted, but some of the finest examples in Britain ofnative pinewoods occur in Glen Tanar, Ballochbuie andMar (Steven & Carlisle 1959). We found it difficult insome cases to distinguish well-established pineplantations from the native woodlands on structuralgrounds alone; in these cases, such woods werecalled native. Two different types of pinewood wererecognized: those which had an open canopy andwere old, dominated by heather and Vaccinium spe-cies (the Pinetum Vaccineto-Callunetum of McVean& Ratcliffe), and those which had a more closedcanopy and a field layer dominated by mosses orbilberry. The closed-canopy type was found only onthe south bank of the river (ie it had a northerly aspect)

trace)

57

58

and only as far down the river as 250 m above OD. Themore open canopy type was found on both north- andsouth-facing slopes and down to 70 m above OD nearBanchory. Both types tended to occur on podzols, andwere found on slopes or terraces above the floodplainrather than on alluvial soils or poorly drained areas.Pinewoods covered 12.6% of the study area (Table 1)and were greatly affected by management.

2.1.2 Conifer plantations

Apart from Scots pine, Sitka and Norway spruce werecommonly dominant. Other species planted exten-sively were European larch, Douglas fir and westernhemlock. Grand fir, noble fir, Japanese larch andlodgepole pine were less frequent, usually occurringas stands within larger plantations. Most plantationswere on land that was previously heather moorland,Scots pinewood or birchwood. Plantations were rarelyfound on former arable ground or in areas subject toflooding. This woodland type occupied 9.9% of thestudy area (Table 1). It was most frequent fromBanchory to Aboyne (21.8% of that area), and it wasthe second commonest type after arable agriculturalland in the low-altitude section from Banchory toAberdeen. Overall, conifer plantations at 5.1% werethe second highest category after arable.

2.1.3 Birch

Both downy and silver birch occur in the Dee Valley,but silver birch is confined to the drier eastern sectiondownstream from Glen Gairn. It grows on free-drainingslopes, whereas downy birch in these sections re-quires sites with a high water table, eg flats by theriver, where it grows with alder and bird cherry. Abovethe mouth of the Gairn, downy birch occurs on bothwet and freely drained soils. Four types of birchwoodwere distinguished by their field layer: (i) birch invadingheather moor (at low altitudes primarily silver birch, egMuir of Dinnet), (ii) dense stands at high altitude, withbilberry dominant as understorey, (iii) birch over aspecies-rich grassland, (iv) moist birchwood (downybirch) with wood horsetail and tufted hair-grass. In thestudy area, birchwood was the dominant habitat fromAboyne to Ballater (25.7%), and this section contained49.1% of all the birchwood (Table 1).

The birchwood category also included a small numberof woods in which birch was mixed with other trees,but birch or birch and rowan predominated. Thesewoods generally had a similar field layer to theherb-rich type of wood (type (iii) above). On north-facing slopes with better soils, they graded intooakwood.

2.1.4 Oak

Oakwood is scarce in the Dee catchment, withsignificant stands only at Drum, Crathes, Aboyne,Dinnet and Craigendarroch. Oaks grow on acid brownearths, on better brown earths as at Dinnet, andsometimes on granite rock ledges and scree. Only thestands at Craigendarroch and Dinnet came within thelimits of the survey. Both were on steep slopes

(generally greater than 25°) above the floodplain.Isolated oaks, including some superb specimens,grew beside the river upstream from Banchory, atQuithel Wood, and above Ballater at Coilacreich. Mostof the trees were pedunculate oak and few weresessile oaks, though several had hybrid characters.

2.1.5 Alder

Alder occurred in locations characterized by thepresence of moving water, ie the banks of the Dee andits tributaries, those parts of the floodplain which had apermanently high water table, and flushed slopesabove the plain. Because most of the alluvial flats onwhich alder woodland might grow were farmed, thelargest stands were found on islands or on flats belowsteep cliffs with difficult access (eg at Quithel). Onbanks with easy access, saplings were cut back to aidfishing. Woodland with a slightly lower water table hadhigher proportions of birch and bird cherry.

The field layer was very variable, depending on soilmoisture status; sedges, great wood-rush, or bram-bles were sometimes dominant, but some woods hadan interesting herb-rich flora. Although alder treeswere often conspicuous, standing in lines beside theriver, alderwood had a very limited extent in the studyarea (0.9%, Table 1). This total did not includeimportant stands alongside the Feugh and Gairn, orsome limited stands at the Dinnet lochs, but didinclude woods with a small proportion of bird cherry,birch, hazel and ash.

2.1.6 Willow

Willows had a more restricted distribution than alderand preferred wetter soils. Abandoned channels,islands and banks of the river were the main habitats.Different species were locally dominant, but goatwillow and grey willow were most common. Otherspecies present included white, eared and bay willowsand osier. The field layer under willows containedmore sedges than in the alderwoods, with bottlesedge, common and common yellow sedges. AtQuithel wood, an interesting mixture of goat willowand silver birch was found. The total area of willowswas 14.6 ha (0.2% of the study area) (Table 1). Overhalf occurred in the lowest section of the river, fromBanchory to Aberdeen (8.7 ha).

2.1.7 Mixed broadleaved woodlandThis category included woods in which ash, wychelm, sycamore or beech were predominant, eithersingly or in combinations. Except for beech, thesespecies require flushed or brown earth soils to thrive,and all had been widely planted in the study area,especially in the lower half of the valley. Wych elmand probably ash are native to Deeside, but beech andsycamore are not. We did not attempt to distinguishplanted woods from natural ones. The ash/wych elm/sycamore woods tended to have a relatively herb-richfield layer with such species as dog's mercury andprimrose. This Woodland was common only in thelower reaches of the valley (Table 1). It was the most

numerous type of semi-natural habitat between Aber-deen and Banchory, occupying 3.4% of the valleyfloor.

2.1.8 Hazel

Hazel occurred both as an understorey shrub inoakwoods and also in pure stands, sometimes ascoppice. It may have been planted widely. It preferredwell-drained situations such as the river banks andbluff slopes. Outside the study area, hazel grewabundantly between Ballater and Crathie on south-facing slopes with rich soils. Within the study area,hazel woodland was scarce (Table 1), and occurredonly downstream from Crathie.

2.1.9 Aspen

Aspen was widespread but very local in the valleyfloor, with only occasional stands of more than a fewstems; 8.7 ha were identified in the survey (ie 0.1% ofthe study area), all in the Ballater to Crathie section.The field layer was usually grassy with meadow-grasscommon.

2.2 Arable and improved grass

This category includes arable land and sown grass-lands. Generally, the land was fenced. Perennialrye-grass and cock's-foot were the main introducedgrasses. Rough grassland which did not appear tohave been improved is described below. Arable andimproved ground occupied 36.6% of the study area(Table 1), 3 times the extent of the next commonestvegetation type. Not surprisingly, none was presentabove the Linn of Dee, and by far the greatest area(1699 ha) occurred in the lowest section of the valleydownstream from Banchory. This was equivalent to49.6% of all arable ground in the study area, and was78.7% of the area of the section.

2.3 Short grasslands and dwarf-herb communities

These communities are subject to grazing and occuron alluvial soils, often where the risk of floodingreduces the potential for reclamation.

2.3.1 Species-poor bent/fescue grassland

This community was characterized by the presence ofbrown bent, common bent, sweet vernal-grass, fine-leaved sheep's-fescue and heath bedstraw. It had thewidest range of any community in the valley, formingthe most important grassland at the river mouth andalso at the Linn of Dee. It occurred on alluvial plains,especially where larger tributaries joined the Dee (egat the junction of the Dee and Quoich). The soilstended to be base-poor thin podzols, well drained butsubject to periodic flooding. Where grazing was light,invasion by trees and shrubs was common. Thecommunity was frequent (Table 1), especially in thesection between Ballater and Linn of Dee; 200 haoccurred around the Quoich mouth.

2.3.2 Species-rich bent/fescue grassland

This community was characterized by the presence ofmany herbs such as self heal, meadow buttercup and

59

wild thyme. It covered a wide altitudinal range (30-370 m), but only fragmentary stands were found in thelowest reaches, and 60.3 ha out of 65 ha of thiscommunity were above Braemar. Species-rich grass-land was generally grazed by cattle or sheep and wasconfined to the river bank or to meadows directlyadjoining the banks on slopes of less than 20°. McVeanand Ratcliffe (1962) suggest that the distribution of thiscommunity in the uplands is related to calcareousrocks. In this survey, it was found on well-drained soilswith a brown earth or acid brown earth, less podzolicthan the species-poor type.

2.3.3 Sheep's-fescue and meadow-grass grasslands

These grasslands differ from the 2 foregoing com-munities in the absence or scarcity of bent grasses.They shared 7 constant species, and were themselvesdifferentiated mainly by the presence or absence ofsmooth meadow-grass and fine-leaved sheep's-fescue. They occurred mostly between Ballater andthe .Linn of Dee, and were of small extent (Table 1).The sheep's-fescue grassland was found in only onesite at Mar Lodge, but the meadow-grass communityoccurred in scattered areas on the river bank onmoderately fertile alluvial soils, confined, except forone site, to the north bank of the Dee. This communitywas often closely grazed. In the upper reaches, it wasfound most commonly with herb-rich heather moor,and in the lower reaches with species-poor bent/fescue grassland.

2.3.4 Dwarf-herb nodum

A dwarf-herb community dominated by thyme andclover species was found ne4 the Linn of Dee. Thiscommunity also occurred to the south of the riverbetween Balmoral and Invercauld Bridge, as a complexwith species-rich bent/fescue grassland (2.2 haaltogether). Both sites were ungrazed by domesticstock. Noteworthy plants present in the dwarf-herbnodum were sibbaldia, moss campion, and lesserclubmoss. McVean and Ratcliffe (1962) associated thiscommunity with high-altitude calcareous outcrops.

2.4 Meadow grasslands

There were 2 communities dominated by relatively tallnon-woody plant species. Neither was grazed but bothwere found under circumstances which limited in-vasion by trees and shrubs.

2.4.1 Tall-grass nodum

A distinctive grassland type with tall grasses dominantoccurred on the river banks, beside islands with poorlydrained alluvium (particularly fine silt), in abandonedchannels and at the edge of some alderwoods.Conditions were either too wet for grazing or therewas protection from grazing. Some tall herbs, eg wildangelica, common valerian and meadowsweet, wereassociated with this community in strips beside theriver. The community occupied only 0.9% of the studyarea (Table 1), but was widespread between themouth of the river and 250 m above OD.

60

Table2. Plant species found on the river bank at Banchory in 1976-79 (sources J Green, P Marren, D Welch, pers. comm.)

*Species present within a 8 m2 relevé

2.4.2 Tall-herb nodum

A tall-herb community, with globeflower and commonvalerian frequent, occurred on drier sites along thebanks of the Dee below 250 m. Grazing by domesticstock was always absent, and the swards were usuallymowed regularly by the fishing ghillies. This corn-

munity was similar to the hay meadows of northernEngland and southern Scotland. Characteristic speciesincluded imperforate and perforate St John's-wort,melancholy thistle and spignel; a full list for one bankat Banchory is given in Table 2. The community wasfound most commonly where the river ran in a rocky

bed and with boulders on the bankside. Some of thebest occurrences were upstream from the bridge atBanchory, at Potarch, and near Abergeldie Castle.Because the tall-herb nodum was confined to the riverbank, it occupied little area (35 ha in total. Table 1),despite being of considerable linear extent.

2.5 Gorse and broom scrubScrub communities, generally dominated by gorse andbroom, have frequently developed from ungrazedgrasslands. These areas occurred beside the riverwhere management had ceased or grazing intensitywas reduced. Given freedom from disturbance, suchsites may develop eventually into broadleaved wood-land. A total of 224 ha of gorse/broom scrub waslocated, most commonly between Aboyne andBallater (Table 1).

2.6 Dwarf-shrub heath

Though heather moor was widespread within thecatchment near the Dee, it was only abundant in theupper reaches of the study area (Table 1). Heathermoor was generally poor in species, particularly herbsand grasses, and required burning for maintenance.On less acid soils, a species-rich variant occurredcontaining bitter vetch, wintergreens and many plantsfound in species-rich bent/fescue grassland. Thisvariant occupied only 59.5 ha, all in the section of riverbetween Aboyne and Invercauld Bridge. It was notablefor the presence of 74 different species in 3 relevés,with an average of 35 species per relevé. Another typeof heath contained much purple moor-grass; hence, itis termed Molinieto-Callunetum. It occurred onwell-drained peats within drier heather moor above370 m, and was grazed by red deer.

2.7 Marsh and swamp communities

Communities dominated by heath rush and mat-grass,together or singly, were found within the species-poorbent/fescue grassland where drainage was poor orlateral water movement occurred. Grazing was usual,preventing successions to heath. As the altitudeincreased, these communities became more exten-sive, but the total area occupied was very small(9.4 ha, Table 1). Similar vegetation occurs quitewidely in the catchment on unimproved farmland.

The sedge nodum containing several Carex speciesoccupied small areas characterized by the presence ofa high water table. This type graded into a willowcommunity in places. At the riverside and in back-waters, this nodum passed into fringing reedswamp(Holmes 1985). The total area was 4.4 ha (0.5%),though the type was difficult to map and this figure isprobably an under-estimate.

Another type of marsh called the soft rush nodum(Sphagneto-Juncetum effusi) occupied 24 ha, with18 ha between Ballater and Linn of Dee. It was foundthroughout the length of the river but rarely in patchesmore than a few m2 in size. The cover of Sphagna

61

reduced as the altitude decreased. It occurred in areasof poor drainage on gleyed soils.

2.8 Shingle

The bare shingle islands, of which Cults is the mostnotable, had a montane and ruderal flora of their own.Cults, for example, contained species washed downfrom high level, eg mountain sorrel and northernrock-cress, and, being tidal, there were also coastalplants such as sea campion and thrift. Other montaneplants washed down the Dee and permanently estab-lished at many locations included northern bedstraw,yellow saxifrage and scurvygrass. The Nootka lupin,introduced from North America in the last century(Marren 1982), and now characteristic of major riversin north-east Scotland, is widespread in the sectionsbelow Braemar.

3 The value, management and conservation of thevegetationA complete description of the vegetation of thefloodplain of a river has not previously been attemptedin north-east Scotland. We are not aware of anystandards for comparison. The Dee valley is partic-ularly suitable for such a study as the width of thefloodplain rarely exceeds 1000 m. Determining theboundaries of the study area posed some problems asthe valley floor is bounded by a series of river terraces(Maizels 1985), and a decision as to which of these totake was sometimes arbitrary. In other cases where asite straddled the arbitrary limit set, eg Quithel Wood,it was also difficult to decide exactly where to site theboundary of the study area. However, only the amountof arable land would be significantly altered bychanging the boundary of the area, and we think thatthe figures provide a good guide to the relativeamounts of semi-natural vegetation on the valley floor.

The semi-natural vegetation of the floodplain of theDee is an important resource. Permanent vegetationreduces soil erosion on the valley floor; lupins andother plants stabilize shingle; and alders protect thebanks, slowing down the rate at which the riverchanges course. Many of the woods are used forshooting, particularly of pheasants, and somestretches of grassland and dwarf-shrub heath providegood grazing for domestic stock. In addition to itspurely economic value, the vegetation is important asa reservoir of wild plant species and animals, and as amain contributor to the appearance of the landscape.

The catchment of the Dee contains nationally rareexamples of pinewoods, birchwoods and heathermoors with associated wildlife, and on the valley floorthere are vegetation types scarce elsewhere in theGrampian Region, particularly alderwoods, mixedbroadleaved woods, meadow grasslands and tall-herbcommunities. To a naturalist, perhaps the most in-teresting vegetation type is the tall-herb community,which occurs in 0.4% of the study area. This vegeta-tion probably owes its existence to the association ofrich soils (derived from outcrops of relatively base-rich

62

rocks) with the mowing of the river banks for ease offishermen. If the mowing were stopped, the compos-ition of this vegetation would soon change. Despitefluctuations in income from the salmon fishery, wehope that the practice of mowing will continue. Theattention of the Nature Conservancy Council is drawnto the desirability of preserving these habitats.Already, the diverse wood at Quithel is an SSSI,together with 3.5 km of bank at Potarch, whilstBallochbuie pinewood SSSI and Dinnet oakwood andMuir of Dinnet NNRs lie partly within the Dee'sfloodplain.

It is paradoxical that for the maintenance of botanicaland zoological interest there should be differentpriorities in the management of the river banks. For theconservation of otters and some invertebrates, thepreservation of scrub on the banks is paramount(Jenkins 1981; Davidson et al. 1985). On the otherhand, to maintain grassland, it is necessary to removescrub and mow or graze. Fortunately, compromise ispossible. There is plenty of river bank with acid soilsfor otter havens and the current recommendations toNCC for the specific management of parts of SSSIs forotters are to preserve scrub on only about 50 m inevery 500 m (Jenkins pers. comm.).

Mixtures of marine and alpine vegetation on theshingle islands in the lower part of the valley areunusual but not unique. They are also found on theRivers Tay and Spey in eastern Scotland. Nonetheless,scientifically, they are an important part of the veget-ation of this river system.

Few would challenge the value of trying to maintainthe semi-natural vegetation, but many threats face it.As already mentioned, grasslands, meadows andheaths below the altitude of the treeline will revert toscrub and woodland if they are not grazed, cut orburnt. Conversely, woodlands will not survive ifgrazing by sheep, cattle or red deer occurs at too greatan intensity. Stock reductions or fencing will beneeded in such cases. It may seem perverse to checksuccessions both to and from woodland, but for manyplant species entry to new habitat even little distantfrom present communities seems difficult. Moreover,woods often cannot regenerate beneath their ownshade, and adjacent unwooded ground on whichregeneration can occur should be set aside (andfenced off) to perpetuate the tree species in the wood.

More direct threats arise from changes in land use orfrom the introduction of more intensive forms ofmanagement. The most important threats include thefollowing.

i. Improvement for agricultural purposes of theremaining semi-natural grasslands, by drainage,liming or re-seeding. This is particularly importanton the floodplain.

ii. Conversion of low-yielding broadleaved wood-

lands composed of native species into •higher-yielding coniferous woods.

iii. Planting riverside moors with commercial trees.

iv. Work to stabilize the river course and to reinforcebanks, which leads to the destruction of shinglebars and islands and their natural vegetation.

There is one activity which seems totally unnecessary,namely the disposal of waste over the edge of steepbluffs and into abandoned channels. Some of the mostnotorious examples of this activity concern urbanwaste, but it is common for farm rubbish and stones tobe dumped into abandoned channels and over theedge of beautifully wooded bluffs.

To protect the important vegetation types, moreinformation on variation in composition between differ-ent sites is needed. Given that knowledge, the bestexamples of vegetation can be selected and appro-priate management regimes recommended. The sur-vey of semi-natural vegetation within the GrampianRegion organized by Miss E Woodward will, whencompleted, help decide the scarcity value of the Deevalley communities. Some judgement must also bemade about the extent of the catchment which shouldbe retained in semi-natural vegetation if river quality isto be maintained. The need for more data must not,however, delay action on the ground to protect thebest of the vegetation. There are 4 stages to suchprotection.

i. Agree management objectives for particular siteswith conservationists and owners.

ii. Provide general information to managers,perhaps in the form of guidelines about desirablemanagement methods.

iii. Safeguard key areas by statutory notification ofareas as SSSIs to enable the protection of theWildlife and Countryside Act to be given.

iv. Provide financial support to encourage the con-servation of natural and semi-natural vegetation.

4 ConclusionIt is to be hoped that one successful achievement ofthis symposium will be the widespread disseminationto owners, occupiers and other land managers ofknowledge of the value of the riverside land for natureconservation. In the longer term, the conservation ofthe vegetation cannot be considered in isolation fromthe conservation and management of other biologicalresources in the river, and it may be necessary tocreate an appropriate forum in which the wide range ofissues involved can be discussed constructively.

5 Summary5.1 This paper describes the vegetation of the valley

floor, which is distinguished as the area where

river processes and management related to theriver both operate. The vegetation on the valleyfloor was closely related to river processes,landform and management.

5.2 Twenty-three different vegetation types are de-scribed.

5.3 Mixed broadleaved woods, alderwoods,meadow grasslands and herb-dominant veget-ation were the most significant of these veget-ation types because, in the Dee catchment, theywere virtually confined to the valley floor.

5.4 The survival of these vegetation types isthreatened by land improvement, commercialforestry and bank engineering. To protect thesetypes of vegetation, land managers need to beinformed of their value, and appropriate systemsof management need to be drawn up jointly withwildlife conservationists.

6 AcknowledgementsWe thank the landowners involved for permission tovisit their estates.

63

7 ReferencesBirse, E. L & Robertson, J. S. 1976. Plant communities and soils ofthe lowland and southern upland regimes of Scotland. (Soil survey ofScotland monograph.) Aberdeen: Macaulay Institute for Soil Re-search.

Davidson, M. B., Owen, R. P. & Young, M. R. 1985. Invertebratesof the River Dee. In: The biology and management of the River Dee,edited by D. Jenkins. 64-82. (ITE symposium no. 14.) Abbots Ripton:Institute of Terrestrial Ecology.

Holmes, N. T. H. 1985. Vegetation of the River Dee. In: The biologyand management of the River Dee, edited by D. Jenkins, 42-55. (ITEsymposium no. 14.) Abbots Ripton: Institute of Terrestrial Ecology.

Jenkins, D. 1981. Ecology of otters in northern Scotland. V. A modelscheme for otter Lutra lutra L. conservation in a freshwater systemin Aberdeenshire. Biol. Conserv., 20, 123-132.

McVean, D. & Ratcliffe, D. 1962. The plant communities of theScottish highlands. London: HMSO.

Maizels, J. K. 1985. The physical background of the River Dee. In:The biology and management of the River Dee, edited by D. Jenkins,7-22. (ITE symposium no. 14.) Abbots Ripton: Institute of TerrestrialEcology.

Marren, P. R. 1982. A natural history of Aberdeen. Finzean: RobinCallander.

Steven, H. M. & Carlisle, A. 1959. The native pinewoods ofScotland. London: Oliver & Boyd.

64

Invertebrates of the River DeeM B DAVIDSON*, R P OWEN* and M R YOUNGt*North East River Purification Board, AberdeentDepartment of Zoology, University of Aberdeen

1 IntroductionThis paper concentrates on the larger, bottom-livinginvertebrates of the River Dee. They are a vital link inthe food chain, being essential food for economicallyimportant species such as the salmon, and leadingultimately to those with significance for conservationsuch as the otter and dipper. Furthermore, they aregenerally recognized as being excellent indicators ofthe 'health' of a river and, because they form anintegral part of regional and national monitoringschemes, comparisons can be made easily with otherriver systems. They show the essential characteristicsrequired by organisms which are useful as 'indicators'of the state of a river, that is they are abundant, easilycaught and identified, fairly sedentary, show a widerange of tolerances, and are present throughout theyear. They integrate temporal variations in river waterquality in a way no other features do, and so evenshort episodes of pollution, which might be missed byinfrequent chemical analysis, are detectable by study-ing them.

In this study, the invertebrate data have been looked atin 2 ways. First, the communities of animal species ateach site have been compared, using associationanalysis, and their similarities or. differences relatedsubjectively to environmental factors. Second, certainindividual species have been selected which showdifferent distribution patterns within the catchmentand the reasons for these differences have beenexplored. Additionally, in order to demonstrate dis-tribution differences on a smaller scale, the environ-mental preferences within one site of several mem-bers of the ephemeropteran (mayfly) family, theHeptageniidae, have been investigated.

EMA

Figure 1. River Dee catchment sampling sites

ALLATER

A diversity index and a biotic score system which arewidely used in pollution assessment have beenapplied to the data to demonstrate the effects ofpollution and enrichment at certain sites, and to enablecomparisons to be made with other river systems.

Actual or potential threats to this special river create aparticular need for this type of study. Although many ofthese problems are of a lowland nature, such as fromagriculture, industry and urban centres, there are, inaddition, upland problems such as recreation andafforestation.

This study describes the characteristic invertebratecommunities of the river and its tributaries, andsuggests some explanation of the reasons for theirvariations. It provides an essential baseline for 1983-84, against which future changes may be judged, andallows comparison with other rivers, most of which aremore affected by man's activities. It also enables anassessment to be made of the effect of currentproblems, and suggests actions needed to prevent anydeleterious changes in the future.

2 Study area and sitesSamples were taken at 30 sites for this study, 10 onthe main river and 20 on various tributaries in theupper, middle and lower parts of the catchment(Figure 1).

The sites (Table 1) are situated in a variety of land useareas: for example, the upper tributaries drain theGrampian Mountains with high-altitude deer forest;the intermediate ones drain hill-farming areas, grousemoor and coniferous forest; and the lower ones drainagricultural land, industrial sites and urban areas.

BOYN.BANCH R

ABERDEEN

Table 1. Sites sampled for invertebrates in the catchment of the Dee

The tributaries vary in catchment area from the Ardoe,at a little over 4 km2, to the Gairn at 148 km2, althoughmost are over 40 km2. The majority of the tributarysites were located close to the main river, but somewere placed for the purposes of monitoring specificpollution threats in various parts of the catchment: theupper Clunie is near the ski-ing areas of Glenshee, theLogie is downstream of a sewage works' dischargebut above the Muir of Dinnet National Nature Reserve(NNR), and the Leuchar is the outlet stream of thehighly eutrophic Loch of Skene.

The precise location of each site was carefully chosenso that the substrate types were as similar as possible,with eroding beds and turbulent water flow. However,amongst these 'riffle' sites, there were inevitabledifferences due in part to discharge regimes and alsoto geological variation. In the main river, substratesvaried from well-distributed, evenly sized cobbles andgravels at Braemar, Silverbank, Milltimber and Aber-deen, to fairly heterogeneous compositions of bedrockand boulders with pockets of gravels and sands atPolhollick and Potarch. Many of the tributary sites hadsimilar substrates of bedrock with boulders and aheterogeneous distribution of gravels. These includedthe large watercourses of the Gairn and the Muick, theintermediate ones such as the Tanar, the Canny, theFeugh and the Dye, but also some small ones like the

65

Sheeoch and the Ardoe. Most of the other tributarieshad mixtures of cobbles, pebbles, gravels and sands invarious proportions, although the Kinnernie, for exam-ple, had a rather homogeneous combination of finegravel and sand.

Some of the tributary locations were peculiar, in thatthey were the immediate outlet streams for areas ofstanding water; the Dinnet, the Leuchar and the Ardoedrain the Lochs Davan and Kinord, Loch of Skene andArdoe Dam, respectively.

3 Methods3.1 Sample collection

Samples of bottom-living invertebrates were collectedusing a hand-net and a standard 'heel-kick' technique(Furse et al. 1981). After each site visit, the animalscollected were preserved and hand-sorted from theaccompanying substrate material in the laboratory.Identification was taken to species level in most cases,but to genus and family in a few groups. The level oftaxonomic differentiation was consistent with mostother recent studies on freshwater macro-invertebrates (Jenkins et al. 1984; Clare & Edwards1983; Scullion et al. 1982).

Apart from one site, all the locations were sampled inApril, August and October 1983, and in April/May and

66

August 1984. These dates were taken to correspondwith spring, summer and autumn seasons. The onlyexception was the Sheeoch, which was not sampledin summer 1983. The last 3 sample sets were used toderive a diversity index and, for this purpose, theinvertebrates were also counted.

3.2 Classification and ordination techniques

Groups of similar sites within the river and tributariesand then groups of species with similar preferenceswere identified using TWINSPAN analysis. This is apolythetic, divisive hierarchical classification techniquedevised by Hill (1979b), which splits up the total groupof sites into sub-groups based on their shared speciesand arranges them along a gradient depending on thedegree of similarity between their species compo-nents. The dendrogram produced can be used as a'key' for placing sites in their correct group byidentifying the most characteristic 'indicator' speciesfor each group. No exact statistical guidance is givenas to the level at which division should be stopped but,generally, the earlier the division, the more informationit conveys.

The next stage in the analysis was an ordination whicharrayed the sites along various axes, with succeedingaxes reflecting less and less of the overall variation.The technique used was Detrended CorrespondenceAnalysis, ie DECORANA (Hill 1979a). The clustersproduced by this method do not correspond exactly tothe single axis TWINSPAN dendrogram groups but, ifaxis 1 of DECORANA reflects most of the variation (asis often the case), then correspondence is quite good.

3.3 Diversity index

Environmental stress frequently reduces communitydiversity (Hellawell 1978), and changes in communitystructure are often assessed by a diversity index.Many lotic invertebrate studies have used theShannon-Weaver index (Shannon & Weaver 1963), soit has been used in this study in order to facilitatecomparisons with other rivers.

This index relates the numbers of each individual taxonto the total numbers of animals in the community andtakes the form:

d = — (n/N log2 n/N)

where n, = the number of individuals in the ith taxon,and N = total number of individuals in the sample.

The Shannon-Weaver diversity index was calculatedfor each of the counted samples, ie for each site forautumn 1983, and for spring and summer 1984.

3.4 Biotic score

Certain indices have been developed which make useof the varying tolerances of different species topollution, and many assign 'scores' which reflect thesepollution tolerances. However, not all of these indicescan be applied to a wide variety of river types.

The Biological Monitoring Working Party (BMWP)score system was developed with the aim of nation-wide applicability in assessing the biological conditionof rivers (BMWP 1978), and has recently been testedon 41 rivers throughout the UK, including the Dee(Armitage et al. 1983). Because this system facilitatesthe placing of the Dee in a British context, it seemedlogical to use it in this study. The version used is theAverage Score Per Taxon (ASPT) which allows thecollection of more information with less effort andreduces seasonal variation (Armitage et al. 1983).

The BMWP (ASPT) score has been calculated on thecumulative species list for each site over the 5surveys. This procedure tends to maximize the scoresand integrate seasonal variation and, for the purposesof this study, presents a broad view of the biologicalconditions pertaining during the study period.

4 Results4.1 Species lists

The species lists for each of the 30 sites used in thisstudy, which formed the basis of all subsequentanalysis, are combined in Table 2 to give a total list forthe whole catchment (the raw data are available fromthe North East River Purification Board). The list is notexhaustive and includes only those taxa encounteredin this study; other species undoubtedly occurred atother sites.

4.2 Classification and ordination analysis

The dendrogram in Figure 2 displays the result of theTWINSPAN analysis of the sites/species matrix de-rived from the species lists and, at the fourth level ofdivision, has produced 10 groups of sites associatedby their similar invertebrate communities and arrayedalong a similarity gradient.

The first 4 groups (A-D) included all the main riversites, and geographically adjacent sites were placedtogether. The exceptions were sites 6 and 10 (Potarchand Aberdeen). Most of the major tributaries such asthe Muick, Tanar and Feugh were grouped with themiddle main river sites, while the Gairn and lowerClunie were placed with the 2 highest main river sites.

The central part of the dendrogram (groups E-G)encompassed the richer tributaries of the middlecatchment, which drain areas including agriculturalland, and this description could probably include theCrynoch at the lower end of the catchment. Theanomalous site in this group was the Clunie atCairnwell (site 11), below the Glenshee ski develop-ment.

The remaining portion of the dendrogram (groups H-J)separated off the Leuchar/Culter sub-catchment, theDinnet and the Ardoe, all of which are the outletstreams of lochs or impoundments. The Kinnernie wasincluded in this group and is the inlet stream of theLoch of Skene, below the village of Dunecht.

A similar procedure was followed for the speciesassociations, producing the TWINSPAN dendrogram inFigure 3 which relates to the groups listed in Table 3.Again, the analysis was terminated at level 4, and atthis point there were 10 groups of associated species.The first division split the species into 2 majorgroupings: A-G containing 85 species, and H-L theremaining 40 species. Groups H-L can be defined,broadly, as including those species which occur at themore 'enriched' sites and, in particular, those whichlive in the outlet streams of lochs and ponds.

Table 2. Species found at the sample areas in the Dee catchment

PLECOPTERA (STONEFLIES)

Taeniopteryx nebulosa (L.)Brachyptera putata (Newman)Brachyptera risi (Morton)Protonemura praecox (Morton)Protonernura montana (Kimmins)Protonemura meyeri (Pictet)Amphinemura sulcicollis (Stephens)Nemoura cinerea (Retzius)Nemoura avicularis (Morton)Nemoura cambrica (Stephens)Leuctra geniculata (Stephens)Leuctra inermis (Kempny)Leuctra hippopus (Kempny)Leuctra fusca (L.)Capnia bifrons (Newman)Per lodes microcephala (Pictet)Diura bicaudata (L.)Isoperla grammatica (Poda)Dinocras cephalotes (Curtis)Perla brPunctata (Pictet)Chloroperla torrentiurn (Pictet)Chloroperla tnPunctata (Scopoli)

EPHEMEROPTERA (MAYFLIES)

Baetis scarnbus EatonBaetis vernus/tenax Curtis/EatonBaetis rhodani (Pictet)Baetis muticus (L.)Centroptilum luteolum (Muller)Rhithrogena sernicolorata (Curtis)Heptagenia sulphurea (Muller)Ecdyonurus venosus (Fabricius)Ecdyonurus torrentis KimminsEcdyonurus dispar (Curtis)Leptophlebia marginate (L.)Paraleptophlebia submarginata (Stephens)Ephemerella ignita (Poda)Ephemerella notata EatonEphemera danica MullerCaenis rivulorurn Eaton

TRICHOPTERA (CADDIS FLIES)

Rhyacophila dorsalis (Curtis)Rhyacophila obliterate McLachlanGlossosorna spp.Agapetus spp.Philopotarnus rnontanus (Donovan)Wormaldia spp.Plectrocnernia conspersa (Curtis)Polycentropus flavornaculatus (Pictet)Hydropsyche pellucidula (Curtis)Hydropsyche angustrpennis (Curtis)Hydropsyche contubernalis (McLachlan)

67

Groups A-G represent the characteristic communitiesof the Dee system and, because of the generaluniformity of the sites used in the analysis, definingthese associations in terms of environmental factors isa much more difficult proposition. However, certaingeneral statements can be made: groups A and Binclude species characteristic of the smaller tributarieswith slower flows and gravelly substrates; group Ccontains species which occur in the main river and thelarger tributaries; groups F and G are associations ofspecies which are widespread within the catchment.

Hydropsyche siltalai Doh lerHydropsyche instabilis (Curtis)HydroptilidaeDrusus annulatus StephensEcclisopteryx guttulata (Pictet)Limnephilus spp.Anabolia nervosa CurtisPotamophylax latipennis (Curtis)Potamophylax cingulatus (Stephens)Halesus radiatus/digitatusStenophylax lateralis/sequaxAllogamus auricollis (Pictet)Odontocerum albicorne (Scopoli)Athnpsodes spp.Goera pilosa (Fabricius)Silo pallipes (Fabricius)Lepidostoma hirtum (Fabricius)Lasiocephala basalis (Kolenati)Brachycentrus subnubilus CurtisSericostoma personaturn (Spence)

DIPTERA (TVVO-WINGED FLIES)

Tipula spp.Hexatoma spp.Pedicia spp.Dicranota spp.Elaeophila spp.Taphrophila spp.PsychodidaeCeratopogonidaeChironomidaeSimulidaeHernerodrornia spp.Clinocera spp.AthenX ibis (Fabricius)Lirnnophora spp.

MEGALOPTERA (ALDER FLIES)

Sialis lutaria (L.)

COLEOPTERA (BEETLES)

Brychius elevatus (Panzer)Ha liplus spp.Potamonectes depressus/elegansOreodytes sanrnarki (Sahlberg)Platarnbus maculatus (L.)Agabus spp.Helophorus spp.Hydraena gracilis GermarElmis aenea (Muller)Esolus paralleleprPedus (Muller)Limnius volckrnari (Panzer)Oulirnnius tuberculatus (Muller)

68

Table 2.—continued

HEMIPTERA (WATER BUGS)

Sigara concinna (Fieber)

MOLLUSCA (WATER SNAILS)

Potamopyrgus jenkinsi (Smith)Lymnaea peregra (Muller)Physa fontinalis (L.)Planorbis albus MullerPlanorbis contortus (L.)Ancylus fluviatilis MullerSphaerium spp.Pisidium spp.

HYDRACARINA (WATER MITES)

Unidentified spp.

MALACOSTRACA (WATER SLATERS & SHRIMPS)

Asellus aquaticus (L.)Asellus meridianus RacovitzaGammarus pulex (L.)Gammarus zaddachi Sexton

Sp 122 — Tubificidae

Sp 3 + B. risi Sp 4 + P. praecox

Sp 69 — HexatomaSp 20 — P. bipunctataSp 52 + D. annulatusSp 78 + Hemerodromia

HIRUDINEA (LEECHES)

Theromyzon tessulatum (Muller)Glossiphonia complanata (L.)Helobdella stagnalis (L.)Erpobdella octoculata (L.)Trocheta bykowskii Gedroyc

NEMATODA (ROUND WORMS)

Unidentified spp.

TRICLADIDA (FLATWORMS)

Polycelis tenuis/nigraPolycelis felina (Da lyell)Crenobia alpina (Dana)

OLIGOCHAETA (SEGMENTED WORMS)EnchytraeidaeLumbricidaeLumbriculidaeTubificidaeNaididae

GoRpiolDEA(HORSEHAIR WORMS)

Unidentified spp.

Sp 4 + P. praecox

30 sites

Sp 5 + P. montana

Figure 2. TWINSPAN classification of sites (cf Table 1 for site numbers)

Sp 3 — B. risiSp 104 + Pisidiurn

Sp 6 — P. meyeri

Sp 18 — I. grammatica

610(A)

789

(3)

345

1518232425(C)

12

121314(D)

1619202122(E)

29(F)

11(GI

172830(H)

26 27(J)

Table 3. TWINSPAN species groups

A

Figure 3. TVV1NSPAN classification of species (cf Table 3 for details of groups)

69

70

As expected, the site associations produced by DE-CORANA (Figure 4) were essentially the same as inthe TININSPAN dendrogram (Figure 2). Most of themain river sites were closely associated and groupedwith the larger, more upland tributaries. The othermain association included the middle and smaller,lower catchment tributaries which tended to besomewhat richer and less torrential.

200

100

ABERDEEN

0 Leuchar

0 Dinnet

•

o Ardoe

0o

Lower &middle tributaries

Main river &upper tributaries

• — Main river

0 — Tributaries

0 Upper Clunie

100 200Axis 2

Figure 4. DECORANA ordination plot (axis 1 vs axis 2)

Five sites can be picked out from the diagram as beingsomewhat different on the basis of their invertebratecommunities. These are the main river at Aberdeen(site 10), the Leuchar (site 27), the Ardoe (site 30), theDinnet (site 17) and the Clunie below the ski develop-ment (site 11).

4.3 Relative abundance of Plecoptera (stoneflies) and Ephemerop-tera (mayflies)

Figure 5 shows the numbers of different species ofPlecoptera and of Ephemeroptera which occurred ateach of the 10 main river sites in 1983 and 1984. Ingeneral, there was a tendency for the number ofmayfly species to increase downstream, whereas thenumbers of stonefly species remained fairly constantto the tidal limit (just upstream of site 10), with theexception of site 6 at Potarch.

4.4 Distribution of species

Figures 6i-x show a series of sample sites, indicatedby open circles, along a linear river (double line) and ona number of tributaries (single lines). The positiveoccurrence of a particular species at a site is shown bya filled circle. The information used in plotting thesediagrams came not only from the present study, butalso from Davidson and Young (1980).

These maps show a range of spatial distributionpatterns within the Dee catchment, from the ubiquit-ous species, through those more selective animalspreferring tributary or main river sites, to those specieswith restricted distributions requiring specializedconditions.

12

10

4

6

10

122 4 6

Main river sites

8 10

Figure 5. Relative numbers of species of Plecopteraand Ephemeroptera at main river sites

The riffle beetle Elmis aenea (Figure 6i) occurred atvirtually all sites within the catchment, whereas themayfly Heptagenia sulphurea (Figure 6ii) and thedipteran fly Atherix ibis (Figure 6iii), although wide-spread and well distributed in the main river, tended tobe restricted to the longer tributaries.

The 2 common species of flatworm (Tricladida) in thecatchment were Polycelis felina and Crenobia alpina.These had overlapping ranges but quite differentdistributions. P. felina (Figure 6iv) occurred throughoutthe main river and the smaller tributaries but not in thelarger and more oligotrophic streams, such as theQuoich, the Muick and the Tanar. C. alpina (Figure 6v),on the other hand, was present in the upper part of themain river, overlapping with P. felina (but not neces-sarily at the same season), and in the more uplandtributaries.

The caddis fly (Trichoptera) genus Hydropsyche wasrepresented in the catchment by 5 species which hadcontrasting distributions and which illustrate howanimals which are superficially identical share theoverall habitat. Hydropsyche siltalai (Figure 6vi) and H.pellucidula (Figure 6vii) had very similar distributions,both being well spread throughout the catchment. H.instabilis (Figure 6viii) was completely absent from themain river, but was found in a number of tributaries,while H. contubernalis (Figure 6ix) was restricted tothe middle and lower parts of the main river. The fifthspecies, H. angustipennis (Figure 6x), had the mostrestricted distribution, occurring only in the Cultersub-catchment at 2 sites, and in the main river at onesite. All of these sites were downstream from theeutrophic Loch of Skene.

Figure 6i-x. Distribution of selected species at sample sites in the Dee catchment, 1984

0 Species absent

• Species present

EyBaddoch

Clunie

Ca !later

Source Source

0 00 Species absent

• Species present

0 Lui

0

BRAEMAR

QuoichSlugain

EyBaddoch

Clunie

Cal later

0

0

• •

BRAEMAR

• ABERDEEN • ABERDEEN

Mouth Mouth

Figure 6i. Elmis aenea Figure 6ii. Heptagenia sulphurea

Lui

QuoichSlugain

71

72

Mouth

• ABERDEEN • ABERDEEN

Mouth

Figure Atherix ibis Figure 6iv. Polycelis feline

Muick

•

•BALLATER

•

•

Gairn

TullichMuick

•

•

BALLATER•

Gairn

Tullich

• Di nnet/Log ie • Dinnet/Logie

Tanar •• ABOYNE

Tar land

Tanar •ABOYNE

Tar land

• Dess• Dess

• •

Canny/Beltie Canny/Beltie

Feugh Feugh •

Dye DyeSheeoch • Sheeoch •

Crynoch •Culter/Leuchar

Crynoch •Culter/Leuchar

Mouth Mouth

Figure 6v. Crenobia alpina Figure 6vi. Hydropsyche siltalai

73

Muick

o

0

BALLATERGairn

TullichMuick

•

•

BALLATERGairn

Tullich

•

Dinnet/Logie Dinnet/Logie

Tanar 0 Tanar •0 ABOYNE • ABOYNE

Tar land Tar land

oo

Dess 0•

Dess

Canny/Beltie Canny/Beltie--0---0—

Feugh Feugh 0

Dye DyeSheeoch 0 Sheeoch 0

Culter/Leuchar Culter/Leuchar

74

Mouth Mouth

Figure 6vii. Hydropsyche pellucidula Figure 6viii. Hydropsyche instabilis

Source

0

Source

00 Species absent 0 Species absent

• Species present • Species present0 0

Ey•

LuiEy

0Lui

Baddoch 0Quoich Baddoch 0

QuoichClunie • Slugain

Clunie SlugainBRAEMAR BRAEMAR

Cal later Cal later

• 0

• Gairn 0 GairnMuick BALLATER Muick BALLATER

Tullich Tullich

0

Dinnet/Logie Dinnet/Logie

Tanar 0 Tanar 0• ABOYNETar land

0 ABOYNETar land

0 Dess0 Dess

• 0

0-0-- Canny/Beltie Canny/Beltie-0-0—

Feugh Feugh 0

Dye DyeSheeoch 0 Sheeoch 0

Crynoch •Culter/Leuchar

Crynoch 0Culter/Leuchar

Mouth Mouth

Figure 6ix. Hydropsyche contubernalis Figure 6x. Hydropsyche angustipennis

75

Source

0

Source

00 Species absent 0 Species absent

• Species present • Species present0

Ey0

Lui Ey0 Lui

Baddoch 0Quoich Baddoch 0 Quoich

Clunie Slugain Clunie 0 Slugain

BRAEMAR BRAEMAR

Cal later Cal later

0 0

• Gairn 0 GairnMuick BALLATER Muick BALLATER

• Tullich Tullich

0 0

0 Dinnet/Logie 0 Dinnet/Logie

Tanar • Tanar 00 ABOYNE 0 ABOYNE

Tar land Tar land

0 Dess0 Dess

• 0

0 0 Canny/Beltie Canny/Beltie0 0

Feugh 0 Feugh 0

Dye DyeSheeoch 0 Sheeoch 0

Culter/Leuchar Culter/Leuchar

edrren:

penciLowe- Dee Fr. Dv— s

Indiy:dual specimens were coHected measured andcounted on stones of differing sizes, at d:fferentdepths and at different current speeds The density ofeach species under different conditions :s recorded inTable 4, where it can be seen that each showed acharacteristic set of preferences However, altnoughthe average size range of each species differedsomewhat, there was no evidence to show that therewas a relationship within each species between sizeand any of the environmental variables

These differing preferences suggest that there was, ingeneral, only limited overlap between E. dispar aridsemicolorata and between them and the others,whereas E. venosus and H sulphurea, which wereusually separated, nevertheless occurred together onmedium and large stones, at medium depths and atmedium current speeds (Figure 71

Stomach content identification and examination of themouthparts show a further differentiation in that Hsulphurea preferred detritus, whereas the otherspecies used diatoms and other algae There wasslight evidence that H. sernicolorata has rnouthpartsmore fitted to scraping stone surfaces rather thangeneral foraging, but this is unsubstantiated

4 6 Onif rsity index

Table 5 lists the values of the Shannon-Weaver indexfor each site and for each of the 3 surveys that wereused for this calculation The mean diversity index isalso shown for each site, and these have been plotted

Aberdeen The highest mean diversity indices werefound on the Gairn and Curter. In many cases, thetributary sites hravdsimilar mean diversities to theadjacent main

er,even at the unusually high

E. venosus

Ecdyonurusdispar

Rhithrogenasemicolorata

5C.0 10 15

Heptageniasulphurea

Currentspeed Im s 1)

Stone sizeapprox area cm7

Figure 7 Habitat preferences of 4 species of Epheme-roptera larvae from the lower Dee in September 7977

5

t, 4

- 3

i. d

O

7

6

5

4

Table 5. Shannon-Weaver diversity index and Biological Monitoring Working Party score (average score per taxon)for sites in the catchinent of the Dee

20 40 60 80 100 120 140

Distance (km)

Figure 8ii. Biological Monitoring Working Party score(average score per taxon-ASPT)

• - Main river

o - Tributaries

77

diversity site at Milltimber, just above which both theCulter and Crynoch are confluent with the Dee. AtSilverbank, the mean diversity index was somewhatdepressed but the Canny, Feugh and Sheeoch, whichjoin the Dee in this vicinity, remained significantlymore diverse.

Seasonally, the diversity indices varied to some extent,20 40 60 80 100 120 140 although on the main river sites the differences were

Figure 8i. Shannon-Weaver diversity index (d) neither large nor consistent between the 3 surveys.Amongst the tributaries, the diversities appeared, inmost cases, to be greatest in the spring and in some

i. ASPT streams, notably the Dinnet, Dye, Kinnernie, Leuchar8

and Ardoe, there were quite large seasonal variations.

4.7 Biotic scores

The scores for the main Dee sites (Table 5, Figure 8ii)remained consistent through most of the river'slength, except for a slight dip at Silverbank, and then agradual decline was evident at the last 2 sites. Some ofthe tributary sites had much lower BMWP scores thanwas normal for the catchment. These included theupper Clunie, the Dinnet, the Leuchar and the Ardoe.

78

5 Discussion5.1 Historical modifications to the Dee catchment

Although the Dee is relatively unaffected by man'sactivities compared with other British rivers, thecatchment has changed with the'destruction of someupland pine forest and deciduous woodland along thevalleys. Extensive drainage, especially in the last 150years, has resulted in the loss of large areas of wetlandand of lochs, such as the Loch of Auchlossan, nearLumphanan, which were important for wildfowl.

These historical changes must have altered the patternand composition of the invertebrate communities inthe Dee catchment, probably with a reduction inhabitat diversity and restriction of the ranges of certainspecies.

However, the Dee has probably retained a greaterproportion of natural communities in comparison toother British rivers, especially in the lower reaches,which are usually the parts most influenced by man.

5.2 Comparisons with other British rivers

The importance of the Dee has been put in a nationalcontext by the Freshwater Biological Association intheir River Typing Project. Wright et al. (1984) pro-duced an ordination of 41 British rivers (with 268 sitesin total) and concluded that axis 1 of their ordination(with a total score range of 0-275) distinguished bestbetween river types. Most rivers appeared to have alimited range of axis 1 score and, in general, riverswhich had all sites with scores of less than 100occurred in Scotland and northern England, whereasrivers with axis 1 scores greater than 100 occurred inEast Anglia, the midlands, east and south-east Eng-land. The Dee was placed at one extreme of thisanalysis with all sites having scores of 50 or less, andall the Dee sites were described as typical of uplandrivers.

The Don, rising only a few km from the headwaters ofthe Dee and also entering the sea at Aberdeen, has amore agricultural catchment and this is reflected in itsinvertebrate communities. Although the lowest sec-tion of the Don is as torrential as the Dee and shouldsupport an interesting and diverse invertebrate com-munity to its estuary, it is, unfortunately, adulteratedby industrial and domestic effluents, leaving a res-tricted pollution-tolerant fauna.

5.3 Groups of sites and anomalous sitesIn this study, TWINSPAN and DECORANA demons-trated that there were clear geographical factorsoperating to produce a gradient of community types.The catchment can be divided up conveniently, usingthese techniques, to produce 2 main groups of sites.These are: (i) the main river along with the largerupland tributaries; (ii) the smaller, mid- and lowercatchment tributaries.

Within these groups (Figures 2 & 4), there was adistinct gradient along which the sites were arrayed,

reflecting the major physical and chemical character-istics of the catchment. In general, there was atendency for a slight increase in the concentration ofnutrients down the catchment, along with a reductionin current speed and an increase in temperature. Theinvertebrates responded to these changes and thecommunities became slightly more diverse down-stream at the richer, less torrential and warmer sites.

This general pattern was over-printed by other factorsrelated to land use. For example, most of the streamsin the second group drain some agricultural land andthe invertebrate communities showed signs of furtherenrichment.

Outside these 2 main groups were other sites whichappeared to have unusual invertebrate communitiescompared to geographically adjacent locations. TheLeuchar (site 27) is only a short distance below theLoch of Skene, a large 'lowland' loch, surrounded byagricultural land and receiving drainage from a numberof small settlements. As a result, the loch is eutrophic,with annual blooms of the cyanobacterium Aphan-izomenon (Owen 1983). The discharge from the lochwas rich in nutrients and suspended matter, asreflected by the invertebrate community which in-cluded many detritus- and filter-feeding organismssuch as the caddis Hydropsyche angustipennis and themollusc Sphaerium corneum. The Leuchar becomesthe Culter near the confluence with the Dee, and it hada similar species association but better water quality,due to self-purification in the intervening 9 km anddilution by other tributaries. The Ardoe and the Dinnetdrain a small pond and 2 lochs (one slightly meso-trophic) respectively, and they showed intermediatepositions between the Leuchar and the main group ofmid-catchment tributaries (Figure 4).

The upper Clunie at Cairnwell (site 11) was wellremoved from the main body of sites in the ordinationplot (Figure 4). In the TWINSPAN dendrogram (Figure2), it was separated off in group G, away from theother Clunie sample sites in group D. As it is thehighest sampling point in the catchment, it might beexpected to be a little unusual. However, the situationis complicated and, although the burn flows offrelatively calcium-rich rock, it also receives drainagefrom the expanding ski-ing development at the Cluniewatershed, including oils from vehicles and machineryand also septic tank discharges. The species list forthis site included upland stoneflies such as Proto-nemura praecox and P. montana, but had only 24 taxain all. The mean diversity index was the lowest of allthe sites studied and the biotic score was also muchlower than for the other upland sites, being similarinstead to the enriched sites of the middle tributaries.Unofficial camping in this valley is a potential prOblem,especially during low flow periods, with large numbersof people using the bankside areas of the Clunie. Withno facilities for waste disposal, there is a strongpossibility of pollution although, as yet, the invert-ebrates below this area have shown no signs of stress.

The main river at Aberdeen (site 10) was somewhatunusual relative to the other main river sites becauseof its tidal nature and the occurrence of speciesconfined to this site, such as the essentially estuarineshrimp Gammarus zaddachi, and the snail Physafontinalis, which was also found in the lower Don.Nevertheless, the site also had a good selection ofstoneflies, mayflies and caddis, all characteristic oflarge upland rivers.

5.4 Diversity indices and biotic scores

A measure of the diversity of a community may beuseful where it is highly stressed and a small numberof pollution-tolerant species become dominant.However, where mild pollution causes enrichment ofan oligotrophic watercourse, any loss of speciesintolerant to enrichment is, in general, more thancounterbalanced by an influx of species that are unableto survive in oligotrophic waters and are, instead,favoured by extra nutrients. This increased diversity isnot necessarily a desirable feature as it indicates adeparture from more natural conditions. An oligo-trophic river such as the Dee should show lowdiversity indices. The diversity index is, therefore, not avery useful tool in assessing the potential threats tothe Dee catchment.

The biotic score relies on the presence or absence ofpollution-tolerant or pollution-intolerant animals toassess the pollution stress at a site. However, bioticscores are often too crude in differentiating betweenthe tolerances of species to mild enrichment inoligotrophic systems (eg amongst the genus Hydro-psyche). It is, perhaps, more useful than the diversityindex, and, indeed, BMWP detected pollution stresson the Clunie, Dinnet, Leuchar and Ardoe, but againhad fairly restricted application in the catchment as awhole.

5.5 Species distributions

5.5.1 Variation between sites

The various analyses showed that the most importantfeature of the Dee system was the general uniformityof the invertebrate communities with characteristicupland and oligotrophic species remaining even in thelowest areas.

Although it was possible to demonstrate a change inthe relative numbers of species of stoneflies andmayflies from source to mouth, both groups were stillwell represented at the tidal limit (Figure 5). Thesegroups are characteristic of torrential stream faunas,with the stoneflies preferring colder and better oxy-genated water and, therefore, being dominant in theupper catchment.

The distributions of particular species within thecatchment help to illustrate the various environmentalfactors which act on freshwater invertebrates andwhich are important from the river management pointof view.

79

Most species appeared to have fairly wide tolerancesand could be accommodated in an appropriate niche atmost sites in the Dee catchment. However, certainspecies were more restricted in distribution; forexample, the mayfly Heptagenia sulphurea and thedipteran Atherix ibis preferred sites on the main riverand large tributaries and probably require a relativelyconstant temperature regime and are adapted towithstand fairly strong currents.

The 2 common flatworms in the catchment, Polycelisfelina and Crenobia alpina, showed contrasting dis-tributions (Figures 6iv & 6v) controlled by temperature(Hynes 1970). They provide a classic example ofinterspecific competition and zonation. C. alpina wasrestricted to the upper main river and larger tributaries.It requires colder water to complete its life cycle thanP. felina, which was well distributed in the main riverand smaller tributaries. Both species migrate up- anddownstream (Kaushik & Hynes 1971) to feed andbreed, and consequently there is an overlap in theirranges but probably with a fair amount of temporalseparation.

The breeding range of C. alpina has probably beenreduced with the removal of bankside shading treeswhich keep down the summer maximum tempera-ture. This change may also have affected otherspecies.

In recent years, it has become possible to identifyreliably various species of the caddis genus Hydro-psyche. It is interesting to consider how such a groupof superficially identical animals share habitats andexhibit different environmental requirements, es-pecially as they are often lumped together in onecategory in pollution indices, such as the BMWP scoresystem.

H. siltalai and H. pellucidula had similarly widespreaddistributions, extending to the tidal limit. Thesespecies, however, occupied different niches becauseof a slight temporal stagger in their life cycles,resulting in H. pellucidula over-wintering at a larger size(instar 5) than H. siltalai (instar 3) (Macan 1981).Consequently, the larvae of H. siltalai can live insmaller crevices; they spin finer nets for filter feedingand so catch a different range of food.

H. instabilis was restricted to certain tributaries,probably related to a requirement for low temper-atures and high dissolved oxygen concentrations. Incontrast, H. contubernalis was confined to the middleand lower sections of the main river where thetemperature was higher and less variable.

The fifth species in the catchment, H. angustipennis,was restricted to 2 sites on the Leuchar/Cultersub-catchment and one main river site below itsconfluence with the Dee, presumably responding tothe rich discharge from the eutrophic Loch of Skene.

80

H. angustipennis is known to occur below lakes andponds and to be tolerant of low oxygen concentrationsand high temperatures. One could speculate that H.angustipennis was restricted, originally, to a shortstretch below the loch outlet and, as eutrophicationproceeded, it was able to extend its range down-stream. In 1983-84, it was found regularly in the mainriver at Milltimber, probably because of larvae driftingdownstream from the Culter Burn.

5.5.2 Variation within sites

Mayfly larvae of the family Heptageniidae are commoninhabitants of oligotrophic rivers and more than onespecies often occur together in the same area. Theyare similar morphologically both in general body shape(flattened to avoid the current) and, for most species,in mouthparts (suggesting a similar diet). This similarityleads inevitably to the question of whether theycompete for resources or have some means ofavoiding such competition. The 4 species whichinhabit the Dee at Durris have different preferences fordepth, stone size and current speed (Mac Gowan1978). Their separation is not complete but is substan-tial, and suggests that to a large extent they avoidcompetition. The fact that some overlap remains is nottoo surprising, especially in the light of such studies asReynoldson's (1975), and, in any case, the rather crudeassessment of habitat variables and the restriction ofthe data to one season may have masked their exactpreferences, and so under-estimated their separation.One species, Heptagenia sulphurea, appeared toinclude more detritus in its diet than the others whichconcentrated on algae, but a more substantial studywould be needed to confirm this hypothesis.

These varied micro-habitat preferences were estab-lished in an area which would generally be regarded asuniform, compared with the normal range of rivervariation, and emphasize the fact that generallyaccepted characterization of lotic habitat types israther crude. Our general conclusion in 1983-84 thatthe Dee was largely homogeneous must be temperedby this observation.

5.6 Threats from human activities

Aquatic and terrestrial vegetation is important in theecology of temperate river systems (Dawson 1978;Peterson & Cummins 1974). Water weeds increasethe diversity of a stream, providing additional habitatsand food, either directly or, more usually, as asubstrate for epiphytic algae. The Dee, however, haslittle aquatic vegetation other than mosses, whichmakes the role of the bankside trees all the moreimportant for the conservation and management ofthis important and delicately balanced river system;their removal from river banks may cause changes inthe distribution of aquatic invertebrates. Banksidetrees provide shade, probably reducing the amount ofepilithic algae, but also reduce the temperature and sofavour upland invertebrates. Broad leaved trees alsoprovide much allochthonous material in the form of

leaves and terrestrial invertebrates dropping into theriver, and so shift the community balance from thealgal scrapers and grazers (especially mayflies) to theshredders and detritivores (such as certain of thecaddis). Fish also benefit directly from the extra insectfood dropping from the trees.

A good mix of bankside vegetation helps to maintain adiverse and balanced clean-water invertebrate com-munity, better able to cope with mild pollution.

Perhaps the most obvious symptom of man's influ-ence on the Dee catchment in 1983-84 was the annualalgal bloom on the Loch of Skene. This was probablythe result of nutrient enrichment from agricultural anddomestic discharges, and it is important that a similarfate does not befall the lochs on the Muir of DinnetNNR or in other parts of the catchment.

Other types of human activity pose a variety of threatsto the invertebrates of different parts of the catchmentand even the high-altitude animals could suffer fromdevelopments such as the ski complex on the Clunieand potential problems from large-scale camping onareas with no sewerage.

In 1983-84, the upland areas were largely managed asopen moorland for deer or grouse but a change in useof plantable land to commercial forestry might causechanges in the flow regimes of local streams. Runofffrom commercial forests would probably be less thanfrom moorland or more open native pinewoods (Calder1985) and this water might be more acidic. Anyincreased acidity would reduce the numbers anddiversity of the local invertebrate communities withthe diminution of many species of stonefly, mayfly andcaddis. Subsequent felling of the trees could thenincrease runoff, with the possibility of soil erosion andincreased sediment load in the receiving streams.

Another upland problem (but extending to other partsof the catchment) is the use of pesticides for sheepdipping (Bowen et al. 1978; Doughty 1984). Manysheep dips are located in remote areas with unsatis-factory methods of disposal, often to poorly con-structed soakaways. Spillages to watercourses un-doubtedly occur, sometimes resulting in the elimin-ation of invertebrates and with the possibility ofaccumulation of toxic chemicals in the substrate andfood chain.

Lower down the catchment, especially around towns,domestic refuse disposal has recently become aserious problem, resulting in the use of quite unsuit-able sites close to, and in one case (near Aboyne) over,streams. This situation can result in severe contam-ination of the watercourse by a leachate rich in organicsubstances, nutrients and metals such as zinc andiron, producing growths of sewage fungus and de-posits of iron and causing major changes in the fauna.

Plate 1. Source of the River Dee at 1220 m (4000') in the alpine environment of the Braeriach plateau(Photograph J B Hepburn)

Plate 2. The upper stretches of the River Dee, here flowing out of the Cairngorm mountains through the LairigGhru pass, have seldom been studied by biologists (Photograph I Strachan)

Plate 3. In January and February, the river is often reduced to a narrow channel for much of its length.Temperatures as low as —27.2°C (10 January 1982) have been recorded at Braemar (Photograph J B Hepburn)

Plate 4. Panoramic views like this of the valley are often obscured by regeneration and afforestation at theroadside. The autumn tints show birch to be fairly widespread on the valley floor here west of Coilacriech, withlarch/pine mixtures in the middle distance (Photograph N Picozzi)

Plate 5. Lochs Davan and Kinord (the lower loch) form the central part of the Muir of Dinnet NNR. These lochsare one of the main rearing areas for otters in Deeside, and an important roosting area for greylag geese inautumn (Photograph D Jenkins)

Plate 6. Floodplain deposits overlying coarse cobble beds of fluvioglacial origin. Floodplain deposits compriseboth fine sandy alluvium and coarse pebble beds deposited during large flood events (Photograph J K Maizeis)

hil

idge

Hill ofAitnoch'

Acimahanne

.L chindorhedge 1587

Carn ituighChorroch

of SCkur

why

promPillin Cullachie

•1465

entridge

Thus

Aundo

Auchgouriah Tulloch

Castle• 2366

Corn Hill• 243Eilrig

(BEINN MAUIGDBH) berryCATRN TOOL Cairngorrn

CMWAI y-SABHAIL)GAMS NATURE RESERVECdrn

The Devil's .33163303 a'Mhairn

lonadht3651Mar

CSBents1,3795.,

Bhrotoin Seer Mor• 2266

FOREST

BRAERIACH

%UGH RIABHACHW ouNrAINS

BEN mad:4A 37884149'

MeaddBhuachaille

2654The Queen's Forest

. Monotone' Lodge

243$Sta no

GO° h-lolaire

•2704

tie

3 ice 1,stoned,

..- •' 605

_

BeinnWar manage

545. •GRANtOWN-

ON-GREY

Craggan •DulnainBridge

CreedLiath•1173Carnerory •Catiartown

CAM% LGORM 1'tOe4^,

CAIRNGORM..19983

Damna'Chariainn

Stone

Cairn Geldie2039

Dell „WO,

hy Fdfrest

• 2377

'fridge Shaman

Lettach

Bynack3576 •

More'

Corn.7676

bath

1800Corn n'a thine

Auchna gallin

I.

cornBheadhair

2636

2932

6

OF MA

SgarGooithe

• ForestLodge Braesof

Abernethy

Corn', 4 3298°Bhoc*E- • '

4846Meall're Choire

Ahmhhe

Lettoch

ii5s."5330

Ot:0(10.°°\.°

Dorback'-todge

2692Gee/Charn

224

'2

3Q51Beinn . c

Bhreac9 lÇk15

1

Beind• 3424

luthorn‘Mhhr

..!• 2369Creagano'Chaist..

18.63742Baddoch Tornon

t-Suidhe Mhair

CornnoFeannaige • 2395

2338The B;izach

Forest of Glenavon

The Colonel'sBed

,4s

OC"'

Mons ofDalvey

Br'drowofC

3377 CarnT,Carn rin Righ

Mdr tog!"3194

Tulaicheon

7757Cord noFarraidh

Stab ant-Sluichd

362

7641

Ben Gulahin• Dalmunzie

carea 55

Corn noDrachoide

A 2681

41Ia !Welch

Baddocht2908

Sghi'Mar

ScootmoreForest

if

. Bride ofdvie Avon'eti,

f 982

Corneadhanach

CnapCheochan Aitinn

' 7337

2431 •Big

Gorvoun

A V °

• 3e43

2470den

an Dairnh

Tomintoul Milton

28•30Creogan

Dail Bheog

Carn• 1795

Liath

23.32CraigVeann

Corn• 1866

Dolimh

Auchallater

••'• r*-13719'aiTnhwe ell gehvoilw's

GLAS MAOLCeenlochan

Forest3238. =

CreagLeacach

2613Monarnenach

aik

Bellehiglesh

t CSShenval

Knockandhu

Blairnamarro

Cairnis Muldonich

165/Toarnaooulis

ClashIloirB"ves o lerdivet

Chaneltawn

• 1924

Corn. 259

Liath

29,53Cularcloch

• 2954'Finalty

Hill

*DelnadamphLodge

23'10

.2633„,, Brown Cow

Hill

CastleSallochbuie Forest

CockBridge

Balmoral Forest

Cairnv.•,.„1•,4Bonnoch„greod

326&'irn'

MeikleBeatshach Can."/

1867

BEN RINNES2;59

Favilla

Achnastank

CorryhobbieHill2561

Corgarlf

2224Cam LeacSoighdeir

CairnBaddoch

24.42

Glen! iddichLodge

2541.

GlenfiddichForest

Corn ant-Suidhe Blackwater Forest

2401

7029Cairndrallan

e 2S1.5e6ae 4

GienBreHaoglochCorn• 2600' ' Invec;;;;75

Ealasoid GoalCharn2207 .

Colnabaichin

Tornahaish

Gealloig

Hirt2438

Coilecnech

2726.Lair of

28.15. AldararreGlenda!!Glendon ttadge 2868Forest ,-

'''!Braedownie •

2483 Clo

Cairn inks

• 2129Cairn

of Barns

Round• [9,52

Hill

Hill ofThreetStones

/C65

g.40^S Badenyon

Corn.2639Lodge

GlenbuchatMor

21p LodyleoMossHill

1448

CairnMona Gowan

2456

Crighe

Mead'....- C thia

nMPC

ni;al ia

114 A , R 'At': Birkhall

AteetrlgeldieGoad20" err B

Castlec

.xligercaeld,,-/ -.Creed

9t6G9/I

a ,,..

The Coylesuick"56Braemar L------

C•of M

Aucholzie.'-a-

Bridgeof Germ

BlackHillof Mork Easter

.2931Balloch

ostid

TheScalp

Bndgend

Aldunie

Kirkton ofGlenbuchat

Bellab

Mdon

CandacramHo!

Tornashean, ForestCraig 9F" HillockheodBunzeoch

Morven

Peter'sHill

3043•

Mayor

Muir leseres

Milton oftollich

BallaterPannonich

•House ofHill

Glenmuick

• 2289

32827

Mount Keen-.080

Conaihcraig23,65'

LOCHNAGAR Fashedach

Creag non c Spinal ofGleomuickWhite Alaunth• 2736

Gabhar

Ben Taman

Wheen

Invorharroch

Creagon• 2033

Eunan

A

Cairn• 7293

G100 Lee

Mackie• 269'

Cairn

Cabrach

Haugh ofGlass

Succ

Tam/aeon

1366 'The Buck

Coldsgtionne

Glend4HodCulblean

Glascorrie

044hitd Hill

Mount Of:08

Haddoch

Rinmore

Glenkindi

Tau

Boultenstt

2274 limermarlModawee 110

Inchgrundle

7424Cr4s

74.10

Cr

21

G0°

RIMMI

Croft

T H B$Bailiesward

vieCoynachie

arrochLst

indarroch

1495

le ...Zh re

Lumsden

Scar Hill984

119

&Amid

Kenneth on

Rhynie

Co to n

1'747

Caul!

Aboyrie

•1248Mortlich

irsemore

Tullynes, le

Bus

Marywell

•Auchlossan

MidBettie

Easter7ch

Douneside To aveen15.62Cre k s

Craiglich Finitdortfk DruroIe

Fisherford

Rothmaise

a ir nh ill

Kir tono ayne

Le lie Kirkton

InchArdoyne

G A R I „ l'ila H

ea, •Westh II

Oyne

Pittadrie 1-1:44

pidRayne

.k.

Whiteford,

-Ipeofitch'im°

1

Durno

Pitc p e

Diti.0

...D1 ., .:

Ha Io.' Fingask •

B In

o t.,s, .. X, 7, ,...t I e 68:7a4 r3 r.a.. zBarra ill o

'''' 92o :

three KI '

Mel m,3. Ho Auquhorthies, PI ....,

of Bou e

- f

ouna Castle

ri:,

ethm

OLOMELD UM, ,,,_...„.„..-..•-kenpi,tt.m, edde

docb

C irnbrogieMill ofKin o die

it'i"hitirtge5in.HSL

ttoncrook

,--'1 <

Pe mu k

.

Esslamont

.Cul er len

.2.5

GaAr"ffleck , Tillygret Tillyco h e

Minna

Criihond

•

77,9 I N V, gi ,Ui iflh I ETillykerrie999 Stone

- ' Eircl,Straloch m550 o

A o.,lddler

Auqua hies . Home °f • Manar m ate C...t;s11

„Hoo, Burnhervie\,........,..../

Crate —zKinmuckMiddleto As

wma:li

8 oH11

BlairdaffHi f1/4quhythi

ovenstone

k tErik fir .. ai;:thBaalb tchahHo Ca rnp k

\---,../ „.......$.;e-cb‘tins J

•G t,1

Ulrnmligairo

M or of

Clunyv=\&lick Castle...--)Hill .608

rilhead„

afichie

ssymusk

earfl

<1• Castle

rase, Lauchintilly

HILL OF FARE5:45

tannins » IT E411illtowCraft&

rathens

New ea

Toc er Meikle

Wartle

RothinIfOrman

901Barrnekin

HillEch

Midrnar Forest

Cromwn Barthel

f

Cha elSt Katherines -----

Fog R.) rl rd

ol .1 HtilC

---.1

\---------4804 - cn,e , _,, Balgovei Ctibss of Jackston

Jackstown7

C iodate

° Kemnay

AchathiSyknent

9,,„

OldKinnerni eon

Landerb ry

SouthKirkten

KIN ORE

G rlogie

Him

Myrebird

Leylodge

Callyaiu

rient ou

West Cul erleyraigton

Hardgat

usti P rculterMans of

DrumumoakV

OvertonTyrebaggerl

Hill Aberdeen871 Airport/

triLkie 58BA A ry.To

974k n of

urris

Crossroads\ID" ld'Upper

6rAlgejcagn

fiumh°4°. N hedeBorrowfietilDam's Forest

11241 872(in• khornran-man-earn eikle Corewe

Hill

Harkin!:

Hatton of Gotha IRiney

Au hlee

hmor

El Ink

nK arrachie

S ney ood

Hilton

Belhehie

Pottertop:.

Tap a Noth

'848MtltonNoth

rent

mutt

Kildrummy

(titr Milltown

nnahard

UltimoTawie

Tarland

0

2031Presseildye

Dinnet

fan/Ho

al et

et'

Knockespiock/,

1599

CORREEN HILLS

LordCthArur'sairn1699

acn White Hill

0.118.S.N

Denendlf

Thomastown

Crofts ofN7tongarry

'oftShanquhar

T TillymorgoHill of

Crofts of GLons"o OOO

249McMinn

I 53 Kirkto ofClashindarrochHin of CulsaliandForest

FoudlondGartly1425

Cuff' KnockandyHill Panay

eith

rdlair

•vDuncanston

Aulto

Coldwellst362

WhitehaughSineForest

Mulletown •

1)1724 "...r

arnferg 91 1325kl.' Laniahip Parcie •

Forestof Bi....iie„------,_____<N,

1878Craigniahandle Ballochan

PeeT5„H,

Ke

8°17°1? Montgarrie

WE OIF ALFORDCSVAlford

Auchintoul

Aslou..

CorrennieMoor

Kintecher 1621

BenociLhallie

2433 Co 1c9440illCat

• .,Jkludlee.2252 HilLof

Brocks - '2028Dominic - Mount

2554Battock

of• 593.

• 1648 Fingory ,•1518 Cairn \ielieie o'Mount'

Mindenodge

fl 7.85G SturdyHill

s1986Bulg2273

WeseKnock

HunthillLodge.

1759

2224

Hill ofWirren

1321

Auchleven

Gannoch*N

Badencetts,i,.toetin

Ythanwells Newton RothiebrisbanRedhill

tleFonhes

LicklyheadCastle

Whitehouse

ynturk Kirkton

Tillyfourie

RorendlnCairn

William1468

PidichieForest

Pitmunie

CorrennieForest

1

Glen Dye

1125Melancart

Thainsto

Aucruln

I '2%,\„?

,7.93 I698Bennochie •

Benachie Forest13:10

Militone'Balsam

SMilton of .8a khead \Corsindae

Bockhill ofTrusta...ch Brideaentr:yf

Blackhall 'BANC

Blackhall ForestAuc

ScoltyTillygarmond

Strachan

Fasque

Venetia'Brig

Craig of +•1042

Delfro

Mount Shade1662.

GlendyeLodge."'

Hill of 14,10Bridge of141 Edendocher Dye Tipperweir

14.68

Kerloch175.4

1207.

Lochton

I 2.321Mongour

1052Hill of•Trusto

292Fetteresso Forest

1

LeachieHill

Gt

GoyleHal

erscha72

Mains

lenbili526*Drumtochty

Glenfarquhar

ForestLodge Dellavaiffi

Druloacsha .Hile

k‘ FicellOstov ,A, Auchenbl

Glensaugh toV(0.

soGt,(k\-Cai

RENCEM

GarY...°A

i/L--"'"WoodheadB 9005

Fyvie Crofts o. Hadd thlick498 Orumwhindle

UnionCotthe

Mergie Or 951

•kirk ownklifFettiffesso

• STONEHAVEN

.Bun Fm

F dercairn

%Mt) odti

le Fidd57

ulloch

Carrnont

Buckiburri-ItfOr.

NortlifMld : talt oi

0,,--tfjP4,144‘bC Mastri k-

. i.)filtgE12 00 t' `j C5>

an lrstrrld >

Witer Blacktop

6

78 Biala d - Cults_- .7

ii,iitJtine of a,Er Soul

C \C\ s • ayCove

chd-r1\_I

Findontle•ff';,:indonNess

((Tether Village

nef

cJtanrot

Todhead Point

Garton Point

Downies

ewthhill

Doonie Point

yUoldwe

Bno6hifii

LLON,

ABElen

Hare Nes

INVERBERVIE Reproduced frorn the 1985 Ordnance Survey1250,000 Routemaster reap with the per-Knox

Hill mission of the Controller of Her Majesty'sGordo Stationery Offpce, Crown copyright reserved

;'nnollie Neo,

a

ckd:*k

he of I

, Agol

'1,4. •-

Plate 7. The new raw water Inchgarth Reservoir at Cults also has an important potential for recreation nearAberdeen (Photograph D Glennie)

Plate 8. The harbour at Aberdeen retains its importance as a fishing port. In recent years it has attained majorstatus as a centre for vessels servicing the oil industry (Photograph N Picozzi)

s

• 'al 7fl...

sok ",•-194

»04--s

•

1956

1968

Plate 9. River Dee at Morrison's Bridge, Cults, geometnc changes from 1946-68(Crown copyright reserved)

Plate 70 Constricted bedrock channel nea. Invercanny water works (Photograph Grampian Regional Council)

Plate 71. Large meander in the Maryculter—Murtle reach which has developed since the rnid-19th century Pointbar deposits have been dissected by chute channels (Photograph Grampian Regional Council)

Domestic sewage, on the other hand, causes fewproblems in the main river because of high dilution.Below Banchory (at Silverbank, site 7), however, therewas some evidence of a change in water qualitycaused by the sewage discharge, possibly reflected bythe biotic score (Figure 8). Aberdeen does not dis-charge its sewage to the Dee but disposes of it to theadjacent sea. The expansion of the city has seen themovement of some industry to rural areas, resulting inthe contamination of small watercourses with a widevariety of substances, because of inadequate drainageor treatment provisions.

In order to meet the increased demands for domesticand industrial water supplies, the water abstractioncapacity has recently been increased in the lower partof the river, and during periods of low flow will removea significant proportion of the river's volume. It is notclear how this will affect the invertebrates, but lowerriver levels in the summer may leave sedentaryanimals, like the pearl mussel, stranded.

5.7 Recommendations for conservation

The many threats to the invertebrates, and to the riversystem in general, are all potentially serious, but achange in the trophic status of the catchment bynutrient enrichment from agricultural and forestryfertilizers and domestic or farm wastes is possibly themost important and immediate. The current lownutrient status and the relative lack of disturbance ofthe Dee catchment must be preserved in order toretain its unique series of invertebrate communitiesand to protect the present ecological balance. To doso, the following actions are necessary.

i. Retention and extension of the deciduous treefringe along the river banks, especially in uplandareas. This would particularly benefit the invert-ebrates, but would also provide habitats for birdsand terrestrial insects, shade and feeding for fish,and cover for otters.

ii. Rural commercial developments such as thesmall industrial estates and the ski-ing com-plexes should be subject to stringent planningcontrols and these should be enforced. Unofficialdevelopments must also be controlled. Thequality of water, especially in the headwaters,must be protected.

iii. The effects of water abstraction, especially atlow flows, must be monitored and further rivervolume reduction resisted if this would seem tobe deleterious.

iv. An integrated land use policy should be de-veloped for the catchment to harmonize man'sactivities and to prevent sudden unsympatheticchanges which could be harmful to invertebratesand to other interests, such as fishing, conserv-ation and tourism.

81

6 Summary6.1 Thirty sites, distributed along the River Dee and

its major tributaries, were sampled for invert-ebrates, using a standard 'heel-kick' technique,on 5 occasions in spring, summer and autumn1983 and in spring and summer 1984.

6.2 In total, 125 taxa of invertebrates were identifiedand the resulting site species matrix was thenanalysed.

6.3 Two-way Species Analysis (TWINSPAN) andDetrended Correspondence Analysis (DECOR-ANA) identified groups of similar sites andspecies. Generally, those sites subject to enrich-ment, either natural or associated with man,were separated from the majority of sites, whichwere more oligotrophic, and these latter thensplit into a group containing lowland tributariesand one including the main river areas and themajor, middle and upland tributaries. The analysisidentified anomalies, including one upland tribut-ary site which is known to be stressed artificiallyand one which was associated with a pollutedlowland tributary.

6.4 Some species groups reflected identifiable hab-itat preferences, such as those requiring nutrientenrichment or a gravel substrate, but othersseemed to be generally associated with thestony substrate usual throughout the Dee catch-ment.

6.5 As expected, Plecoptera (stoneflies) andEphemeroptera (mayflies) predominated at allexcept the enriched sites; but it was found, atleast amongst the mayfly family Heptageniidae,that each species has its own characteristicmicro-habitat preferences.

6.6 The analysis showed examples of species with arange of distribution patterns. Some were foundgenerally throughout the catchment, others wererestricted to either upland tributaries, enrichedlowland sites, the main river or other suchsub-sections.

6.7 Diversity and biotic indices indicated a generaluniformity within the catchment, except for afew tributary sites as a result of enrichment orstress.

6.8 It is clear that the Dee catchment is generallyuniform and apparently unmodified; however,certain areas are different and for each of these itis possible to assign a cause. In 2 cases, this isthe presence of a loch upstream of the samplesite, but in most others the effect is due to man'sinfluence, either in the form of agricultural ordomestic effluents or as a result of recreationalactivities.

82

6.9 Rivers as unmodified as the Dee are rare andworthy of conservation. To achieve this, man'sinfluence must be monitored and controlled asnecessary. Invertebrates are ideal as monitorsand the current study provides a baseline forfuture studies on this river and for comparisonwith other, more modified, rivers.

7 AcknowledgementsThe "authors wish to thank Donald Paterson of theDepartment of Plant Science at Aberdeen Universityfor his assistance with data analysis. We are alsograteful to Ian Mac Gowan for the use of his data onhabitat preferences in Ephemeroptera.

The opinions expressed in this paper are those of theauthors and are not necessarily those of the North EastRiver Purification Board.

8 ReferencesArmitage, P. D., Moss, D., Wright, J. F. & Furse, M. T. 1983. Theperformance of a new biological water quality score system basedon macroinvertebrates over a wide range of unpolluted running-water sites. Wat. Res., 17, 333-347.

Biological Monitoring Working Party. 1978. Final report: assess-ment and presentation of the biological quality of rivers in GreatBritain, December 1978. London: Department of the Environment,Water Data Unit.

Bowen, H. M., Cutler, J. R. & Craigie, I. 1978. A survey ofsheep-dip usage in Scotland. Edinburgh: Department of Agricultureand Fisheries for Scotland.

Calder, I. R. 1985. Influence of woodlands on water quantity. In:Woodlands, weather and water. Proceedings of the EnvironmentDivision symposium, Edinburgh, 1984, edited by D. J. L. Harding &J. K. Fawell. London: Institute of Biology.

Calder, I. R. 1985. Influence of woodlands on water quantity. In:Woodlands, weather and water. Proceedings of the EnvironmentDivision symposium, Edinburgh, 1984. London: Institute of Biology.Clare, P. & Edwards, R. W. 1983. The macroinvertebrate fauna ofthe drainage channels of the Gwent Levels, South Wales. Fresh-water Biol., 13, 205-225.

Davidson, M. B. & Young, M. R. 1980. A seasonal invertebratesurvey of the River Dee (Aberdeenshire) and its main tributaries.Banbury: Nature Conservancy Council. (Unpublished.)Dawson, F. H. 1978. Aquatic plant management in semi-naturalstreams: the role of marginal vegetation. J. environ. Manage., 6,213-221.

Doughty, R. 1984. Water pollution resulting from sheep-dipping inScotland, 1979-83. Association of Directors and River Inspectors forScotland (Biology Section). (Unpublished.)

Furse, M. T., Wright, J. F., Armitage, P. D. & Moss, D. 1981. Anappraisal of pond-net samples for biological monitoring of loticmacroinvertebrates. Wat. Res., 15, 679-689.Hellawell, J. M. 1978. Biological surveillance of rivers. Stevenage:Water Research Centre.

Hill, M. 0. 1979a. DECORANA--a FORTRAN program for de-trended correspondence analysis and reciprocal averaging. Ithaca,N.Y.: Section of Ecology and Systematics, Cornell University.Hill, M. 0. 1979b. TVVINSPAN-a FORTRAN program for arrangingmultivaria'te data in an ordered two-way table by classification of theindividuals and attributes. Ithaca: N.Y.: Section of Ecology andSystematics, Cornell University.

Hynes, H. B. N. 1970. The ecology of running waters. Liverpool:Liverpool University Press.

Jenkins, R. A., Wade, K. R. & Pugh, E. 1984. Macroinvert-ebrate-habitat relationships in the River Teifi catchment and thesignificance to conservation. Freshwater Biol., 14, 23-42.Kaushik, N. K. & Hynes, H. B. N. 1971. The fate of the dead leavesthat fall into streams. Arch. Hydrobiol., 68, 465-515.Macan, T. T. 1981. Modifications of populations of aquaticinvertebrates and the quality of the water. In: Dynamique depopulations et qualité de l'eau, edited by H. Hoestlandt, 161-192.Paris: Gauthier-Villars.

Mac Gowan, I. 1978. The microdistribution of four species of mayflynymphs in the River Dee. BSc thesis, Aberdeen University.Owen, R. P. 1983. Causes and effects of nuisance populations ofphytoplankton in the Loch of Skene, Aberdeenshire. Aberdeen:North East River Purification Board.

Peterson, R. C. & Cummins, K. W. 1974. Leaf processing in awoodland stream. Freshwater Biol., 4, 343-368.Ratcliffe, D. A. 1977. A nature conservation review: the selection ofbiological sites of national importance to nature conservation inBritain. Vol. 1. Cambridge: Cambridge University Press.Reynoldson, T. B. 1975. Food overlap of lake dwelling triclads in thefield. J. Anim. Ecol., 44, 245-250.

Scullion, J., Parish, C. A., Morgan, N. & Edwards, R. W. 1982.Comparison of benthic macroinvertebrate and substratum compo-sition in riffles and pools in the impounded River Elan and theunregulated River Wye, mid-Wales. Freshwater Biol., 12, 579-595.Shannon, C. & Weaver, W. 1963. The mathematical theory ofcommunications. Urbana: University of Illinois Press.Wright, J. F., Moss, D., Armitage, P. D. & Furse, M. T. 1984. Apreliminary classification of running-water sites in Great Britainbased on macroinvertebrate species and the prediction of commun-ity type using environmental data. Freshwater Biol., 14, 221-256.

Vertebrates, except salmon and trout, associated with theRiver Dee

D JENKINS* & M V BELLt*Institute of Terrestrial Ecology, BanchoiyI-Institute of Marine Biochemistry, Aberdeen

1 IntroductionThis paper gives an exploratory account of the wildvertebrates of the River Dee, its tributaries, lochs andassociated marshes and other wetlands. We do notattempt a comprehensive survey as no such surveyhas been done, and most information is qualitative.More survey work would be useful, and the paperdraws attention to some obvious gaps in knowledge.We also emphasize some of the requirements of thevertebrates, and indicate potential conflicts betweentheir conservation and other interests.

2 FaunaAs mammals and birds, especially, are mobile, it isdifficult to define those species clearly associated withthe Dee and those not. In the case of waders and gulls,some of which nest on river shingles and nearbymarshes, we also mention groups on nearby hills asthese birds are part of the Dee populations. In general,the vertebrates associated with the Dee are notunusual or important. The main exceptions are theotter and the long-eared bat.

2.1 Fish

Small numbers of sea lampreys occur in the Dee.Adults migrating from the sea to spawning grounds inthe lower reaches of the river are obvious in June andJuly. Sea lampreys are unlikely to occur in thetributaries but, although they have never been re-corded, the young must occur in silty areas in thelower river. The river lamprey probably occurs in theDee in small numbers, with a small run each autumnand spring. Adults run into both the Don to the northand the North Esk to the south.

In contrast, the brook lamprey appears to be commonin the Dee, especially in the lower and middle reachesbut also in many tributaries, including the Girnock andDinnet Burns. In the Dinnet, W Irvine collectedapproximately one lamprey every 10 m when electro-fishing in 1967.

As in many other parts of Scotland, rainbow trout havebeen introduced to several lochs within the Deecatchment for sport fishing. No stocks became estab-lished and any fish occurring in the Dee can only bestrays.

Pike are common in several Deeside lochs (eg Davanand Kinord (Treasurer 1983)), and occur in smallnumbers in slow-flowing parts of the lower reaches ofthe main river. They also occur in some smaller

83

tributaries of the Dee. In 1967, W Irvine found aboutone specimen every 120 m during electro-fishing.

Minnows are common, occasionally abundant, in thelower and middle reaches of the Dee and in many of itstributaries. Spawning occurs in May and June, and frycan be seen in large numbers along the edges of poolslater in the year.

Because of its bottom-living habit, the stone loach isbelieved to be less abundant than it actually is. In fact,it is quite common in the lower and middle reaches ofthe main river and in some of its tributaries.

Eels are common in most parts of the Dee systemwhich are accessible from the sea. Electro-fishing inthe Dinnet in 1967 (W Irvine) revealed about one eelevery 5 m. Jenkins and Harper (1980) showed thateels were a major item of the diet of otters in the area.Three-spined sticklebacks are found in some of themore slow-flowing parts of the Dee and its tributariesand in some lochs. Several lochs in the Dee valleycontain populations of perch (see Treasurer 1983). Thisspecies also occurs in some numbers in the slow-flowing parts of the lower reaches of the River Dee.

Young flounders ascend from the sea to the lower andmiddle reaches of the Dee and may spend some timethere before returning to the sea as young adults.However, flounders were uncommon in the Dee nearBanchory in 1925-30 (W J M Menzies).

2.2 Amphibia and reptiles

Frogs and toads are widespread. At high altitudes.,some tadpoles still occur in autumn and are unlikely todevelop further over winter (A Watson), so either theymetamorphose in their second year, or die. Largenumbers of frogs and toads spawn at some ponds andlochs, eg at Glen Tanar and Loch Kinord. At Kinord, atleast 18 common buzzards were counted preying onspawning amphibia in April 1983 (J Parkin).

Slow worms, viviparous lizards and adders are welldistributed along the Dee in the valleys, glens, andlower moors, as well as on river banks (Taylor 1963).Taylor records the smooth newt from Braemar in 1950and Morrone in 1956, but recent records, whenchecked, have turned out to be palmate newts (MYoung). Taylor had no records of the great crestednewt in Deeside. However, it is apparently notuncommon in the Moray basin, and it would beworthwhile to look for it in deep ponds over better

84

soils in Deeside. The palmate newt is common inpools and wet places along the Dee.

2.3 Birds

2.3.1 Waterfowl in mid-DeesideThe largest concentration of waterfowl in mid-Deesideoccurs on Lochs Davan and Kinord at Dinnet. Data areavailable for 2 series of winters, 1955/56-1962/63 and1981/82-1983/84 (Table 1). Apart from mallard, thenumbers of which have remained roughly constant, allother species have increased. Mean monthly counts ofwigeon, for example, in 1981-84 were 82-223 com-pared with 6-162 in 1955-63. Mean March numbers oftufted duck and goldeneye, in addition, are now over60, compared with 15-20 only 21 years ago. Teal,pochard and goosander also now occur regularly insmall numbers. These larger numbers in recent yearsfollowing protection give encouragement to the effortsof wildlife conservation. In some years, whooperswans occur in large numbers for several days orweeks (over 110 in 1971 and 260 in 1972 (J A Forster))and up to 87 were present in autumn 1983. Other duckspecies recorded at the Dinnet lochs include gadwall,pintail, shoveler, shelduck, scaup, long-tailed duck,common scoter, red-breasted merganser and smew.A few species nest each year, including mallard, teal,wigeon, tufted duck, and also coot and moorhen. Fewyoung are reared, however, perhaps because ofpredation by pike. This predation has often been seenbut its effect on the duck populations is unknown.

'Thriving breeding populations' of waterfowl werefound till recently at the Loch of Aboyne (Nethersole-Thompson & Watson 1981), but, unlike the situation atprotected Davan and Kinord, this is no longer the caseat Aboyne, probably because of disturbance by recrea-tion (see below).

Greylag geese started roosting at the Dinnet lochs inthe early to mid-1970s and numbers have increasedrapidly since. Numbers appeared to have stabilized ataround 4000 by 1979, but in the last 2 winters peaks of10 000 and over 11 000 have been recorded. Up to1500 can be seen at Auchlossan throughout thewinter, but in autumn most feed on arable fieldstowards Donside as far as Muir of Fowlis andGlenkindie, if not further. They usually leave when the

Table 1

stubbles are snow-covered or frozen in late November-December and only small numbers occur in spring,favouring areas at Auchlossan/Ballogie, Torphins andthe Howe of Tarland. Up to 1500 sometimes roost onriver shinale at Ballogie in winter when the area isundisturbed.

Small numbers of pink-footed geese occur at theselochs and elsewhere, mainly in spring (eg 76 on 5February 1977), and there is an exceptional record of780 at Tarland in April 1979. Up to 6 barnacle geese, 4Canada geese, 3 snow geese and 12 bean geese havebeen seen at Davan. Bean geese have also beenrecorded at Aboyne, and a white-fronted goose at leastonce at Auchlossan (1982) and once at Birse (1985).

Loch Davan is one of the northernmost sites forbreeding great crested grebes in Britain. They suc-cessfully fledged young here in some years up to 1976when 3 pairs were present, but there has been onlyone pair each summer since 1977. Nesting wasattempted in some of the later years (not 1983 and1984), but no young have fledged. They formerly bredat Auchlossan (per J Parkin) and Aboyne, and wereregular at Kinord up to the early 1970s; however,nesting has not been recorded at Kinord since the1940s (see Nethersole-Thompson & Watson 1981).

The Dinnet lochs are the centre for wigeon onDeeside; about 30 pairs breed between Ballater, GlenTanar and Aboyne up to at least 220 m on the hills.Mute swans breed on both Davan (up to 4 pairs) andKinord (up to 3 pairs, but none in the last 3 years). Theonly other sites with mute swans are the small lochs atBraeroddach and Aboyne. None breeds on the Deecompared with many pairs on lower Donside, probablybecause the Dee is too shallow, fast flowing andnutrient-poor to allow the growth of aquatic plants onwhich they depend.

The Dinnet lochs and the surrounding marshes alsosupport a few pairs of water rails; the only other sitesin Grampian to do so are Loch of Leys and Loch ofStrathbeg. There are several sedge warblers at bothlochs, and occasional grasshopper warblers. Wadersinclude redshank, now scarce in Deeside due todrainage of wet places; dunlin sang at Kinord up to theearly 1970s.

Mean mid-monthly wildfowl counts at Lochs Davan and Kinord for the winters from 1955/56-1962/63 (inbrackets) and 1981/82-1983/84, plus the maximum count for each species within each period. Coots werenot counted in the earlier period

Sept Oct Noy Dec Jan Feb March MaxMallard 301 (153) 309 (305) 318 (420) 425 (638) 130 (279) 317 1178) 66 (126) 680 (18701Teal 6 (0) 25 (0) 37 (0) 9 (0) 0 (0) 6 (0) 38 (0) 110WigeonTufted duck

12153

(6)(2)

21388

(85)(7)

19367

(162)(18)

22339

(141)122)

1016

(103)18)

8274 ((57)61 17169

(38)(20)

600116

(650)(130)Pochard 7 (0) 25 (0) 13 (0) 9 (0) 0 10) 5 (0) 12 (0) 37Goldeneye

Goosander11

(0)(2)

2516

(0)(0)

3011

(0)(21

207

(0)(2)

62

(6)(2)

25 13 rel6718

(1)51(2 7847

(40)(15)Mute swan 17 (4) 22 (5) 24 (4) 23 (3) 13 (1) 12 (2) 15 (2) 27 (16)Coot 116 150 187 217 30 38 68 350

2.3.2 Gulls associated with the Dee

Colonies of black-headed gulls and .common gullsassociated with the Dee are shown in Figure 1. Mostblack-headed gulls are found on the agriculturally richerground of middle and lower Deeside. They nest inmarshes or marshy loch margins and their numbersmay be limited by lack of suitable nesting habitat. Thetotal Deeside population numbers about 3000 pairs, ofwhich by far the largest colonies are at Dinnet(1000-1500 pairs) and at Leys (about 1000 pairs). Mostof the population of 150-300 pairs of common gulls arefound in upper Deeside where they nest on rivershingle and nearby vegetation. There are 3 small hillcolonies lower down the valley at Morven, Kerloch andDurris, and a few scattered pairs elsewhere. In recentyears, there has been a large increase in common gullsand black-headed gulls in the Braemar area, which isassociated with an increase in tourists, as many gullsnow feed in roadside lay-bys during the summer. Inspring, there is a large roost of up to 10 000 black-headed gulls on Loch Kinord. The herring gulls fromthe Aboyne rubbish tip and elsewhere roost on Kinord.

2.3.3 Birds of the hill streams and lochs

Loch Muick is little used by wildfowl. A few goosan-ders may be seen there, and there are commonsandpipers and grey wagtails round the shore. LochCallater has a small black-headed gull colony of about50 pairs and a few common gulls and commonsandpipers. Besides gulls, one can find mallard, teal,goosanders, common sandpipers, dippers and greywagtails up tributary rivers, streams and lochans highinto the hills. Dippers have been recorded at LochEtchachan (900 m) on several occasions. Herons canalso be found fishing in places like Glen Derry andChest of Dee. Scattered pairs of greenshanks breed onthe upper Dee or in association with small uppertributaries. Above Braemar, there were 4-7 pairs in7-12 sites checked in 1969 and 1971-73. On the Dee,greenshanks are on the edge of their range, and not allknown breeding sites were used in every year. In

0 Common gull• Black-headed gull

0

o 0-50 pairs0 51-200 pairs0 201+ pairs

0

0

Figure 1. Gulleries associated with the River Dee in 1984

85

about 40 breeding attempts, only 7 pairs produced oneor 2 flying young and, overall, the average breedingsuccess was about 0.6 young per pair per year (RMoss, in Nethersole-Thompson & Nethersole-Thompson 1979). In mid-Deeside, at least one pairreared young in some years in the period 1978-82.

2.3.4 Birds of the middle and lower Dee

Oystercatchers are confined to the river when theyfirst come to Deeside in February and March, and,atthis time they roost on river shingles (over 2000 birdscounted between Ballater and Peterculter), but mostdo not nest there. Other waders, including curlew,lapwing, redshank and greenshank, use the river as amigration route. In late autumn and winter, greatblack-backed and other gulls, herons and occasionallybuzzards eat dead salmon (there may be as many as5000 dead kelts on the river (W Irvine)), and many gullsfollow the line of the river, flighting to and from theirroosts on the lochs. In autumn and winter, a variety ofbirds use the reed beds and marshes as roosts, themost notable being up to 14 hen harriers (including abird wing-tagged in Orkney). Swallows, sand martins,pied wagtails and starlings also roost there at varioustimes. Marsh harriers have been recorded at theDinnet lochs and at Braeroddach.

Apart from about 5 nests near Braemar, knownheronries are below Ballater. Herons frequently movetheir nesting sites if disturbed, but counts between1971 and 1981 showed 10 colonies with 4-10 (average6.4) nests per colony along the Dee. For some of thesecolonies, there are data for only 1-2 years, but thepattern (Figure 2) shows a relatively large number ofsmall colonies, closely spaced below Durris andprogressively further apart upstream to Braemar. Twocolonies in mid-Deeside had most nests (on average10 each), but there was no obvious trend in averagecolony size in relation to the course of the main river.The birds feed mostly in tributaries and ditches,presumably mainly on eels and small salmonids (MMarquiss).

km

•

10

•

•

•3

•

86

• Heronry

Figure 2. Heronries associated with the River Dee in 1980-84

Goosanders on the river were counted on 25 March1984. Results were: Aberdeen-Banchory (c 29 km),18 drakes, 16 ducks; BanchorytChest of Dee(c 81 km), 20 drakes, 16 ducks, plus one unsexed. Onthe lower stretch, the density was 1.2 goosanderskm-1, compared with 0.46 birds km-1 on the upperstretch. One red-breasted merganser drake was seenbetween Aberdeen and Banchory. Individuals of thisspecies, perhaps more commonly single drakesbeyond Banchory, are seen throughout the length ofthe river, and nests have been recorded.

Table 2. Densities of breeding birds 10 km-1 of the River Dee, Aberdeenshire, in June-July 1984

Table 3. Habitats occupied by birds on 3 tributaries of the River Deein June 1984: Tanar 17.5 km), Gairn (9.5 km) and Muick(8.5 km)

km

10

Special walks were done to count birds on the middleDee between Invercauld and Banchory, and on theTanar, Muick and Gairn, in late June-early July 1984(Tables 2 & 3). Spring 1984 was unusually dry and thelow water level in the river exposed more shingle thanusual. The main feature of the data is the low densityof birds. Even the most numerous, the commonsandpiper, reached a maximum density of only 4.4birds km-1 and averaged 1.8 over nearly 70 km.Common gull was next most numerous, averaging1.1 km-1, and the other species ranged between 0.2

km-1 and 0.9 km-1. The highest densities of commonsandpiper and pied wagtail were at Balmoral-Dalraddie, of common gulls at Polhollick-Ballater, andof grey wagtail at Balmoral-Polhollick. Overall, densit-ies were similar on the short stretches walked on thetributaries, with common sandpiper and grey wagtailmost numerous (0.8 km-1 and 1.0 km-1). More dip-pers are seen on the Dee in winter than in summer,often singing on ice. Presumably they move downfrom the tributaries when these freeze in winter.

Table 4. Adult birds (%) seen in different habitats on 68.8 km of River Dee in June-July 1984, and trends* inpreference or avoidance of these habitats

i. Habitats classified by vegetation on bank

ii. Habitats classified by river characteristics

" Trend towards avoidance

TTrendstowards preference )1 1 1, P <0.001; 1 1 , T P <0.01; 1.1 P <0.05; NS = Not significant

'The units for analysis are individual birds; in the case of common tern and common gulls, there were loose flocks of 9 (tern) and 5,10, 3, 15, 3, 10, 2, 5 (gull), each of which is treated here as 1.

cumulative binomial probabilities

In the walks done in June-July 1984, the river bankhabitats were broadly classified as open (moor orfields), wooded, or half-open (ie one bank open, onebank wooded), with regard to the main habitat featuresat the point where each bird was seen. The propor-tions of these different habitats were measured overthe whole stretch. The character of the river wasfurther classified according to rate of flow as 'exten-sive shingle' or 'open water from bank to bank' or as'rapids with large boulders in the bed of the river'.Some trends for preference or avoidance shown inTable 4 are statistically significant.

Common sandpipers, grey wagtails, dippers, commongulls, pied wagtails and probably common ternspreferred open habitats; and common sandpipers,common terns, mallards, grey wagtails, common gullsand pied wagtails avoided wooded banks. Except forthe dipper and common sandpiper, all species prefer-red stretches where there was open water from bankto bank, and all, except perhaps common sandpipers,avoided extensive shingle. Dippers preferred rapids.Common sandpipers showed no significant trends.Grey wagtails and pied wagtails were similar in theiroccupation of habitats, but grey wagtails showed astronger trend than pied towards preference for openor half-open habitats and towards avoidance of woods.Twice as many grey wagtails were seen as pied, and itis interesting to speculate whether these 2 speciesmay compete for the same riverine habitats, with thegrey wagtail predominating. These preliminary data

87

suggest that a detailed comparison of the 2 wagtailspecies might be worthwhile.

The proportions of the different habitats on thetributaries (Table 3) were not recorded. However, the 2most numerous species, common sandpiper and greywagtail, were found more often in habitats registeredas shingle than elsewhere; this apparent differencefrom Table 4 should be checked in a wetter springwhen there may be less shingle on the tributaries.Dippers were again more numerous on the rapids thanother species. On these tributaries, grey wagtails werenearly 5 times as common as pied wagtails.

No kingfisher was recorded in 1984, but they do occurrarely. Ospreys have been recorded regularly in theDee river system since the mid-1970s and occasionallyearlier, with at least 2 nesting attempts, both unsuc-cessful. Single birds have been seen from Crathes toBallater, most frequently within a few km of Dinnet.The contrast between the Dee and Spey is notable. InSpeyside, there are several successful nests eachyear, but none in Deeside.

Small numbers of goldeneyes and goosanders areregular inland on the Dee in winter. On the Dee inAberdeen, the winter population of wildfowl includes100-150 mallard, up to 25 goldeneyes and 10-20goosanders. Small numbers of mute swans occur,together with occasional tufted ducks, pochards,long-tailed ducks, smew, mergansers and eiders. In

88

hard weather when the lochs freeze over, the unfrozenriver at Aberdeen can be important for waterfowl, andin January 1979 it held up to 200 tufted ducks, 4 scaup,88 pochards and 100 goldeneyes for a few weeks.Goosanders occasionally moult there in summer; it isunusual to find this shy species so close to people.

2.4 Mammals

Besides seals at the river mouth (excluded here), andbats (see below), only 4 species can properly beregarded as riparian, let alone aquatic. We exclude roedeer, which are often seen in reed beds, hares andrabbits. The 4 main species are water shrew, watervole, mink and otter.

2.4.1 Insectivores and rodentsThe status of water shrew and water vole is uncertain.There are 4 records of water shrews for Deeside for1964-69 at the Biological Records Centre, ITE, MonksWood Experimental Station (H R Arnold). In addition, RHewson and T Healing trapped a water shrew in 1976at Blackhall, Banchory; R J Harper saw one at a marshat Loch Davan in 1976 and 1977; and M P Harris sawone near the Canny in 1981. Corbet and Southern's(1977) description of the habitat used by water shrewsis 'by clear, unpolluted streams wherever there iscover', and trapping by the Dee and its tributarieswould probably show that the water shrew is morewidespread than these few records suggest.

Sim (1903) stated that water voles were abundant inmost streams in Deeside. The Biological RecordsCentre has 3 records for the middle Dee and another 2from above Braemar. They are local but widespreadalong streams in the hills, though numbers fluctuate. JOswald reports that water voles occur on the Tanarbeat of Dee wherever there are patches of rushes, butno signs were found on the Dee by Jenkins' team in1974-79 during work on otters. However, they werefound on the Dinnet near its source in Loch Davan, andwere also trapped there in 1977 by R Balharry. H Kolbtrapped water voles in 1972-73 at Moss-side,Strachan. The distribution of water voles may bepatchy, in dense vegetation and, perhaps especially, inplaces with over-hanging cover, small deep pools andareas with good grass on stream banks or ditches,rather than along rivers (R Hewson). Although wide-spread, signs of brown rats on stream banks in1974-79 were mostly found near farms and houses.

2.4.2 CarnivoresThe first records of wild mink breeding in Aberdeen-shire were in the early 1960s (Cuthbert 1973). They arenow widespread in north-east Scotland (Corbet &Southern 1977), occurring throughout the Dee valleyup to at least 300 m, and have therefore become anintegral, naturalized part of the fauna. Mink mainlyfrequent good cover on stream and loch banks, butthey also use stone dykes and may be seen one km ormore away from water.

On the Glen Tanar estate bank of the Dee (36 km), 39were killed in 1971-83. J Oswald considered that therewere no more than 1-2 families of mink on this stretch.Elsewhere, they have not been systematically trapped,presumably because numbers are low. Few signs ofmink were found during walks of the river and itstributaries in 1976-78 during research on otters (Jen-kins 1980). Indeed, places with mink signs were oftenseveral km apart. In 1977, no trace of mink was foundon 9.5 km of the Gairn in January-June or on the upperGairn in May-July, or on 7.5 km of the Tanar inMarch-November; only scattered signs were found onoccasional visits fothe Feugh, but they occurred up toat least 8 km from the Dee. However, on 5-6 km ofriver banks at Park, 54 mink were trapped in 1981-84,and 60 at the mouth of the Dinnet in 1972-75.

In April-November 1975, 23 mink were live-trapped,marked and released in the 8 km banks of the Dinnetlochs. Seven marked mink were recaptured (one 5times) 1-31 days later, near where they were firstcaught. Adult males were caught in 8 different places,and there were at least 6 family parties. Theseobservations suggest that, at least in 1975, mink mayhave been more numerous on these lochs than onmuch of the nearby main river and other tributaries.Few have been killed on the National Nature Reserve(NNR) at Muir of Dinnet subsequently (14 in 1982-84),not because they are scarcer than in 1975 so much asbecause less manpower is available for trapping them.

In 1969-70, mink scats from the Sheeoch, east ofBanchory (Figure 1), contained mainly salmonid fishand eels, with no remains of game birds (Cuthbert1979). In 1975-78, the main items in mink scats frommid-Deeside were mammalian, mainly rabbits,throughout the year (Jenkins & Harper 1980). Inneither of these studies was there any suggestion thatmink predation was important to human interests, andtrapping them on an NNR may cause more disturbancethan the animals' predation warrants. The main con-clusion is that mink are probably widespread butrelatively scarce on most parts of the middle Dee andits tributaries and, though a potential nuisance, areunlikely to be a major problem to landowners or natureconservationists.

Jenkins (1980) and his colleagues studied the distribu-tion of otter families in 1974-79 on the Dee aboveAboyne. About 9 otter families were thought to havebeen reared in 1977-78 on 26 km of river plus 124 haof lochs near Dinnet. In 1974-79, 1.4 yobng werereared, on average, per successful family, so thatabout 13 otters may have been reared on this stretchin 1977-78. Otter families were also seen on the Tanarand Feugh, and Jenkins was told of otter families on 4other tributaries. Most signs of otters were foundwhere the bankside vegetation provided good cover(Jenkins & Burrows 1980; Bas et al. 1984); and thenational decline in otters is attributed partly to loss ofsuch cover, partly to increased disturbance due to

increased riverside recreation, including angling andpleasure boating, and partly to the greater use of agro-and other chemicals which later get into river systems(O'Connor et al. 1977). Aboyne-Polhollick may be thebest stretch of the Dee for otters. Signs of them werescarcer further upriver, and the agricultural and popul-ated stretches of the lower river would not beexpected to provide suitable habitat for them.

The Dinnet lochs were important rearing areas forotters, with up to 2-3 families at once, staying for 7-12months. The young then moved to the nearby river. In1984, otters should probably be regarded as the mostimportant wild vertebrates on the Dee, with the Dinnetarea vital for their conservation.

Their main food was eels in the Dinnet lochs, and eelsand salmonids in the rivers (Jenkins & Harper 1980).Most eels and salmonids taken were small (eelsmostly 23-32 cm, salmonids mostly <13 cm). On thetributaries, more spraints in autumn were associatedwith a higher proportion of undigested remains ofsalmonids. Presumably more otters were in thetributaries in autumn than at other seasons, perhapsbecause they followed spawning fish. However, ottersoccurred in the Feugh in January 1982, when onekilled a moorhen. Care should be taken when interpret-ing evidence from numbers of faeces because ottersmay use spraints for marking their ranges, and mayvary this practice seasonally, so a reduction in thenumber of spraints found may not necessarily indicatea change in the number of otters present.

One to 3 otters were found dead in Deeside each yearfrom 1978-84, in addition to the families of youngwhich disappeared (see Jenkins 1980). Eight of 11known deaths were due to road accidents, though atleast one animal was ill beforehand. Most deaths werein January-June. Otters are not now known to bedeliberately killed by man on the Dee.

• Long-eared• Piprstrelle• Daubentons

•

Roosts in same house

Figure 3. Known roosts of bats in Deeside in 1984

1 !)

••••

••

•

0 10

km

89

2.4.3 Bats

The pipistrelle bat, Daubenton's bat and brown long-eared bat are common north of Perth in the summer(Racey & Swift 1982), the noctule and parti-colouredbat are rare vagrants, and Natterer's bat occursoccasionally. The 3 common species are frequentlyfound roosting close to rivers or standing water, wherethey can easily find food. Long-eared bats may beunusually common in the mid-Dee valley because ofthe abundance of Lepidoptera which are gleaned fromriverside trees by hovering bats (Swift & Racey 1983).Daubenton's bats, which hunt over rivers and openwater (Swift & Racey 1983), were found in a singleroost near the Dinnet lochs; they depend on chiron-omids, caddis flies and other insect fauna associatedwith these areas of water.

Most (>95%) reported bat roosts in Deeside are inoccupied dwelling houses within 4 km of the riversystem (Figure 3). Few roosts have been reportedupstream of Ballater, but, of 21 nursery roostsreported downstream in summer 1984, 13 were oflong-eared bats, one of Daubenton's bats, and 7 ofpipistrelle bats. No roosts of long-eared bats havebeen recorded downstream of Banchory. Above Ban-chory, roosts of this species outnumber all others(Figure 3). Numbers of roost occupants fluctuate insummer but no movement of bats between roosts hasbeen recorded, so that the proportion of known roostsmay be low. A single instance of long-eared batsroosting in a tree has been reported in Deeside (BMitchell).

It is conservatively estimated that in summer the Deevalley (downstream of Ballater) contains 1000-10 000adult female pipistrelle bats, 100-1000 adult femalelong-eared bats, and 10-100 adult female Daubenton'sbats. Solitary adult males are encountered only infre-quently and are impossible to census. Without theriver and its associated trees, the density of batswould probably be much lower.

90

The pipistrelle bat is evenly distributed along thevalley, where it forages among riparian trees. Thesebats may travel up to 5 km from the roost each night(Racey & Swift 1985), and eat mainly caddis flies andNematocera (Swift et al. 1985).

Virtually nothing is known of the wintering populationsof these bats, mainly because of their reduced activity.Most Scottish bats may migrate south in winter (Racey& Swift 1982).

3 Conservation of the vertebratesFew if any of the species described here are at risk atthe moment. However, a species which is apparentlysecure now may become vulnerable to potentialthreats (eg from recreation) in the future. The securityof the vertebrates of the river depends partly on theinterest, tolerance and goodwill of the landowners,and partly on active voluntary and statutory conserv-ation.

The continuing increase of wintering and moultingwaterfowl at Muir of Dinnet NNR following protectionis associated with a new freedom from disturbance.This protection has led to increased opportunities forlocal bird-watchers.

Other recreational uses of the Dee system includeangling, canoeing, water ski-ing, walking and picknick-ing on the river bank, and casual and organizedcamping on stream bank and lochside. Wind-surfinghas recently begun on the Loch of Loirston, which isnear the Dee though hardly a part of the river system.It is nonetheless a good example of a potential threat;the regular winter duck population of this loch includesnot only mallards, wigeon and tufted ducks, but also aroost of up to 100 goldeneyes and 20 goosanders, anda flock of up to 41 whooper swans. In 1982/83 and1983/84, these swans were driven off by wind-surfers.

The present wildlife of the Dee has co-existed withangling, agriculture and forestry for a long time.However, trends are changing with the income fromangling now setting off losses from other game, andwith perhaps less money available for the mainten-ance of river banks. This maintenance sometimesinvolves powerful machinery, including not only bull-dozers which quickly convert drains into canals butalso power-saws and a risk of the removal of wholetrees, instead of branches as previously with ahand-saw. On some tributaries especially, river bankfences are decaying and not always being replaced,and this leads to grazing on the tributary bank andfewer trees regenerating. This situation may be offseton the Dee by a decrease in the number of smallhold-ers, resulting in less intensive grazing on the riverbank (J Oswald). Our problem is not only that few ofthese changes have been measured, but that differentconservation interests require different land uses.

For wildlife, the Dee is of prime importance for otterswhich should feature prominently in any plans for

wildlife conservation. Otters require good cover forshelter on river banks, especially near pools withmuddy bottoms where there are eels. These poolsmay be good for angling, requiring some tree clear-ances to make room for casting. Grassy banks may bealso preferred by botanists because short cover mayfavour flowers. A new suggestion is that conifer treesshould not be planted up to the river's edge because ofaccentuating problems of acidity in the river from acidrain, yet one of the best places for seeing.young ottersabove Aboyne is where the trees come right to theriver's edge with the path set well back into the forest.This place is interesting because, despite the trees, itis good for angling, with small areas cleared of trees sothat casting is easy in the best places for catching fish.There is little shingle, so there is no conflict betweenthe needs of otters and riverside birds, and it is a goodexample of the development of forestry, angling andotter conservation to the benefit of all three. Otherplaces could be developed actively to foster severalinterests, perhaps, for example, angling, picnic sites,access for canoeists and drinking places for domesticstock.

Apart from the loss of riverside trees on which thefood of long-eared bats in particular depends, the mostimmediate threats to bats near the Dee valley areincidental destruction of bat roosts during houseimprovements, remedial timber treatments to killwood-boring beetles, and deliberate attempts byhouseholders to remove bats. A more long-term threatto bats would be greater use of agricultural insect-icides on new farm crops, which would reduce theirinsect prey as well as resulting in the accumulation ofsuch insecticides by the bats. Insecticides have beeninferred as important in reducing bat populations inwestern Europe (Stebbings 1982), and it is importantto recognize this potential area of conflict betweenagriculture and wildlife conservation. Because suchinsecticides are seldom used in north-east Scotlandand the local bat populations are apparently healthy,Scottish river valleys such as the Dee with large areasof riverside trees may represent sites of nationalimportance for bats, and Deeside farmers should beaware of this fact.

In addition to the situation at the Loch of Loirston,recreation and wildlife on the Dee are also in conflict atthe Loch of Aboyne. In the case of the Loch ofLoirston, it is to be hoped that wind-surfers could bediverted elsewhere. Water ski-ing at the Loch ofAboyne has led to the loss there of most breedingducks and coots, though mute swans still breed, andan otter was seen there at night in August 1984. AtAboyne, recreation would seem to be a reasonablepriority, and the loss of the waterfowl presumably hasto be tolerated by bird-watchers. On the river, canoe-ing at present mostly occurs near bridges, where otherdisturbance is likely anyway. As canoeists extend theiractivities, however, it is to be hoped that they will notland on islands or in otter havens (see below), and

agreement is desirable on the optimum use of eachsection of the river system.

Despite the apparent lack of serious conflict so far,there is still a need for conscious management forwildlife conservation as a major use of the riversystem. The main threats to all wildlife are likely toarise from loss of habitat, and the dangers of disturb-ance, if cover is removed, should not be under-estimated. Wildlife conservation at the Dinnet lochs isensured by a statutory NNR, managed jointly by theNature Conservancy Council and the landowner. ThisNNR includes both lochs but very little of theirassociated waterways and none of the Dee itself. Onlya small part of the watershed of the lochs and none oftheir outlets are included in the protected area. Marren(1984) describes the importance of the purity of thewater of Loch Kinord, thereby emphasizing the desira-bility of careful control of the management of thewhole watershed.

Other Sites of Special Scientific Interest (SSSIs) in theDee system include parts of streams in Glen Tanar,within the conserved forest system, but with neitherthe source nor much length of the streams formallyprotected. Quithel Wood at Ballogie and Crathie Woodnear Braemar are other SSSIs on the main river bank,together with the drained Loch of Park nearby. TheLoch of Skene is also an SSSI.

However, apart from the montane Cairngorms NNR,which includes the source of the river, and 3800 m ofboth banks at Potarch, mostly treeless and chosen forits interesting flora, no part of the Dee is includedwithin an SSSI for its intrinsic interest as part of theriver system.

The Department of Planning in Grampian RegionalCouncil has gone to much trouble in defining 'Sites ofInterest to Natural Sciences'; these SINS carry apre-emption for natural history use in planning, buthave no statutory weight if there is a conflict. Anotherrecent development is an otter haven of 7.25 km ofone bank of the middle Dee. This is an importantinitiative between the landowner and the ScottishWildlife Trust, embodying a formal commitment tomaintain existing bankside vegetation, to fence offother areas at present without trees, and to plantnative species in order to help preserve or increasehabitat suitable for the Dee's most important wildanimal. A similar agreement on a tributary betweenanother landowner and the Friends of the Earth seemsto be in abeyance, though the landowner has erected anew fence to exclude stock from a stretch of riverbank. These initiatives are much to be encouraged,and it is hoped they will be followed elsewhere.

4 DiscussionThe relatively sparse fauna of the Dee might beregarded as typical of a nutrient-poor, short, fast-flowing, shallow river rising in granite mountains. In its

91

upper reaches, it runs mostly over acid rocks, and,though it sometimes flows slowly with a lesser fall, itis inhospitable because it often freezes in winter. Itflows over richer rocks in the Dinnet stretch and isfurther enriched by nutrient runoff from agriculturalland below Banchory. A poorer fauna above Ballatermight be expected to become progressively richer ormore numerous towards the sea, but this simplepicture is complicated by loss of natural habitats, bydisturbance, and perhaps by the effects of sewagepollution and eutrophication, fortunately slight at pre-sent.

In view of the importance of the Dee as a salmonidfishery and in scientific terms, it is surprising that solittle is known about its fish. Other than salmon andtrout, basic knowledge on the status and distributionof fish in most parts of the system is almost entirelyabsent. It is possible, for instance, that several speciesmay occur in the area but have not been mentionedhere. The arctic charr is known to occur in lochs to thewest, north and south of the Dee catchment, andseveral lochs in the upper Dee valley (eg LochEtchachan, Lochan Uaine) would seem suitable for thisspecies. One of the urgent scientific requirements,therefore, is a thorough survey of fish populations inthe area, as is equally true of several other importantriver systems in Scotland (Maitland 1972). The gap inknowledge is also relevant to some important appliedareas, as well as preventing further scientific under-standing of the river system. It has been shown(Maitland 1977) that several fish species (some ofthem previously absent from Scotland) are graduallymoving northward in distribution (eg dace and ruffe).Some of these may cause problems in salmonidnursery streams in the future, but it will be impossibleto make any assessment of their impact there or todevelop control measures, unless we have basicinformation on the fish communities which are presentnow.

Acid precipitation is a topical and relatively new form ofpollution which appears to be affecting fish populat-ions in some Scottish lochs. Because of the extensiveareas of granite within the Dee catchment, and thefact that much of the rainfall there is fairly acid, it islikely that some sensitive fish (eg arctic charr) havealready been affected. However, because of ourignorance of the status of this and other species, wemay never know if this has been the case.

For other wild vertebrates, most conservation interestfocuses on the otter and on bats, both of which arerare in most agricultural habitats but which still occur innatural or semi-natural habitats on the Dee, and forwhich the main requirements are absence of disturb-ance and the preservation of riverside vegetation. Inthe case of birds, the main conflicts arise from thedestruction of habitats through drainage (eg one of thehen harrier roosts at Dinnet is at risk), or disturbance,including actual killing by gamekeepers (eg shooting

92

goosanders or poisoning birds of prey), and bybird-watchers. There is a risk that publicity for rarebirds can result in over-exposure; this risk must usuallybe faced openly as one of the grounds for wildlifeconservation is for people to enjoy seeing wildlife.

An aim of this conference is to identify areas of actualor potential conflict. Naturalists, anglers and canoeists,for example, should discuss whether there are parts ofthe Dee where bankside vegetation should be pre-served and recreation limited, and other parts wherewildlife is less at risk, where other interests shouldhave priority, and where recreational activities canflourish. In principle, the middle section of the Deemay be the best for the wildlife at present, at least incomparison with places where the Dee flows throughfarmland. However, if riverside woods of birch, willowand alder could be re-established on the upper Dee,the vertebrate fauna there would probably be richerthan it is now, especially on the flatter stretches wherecarr might develop and encourage more waterfowl tostay in summer.

The need for safeguards in management of the riversystem as a whole has apparently not registeredstrongly in the NCC's scientific appraisal of the wildliferesource of Grampian Region, which is surprising inview of the loss of the Lochs of Auchlossan, Leys andPark. Even on the NNR at Dinnet, both eastern andwestern parts of the Muir have recently been drained,and it is surprising that such a small part of the habitatsknown to be used by otters has official protection. TheDee is important nationally for the conservation ofotters. Organized camping on the edge of Loch Kinordin the middle of the NNR also seems incongruous toconservationists. It is a potential cause of disturbanceand should be stopped. Other possible active steps bythe NCC include the consideration of parts of the riveras SSSIs.

The symposium highlights the desirability not only ofan agreed code of conduct between users of the riverand its banks, but also of a re-assessment of aspectsof the conservation of vertebrates of the Dee and theirhabitats, and of the responsibilities of the NatureConservancy Council.

5 SummaryThis paper describes some wild vertebrates which areassociated with the river system of the Dee, includingnot only the main river but also its tributaries, lochs andassociated marshes.

Little is known about the non-salmonid fish of the Dee,and it will be impossible to assess their impact withoutbasic information on the fish communities which arethere now.

Vertebrates sensitive to environmental pressures in-clude otters and bats. The Dee is an important river for

the conservation of otters, while long-eared bats areespecially characteristic of the middle Dee. Birdsnesting on the Dee characteristically occur at lowdensities, reflecting the low productivity and fast rateof flow of the river. The most numerous of thebreeding river birds are common sandpipers and greywagtails, reaching densities of about 2-4 km-1 butmostly less. Non-breeding waterfowl are much moreabundant, with lochs and river shingles important asroosts for geese and oystercatchers. Most duckspecies wintering at the Dinnet lochs are increasing.

The main features identified as affecting vertebrates inthe Dee include: (i) the low productivity and fast flowof the upper river (poor for wildlife); (ii) bare banks,especially on the upper stretches of both main riverand tributaries (often in association with heavy grazing)and in the richer agricultural stretches (poor forwildlife); (iii) conifer woods above Aboyne and de-ciduous woods and fringes between Banchory andBallater, providing shelter on the river bank, with thebirches and alders probably good for Lepidoptera asfood for bats; (iv) the pure Loch Kinord and ratherricher Loch Davan, and their important outlets andassociated marshland. The middle parts of the riversystem are richest for vertebrates at present. Thissituation could change with management aimedspecifically to improve habitats for vertebrates. Thereis also a need to integrate wildlife conservation withpressures from recreation (water ski-ing, canoeists,wind-surfers, campers, picknickers) on lochs andelsewhere, and to integrate the requirements offishermen and of farmers with the scientific aims ofwildlife conservation and with satisfying the pleasurethat people get from watching wild animals in semi-natural habitats.

There are at present few conflicts on the Dee, and thismay make it especially easy to have discussions on acode of conduct for river users, including farmers,foresters, anglers, and naturalists, as well as thoseseeking recreation. Formal conservation of the river bythe Nature Conservancy Council is minimal, and wesuggest that a scientific appraisal of conservationneeds for the whole river system would be timely inorder to identify sensitive areas and avoid conflict inthe future.

6 AcknowledgementsWe thank many friends for help, especially PeterMaitland, John Speakman and Paul Racey for thesections on fish and bats, Dan McCowan and Rose-mary Harper for work on mink and otters, Andrea vander Berg, Anneke Stolte and Addy de Jongh whocounted the birds in June-July 1984, and others,including those named in the text, for the use of theirnotes. We are grateful to the Deeside landowners andto the Nature Conservancy Council for permission tovisit their lands, and to several colleagues, particularlyAdam Watson and Don French, for comments on themanuscript and help with statistics.

7 ReferencesBas, N., Jenkins, D. & Rothery, P. 1984. Ecology of otters innorthern Scotland. V. The distribution of otter (Lutra lutra) faeces inrelation to bankside vegetation on the River Dee in summer 1981. J.appl. Ecol., 21, 507-513.

Gorbet, G. B. & Southern, H. N. 1977. Handbook of Britishmammals. Oxford: Blackwell Scientific.

Cuthbert, J. H. 1973. The origin and distribution of feral mink inScotland. Mammal Rev., 3, 97-103.

Cuthbert, J. H. 1979. Food studies of feral mink (Mustela vison) inScotland. Fish Manage., 10, 17-25.

Jenkins, D. 1980. Ecology of otters in northern Scotland. I. Otter(Lutra lutra) breeding and dispersion in mid-Deeside. Aberdeenshirein 1974-79. J. Anim. Ecol., 49, 713-735.

Jenkins, D. & Burrows, G. 0. 1980. Ecology of otters in northernScotland. II. Analyses of otter (Lutra lutra) and mink (Mustela vison)faeces from Deeside, NE Scotland in 1977-78. J. Anim. Ecol., 49,737-754.

Jenkins, D. & Harper, R. J. 1980. Ecology of otters in northernScotland. II. Analyses of otter (Lutra lutra) and mink (Mustela vison)faeces from Deeside, N.E. Scotland, in 1977-78. J. Anim. Ecol., 49,737-754.

Maitland, P. S. 1972. A key to the freshwater fisheries of the BritishIsles, with notes on their distribution and ecology. (Scientificpublications no. 27.) Windermere: Freshwater Biological Associa-tion.

Maitland, P. S. 1977. Freshwater fish in Scotland in the 18th, 19thand 20th centuries. Biol. Conserv., 12, 265-278.

Marren, P. 1984. The natural history of Loch Kinord. Deeside Field,1983, 11-16.

Nethersole-Thompson, D. & Nethersole-Thompson, M. 1979.Greenshanks. Berkhamsted: Poyser.

93

Nethersole-Thompson, D. & Watson, A. 1981. The Cairngorms:their natural history and scenery. Perth: Me !yen.

O'Connor, F. B., Sands, T. S., Barwick, D., Chanin, P., Frazer,J. F. D., Jefferies, D. J., Jenkins, D. & Neal, E. G. 1977. Report ofthe joint NCC/SPNC Otter Group. London: Nature ConservancyCouncil/Society for the Promotion of Nature Conservation.

Racey, P. A. & Stebbings, R. E. 1972. Bats in Britain - a statusreport. Oryx, 113, 19-27.

Racey, P. A. & Swift, S. M. 1982. Bats in Scotland. Scott. Wildl., 18,11-18.

Racey, P. A. & Swift, S. M. 1985. Feeding ecology of Pipistrelluspipistrellus (Chiroptera Vespertilionidae) during pregnancy and lacta-tion. 1. Foraging behaviour. J. Anim. Ecol., 54, 205-215.

Sim, G. 1903. The vertebrate fauna of Dee. Aberdeen: Wyllie.

Snow, D. W., ed. 1971. The status of birds in Britain and Ireland.Oxford: Blackwell Scientific.

Stebbings, R. E. 1982. Distribution and status of bats in Europe.Natural Environment Research Council contract report to theCommission of the European Communities. Abbots Ripton: Instituteof Terrestrial Ecology. (Unpublished.)

Swift, S. M. & Racey, P. A. 1983. Resource partitioning in twospecies of vespertilionid bats (Chiroptera) occupying the same roost.J. Zool., 200, 249-259.

Swift, S. M., Racey, P. A. & Avery, M. I. 1985. Feeding ecology ofPipistrellus pipistrellus (Chiroptera Vespertilionidae) during preg-nancy and lactation. II. Diet. J. Anim. Ecol., 54, 217-225.

Taylor, R. H. R. 1963. The distribution of reptiles and amphibians inEngland and Wales, Scotland, Ireland and the Channel Isles, withnotes on species recently introduced. Edinburgh & Aberdeen:Nature Conservancy Council. (Unpublished.)

Treasurer, J. 1983. Estimates of egg and viable embryo productionin a lacustrine perch, Perca fluviatilis. Environ. Biol. Fishes, 8, 3-16.

94

Agricultural land use beside the River DeeH L BLACKNorth of Scotland College of Agriculture, Aberdeen

1 IntroductionThis paper is presented by an agriculturist who isconcerned with land improvement, crop and animalproduction, and the well-being of the farming com-munity. It deals with the parishes listed in Table 1. Thearea of agricultural holdings and in crops and grass isgiven in Table 2.

Previous papers have described this pleasant land,some of it quite flat, running beside a wide, fast-flowing river. There are many species of tree whichprovide excellent shelter for livestock. Certain areasare particularly impressive and characteristic, like the

Table 1. List of parishes used in statistical survey, 1972-83

AberdeenAboyneBanchory DevenickBanchory TernanBirseCrathie/BraemarDrumoak

Table 2 Statistics for Deeside parishes (areas in ha) (source: derived from DAFS annual censuses, unpublished)

DurrisGlenmuickKincardine O'NeilMaryculterPeterculterStrachan

birch scrub near Dinnet. Many people travelling alongthis beautiful valley would not classify it as anoutstanding agricultural area, but, nevertheless, it stillhas an extremely important and traditional function insupplying livestock to the nearby lowlands. One tendsto think of the Dee as starting at Braemar, yet at thatstage it is already 42 km from its source at the Pools ofDee, nearly 900 m above OD in the Cairngorms.

This paper considers the following 4 questions inrelation to agricultural land use beside the Dee.

i. What is the land used for today?ii. What determines this particular land use?iii. Has time changed this use?iv. What effect has this land use change had upon

the valley?

2 The principal uses of the landThe valley Of the Dee is relatively narrow, at its widestno more than 11 km and often only one field deep.

1545 471

5081

295

1 066

24 391

104 000

19 72925 737

1984

412

59.229 600

4 440

7 88333 000

1 028

2 022224 000

510

There is, however, a surprising number of types ofland use, including private and commissioned forestry,sport, including stalking red and roe deer, shooting redand black grouse, capercaillies, pheasants, partridgesand brown and mountain hares, and fishing for salmon,sea trout and brown trout. The catalogue of countrypursuits includes hill walking, general sight-seeing ofthe area or specific buildings of importance, bird-watching, canoeing, water ski-ing and pearl fishing.When people are attracted to an area by theseactivities, associated infrastructures are required (egplay areas for children in association with chaletdevelopment and caravan sites, golf-courses, craft andamenity centres and, most of all, hotels and touristfacilities).

It is surprising that a river the size of the Dee does notsupport much industry until the last few km fromPeterculter to the sea, where the remains of thepaper-mills and other mills dependent upon water canbe seen. Small isolated sawmills and estate nurseriesoccur occasionally along the valley. Because of thislack of heavy industry, the Dee is a very clean river andgreat attention is paid to the well-kept sewage farmsservicing the villages on the upper and middle reaches.The river is so clean that water is abstracted atInvercannie as the main source of water for Aberdeen.

Agriculturally, the existing wide range of activities isincreasing. Until the last 20 years, Deeside was solelydependent on quality livestock production, mainlysheep and beef cattle which were produced in theuplands for finishing on the better ground further downthe valley (Figure 1).

/

Livestock mart

Finished lambs

Butcher

Cropping

Figure 1. Movement of livestock in marginal farming

For agriculture, land is divided into the following 4categories:

i. hill land;ii. upland;iii. lowland;iv. the best quality suitable for all crops, including

horticulture.

Hill land predominantly supports animals such as hillsheep or red deer, which require a minimum of labour.Upland areas can produce a limited amount of crop,but permanent and improved pasture predominates.This pasture can support hill cows, suckling one beefcalf per annum. On the same land, sheep are alsoimportant and tend to be Blackface, Cheviot oroccasionally 'crossed' animals such as Greyface. Thecrops normally grown in uplands are those required tofeed the livestock, eg oats, barley, turnips, and grassfor conservation either as hay or silage. The weanedcalves or weaned lambs from these areas movethrough local markets to be purchased by low-groundfarmers for finishing on the better land before slaugh-ter to supply townspeople.

Additionally, the low-ground farmers may introduce alow-ground sheep flock lambing in January and Feb-ruary, for the early market in May and June. These aresold before store lambs from the uplands are bought inthe autumn. Store cattle purchased from the uplandsmay either be finished in winter in courts or, alternat-ively, fattened off grass in summer.

Table 2 shows numbers of livestock in Deesideparishes from 1972-84 and the area of land currently

Hill sheep

Suckler cows

Store lambs

95

96

used for traditional crops. There is considerableadditional scope for new crops such as oilseed rapeand . protein peas, as well as for growing barley formalting. The richer grassland on the better land nearcentres of population supports a few dairy herds. TheDee valley has little of the very best land, butoccasional pockets, as far upstream as Dinnet, cansupport horticulture in the form of nurseries or fruit.

3 What determines these land uses?Obviously the agricultural enterprises are not clear-cut.Many factors affect a farmer's particular choice offarming system. Traditionally, much of Deeside hasbeen tenanted with certain restrictions placed uponthe type of farming allowed. It has tended to be anarea of family farms with the son following in hisfather's footsteps, often following his father's systemof farming. Other important factors include altitude,soil type and associated deficiencies, length of winterand depth of frost, proximity to both market and town,the needs of the population for dairying or horticultureand, specifically to Deeside, the lack of moisture in arain-shadow area.

There is a great diversity of soils in Scotland. There are,for example, more than 700 categories on a recentlypublished 1:250 000 soil map of Scotland. The 5 mainfactors affecting soil formation are parent material,topography, climate, time, and soil fauna and flora; andthese factors have produced considerable diversity inDeeside, as elsewhere.

The main characteristics of Deeside soils are that theyare fairly young (10000 years since the last glaciation),are often shallow, have an indurated or hard sub-soil,are frequently stony, and show great variation overshort distances. •

The main types in Deeside are the following:

i. Alluvial — low-lying in river channel areasCountesswells — stony till, from granite andgranite gneissBoyndie/Corby — formed from fluvioglacial sandand gravel

iv. Tarves — till derived from acid and basic igneousrock

v. Peat — originating in wet climates in areas ofpoor drainage; often blanketing the hills andground in wet hollows.

This list is not exclusive and a variety of individual soilsis found up the valley. Within each group, differencesin slope, natural drainage, depth and other factors addto the complexity, giving rise to a multicoloured soilmap (Bibby 1982).

4 Land use capability beside the DeeThe agricultural potential of the land is determined bysoil, climate and topography. Most of the best land(Appendix 1) is in low-lying areas between Aberdeen

and Banchory (classes 31 and 32), but, inland andtherefore upslope, soils deteriorate rapidly. The lastremnant of 32 disappears at Ballater, and even alluvialsoil is mapped as 42 thereafter. Upslope, the land isclassed as 5, 6 or 7, depending upon severity of slopeand climate, which have an over-riding influence. Ingeneral, these soils are not suitable for intensiveagriculture, and mixed farming predominates. Thegood agriculture practised here depends on physicalfactors (cultivation, drainage and levelling), andchemical factors (addition of lime and nutrients forimproved productivity). The main features of landimprovement have included drainage, fertilizer applic-ations, and correction of trace element deficiencies.

5 Has time changed this land use?Although the soil has not changed specifically, farmershave learned to work it better using more powerfultractors and implements able to deal with difficult,stony soil. Mechanization has helped because theseason is short and timeous sowing and harvesting ofcrops are vital. The move away from hand and horsework to the binder, the tractor, and eventually thecombine harvester have all contributed to a majorreduction in farm labour. This reduction in numbers ofpeople employed on the land has been accelerated bya more lucrative living not only further south but in thenearby cities. The greatest single effect on Deesidefarming, however, appears to have been the recentchange in the economic situation which has affectedall agriculture. It has become more profitable toconcentrate on crop production. Many farmers havechanged from producing weaned calves out of a beefcow valued at between £500-£700 to growing barley.The capital required to produce barley is approximately£250 ha-1, whereas suckled calf production wouldrequire £1,000 ha-1. Spring cereal production onlyrequires capital from sowing in March to harvesting inSeptember, whereas for beef production capital isrequired for the full 12-month cycle. The financialreturn on the capital investment is greater in croppingthan with livestock. We have reached a situation inwhich beef is a luxury product, yet it was a staple foodonly a few years ago.

Because of improved varieties, particularly of barley(eg Golden Promise), it has been possible to growcrops which not only mature very early but are alsosuitable for malting, and can therefore be sold readilyin north-east Scotland. Barley is now grown at 300 mabove OD. The recent introduction of hardy winterbarley varieties, such as Gerbel and Igri, has allowedwinter barley to be grown as far up Deeside asBraemar, giving more than double the amount onewould have expected from spring barley only 10 yearsago.

After the UK joined the European Economic Commun-ity in 1972, government policy encouraged the amal-gamation of smaller farms to make larger farmingunits. This involved either 2 or more of the 20-40 ha

farms which were common on Deeside combining toform one 60-80 ha unit, or an existing large unitabsorbing a small neighbouring farm. The incentive toamalgamate was 2-fold in that the in-goer received agrant for the amalgamation and the out-goer receivedeither a capital sum or a pension for giving up farming.Many landlords encouraged amalgamation because itusually involved a reduced upkeep of steadings.

There has also been a trend over recent years forlandlords to take in hand any tenanted farm whichbecomes available. This was because of a taxationincentive and, as importantly, because of the in-creased value of freehold land as opposed to tenantedland. These in-hand estates often range over greatdistances along the river and require considerableexpertise in managing a multidisciplinary unit involvingsporting, agriculture, forestry, tourism, and the remain-ing tenanted holdings.

Subsidy support used to be a major factor influencingfarmers to produce one particular type of livestock. Inrelative terms, prior to 1972, when the UK joined theEEC and came under the Common Agricultural Policy(CAP), the Hill Cow Subsidy and the Calf Subsidy wereconsidered a necessity to give the upland farmer areasonable standard of living while working in difficultconditions and on demanding soil. Nowadays,although support is still given, it is a much lowerproportion of the enterprise output and hardly enoughto persuade a farmer to continue with a livestockproduction enterprise which demands a high level ofcapital and gives a low level of return.

On the other hand, the sheep regime under the CAPhas had a major influence in encouraging increasedproduction of sheep meat. This regime raised thereturns from sheep dramatically compared to 5 yearsago and has encouraged the expansion of many flocksof hill and upland sheep. It has also stimulated arenewed interest from the low-ground farmer in thepurchase of store sheep for final finishing in spring.Indeed, the renewed interest in sheep has alsoencouraged the upland farmer to consider selling moreof his own lambs to local abattoirs than in the past. Asa result, sheep numbers have increased by 15%, whilenumbers of beef cattle have dropped by 9%. While theintroduction of winter barley varieties has not greatlyreduced grassland areas, some marginal land has beenimproved to carry more livestock and to release someof the better grassland for cropping. Dairying hasalmost disappeared from the middle and upperreaches of the Dee.

6 What effects have these changes in land use hadupon the valley?Family-type farming was typical on Deeside but therehas been a decrease in the population involved inagriculture (Table 2), and sons and daughters havetended to move away. Some have been absorbed intothe tourist industry, working, for example, in hotelsand caravan sites. Very few have been taken on to the

97

sporting side of estates which are now tending to loseclients and to decline. Similarly, forestry has beenunable to support many more employees. Therefore,the population of Deeside is falling.

When an area's major industry, such as agriculture,declines, support industries such as blacksmiths,garages, agricultural merchants and machinery dis-tributors also go into decline. As a result, manyfarmers no longer do business at the local towns ofAboyne, Ballater and Banchory, but by telephone orwhen visiting Aberdeen. In Aboyne, although thelivestock mart is not in immediate danger of closing,reduced numbers of livestock are being sold. Live-stock from the west end of the valley tend to be sold inPerth and from the east end in Aberdeen. Theexpansion in residential housing along Deeside, esp-ecially at Banchory, but also Aboyne and even Ballater,although not directly related to agriculture, is usingarable land for houses, many of good quality. Some ofthe people required to service these towns willpresumably come from farms which may be run on apart-time basis on a more extensive type of agri-culture, but at least the total population of the valleyshould no longer continue to decline.

The main changes in agricultural land use beside theRiver Dee in the period 1972-83 are that the number ofholdings has dropped by 20% and there has been aslight rise in owner-occupation. There has been a bigrise in the barley area of 58% which has mainly comefrom the fall in oats of 80%. There has been anincrease in the amount of grass conserved as silage,which is reflected in the fall in the turnip area of 25%.Although the area of grass has increased slightly, thetotal area under crops remains unchanged. The area ofcrops and grass per holding has risen, because thereare fewer holdings, but not many are sufficiently largeto give a reasonable output and therefore a reasonableincome (Table 2).

There has been a decrease in hired labour, especiallyof full-time hired men, from 549 to 371, a drop of 32%.Other categories of labour are fairly constant, indicat-ing that part-time casual staff have been retained.Statistics for total land in production show an increase,due to a small amount of land being improved as roughgrazing but mainly due to farmers re-designatinggrouse moor as rough grazing and to DAFS re-classifying land; 80% of the farmed land is rough.Although there has been a recent fall in numbers ofboth dairy and beef herds, the total number of beefanimals has only fallen slightly (Table 2). The shortfallin suckled calves is being made up by importingartificially reared calves from dairy herds in England.The drop in intensity of Deeside farming is confirmedby the disappearance of pigs and the survival of only afew poultry.

7 DiscussionAlthough the agricultural land runs very close to theriver bank, it is unusual for the livestock to gain access

98

directly to the river. There are exceptions, however,mainly on the upper reaches where cattle can be seenstanding in the river to get away from flies. It is usualfor the river bank to be fenced off, primarily as a safetyfeature, as thaws and heavy precipitation can causethe river to rise very quickly and animals could bewashed away. Second, there is the need to allowfishermen access to the river along this fencedright-of-way.

Because most of the Deeside soils are freely drainedand with a low rainfall, it is perhaps surprising thatirrigation is not more common. It is suggested that thevalue of crops is not high enough to justify the capitalcost of an irrigation plant, but it is also likely that, ifcropping continues, farmers will consider the benefitsof irrigation. An area for discussion, therefore, will bethe ability to draw water from the river.

Agriculture, as practised in the catchment area, isunlikely to have any adverse effects on the biology ofthe river. Modern agriculture utilizes many agro-chemicals which are important to optimize crop andlivestock output. Because the valley is narrow, theamount of runoff and natural drainage transporting anyof these chemicals from agricultural land is small inrelation to the volume of water in the river. TheAdvisory Service is aware of the potential risk ofpollution, but the continuing vigilance of farmersensures that this will not happen. The manner offarming along the river bank poses no threat to any ofthe other interests previously described. Agriculturelives happily with its neighbours, whether they aresporting interests of fishers needing access to theriver banks, or tourists sight-seeing and travelling overthe rough ground.

8 Summary8.1 Traditionally, the main agricultural land use be-

side the Dee has been stock raising, breedinganimals in the uplands and finishing them on thericher ground further down the valley.

8.2 This land use is changing due to economicpressures which favour crop production, insteadof breeding or fattening cattle. However, sheepnumbers are increasing because the CAP hasencouraged more production of sheep meat.

8.3 The main changes in cropping are the substit-ution of oats with barley, an increase in grasslandconserved for silage, and a few alternative crops.New varieties of early-maturing spring barleysuitable for malting, and hardy winter barleywhich can be grown up to 300 m above OD aresuitable for Deeside.

8.4 The amalgamation of agricultural holdings andreduction in the number of people working theland are continuing, but the average size ofholding is still not enough to maintain a reason-able 'financial output'.

8.5 Agriculturists do not envisage that changes inmechanization, tillage, and the use of fertilizerand agrochemicals will affect runoff from theland or pollute the river. Outside agriculture, theperiod of highest pollution is likely to take placeafter a heavy precipitation or snowmelt when theriver is in spate and brings down its own debris.

8.6 As cropping increases, a demand for irrigation fornew crops beside the Dee should be anticipated.This is a potential source of conflict with thosepeople who think that enough water is alreadyabstracted.

9 AcknowledgementsI am grateful to several colleagues in the North ofScotland College of Agriculture for their help inpreparing this paper, especially Mr C I Lumsden,Regional Director, Dr G Dalton, Economics Division, DrE Farr, Chemistry Division, Mr P Leat, EconomicsDivision, and Mrs M Wilson, Visual Aids.

10 ReferenceBibby, J. S., ed. 1982. Land capability classification for agriculture.(Soil Survey of Scotland monograph.) Aberdeen: Macaulay Institutefor Soil Research.

APPENDIX 1. Summary of guidelines extracted fromthe Land capability classification for agriculture (Bibby1982)

Land suited to arable cropping

Class ILand capable of producing a very wide range of crops.

Cropping is highly flexible and includes the mostexacting crops such as winter-harvested vegetables.The level of yield is consistently high. Soils are usuallywell drained, deep with good reserves of moisture.There are no or only very minor physical limitationstaffecting agriculture use.

Class 2Land capable of producing a wide range of crops.

Cropping is flexible and a wide range of crops can begrown. The level of yield is high but less consistentlyobtained than on class 1 land, due to the effects ofminor limitations affecting cultivation, crop growth orharvesting.

Class 3Land capable of producing a moderate range of crops.

Land in this class is capable of producing good yieldsof a narrow range of crops, principally cereals and

'The limitation types. Soil, site and climate are involved in complexinteractions which affect land use. The 5 principal kinds of limitationare recognized as climatic, gradient, soil, wetness and erosion.

grass, and moderate yields of a wider range includingpotatoes, vegetables and oilseed rape. The degree ofvariation in yield will be greater than is the case forclasses 1 and 2. The moderate limitations requirecareful management and include wetness and slope.

Division 1Land in this division is capable of producing con-sistently high yields of a narrow range of crops(principally cereals and grass) and moderate yields of awider range including potatoes and root crops.

Division 2This land is capable of average production, but highyields of grass, barley and oats are often obtained.Grass leys are more common and reflect the in-creasing growth limitations for arable crops.

Class 4Land capable of producing a narrow range of crops.

This land is suitable for enterprises based primarily ongrassland with short arable breaks (eg barley, oats,forage crops). Yields of arable crops are variable due tosoil type, wetness or climatic factors. Although yieldsof grass are often high, difficulties of production orutilization may be encountered. The limitations mayinclude severe wetness, shallow or very stony soils,moderately steep gradients or moderately severeclimate.

Division 1Land in this division is suited to rotations which,although primarily based on long ley grassland, includeforage crops and cereals for stock feed.

Division 2This land is primarily grassland with some limitedpotential for other crops.

Land suited only to improved grassland and roughgrazing

Class 5Land capable of use as improved grassland.

The agricultural use of land in class 5 is restricted tograss production, but such land frequently plays an

important role in the economy of hill land. Mechanizedsurface treatments to improve the grassland are allpossible.

Division 1This land is well suited to reclamation and is used asimproved grassland. The establishment of a grasssward and its maintenance present few problems andpotential yields are high.

Division 2The land is moderately suited to reclamation and isused as an improved grassland. Sward establishmentpresents no difficulties but limitations may causemaintenance problems.

Division 3This land is marginally suited to reclamation and isused as improved grassland. The land in this divisionhas properties which lead to serious difficulties insward establishment and, although establishmentmay be possible, deterioration in quality is often rapid.

Class 6Land capable of use only as rough grazing.

This land has very severe site, soil or wetness limit-ations which generally prevent the use of tractor-operated machinery for improvement.

Division 1The grazing value is determined by the dominant plantcommunities which may contain a high proportion ofpalatable herbage.

Division 2The moderate grazing value is dominated by poorergrass strains.

The lowmunities

Division 3grazing value is dominated by plant corn-with low grazing values, such as heather.

99

Class 7Land of very limited agricultural value. It has extremelysevere limitations that cannot be rectified.

100

Forestry beside the River DeeH G MILLER* and G TULEYt*Department of Forestry, University of Aberdeen'Forestry Commission, Durris, Banchory

Except it be for fish or treeAe yard o' Don's worth twa' o' Dee

1 IntroductionThe 2 well-known lines of doggerel that dismissDeeside as being good for little but trees seem toencapsulate the frustration of generations of farmersin this valley. Although deep, well-draining soils can befound on the frequent river terraces, the granitic parentmaterial of the region contributes few bases, andproductivity must have been low until artificial fert-ilizers became widely available. Beyond the terraces,slopes are steep and soils thin and stony. Here,extensive forests of oak, Scots pine and birch re-mained long after the clearings that had transformedmost of the British landscape. Problems of transportand distance from markets continued to protect thisresource and, paradoxically, the very size of individualtrees may have saved Deeside forests from the ironsmelters who ravaged vast areas on the west coast.

By the mid-18th century, however, wars on thecontinent and problems in the North American col-onies encouraged the emerging industrialists to look tothe forests of the Dee, Tay and Spey as a new sourceof timber for high-value uses, including ship-building.Then started a period of blatant 'timber mining' thatwas to be the forerunner of the better knownexploitation of the forests of Quebec, Ontario, BritishColumbia, the western US states and Tasmania. As inthese later Cases, riverswere the prime mode oftransport. Floating on the Dee continued until thecoincidence of the building of the Potarch bridge andan Act of Parliament which made loggers responsiblefor damage to bridges and river banks. Now, however,there was an infrastructure of roads and extractioncontinued by horse and ox cart. Then came the fellingsof the 2 World Wars, damage during the first beingparticularly severe, and, finally, the great gale of 1953.

Despite this record of destruction, forestry has stir-vived. Some remnants of the ancient forest remain,generally because of problems of access as at GlenTanar, but also, in the particular case of Ballochbuie,because of royal protection. Unlike the situation in therest of Britain, the areas felled largely returned toforest. In part, this was due to a relatively low sheeppopulation, but also it was in no small measure due toan indigenous and growing tradition of active forestmanagement. This tradition was born on the localestates and survived because of their stability andcontinuity of ownership. It is a proud tradition, a pridethat is justified by the resource that we have inherited.

2 Distribution and nature of the resourceAs watersheds seldom coincide with political boun-daries, details of the forest resource for the Deecatchment per se are not readily available. To a certainextent, it can be assumed to represent much of theresource in the Grampian Region. However, becausethis political entity is new, an indication of the patternof development with time can only be obtained fromthe figures for the East Scotland Conservancy of theForestry Commission. This area includes, in addition toGrampian Region, small parts of Central, Tayside andFife. Certainly, in 1980, the proportion of land areaunder forest in Grampian Region was very similar tothat for East Scotland (Table 1) and, indeed, to that ofScotland as a whole, although afforestation in the restof Scotland had only now brought the overall propor-tion up to the relatively stable value in the east. EastScotland, including Deeside, therefore, has a relativelyhigh and long-established forest area that is expandingonly slowly. However, a particular shift since the warhas been that from private to public ownership,although private ownership remains dominant.

Table 1. Area of woodland and pattern of ownership(source: Forestry Commission census of woodlands andtrees 1979-821

*The East Conservancy of the Forestry Commission

Immediately after the War, many of Britain's forestswere in ruins, to the extent that in 1947 less than halfof Scotland's forest area could be classified as highforest (Table 2). Over the following years, there hasbeen a great increase in the proportion of coniferousforest, an increase that has occurred without adecrease in the proportion of broadleaved high forest.Indeed, it is perhaps somewhat surprising that, even insuch an intensely coniferous area as eastern Scotland,broadleaved species cover about a tenth of the forest

Table 2. Distribution of forest types by regions

*'Otherincludes felled or regenerating forest and scrub

area and represent about one sixth of the standingvolume in high forests (Table 3). Increasing interest inthis broadleaved resource, primarily from an amenitypoint of view but also as a means of ensuring a widthand diversity at the base of the forest industry, bodeswell for its maintenance and continued management.

Table 3. Allocation of high forest among species groups by regions in 1980

Among the coniferous species, the spruces, and inparticular Sitka spruce, dominate the forests of Scot-land (Table 3). Sitka spruce is an astonishingly success-ful species in our maritime climate, being able to growrapidly on a wide range of soils even under veryadverse conditions. As it is also relatively easy toestablish, is amenable to management dictates, andproduces readily saleable material, it is little wonderthat it has come to dominate so much of our forestlandscape. Deeside, however, is quintessentially theland of the Scots pine and, despite the attraction of therival spruces, has remained so. In part, this reflectsgrant inducements to plant native pine close to theorigin of the seed. It might be argued alternatively thatthe available inducement is somewhat parsimonious,for it excludes large areas of Deeside. Here, the

Table 4. Standing volumes in 1980 by Regions

Over-bark (000 rn3)

amenity arguments for planting Scots pine, and theeconomic arguments against, are as great as those forplanting broadleaved species in other more favouredareas. The credit, therefore, must largely go to thesensitivity of the many forest owners in the region.

By the start of this decade, the Grampian Regioncontained nearly one fifth of the standing volume oftimber in Scotland (Table 4). In the Region, as inScotland as a whole, this is a rapidly increasingresource to the extent that the supply of timber fromour own forests will double before the end of thecentury. As round timber is an expensive commodityto transport, sawmilling must be carried out relativelyclose to the forests. Already in Deeside there are anumber of well-established sawmills, and almostwithout exception the managements of these arelaying plans for expansion in response to the projectedincrease in supply. This development will have asignificant impact on both the economy and employ-ment opportunities in the locality. The prospects areexciting and a great vindication of the planning andforesight of the foresters and landowners of the firsthalf of the century.

Some idea of the distribution of the forest resourceswithin Deeside can be obtained from the figures forafforestation and replanting (Tables 5 & 6). Excludingan unusually large single afforestation programme inthe middle reaches during very recent years, it wouldseem that, in absolute terms, forest activity is fairly

Table 5. Rate of new afforestation in the Dee catchment by allownership categories (values in parenthesis are those forprivate woodlands)

101

*Upper - from about Ballater west (including the 5 km east ofBallater)

Middle - area east of upper section as far as BanchoryLower - east of Banchory

102

Table 6. Rate of replanting in the Dee catchment by all ownershipcategories (values in parentheses are those for privatewoodlands)

*For details of partitioning of catchment, see Table 5

evenly spread throughout the valley. Examination ofthe present position in relation to potentially plantableland would suggest the likelihood of a continued slowexpansion in all regions, an expansion that is likely tobe fairly evenly distributed between the public andprivate sectors. It should be emphasized, however,that in this area of relatively stable land use there isunlikely to be any rapid shift favouring the extension offorestry.

3 Forest and riverUnusually large stretches of the Dee and its tributariesare surrounded by coniferous forests, a not unusualsituation in most other northern regions of the worldbut one which appears to invoke some concern in thiscountry, and in this country alone. The fears areseldom well founded or thought out, but neverthelesscannot be casually dismissed.

It is perhaps first necessary to separate the forest perse from the single line of trees along the river bank.The latter, usually broadleaved trees, provide thefisherman with shelter, the river with a source ofreadily decomposable organic matter, the fish with aprotected source of terrestrial insects, and the otterwith banks well stabilized by tree roots. This treeresource should be protected, although it is presentlyageing, and some consideration must be given to itsreplenishment.

The dark coniferous forest, by contrast, seems to havefew advantages. The conifer forest and its managersstand accused of (i) reducing water yield, (ii) allowingsilting of streams, (iii) polluting oligotrophic waterswith phosphatic fertilizers, (iv) accelerated inputS ofnitrate, and (v) increased strearnwater acidification.Although there is some justification for each of theseaccusations, their importance is frequently over-stated. Furthermore, the pollution disadvantages canall be minimized, or even eliminated, by adopting a fewfairly simple management rules.

Forests reduce water yield by increasing the evapor-ation of intercepted water, not, as is sometimesimplied, by increased rates of transpiration. In recentyears, considerable information has become availableto explain this feature. Both transpiration and evapor-

ation require energy to vaporize the water, andsufficient dry air to receive further water vapour.Clearly, if the sun is shining, to provide the energy, andthe wind is blowing, so ensuring that the air around theleaf is continually changed before it becomes satu-rated, then both transpiration and evaporation ofintercepted water will be high from all vegetationtypes. The particular difference between forests andgrassland, however, is that, because forests have amuch rougher upper surface, the wind passing overand through them has a greater turbulence imposedupon it. The resulting eddies and accelerated windsensure both the removal of energy as heat from leavesand the continual replenishment of air within thecanopy with drier air from above. The latter effect hasbeen shown to be most important when the canopy iswet, for it means that there is more rapid evaporationof intercepted rainwater from a forest canopy thanfrom a smooth grass sward, despite the coolerconditions. The same efficient exchange of air shouldmean that transpiration is also higher from forests.However, as stomata close when evaporative stress ishigh, any such effect is prevented, so that transpirationfrom a dry forest canopy tends to be less than fromgrass, reflecting the cooling effect of the air turbul-ence.

The net effect of this enhanced loss from tree foliagewhen wet, and slight reduction when dry, is an overallgreater loss from forests. The extent of this increasedloss depends on the differences in (i) turbulence,(ii) the surface areas of the canopies being compared,and (iii) the proportion of time that the canopies arewet. At one extreme is the difference between coniferand grassland in areas of relatively flat topography.Work by the Institute of Hydrology suggests thatafforestation in these conditions can increase waterloss from 18% to 38%, thus confirming earlierAmerican work at Coweeta. The difference is nar-rowed if any of the 3 factors outlined above arereduced. For example, broadleaved forests, which areleafless in winter, have a lower surface area, heatherand bracken invoke more turbulence than grassland,and the duration of wetness is less on the east coastthan the west coast of the country.

On Deeside, the difference in turbulence is clearlyreduced both by the nature of the hill vegetation andby the fact that the rugged topography itself imposessignificant turbulence on the wind. It becomes diff-icult, therefore, to suggest quite what reduction couldbe expected following afforestation. It certainly wouldnot be greater than 15%, and could be as low as 10%.If a value of 12.5% is assumed, the 4000 additional haof forest planted in the 20-year period 1964-84 wouldhave reduced water yield from the 2100 km2 catch-ment of the Dee by 0.24%. This would seem anunmeasurable change and one that could hardly havehad any economic or ecological significance, partic-ularly in view of the fact that the difference would begreatest during periods of high rainfall.

The silting of streams is much reduced by thepresence of a stable forest, loss of silt being limited tothe periods of ploughing at the start of the rotation andof felling at the end. It is now many years since forestmanagers were first issued with instructions as to howto avoid this problem. There are various componentsbut the essential point is that no drain or plough furrowshould discharge directly into a stream. Rather, itshould be terminated some distance back, thusrequiring the drainage water to spill out across theintervening undrained land, and in so doing deposit itssilt load before reaching the stream. This is a simpleexpedient that has proved to be remarkably effective.

Phosphatic fertilizers have been applied to uplandforests since the 1920s with few, if any, reports ofsubsequent algal blooms in streams or reservoirs.However, a recent very stringent EEC guideline as tothe levels of phosphate in potable waters has invokeda new debate. The phosphate anion cannot betransported any significant distance through mineralsoil before it becomes tenaciously bound by theabundant iron and aluminium in soil. Thus, if drains areterminated well short of streams, as prescribed forremoving the silt load, the drainage through themineral soil of the riparian zone will ensure theeffective filtering out of all phosphate before drainagewater enters the stream.

A somewhat similar process would ensure that drain-age water is purged of high levels of nitrate. In thiscase, however, removal is not by chemical reactionwith the soil, but rather through uptake by vegetationand soil microbes, for these are usually somewhatnitrogen-stressed in upland regions. The fear of suchnitrate pollution relates to the flush of decompositionfollowing the felling of forests, and stems largely froma single American study in which re-vegetation afterfelling was prevented by repeated application of theweedkiller 2,4,5-T. Many subsequent experiments inwhich re-vegetation was allowed to proceed normallyhave failed to suggest any real cause for concern,although some small elevation in nitrate levels is notunusual. This finding, of course, is in contrast to thenitrate pollution that can occur in intensive agriculturalregions, where repeated heavy applications of fertilizernitrate may exceed the retention ability of theecosystem.

The inter-relation of afforestation and streamwateracidification is far from understood. As yet, it appearsto be a uniquely British phenomenon. What is unden-iable is that both in the Trossachs and in mid-Walesstreams flowing through forests have been observedto be consistently more acid than analogous streamsflowing through neighbouring heath or moor. Variousmechanisms have been suggested, but as yet none iswholly convincing. Most explanations involve the inputof pollution-derived acid or sulphate, the latter possiblyreacting with mor humus to provide the mobile strongacid. Whatever the derivation of the acidity, the

103

amounts are not significantly greater than can beexpected to be generated naturally in upland eco-systems, and should be easily neutralized by reactionwith mineral soil. Again, the solution seems to be toensure that drainage water infiltrates through themineral soil of the riparian zone prior to enteringstreams, and drainage schemes should be designed toensure that this infiltration occurs. To reduce thepossibility of sulphuric acid production by the reactionof exchangeable H+ in humus with pollution-derivedsulphate, the vegetation immediately alongsidestreams should be managed to reduce the accumul-ation of acid mor humus. Thus, as a generalization,broadleaved species should be encouraged over con-iferous species in the riparian zone (perhaps 20 meither side of any significant stream). A particularexception is the alder species. The nitric acid produc-tion associated with the nitrogen-fixing ability of thesespecies risks not only the increase of acidity in streamsbut also elevated nitrate levels.

It should be borne in mind that the neutralizationreactions that occur as water drains through soilcontinue between streamwater and the minerals onthe stream bed. Acidification, therefore, is essentially aproblem of small first order streams and the riskdeclines with increasing stream size, to the extent thatthe pH of a river such as the Dee has been shown tovary very little around the circumneutral point.

4 Forest management and the river bankIn areas of poor soils and oligotrophic waters, particul-arly where there is significant input of pollution-derived acidity (acid rain), the forester may have totake certain simple steps to minimize chemical disturb-ance to streamwater. Fortunately, these steps arecovered by 2 general rules, both of which have alreadybeen introduced to routine forest management forother reasons. The drainage guidelines drawn up toeliminate silting of spawning streams, effectively thestopping of all drains well short of the stream bank, willensure that phosphate and acidity are removed fromdrainage water prior to it entering streams. To becompletely effective, measures should be taken toensure that drainage is through mineral soil. Similarly,the encouragement of grass and broadleaved treesalong riversides, originally advocated for reasons ofwildlife conservation, will assist in reducing the loss ofboth acidity and nitrate into streams.

Although research has still to define quite how widethe undrained zone should be under all conditions ofslope and soil, the broad outlines of what is requiredare now clear, and there would seem to be fewproblems in ensuring implementation in futureafforestation schemes. More worrying, perhaps, is themeans of ensuring the continuation of the narrow stripof broadleaved trees that fringe the Dee over manykm, particularly through some agricultural areas.Perhaps the most effective expedient would be tocapitalize on the very real public interest in environ-

104

mental matters and to mobilize volunteer labour forreplanting and subsequent management. This step,however, must be preceded by appropriate discus-sions with the owners and all other parties with validinterests. In preparing planting schemes for suchareas, the use of alder should perhaps be treated withsome caution.

5 SummaryWithin the Dee catchment, there is a long history offorest management and, despite heavy felling over thelast 200 years, extensive forests remain. These are stilllargely dominated by the native Scots pine which hascontinued to be used, despite the financial incentivesin planting spruce. Over recent decades, there hasbeen a steady expansion of forestry in the area andsome further slow expansion can be expected.

Afforestation may bring with it certain adverse effects,including reduced water yield, silting of streams and an

increased loss of phosphate, nitrate and acidity indrainage water. Reduced water yield is the con-sequence of increased evaporation in the turbulentwinds characteristic of forests. However, the in-creased afforestation over the past 20 years wouldonly have reduced the water reaching the Dee by afraction of one per cent. Silting should be controlled inproperly planned forest operations. Similarly, thesuitable planning of drains to utilize the mineral soil ofthe riparian zone for filtering drainage water willprevent loss of phosphate, nitrate and pollution acidityto streams.

There is a need, however, to find a means of ensuringthe replacement and continuation of the narrow stripof broadleaved trees that fringes many parts of theriver.

River engineering and the control of erosion

B B WILLETTSDepartment of Engineering, University of Aberdeen

1 IntroductionThe River Dee is a gravel river. It falls from approx-imately 330 m above OD at Braemar to sea level inperhaps 110 km. Inspection of its bed anywherebetween Braemar and Aberdeen confirms that verylittle material' of sand size or finer is exposed at itssurface. Shingle bars occur frequently, often at a bankof the river, but sometimes forming an island in thestream. However, bulk sampling at the bed revealsmaterial of a wide range of sizes, including sands, andeven silt, as well as the larger gravel and cobbles seenon surface inspection. The river is said to have anarmoured bed. During high flows, finer, more mobile,material is removed from a surface layer which thenserves as a protective crust isolating from directexposure to the flow the finer grains from underlyinglayers, until that crust is disrupted by another suff-iciently large flow.

Because of the phenomenon of armouring, the bed isinactive in all but spate flows, except in localitieswhere bank collapse furnishes a repeated supply ofaccessible fine material. (Bank retreat is not alwaysconfined to periods of high flow.) However, thelarge-scale movements of bed material which producechanges in boundary geometry, and therefore in flowpatterns, occur during flows which are exceeded for asmall proportion of time. During these intervals, a largeamount of material is dislodged and transported. Thisprocess is evidenced by the changes observed instage discharge characteristics at Cairnton gaugingstation (McClean 1934; MacDonald 1975, section5.1.1), and can be detected in comparisons betweenrecords (cartographic or photographic) of the same siteat different times. Plate 9 illustrates changes since1946 at Morrison's Bridge, Cults.

The average gradient of the river taken over three-quarters of its length is approximately 3.5 m km-1.The upper valley having been broadened by glaciermovement and the lower valley's contours softenedby vast accumulations of overlying ice, the river isfree to meander or braid over much of its length. Theriver channel is incised in a valley with floodplains onone or both sides (Figure 1). During the last few

Presentchannel

456..5,00 ‘>06,6?,011,, 0.60 .'6.9;04,GkIroonof,.05,-;0° A.--

Present floodplainand range of earlierchannel positions

O':eoe'r

Figure 1. The river channel in its floodplain (schematiconly and not to scale)

105

thousand years, the river course has probably migratedover at least the width of the present floodplain,where one exists. The tendency to meander dependson channel gradient, grain size, and bank-full dis-charge. It is clear from the behaviour of unrestrainedparts of the river that conditions favour both kinds offeature here and there (where a local gradient obtainsan appropriate value). For the lower valley, with abankfull discharge of perhaps 300 cumecs, a veryrough estimate of the critical gradient for the initiationof meandering would be 1 m km-1 (Leopold &Wolman 1957).

The floodplain has other important functions besidesproviding for natural changes in the channel course. Asits name implies, it is inundated when river flowexceeds what can be conveyed by the channel, with 2beneficial consequences.

i. Flow on the floodplain is slower than in thechannel. Therefore, water spilled on to thefloodplain is delivered later to downstreamreaches of the river than it would be otherwise.The floodplain acts as a temporary reservoirwhich holds back water during peak flows andyields it up when the hydrograph peak is past. Ittherefore diminishes the risk of flooding down-stream.

ii. Although flow velocities on the floodplain arelow, the cross-sectional area of the flow over it isoften very large. Consequently, a substantialproportion of the total flow may be conveyed bythe floodplain rather than by the main channel.The channel velocities are reduced by the flood-plain flow from those which would obtain if thechannel carried the whole flow.

Artificial encroachments on the floodplain may impaireither or both of these functions. Flooding problemsmay then be aggravated downstream of the encroach-ment and/or the river channel may be scoured outlocally, with deposition of sediment downstream, andthus potential gradient changes.

River engineering is done in 3 circumstances:i. when the route or site of engineering work

coincidentally encroaches on the river, eg Mill-timber Bridge, raw-water reservoir at Cults;

ii. when use is deliberately made of the river, eg thewater intake at Cults, riverside wharfage atAberdeen;

iii. when some natural feature of river behaviour issuppressed artificially, eg bank protection toprevent or curb meandering.

Engineering schemes of all 3 kinds are a source ofperturbation of the natural solid transport in the river

106

and, in consequence, of the river geometry. Thechanges produced in the river may be seen as whollygood by all interested parties. It is more usual,however, for some of them to be thought undesirableby some of the people affected. Promoters of suchschemes, therefore, have a neighbourly duty toinvestigate thoroughly the consequences of theirproposals, both those consequences which are objec-tives of the scheme and those which are by-products.Unfortunately, our understanding of river behaviour isincomplete and it is often impossible to predict all theeffects of a proposed intervention. It will be thepurpose of this paper to review very briefly thetechniques available to the river engineer in attemptingto make predictions when considering rivers like theDee. It will also indicate briefly the gaps in what can bedone.

2 River processes and calculation methodsA river is primarily an arterial drain in a system which isgravity-powered. (No engineering which neglects thisfact is competent.) In a particular reach under con-sideration, the flow is set by the hydrology of thecatchment which feeds it and the channel character-istics upstream of it. Water is driven through the reachby its own weight, which has a component parallel tothe channel bed, resisted by traction at the interfacebetween the stream and the bed and banks. Thistraction is built up from the drag exerted between theflow and individual roughness elements, which in thiscase are partially exposed grains of the boundarymaterial (predominantly gravel and cobbles in theDee). It is often expressed in terms of a boundaryshear stress which, of course, varies from point topoint on the boundary.

As the water descends, its potential energy dimin-ishes. It is expended at the boundary in overcomingflow resistance, and within the stream in sustainingthe internal flow structure characteristic of a particularstream geometry. This flow structure is complex,containing both turbulence which is apparently ran-dom, and secondary currents which are more orderlycirculations normal to the flow direction. Both turb-ulence and secondary currents have an importantinfluence on the pattern of stream direction velocity inthe cross-section and, consequently, on the distrib-ution bf boundary shear stress. As this distributiondictates the effect which the river has on its bound-aries, an ideal background to the design of river workswould be a prediction of the flow which includes lateralas well as streamwise components. This predictionrequires the solution of the following set of equations,subject to appropriate boundary conditions:

au av OwContinuity — + — + — = 0 111

ax ay a z

au 0 a a azsMomentum — + — (u2) + — (uv) + — (uw) + g — = 0(2)

at ax ay a z a x

ay a a az,— 2+ —(uy) +— (y) + + g— = 0 (3)at ax ay az ay

Here, t is time, g is the acceleration due to gravity, andu, v, w, x, y, z are defined in Figure 2, z beingmeasured vertically from a horizontal datum plane(notations are defined below). A solution cannot beobtained mathematically but numerical solutions cansometimes be achieved at the expense, naturally, ofsome detail. However, they are time-consuming andexpensive, and it is much more common to usesimpler equations which employ quantities averagedover the cross-section. From their solution can beobtained the average shear stress at the boundary ofeach cross-section. The way in which this shear stressis distributed has to be estimated subsequently fromempirical results for cross-sections of similar shape.

Section 1

z,w

X,U

y,v

aQ a zsContinuity — + B = 0

a x a t

Section 2

Figure 2. Flow in a prismatic channel showing co-ordinate axes for analysis

The equations can conveniently be written thus:

dQ a Q2 az, ProMomentum — + — — ) + gA — + — = 0

at ax A a x

(4)

(5)

Hence, Q is volumetric discharge rate of water, B isthe bank to bank width, A is the cross-sectional area ofthe stream, P is the length of wetted perimeter, To isthe mean boundary shear stress, zs is the height abovedatum of the water surface, and Q is the density ofwater. The derivation of these equations (1 to 5) can befound, for example, in Jansen et al. (1979).

The answers from one of the 2 equation sets (1, 2 and3, or 4 and 5) describe the flow in a channel of fixedshape, ie a non-eroding channel. Yet the magnitude ofthose answers may indicate that the channelboundaries become mobile at the calculated flows. Itis necessary then to take the calculation further,estimating the change in channel shape in response toa suitable period of flow and then re-calculating theflow in the modified channel. The effect of the newflow on the channel produces a further channelmodification, and the flow again changes in responseto it. The calculation must cycle repeatedly betweenthe flow and the sediment movement which producesgeometric changes.

Three steps are needed in each cycle to calculate thechange of geometry. Referring to Figure 2, thedifference between the quantity of sedimenttransported as bed load past section 1 in a given timeinterval and that transported past section 2 is equal tothe accumulation of bed material in the river betweensections 1 and 2 — a continuity equation for sediment.It can be written:

ÖN/ Qs, —CL2 - 9-

SO 8t

D is the size of grains exposed to the flow.

(6)

Os is sediment transport rate, s is the specific gravityof sediment, pb is bulk density of deposited material,and V is the volume of material (using an arbitrarydatum value) between sections 1 and 2.

Values for Os are obtained from a sediment transportformula, of which there are many alternatives. Theyusually have the general form:

05 =f(t0, D) (7)

It remains to determine how the deposition of ac-cumulating material or the removal of eroded materialis distributed in the cross-section. This distribution willproduce the modified shape of which a description isneeded for the next stage of flow calculation. Somefeatures of the distribution may be clear in a particularcase; for example, a length of bank may be revetted sothat erosion is prevented and deposition is unlikely.However, such rules of general application as exist forthe mobile parts of the boundary are tentative atpresent. Consequently, the calculation of shapechange is uncertain, which creates difficulty in in-terpreting the predictions of the whole calculationprocedure.

River engineering cannot stop while the researchneeded to improve calculation methods is completed.Engineers, therefore, have to be particularly careful inselecting 'design conditions' relevant to each schemeand ingenious in applying calculation methods whichdeal safely with those design cases. Nevertheless, adegree of uncertainty usually remains, and for largeschemes it is quite normal to conduct a physical modelstudy to reinforce the design calculations.

3 Application to engineering schemesThe last section outlined the calculations of riverbehaviour which are technically possible. They wereseen to depend upon subordinate calculations of (i) theflow, (ii) the consequent sediment movement, and (iii)the changes in channel shape resulting from sedimentmotion. It was acknowledged that (iii) still presentsunresolved difficulties. Unfortunately, in the particularcase of gravel rivers, the prediction (d) of sedimenttransport is also problematic, and, as we have seen,the Dee is a gravel river. Transport rate formulae havebeen devised for armoured rivers, but they have been

107

insufficiently tested and must be regarded cautiouslyas design tools. Designers, whose clients may outlaylarge sums of money on their advice, rely only oncalculation methods which have a proven record ofsuccess. It is instructive to list methods which fall intothis category, with particular application to gravelrivers.

The depth and flow velocity can be calculated for agiven flow rate in a given cross-section (equations 4and 5). If the expense is warranted and the plangeometry not complex, the distribution of velocity inthe cross-section can be obtained (equations 1, 2, 3),and from it the distribution of shear stress at theboundary. In other cases (which predominate), thedistribution of boundary shear stress can be estimatedby analogy with cross-sections for which soundmeasurements have been made and published. Onthe basis of this knowledge of the flow, all that cansafely be deduced about sediment movement is thesmallest grain which will not be moved at each pointon the flow boundary. For each value of boundaryshear stress, there is a stone size at the threshold ofmotion; larger stones will remain in static equilibriumon the bed (unless made less stable by an unusualshape).

For work on the Dee, this rather limited calculationcapacity is nevertheless very useful. Thankfully, theriver remains close to its natural state, and the objectof most engineering work on it is to prevent undesir-able changes in plan geometry. In most cases, this isdone by stabilizing a bank in its present, or a recent,position without other changes. The appropriatemethod is to protect the bank, possibly after somere-shaping, by placing on it material which is below thethreshold of motion in the design flow condition. Forthis purpose, our limited proven predictive capacity isadequate.

The basis of design of such a revetment in a river likethe Dee is to provide an immobile bank, while allowingthe bed to behave naturally and the channel to conveyits natural (occasional) bed load without impediment.The required stone size can be calculated withconfidence once the expected boundary shear stresshas been determined. The appropriate size will be afunction of the slope of the bank, a steep bankrequiring larger stones than a flatter one. Where theriver course, for example, is curved in plan, the peakboundary shear stress is very much higher than themean (de Vriend & Geldoff 1983). The distribution ofshear stress must be examined for such peaks beforea stone size is selected. The procedures (eg Andersonet al. 1970) for size selection are adaptations of athreshold criterion empirically obtained by Shields(1936).

Having prevented direct attack by the stream by theprovision of immobile surface members, it is neces-sary to prevent loss of fine material from behind them

108

by interstitial flows and by groundwater seepage intothe river course. This prevention can be done bycreating a filter of granular material having increasinglyfiner size with distance from the river towards thenatural bank, or by using a porous fabric under theprimary armour.

It is important to recognize that any artificial change inthe boundary (bed or bank) geometry produceschanged flow distribution at any given flow rate. Bankprotection on an existing line introduces a smallchange which can often be ignored. However, anystructure which projects into the stream necessarilyproduces radical flow changes, and may so redistrib-ute boundary shear stress that an erosion or depos-ition pattern is created which results in a changedstream geometry. In this context, a feature whichprevents flow over a floodplain is a projection into theflow when that flow is at its most aggressive. Presentknowledge does not permit such geometric changesto be calculated reliably for gravel rivers.

4 Engineering on the DeeThe purpose of engineering is to execute decisionsabout the use and management of the river and itsfloodplains. The decisions usually originate from aneed perceived in relation to one function only, andthere seems to be no effective unifying managementagency to reconcile actively the interests of differentfunctions.

The roles expected of the river are to drain the valley,to convey a small amount of domestic and tradeeffluent, to provide water for the public supply, toprovide an important general amenity, and to act as abreeding haven for salmon. The fishing provided by thesalmon is valuable. Its requirements are often inter-preted to be that both the natural discharge pattern ofthe river and the natural river geometry remainundisturbed.

The matters in which engineering is likely to be auseful instrument of management can be dealt withunder the 2 sub-heads: flood control, and constraint ofthe main channel.

4.1 Control of flooding

Flood damage is best limited by restricting the use ofthe floodplain to activities which suffer little fromoccasional inundation, such as grazing and forestry.Arable farming and building development are keptabove the greatest expected flood level of the river. Inthe Dee and Feugh valleys, buildings have beenprudently sited but the temptation to plough land in thefloodplain has not always been resisted.

During the last century, substantial flood embankmentschemes were constructed to exclude flood waterfrom areas of land, so as to fit them for crops otherthan grass in permanent ley. Provided that theembankment height is correctly chosen and that

maintenance is not neglected, this measure provideseffective local protection. However, it does removefloodplain storage and flow capacity. The 19th centuryschemes will, therefore, have aggravated conditionselsewhere on the river downstream of each embank-ment, and modified the local geometry of the riverchannel. Were the device to be generally adopted,floods in the river would become flashier, rising moresteeply and to greater flood levels. It is clear that someexisting schemes have deteriorated for lack of maint-enance. Embankments constructed with horse cartsand hand labour have not been regularly repaired in anera of mechanized farm and construction work! Inconsequence, there are occasional losses of soil,fertilizer and seed when land which is no longerprotected is used for arable farming.

One modern method of flood control is to regulate ariver by damming its headwater and/or those of itsmajor tributaries. The pattern of releases from theresulting reservoirs is then planned so as to avoidundesirable extremes of flow, both floods and droughtflows. This method is high in capital cost and appearsto have been rejected for the Dee. It is not clear whatconstraints, if any, would have to be placed on theoperational procedures in order to avert potentialdamage to the salmon fishings.

4.2 Constraint of the river channel

As noted earlier, the Dee (and several of its tributaries)has, in some reaches, a tendency to meander. Thechanges in plan position which would result fromallowing this tendency to develop freely are usuallyundesirable. Among the consequences would besome of the following.

i. Disruption of established uses of the floodplain(usually loss of pasture at the retreating bank,accompanied by recovery from the river ofagriculturally useless land at the other bank).

ii. Change in property boundaries.

iii. Increase in the sediment load of the river,changes in bed topography and character, andconsequent changes in fish behaviour.

iv. Loss, or loss of use, of buildings or otherstructures. (Morrison's Bridge at Cults is perhapsthe best-known example.)

Changes, of course, can always be prevented in acomparatively small river like the Dee, provided thatthe engineer has a client with the will and thenecessary financial resources. The client will embarkupon a scheme only if the expected benefits to him (orin the case of a syndicate, to them) outweigh the cost.The circumstances of some clients lead to a prefer-ence for capital schemes of protection with a longdesign life; others prefer a regular expenditure onminor works when the nature of the problem permits.Thus, whereas large sums have been invested toprovide protection from river misbehaviour at Mill-timber Bridge and at the new reservoir at Cults, there

are, in contrast, very many private schemes ofcontinuing bank management wherever the river isactively fished.

Landowners are entitled to implement such schemesin order to prevent the loss of land to the river. Theyfrequently do so without regard to the effect on riverregime of the replacement of an erodible bank by ahard one.

The principal source of risk in engineering constraint ofthe river lies in piecemeal consideration of individualproblems. It is common to under-estimate the zone ofinfluence of any intervention, and to confine theexamination of its consequences to too small a lengthof river. This situation arises primarily from thecircumscription of each client's interest. It is inevitablegiven the present fragmented control of the river.Nevertheless, it leaves the river system vulnerable todamage from one or a series of incompletely con-sidered schemes. There is a strong case to be madefor a management agency with a duty to oversee andco-ordinate proposals to interfere with the river.

5 Research needsThe engineering designer needs 2 categories ofinformation: the first is about the flows to be ex-pected, and the second is about the interactionbetween these flows, his structure, and the naturalriver course.

He must know about the flow pattern (the runoffhydrograph) in order to select an appropriate 'designflow'. Its effect on his structure, and the effect of theirinteraction on the river course then become importantconstraints on the design of the structure. Historicalrecords of streamflow provide the best basis forselecting appropriate design flows. Flow records forthe Dee have been kept for many years (since 1929 atWoodend), and therefore the selection of design flowfor a particular purpose is subject to less uncertaintythan is the case on most rivers. Nevertheless, thereare unresolved difficulties which merit research. Twoof them are as follows.

i. Land use in the catchment may change duringthe life of the installation. This change will, inturn, affect the runoff characteristics and impairthe relevance to design of past records. Work isneeded on the influence of land use on runoff sothat estimates of future runoff can be adjustedappropriately.

ii. The engineering scheme is rarely near the rivergauging site. Conventional calculation methodsfor routing the flood from the gauging station tothe site of the proposed works depend onaccurate knowledge of any lateral flows into orout of the river course between the 2 sites. Forflood flows, this calls for better knowledge offlow from tributaries; for drought flow predic-tions, work needs to be done on the interchange

109

between river channel flow and groundwater inthe gravels of the floodplain.

As will have been detected from earlier sections,some of the difficulties in calculating the effect of theflow on the river channel are still unresolved. Kling-eman and Emmett (1982) have found existing sedi-ment transport formulae unsuitable for use inarmoured gravel bed rivers. The methods of estimat-ing local boundary shear stress values are often tooexpensive in one case, and insufficiently versatile (withregard to cross-section shape) in the other. There is noauthenticated method of calculating change of cross-section shape when bed and bank mobility occur. Onecould therefore compile a long list of research needs.Among the most urgent are the following.

i. A more comprehensive library is required of theboundary shear stress distributions which occurin different cross-section shapes.

Overbank flow and its effect on main channelflow need to be better understood.

iii. There is an urgent need for better understandingof, and simulation capacity for, armoured bedactivity, leading to a reliable sediment transportformula for armoured bed gravel rivers.

iv. Mechanisms of bank retreat have been categor-ized (Thorne 1982), but prediction methods havenot been developed.

v. The principles upon which calculations of shapechange should be based are not established. Thecurrent hypotheses about shape change shouldbe evaluated by studying case histories so that asound prediction method can be evolved.

6 SummaryThe Dee is a gravel river with an armoured bed. Inmany places, it has floodplains which permit meander-ing and also moderate downstream flooding. Beforeundertaking engineering work on the river, it isnecessary to predict what changes will be produced inthe geometry of the river channel and its flows.

For a non-erodible channel of reasonably simplegeometry, calculation methods are available for est-imating depth and velocity for a given flow rate. Atconsiderable additional cost, velocity distribution in thecross-section can also be calculated. When the bound-ary is erodible sediment, transport rate for a particularflow condition (characterized by boundary shearstress) can also be calculated from one of a variety ofempirically based formulae. It is then possible tocompute changes in the cross-sectional area of thechannel. Means of predicting the shape changesimplied by these area changes are unproven. How-ever, subject to this last uncertainty, progressivechanges in channel geometry can be calculated by amethod which steps from calculating flow in a givenchannel to estimating the change wrought by thatflow, and then back to re-calculating the flow in thechanged channel, and so on. .

110

Unfortunately, sediment transport processes are lesswell understood for armoured gravel rivers than forsand- and silt-bounded ones. Therefore, for the pre-sent, the knowledge of sediment behaviour to beobtained from a calculation of flow in a gravel river islimited to an estimate of the size of material whichwould be at the threshold of motion at each position inthe channel. This estimate can be used to designprotective layers to prevent the retreat of particular

'sections of bed or bank. However, channel changesproduced in gravel rivers by engineering installationswhich induce substantial flow changes cannot becalculated with any reliability at present.

Engineering work is undertaken on the Dee for clientswhose interest arises from one or more of the roles ofthe river. These roles include drainage, conveyance ofdomestic and trade effluents, water supply, provisionof salmon breeding sites and fish passage to them,and general amenity. The 2 main types of workundertaken are flood protection, and the restraint ofbahk erosion. There is little technical difficulty inproviding locally effective schemes for either purpose.However, the consequences of such schemes are notnecessarily all local. More remote consequences aredifficult to predict and there is little incentive forscheme promoters to pay for any attempt to make theprediction. There is a case for establishing a super-vising agency with a duty to co-ordinate engineeringon the river in the general interest.

Much needed improvement in prediction methods forgravel rivers calls for research on several particularfeatures of behaviour. Some of the urgent ones are

NotationA

Os

— area of cross-section of flow,— channel width (measured at the water surface),— equivalent grain diameter,— acceleration due to gravity,— length of the wetted perimeter of flow,— volumetric flow rate of water,— transport rate of sediment, by volume,— specific gravity of sediment,— - time,

v — orthogonal components of water velocity (u in the stream direction),

V — volume of bed material in deposit in a specified reach,

1— orthogonal co-ordinates of position (x in the stream direction),

zs — height above datum of the water surface,Q — density of water,Qb — bulk density of sediment deposit,

— boundary shear stress (tractive force per unit area)

listed in the preceding section. Perhaps the mosturgent of all concerns the better understanding oftransport processes at armoured beds.

7 AcknowledgementsI thank the Scottish Development Department forpermission to publish Plate 9.

8 ReferencesAnderson, A. G., Paintal, A. S. & Davonport, J. T. 1970. Tentativedesign procedure for rip-rap lined channels. Project report no. 96, StAnthony Falls Hydraulic Laboratory, Minneapolis. (Report 108.)Washington: Highway Research Board.

Jansen, P. Ph., van Bendegom, L., van den Berg, J., de Vries, M.& Zanen, A., eds. 1979. Principles of river engineering. London:Pitman.

Klingeman, P. C. & Emmett, W. W. 1982. Gravel bedload transportprocesses. In: Gravel-bed rivers, edited by R. D. Hey, J. C. Bathurst& C. R. Thorne, 141-169. Chichester: John Wiley.

Leopold, L. B. & Wolman, M. G. 1957. River channel patterns:braided, meandering and straight. (Professional papers 282-B.)Washington: U.S. Geological Survey.

MacDonald, Sir M. & Partners. 1975. Regional water resourcesstudy for North East of Scotland Water Board. Vol 2. Cambridge: SirM. MacDonald & Partners Ltd.

McClean, W. N. 1934. Flow of the River Dee. Engineering, Lond.,138, 338-339.

Shields, A. 1936. Anwendung der Ahnlichkeitsmechanik und derTurbulenzforschung auf die Geschiebebewegung. Mitt. preuss.VersAnst. Wasserb. Schiffb., no. 26.

Thorne, C. R. 1982. Processes and mechanisms of river bankerosion. In: Gravel-bed rivers, edited by R. D. Hey, R. D. Bathurst &C. R. Thorne, 227-271. Chichester: John Wiley.

de Vriend, H. J. & Geldoff, H. J. 1983. Main flow velocity in shortand sharply curved river beds. (Report no. 83-6.1 Delft: Dept of CivilEngineering, Delft University of Technology.

Water abstraction from the River DeeI D BROWNDirector of Water Services, Grampian Regional Council

1 IntroductionThis paper describes the history of the abstraction ofwater from the Dee to supply the city of Aberdeen andits hinterland. It touches briefly on the means oftreating this water and the methods of distributing it.

Most historical records studied refer to measurementsin imperial units, but metric units are used here (Table1).

Table 1. Metric units used in this paper

1 Mega litre = 1000 cubic metres (1000 m3)4.54 Mega litres per day (Ml d-1) = 1 million gallons per day (mgd)1 metre = 3.28 feet1 kilometre (km) = 0.62 miles1 cubic metre per second

(cumec) = 19.01 mgd

2 The first abstraction from the River Dee for publicwater supply purposesThe town of Aberdeen relied on wells, springs and theloch (located near the present Loch Street) for itswater supply until the early 1800s. The population in1792 was 'about 25000, and the daily consumption perhead was about 5.3 litres. The magistrates of thisperiod were concerned about the sufficiency of thesupply, and it appears that a drought in 1826 sparkedoff positive action. Various schemes to enhance thepublic water supply were examined, including thepossibility of using the Loch of Skene, the Burn of

Cairnton Intake

?

• BALI:ATER).

— Park GaugingStation

Glen Dye Intake Woodend Gauging Station

10 20

km

Figure 1. Catchment of the River Dee — water intakes and gauging stations

Culter, the Aberdeen-lnverurie Canal and the Dee. Thescheme selected was based on abstracting waterfrom the Dee approximately 650 m upstream of theBridge of Dee on the north bank (Figure 1).

A General Improvements Act, which covered police,paving, lighting, cleansing, watching, etc, was passedin 1829. This Act included the necessary clauses to getthe scheme under way. It is interesting to note thatmany new house owners who had recently dug theirown wells objected to the new scheme.

Initially, water infiltrated into a tunnel constructed inthe gravels adjacent to the river. This tunnel led to a3.7 m diameter well which, in turn, was connected bya 600 mm diameter cast iron pipe to a steam pumpinghouse. The pumping house stood at the north-east endof the Bridge of Dee and the pumps were driven by 2Watt and Boulton steam engines. The site is nowoccupied by tenement buildings. From the pumpinghouse, water was conveyed by a 380 mm diametercast iron main along Holburn Street to a reservoirincorporated in the fine building still located at 478-484Union Street. Round about 1850, the 'infiltration'tunnel was extended to the river to allow directabstraction.

The abstraction was limited to a 'half inch abstractionfrom the surface of the water' which appears to have

Bridge of DeeMorrison's Bridge Intake Intake

(disused)

BANCHORY

,ABERDEEN

•sie MichelinIntake

1

111

112

equated to a maximum of 5.7 MI per day (5.7 MI c1-1).In the early years of the scheme (the 1830s), pumpingwas only necessary for about 6 hours each day. By

. 1851, the population had increased to 72 000 (Figure 2)(Anon 1973), and pumping was necessary for 22 hourseach day, approximately 71 litres per head per day.The well house still remains and can be seen on thewest side of the now disused Garthdee Works.

.2

1

240 000

220 000

200 000

180 000

160 000

140 000

120 000

100 000

80 000

60 000

40 000

20 000

1790 1810 1830 1850 1870 1890 1910 1930 1950 1970 1990

Figure 2. Population of the city of Aberdeen 1790-1981

3 The first Cairnton abstraction schemeIn 1855, the Town Council engaged Mr JamesSimpson of London, then President of the Institutionof Civil Engineers, to investigate and report on how thepublic water supply could be enhanced. Mr Simpsonreported in September of the same year and listedseveral alternative schemes, including abstractionfrom the Dee at Potarch, from the Dee at Cairnton(approximately 6 km upstream from Banchory), fromthe Don at Paradise (north of Monymusk), and fromthe Denburn.

It appears that the Council took some time to decidewhich scheme should be adopted, but in 1861 it wasfinally agreed that the new water supply would befrom the Dee at Cairnton. It is recorded that there wasmuch opposition to the Dee schemes, principallybecause of cost.

In 1862, an Act was passed which allowed 27.2 MI c1-1to be abstracted at Cairnton. The new scheme,construction of which started in 1864, consisted of aside intake, a tunnel to the treatment works atInvercannie, the provision of a circular settling tank and2 circular slow sand filters. A brick aqueduct wasconstructed to convey the filtered water to a newreservoir at Mannofield within the city. This schemewas opened by Queen Victoria in 1866.

By 1871, the population had increased to 88 000(Figure 2) and the consumption had risen to 14.2 MIc1-1 (Figure 3). This gave a consumption of 159 Ihead c1-1, more than double that of 1851. In the1870s,

1790 1810 1830 1850 1870 1890 1910 1930 1950 1970 1990

Figure 3. Rates of abstraction of water from the RiverDee and the allowable abstraction rates, 1829-1980

big advancements were made in sanitary arrange-ments, and just after the opening of the Invercannieworks a main drainage system was constructed in thecity.

4 Changes in the location of the intakeBy 1885, the population had increased to about120 000 and the consumption was now approachingthe 27.2 MI c1-1 allowable. An Act was passed in thatyear which allowed the abstraction to be increased to36.3 MI d-1. To enable the extra water to be passed tothe treatment works, it was necessary to re-site theintake to a point 120 m upstream. The Act also allowedfor additional service reservoirs to be constructed inthe city at Mannofield and Slopefield, and a steampumping station at Cults. These works were carriedout under the direction of Mr William Boulton, BurghSurveyor.

Nothing has been said about the quality of the waterso far and it must be assumed that the civic fathersand officials considered the Dee to be a relatively cleanriver. However, the 19th century also saw big improve-ments in scientific and analytical techniques, and in1891 some concern was expressed about the qualityof the water, with particular emphasis on pollution ofthe river above the intake. A Mr Gale, Civil Engineer,was appointed in 1891 to report on these matters, andin 1893 an Act was passed which allowed, amongstother things, the construction of sewage farms atBraemar, Ballater, Aboyne and Kincardine O'Neil.

5 The possibility of abstracting from the River AvonIn the early 1900s, abstraction was exceeding thestatutory 36.3 MI d-1 quite frequently, and during thewinter of 1910 the maximum daily demand hadpeaked at 45 MI c1-1. Clearly, the existing system hadto be expanded or a completely new source found.The Burgh Surveyor at that time, Mr William Dyack,and a local Civil Engineer, Mr George G Jenkins,submitted a report on the possibility of abstractingwater from the Avon, a tributary of the Spey, andconveying it to Aberdeen. In 1909, the Town Council

promoted a Bill for the supply of 90 Ml ic1-1 from theAvon, but this was opposed by the railway companies,land and property associations, 'private citizens andmanufacturers. Phase 1 of the scheme, which wouldhave provided 54 MI d-1, was to cost £1,068,000, andit seems that, once again, cost was the main objection.The scheme never got off the ground and in 1910 aCommittee was appointed to look at the Dee again.

In 1912, a report was prepared by the then City WaterEngineer, Mr Cecil H Roberts, and Mr C P Hogg, CivilEngineer, which set out 2 alternative schemes.

i. An extension of the Cairnton and Invercanniescheme to increase the abstraction by 9.1 MI c1-1to 45.4 MI d-1, incorporating, among otherthings, slow sand filters and an additionalaqueduct to Aberdeen.

ii. A new scheme based on an impounding reser-voir on the Dye at Tillyfumerie, with mechanicalfilters and an aqueduct to Aberdeen.

The estimated development costs, including capitaland the capitalizing of annual running costs of bothschemes (which allowed for 38 MI d-1), were Dee£551,900 and Dye £481,500.

The saving of £70,400 by adopting the second schemewould have been substantial in 1912, and the reportrecommended the Dye scheme. This scheme had theadded bonus that much of the city could be suppliedby gravity. It is interesting to note that in this report thepopulation forecast for 1951 (approximately 40 yearson) was 237564. The actual population in 1951 turnedout to be about 183000. There were strong feelings inthe city about the future of the water supply and anoutspoken champion of the rejected Avon schemewas a Thomas Jamieson, the Public Analyst forAberdeen, who spoke about the 'foul' Dee.

However, in February 1915, the Town Council decidedto proceed with an extended Dee scheme, and thenecessary Parliamentary powers were obtained in1916. Because the First World War was in progress, astart to construction was delayed until 1920.

Although not strictly affecting the abstraction from theriver, the following information may be of interest. Inthe autumn of 1912, there was an outbreak of entericfever in the city. The outbreak was at first attributed tothe water supply, but was later believed to have beendue to infected milk. In 1913, a lime sterilizing plantwas constructed along with 5 slow sand filters and aspiral baffle wall in the existing circular settling tank.

6 The present Cairnton abstraction schemeThe 1916 Act allowed abstraction to be increased from36.3 MI d-1 to 50 MI d-1, and it was necessary toconstruct a completely new intake in order to copewith the extra abstraction rate. This intake was locatedabout 70 m upstream of the 1886 intake, and is still the

113

intake in use today at most times in the year. Theintake is a side intake, ie it is on the side of the riverand is basically a reinforced concrete box belowground level measuring approximately 11.0 m x10.0 m overall. The superstructure is in masonry andthe building has a flat roof. Three openings approx-imately 2.4 m wide are provided on the river side ofthe intake. The flow through each opening is controlledby a rectangular penstock 0.8 m high x 1.2 m wide.The openings are protected by vertical bars andpressed galvanized steel screens with 6 mm open-ings. Two sets of copper screens were originallylocated in the inlet chamber but these were eventuallyreplaced by 0.5 mm mesh stainless steel screens. In1981, these screens were superseded by a self-cleaning drum screen with 1 mm apertures. Thisscreen is located at the actual water treatment works.The outer mesh screen at the intake is cleaned byimpinging air from an air blower on to the outer face.The operation is controlled by time clock and cleaningtakes place 3 times a day. From the intake, the water isconveyed to the original tunnel by a 1.5 m diameterconcrete conduit cast in situ. The tunnel and theconduits at either end can convey a total of 91 MI d-1from the intake to the treatment works.

Construction of the intake and conduit started in 1922and was completed in 2 years. During all this time, thewhole of the supply to the treatment works had to bepumped from the river.

The new lnvercannie/Cairnton scheme was formallyopened by HRH Princess Arthur of Connaught on 30September 1926.

7 Pumped abstraction at CairntonThe 1926 intake was able to supply water by gravityfrom the river during the 1920s, 1930s and 1940s, butby the 1950s sufficient water could not be taken intothe conduit/tunnel by gravity when the river level"dropped. In 1960-61, a permanent electrically drivenpump was installed at Cairnton at the old 1886 intakehouse. It is a 760 mm diameter low lift axial flowpump. When this pump is in use, it is necessary toshut the 1926 inlet penstocks. In the past few yearsthe pump use has increased, and in 1984 up until 20September it had been in use 154 days out of 264. Thisinlet is also screened by a bar screen and a 6 mmmesh screen.

8 Increase in allowable abstraction ratesBecause of the increase in water consumption in the1930s, the Aberdeen Corporation (Water, Gas, Elec-tricity and Transport) Order Confirmation Act 1937allowed the abstraction to increase to 68 MI d-1.However, the abstraction was limited to 63.6 MI d-1until certain sewerage works between Peterculter andthe City had been constructed by the then AberdeenCounty Council and Aberdeen Corporation. The Aber-deen Corporation Order Confirmation Act 1955allowed Aberdeen Corporation to abstract 68 MI d-1 at

114

any time when the total flow in the river was equal to,or exceeded, 120 million gallons per day (mgd) or 6.3cumecs measured at Cairnton. When the river flowdropped below this level, the abstraction rate had to bereduced to 63.6 MI c1-1.

9 The new River Dee abstraction schemeWith the discovery of oil and gas resources in theNorth Sea in the 1960s, it was evident that Aberdeenwould become a base for the necessary on-shore andoff-shore operations. This would mean an influx ofpeople seeking houses and a development boom wasforecast. Although the former Aberdeen County Coun-cil, and their water authority successors of 1968, theNorth East of Scotland Water Board, had a largescheme in being for supplying the small towns andvillages north and west of Aberdeen, it was clear thatmore water had to be found for the city and itssuburbs. The Board commissioned Sir M MacDonaldand Partners, Consulting Engineers, Cambridge, toundertake a water resources study of the whole of theBoard's area. The results of this study (MacDonald &Partners 1975) were produced in the year of the localgovernment re-organization in Scotland, and it was leftto the Board's successors, Grampian Regional Council,to put the consulting engineers' recommendationsinto practice. The main recommendations were that:

i. additional abstraction should take place from theDee at Cults;

ii. temporary additional abstraction should takeplace at Cairnton until the Cults scheme wascompleted;

iii. water should be abstracted from the Spey tosupply the northern coastal belt.

Scheme (i) included a reservoir on the Dye at Tilly-fumerie.

During the laste 1970s, extensive discussions wereheld with the fishing interests of Dee (the Dee DistrictSalmon Fishery Board and the Dee Salmon FishingImprovement Association) and other interestedbodies. Water Orders were obtained in 1978 and 1980(Grampian Regional Council (River Dee) Water Order1978 and the Grampian Regional Council (River Dee,MorriSon's Bridge Intake) Water Order 1980), whichallowed the extra abstraction at Cairnton and the newabstraction at Cults near Morrison's Bridge. Afterdiscussions with fishery interests, it was decided thata reservoir on the Dye was not desirable. The 1978Order allows a total abstraction at Cairnton of 91 MId-1, but, when the flow in the river measured at theNorth East River Purification Board's gauge atWoodend (about 2 km upstream of Cairnton) fallsbelow 680 MI c1-1 (7.9 cumecs) for 7 consecutivedays, hosepipe restrictions in areas served by the Deehave to be introduced. When the river flow exceeds770 MI c1-1 for 7 consecutive days, the restrictions canbe removed. After the end of 1985, the allowableabstraction will fall to 70 MI d-1 and the hosepipe rules

will no longer apply. The hosepipe restrictions apply tothe washing of private motor cars and watering privategardens.

The 1980 Water Order allows for a total abstraction of75 MI c1-1 from the river at a point approximately550 m downstream of Morrison's Bridge (the Shakin'Briggie). To allow fish to 'run' during low flows, theOrder incorporates a provision that all abstraction willcease for a period of 6 hours each day when the riverflow at Park (near Drumoak) is 640 MI d-1 or less. Theactual times of shut-down will be determined by theDee District Salmon Fishery Board. In addition, anagreement exists with Aberdeen Harbour Board thatthey will suspend their sweep net operations whenthe flow at Park drops below 454 MI d-1 (5.26cumecs). Compensation will be paid by the Councilduring this suspension period. The interruption topumping and the suspension of net fishing were aresult of the discussions with the fishing interests.

10 The new river intakeSir M MacDonald and Partners were chosen to designthe new scheme, which includes the intake, a pump-ing station, a bankside raw water storage reservoir,rising mains and a treatment works located at Mann-ofield within the city.

The intake is a bed intake, ie the actual intake islocated below the bed of the river. A bed intake waschosen after examination of the performance of theCouncil's bed intake on the Deveron at Turriff and aftertrials with a pilot intake on the Dee at the site of theproposed operational intake. The bed intake has anumber of merits.

i. It leaves the river bank unchanged as abstractionis through river gravels into boxes constructedbelow the bed of the river.

ii. The bed of the river is almost unchanged, thusbenefiting the fishery interests.

iii. The alternative side intake would have to be verylong in a shallow river such as the Dee.

The intake consists basically of 6 adjacent concreteboxes, each measuring 2.75 m x 2.0 m internally. Astainless steel distributor plate with 2240 10 mmdiameter holes is located 400 mm above the floor ofeach box. This supports 100 mm of washed rivergravel, 300 mm of 12-19 mm crushed granite and350 mm of 20-50 mm washed river gravel, all sur-mounted by 150 mm of 50-100 mm cobbles.

The approach velocity at full allowable abstraction is27 mm s-1, two-thirds of the value tested in the pilotintake; 400 mm diameter ductile iron pipes lead fromeach box to the low lift pump sump located in thepumping station on the north bank. Valving is arrangedsuch that each intake box can be backwashed at ratesup to 150 mm s-1 from the high lift delivery mainbetween the station and the treatment works locatedat Mannofield within the city.

The intake is not intended to serve as a filter, but ratherto provide a means of reducing the approach velocityto such a value as to create no problem to fish. It is,however, essential to ensure that the intake does notbecome blocked and is stable under all conditions ofriver flow. The pilot intake was operated for a period ofabout 2000 hours and parallel tests were carried out atAberdeen University upon a pilot intake built in alaboratory. Backwashing at rates within the range100-115 mm s' was found to be effective withoutthe use of air. During periods of low river flow, the rateof intake blocking should be very low, but at high flowsbackwashing is likely to be required every few days. Aproblem experienced at the Deveron intake at Turriffwas associated with short-circuiting during back-washing, which led to a blow-out of bed material. Thisproblem was overcome by the use of a distributionplate, and an even flow distribution during back-washing should be assured under all conditions.

The raw water reservoir incorporates a flood reliefstructure such that in flood conditions water can enterthe reservoir through non-return valves. The ingress offish is prevented by mesh screens with 25 mm x12 mm openings.

11 The Michelin Tyre Company Ltd intakeIn the early 1970s, the Michelin Tyre Company Ltddecided to set up a factory in Aberdeen and requiredapproximately 10 MI c1-1 of untreated water. TheNorth.East of Scotland Water Board promoted a WaterOrder in 1972 which would allow the Board to abstract9.84 MI c1-1 solely from the Dee for supplying theMichelin Factory located at Redmoss on the south sideof the city.

The intake was designed by Robert Cuthbertson andPartners, Consulting Engineers, Edinburgh, and islocated on the south bank of the river approximately620 m upstream of the Bridge of Dee. It is a bed intakeand consists of 3 concrete boxes sunk into the bed ofthe river. The boxes are 2 m square and about 850 mmdeep. Each box is covered by 4 stainless steel meshtrays which support a 225 mm layer of gravel sur-mounted by 300 mm of broken stone. A pipe leadsfrom each box to the pump well located on thebankside. Two pumps are housed in the pumpingstation, one of which is a reserve pump. The water ispumped to the Michelin Factory where a proportion ofit is used for processing and cooling water. Althoughallowing a total abstraction of 9.8 MI c1-1, the WaterOrder also calls for 80% of this water to be passedback to the river via a return main from the factory tothe river. The outfall of this return main is justdownstream of the pumps. Backwashing of the intakeboxes can take place by feeding water from the abovereturn main to any one of the 3 abstraction pipeslocated in the river.

In November 1984, the average daily abstraction wasonly approximately 1.4 MI c1-1, approximately 14% ofthe allowable abstraction.

12 Other water supplies derived from the DeecatchmentThis paper has concentrated on the major abstractionsfrom the Dee for public water supply. However, thereare a number of settlements which derive their watersupplies from the Dee catchment, and these are listedbelow with their average consumption.

Settlement Tributary

BraemarDinnetBallaterAboyneKincardine O'NeilLumphananTorphinsTarlandBirseGlendyeCrathie

BanchoryLope ColdstoneCod

UnnamedSpringsRiver GairnSprings/A/It Roy burn(Cairnton)(Cairnton)(Cairnton)SpringsSpringsWater of DyeSprings/Wen at bank ofRiver Dee (Emergency)(Cairnton)SpringsSprings

A proportion of this water is returned to the Dee.

115

Average dailyconsumption

(m3 cl-')

29016

520480

1596

4500

22

553

13 River quality as it pertains to public water supplyFrom the point of view of its suitability for public watersupply, the Dee at Cairnton is a high quality source ofwater which requires only limited treatment beforebeing supplied to the consumer.

The river at this point contains no harmful or objection-able chemicals in significant amounts. There is littlemanufacturing industry, and therefore there are noproblems from industrial effluents. The principal aimsof treatment are to ensure the absence of faecalbacteria in the product water, to prevent silt broughtdown the river during high flows from entering thesupply system, and to guard against the unexpected,eg spillage from tankers.

Treatment consists of settlement, slow sand filtration,chlorination and lime dosing. In slow sand filtration, thewater flows by gravity through large beds of fine sand.The community of micro-organisms which becomesestablished in the filter greatly reduces the numbers offaecal bacteria in the water. This is then chlorinated toensure disinfection. As humic substances are notreadily biodegraded, they pass through treatment,which accounts for the pale yellow colour observablein Aberdeen water. The final treatment step is to addhydrated lime to the water in order to reduce itscorrosive effect on lead pipes which have been widelyused in the past as service pipes connecting theconsumers' taps to the water mains.

14 Factors affecting the quantity of water available forabstractionIn recent years, scientists and engineers have beeninterested in the effect of evaporation from forests incatchment areas and the effects of climatic change on

116

runoff. It is not known if any such investigations havebeen carried out in the Dee catchment, but a recentstudy of hydrological records from the 1930s to dateindicates that no dramatic changes have taken place.There has been a claim from some quarters thatprolonged periods of low flow have been increasing,but there is no evidence to sustain this claim.However, if the resources were made available itmight be interesting to study the matter of evaporationfurther.

This paper has concentrated on the use of the Dee forpublic water supply purposes. The Regional Councilrecognizes, however, that the river is used for otherpurposes and that is why all interested parties wereconsulted and had a chance to air their views beforethe Cults scheme was progressed. No formal objec-tions to the Water Order were made and therefore noPublic Inquiry was necessary. The Council retained 2fishery consultants to advise on fishery matters, and,although allowable abstraction is many times thatallowed in the 1800s, the Council are confident thatthis level can be sustained without damage to the floraand fauna of the river.

15 Summary15.1 The Dee is the source of the water supply for

Aberdeen and its suburbs. The river's tributariesalso supply water to Braemar, Ballater, Aboyneand north Kincardineshire. Other villages in theDee catchment are supplied from springs with-in the catchment.

15.2 The first Aberdeen supply from the Dee wasapproved by Parliament in 1829 and the abstrac-tion took place 650 m upstream of the Bridge ofDee in Aberdeen. The allowable abstractionwas 5.7 MI d-1.

15.3 By the 1850s, the quantity available was notsufficient to meet the demand of an expandingpopulation in Aberdeen. In 1862, an Act waspassed which allowed a new water supplyscheme to be constructed with an intakelocated at Cairnton, about 6 km upstream fromBanchory. The treatment works were con-structed at Invercannie about 2 km east ofCairnton. The allowable abstraction was27.2 MI d-1.

15.4 Because of mounting consumption, an Act of1885 allowed the'abstraction to be increased to

36.3 MI d-1. The intake was relocated 120 mupstream.

15.5 Again to meet increasing demand, an Act of1916 allowed the abstraction to be increased to50 MI d-1. Because of the First World War, itwas not until 1926 that the works necessary tocope with this quantity were completed. Theseincluded a new intake about 70 m upstream ofno. 2 intake.

15.6 In 1937, an Act allowed the abstraction to beincreased to 68 MI d-1 under certain condi-tions. It was modified by another Act in 1955.

15.7 To enable abstraction to take place duringperiods of low flow, a pump was located in theno. 2 intake in the early 1960s.

15.8 In 1972, the North East of Scotland WaterBoard promoted a Water Order to enable9.84 MI c1-1 to be abstracted from the Dee620 m upstream of the Bridge of Dee, to supplythe Michelin Tyre Company Ltd located atRedmoss in Aberdeen.

15.9 The coming of the oil industry to Aberdeencalled for even more water, and a Water Orderof 1978 allowed the abstraction at Cairnton tobe increased to 91 MI d-1 until the end of 1985.Certain conditions were imposed.

15.10 To meet the forecast demand into the earlyyears of the 21st century, a new abstractionscheme based on abstraction at Cults waspromoted during the 1970s. The authority toabstract water was confirmed by a WaterOrder in 1980. The total allowable abstractionat Cults will be 75 MI d-1. Abstraction will berestricted to 18 hours each day when the riverflow measured at Park drops to 640 MI d-1.Sweep netting at Aberdeen Harbour will besuspended when the river flow drops to454 MI d-1.

16 ReferencesAnon. 1973. Census 1971 Scotland — county report: Aberdeen city.Edinburgh: HMSO.

MacDonald, Sir M. & Partners. 1975. Regional water resourcesstudy for North East of Scotland Water Board. Vol. 2. Cambridge: SirM. MacDonald & Partners Ltd.

Fishery management on the River Dee

J OSWALDGlen Tanar Estates, Aboyne

1 IntroductionI write this paper as someone who is directly involvedwith the day-to-day management of 23.6 krii of salmonfishing on the Dee between Aboyne and Ballater. With34 years of experience of the river, to me it is morethan an academic feature on the map. It is a living thingthat we must care for, through the hard frosts ofwinter, the spring spates, and the droughts of sum-mer. Salmon fishing by rod and line has a position on amarket. Giving a good service to the fisherman byhaving a well-maintained section finds the best out ofthat market. Naturally it is necessary in salmon fisherymanagement to have a plentiful supply of salmon, andalso to have good access tracks, nice fishing huts andwell-trimmed pools in which to fish. There is also thewildlife consideration, from freshwater mussel to thehumble dipper, whose lives are entirely dependent onthe river.

The way we conserve all the things we like for futuregenerations must be high on the order of priorities.Many things that concern the fisherman are nowcompletely out of his control, and we have to rely onthe good offices of the people in positions of authorityand power to use their wisdom when dealing withsuch matters. I hope the following may help to givethem a better understanding of some of the practicalproblems.

2 HistoryThe Dee is one of the most prolific rivers in the UnitedKingdom, and probably holds more salmon and seatrout per unit volume than any other. It has a course ofsome 140 km (87 miles) to the sea, starting high in theCairngorm Mountains and entering the sea throughthe harbour at Aberdeen. There is virtually no sewagepollution in the river, which has supplied fresh water tothe city of Aberdeen for many years. Some of theearliest settlers in Britain chose a site betweenAberdeen and Banchory for their encampments, inplaces where no doubt a good source of food wasavailable from the river nearby, in the form of salmon.In the last century, farm workers would not take up acontract of employment 'fee', unless guaranteed thatsalmon would not be served more than twice a week.

Various methods evolved for catching salmon. The netwas the most efficient, but the wattle fence trapwould also have played a big part in medieval times.Although traces of these traps have long disappearedfrom the Dee, they have been dug up on other slowerand muddier rivers, and there is good reason to believethat they were used on the Dee. Various types ofbarbed spears or multi-toed forks called 'leisters' werecommon and easy to make. They were quite heavy

117

and had a long shaft, with a spear-head made fromwrought iron with 4-6 toes, barbed on both sides sothat the fish was well impaled. (In 1850, just aboveBalmoral bridge, Queen Victoria and Prince Albertwitnessed an exhibition of salmon leistering. 'It had avery pretty effect; about 100 men wading through theriver, some in kilts with poles and spears very muchexcited' (Victoria, Queen 1868)1 In the tributaries, thefarm folk used the gaff, a simple steel hook, as theirmain weapon. In medieval times, virtually any methodof getting salmon from the river was acceptable, asthey were so abundant.

Fishing for salmon as a 'field sport' probably started inearnest around the mid-17th century. A trooper inCromwell's army, Richard Frank, recognized thepotential while stationed in Scotland, and after beingdemobbed returned in 1650 to fish for salmon. Withthe ever improving transport systems of road, sea andrail, by the middle of the 19th century salmon fishinghad become a popular recreation in Scotland. At thisperiod, the river was netted for the 26 km up toBanchory by a series of sweep nets.

The demand for rod and line fishing increased everyyear, but a decline in the number of salmon reachingthe spawning grounds in the upper regions of the riverwas attributed to over-fishing by the nets betweenBanchory and Aberdeen. The District Board could notundertake the removal of the nets, so the proprietorsabove Banchory formed an Association (Dee SalmonFishing Improvement Association) for this purpose. Onthe initiative of the Marquis of Huntley, the DeeSalmon Fishing Improvement Association was formedin 1872. Its purpose was to control and lease the netfishing, and it was very successful.

Commencing at Banchory, 27 km distant from themouth, the river nets were leased and removed asfunds permitted. In the first 10 years of its existence,the Association spent nearly £4,000, until there wereno nets in the river above the railway bridge atAberdeen. In 1882, the river apparently yielded a richharvest, for in that season upwards of 5000 salmonand grilse (fish returning to the river after spending onewinter at sea) were taken by the rods. In 1902, thecollective value of the angling rents produced 8 timesthe income paid by the netting rents in 1871.

From that time, the Dee has never looked back. It hashad its bad years and good ones, like all other rivers,with the difference that a bad angling season on theDee means a nearly absolute dearth of sport on anyother river. By degrees, however, the increase of netson the coast began to have its effect. As early as the

118

turn of the century. Grimble (1902) wrote that 'theAssociation must now turn their attention in thatdirection'.

However, despite the coastal nets, salmon fishingremained fairly constant and predictable until the1960s, when a localized salmon fishery in westGreenland developed into a significant autumn inshoregillnet fishery, exceeding 1500 tonnes (t) by 1965. Anoffshore drift net fishery also developed from that timein Greenland and the combined annual catch exceeded2600 tonnes in 1969-73, when regulations limiting thequantities of salmon caught in the fishery wereintroduced by the North West Atlantic FisheriesCommission (Association of Scottish District SalmonFishery Boards 1977).

From 1968, a small long-line fishery for salmondeveloped off the Faroes, extending up to 112 kmnorth of the islands. Catches had built up to 40 t yr-1by 1977. Vessels from Denmark then became in-volved, and catches subsequently increased andpeaked at about 1000 t yr-1 in 1981. In the early phaseof this fishery, until the late 1970s, most of the catchconsisted of one sea-winter fish. More recent observ-ations made on commercial vessels have suggestedthat about 90% of the landed catch was of 2 or moresea-winter fish and that only about 2% had spent onewinter in the sea. From 1982, the Faroese governmentagreed to a voluntary quota system, with a total catchlimit of 750 t yr-1 in 1982, reducing to 625 t in 1983(Shearer & Clarke 1983).

Two other fisheries are known to exploit salmonstocks of Scottish origin, from the capture of adultstagged as smolts. These are the Irish drift net fisheryand a somewhat similar fishery operated in the NorthSea off the coast of Northumberland and, to a lesserextent, Yorkshire. The catching power of both thesefisheries increased in the 1960s with the introductionof nets manufactured from synthetic twine. Bothfisheries catch mainly grilse. The influence of theNorthumbrian fishery is confined primarily to the eastcoast of Scotland, between the Rivers Bervie andTweed, but there is a lack of information regarding thearea affected by the Irish drift net fishery.

In 1962, a legal drift net fishery operated off theScottish coast. This method of fishing was banned in1963, but illegal fishing using similar techniquescontinues and none of the catch appears in thenational statistics.

By the end of the 1960s, fishermen on the Dee werehaving great difficulty in catching salmon on rod andline. Salmon were being caught in February withdamage to fins and recent signs of being tangled innets. The disease UDN (ulcerative dermal necrosis),which arrived in Scotland in 1966, also started to takeits toll. A combination of this disease and drift nettinglowered salmon stocks to rock bottom, and they have

been struggling to reach anywhere near their formerlevel on the river as a whole ever since. The decline isshown clearly by data from a salmon trap in theGirnock, a tributary of the Dee above Ballater (Table 1).To help combat the decline, a salmon hatchery wasbuilt at Dinnet in 1964, but its use has not increasedcatches markedly.

Table 1.Numbers of adult salmon trapped experimentally when

3 Preventing poachingWhile control of the fishermen is kept in the hands ofthe riparian owners, the policing of illegal fishing fallsto the Dee District Fishery Board. Its team consists ofan inspector and 7 water bailiffs who patrol the river,mainly in the dark, and have radio-equipped vans forquick communication. The use of the trammel or gillnet seems to be the favoured illegal method forpoaching salmon from the river. In low water con-ditions, 'Cymag' gas is sometimes used, and kills allfish without distinction. Over the years, manyattempts to foil net poachers with various types ofanti-net hooks have been made. These are calledcleeks, anchors or huds; most are of the 3-hook typewith inside barbs. Some have been loaded into bigrocks, and others are embedded into concrete blocks,many of which are upturned or broken off. A revival ofthe anti-net hook would certainly trap many of thepoachers' modern nylon nets, which are so light andstrong that they can be carried in a rucksack. Thecombination of nylon net and the wet suit has madeillegal fishing so much easier than yesteryear, when italmost needed a horse and cart to carry a wet cord net.The main powers of the water bailiffs are set out in theSalmon Fisheries (Scotland) Act 1868 and the Salmonand Freshwater Fisheries (Protection) (Scotland) Act1951. A close season extends from 1 October to 31January.

4 Rod and line fishingIn February and March, spinning lures such as 'Dev-ons' and 'Spoons' are allowed. From 16 April onward,

it is 'fly only' for those fishers who adhere to theguidelines of the Dee Salmon Fishing ImprovementAssociation, during high water. Water gauges on mostbeats define the level where the change from fly tospinning takes place. Prawn and worm fishing isfrowned upon, and on most beats totally banned. Astrict control is kept on the number of rods allowed tofish at any one time, and this is supervised by amember of estate staff (eg factor, gamekeeper orghillie). A detailed and accurate record of all salmonand sea trout caught is kept by the fishing proprietor,and is used by the assessors to determine the rates inthe 'Valuation Roll for the Grampian Region ValuationArea'. An important contribution is made by thesalmon fishery to the Grampian Region.

In the early days of salmon fishing, rods were made ofheavy greenheart wood and could measure 5 m. Reelswere made of brass with heavy linen lines, and flyhooks were dressed with exotic bird feathers andcould measure up to 10 cm in length. They were joinedto the line by a gut cast. The method was known as'sunk fly' fishing, and continued to the 1930s. A newtechnique was then developed called 'greased line'fishing which involved using a light 'split-cane' (bam-boo) rod, and tiny flies of some 2 cm in length dressedvery lightly with a dull game feather attached by gut toa dressed linen line. This line had to be dried 3 or 4times a day and greased with a water repellent, so thatit would float on the surface. It was very effective inwarm water and more than doubled the catch from thesunk line method. Sandy Thompson, a ghillie fromAboyne, described to me in 1951 the start of thegreased line in the 1930s, '1 niver saw salmon tak' a flylike it; they rose, showing head, back and tail an' thena pause, a great rug, an' they were on'.

This is the current method, but the equipment hasadvanced. The early 1950s saw the introduction ofnylon for casts, together with fibreglass rods and linesmade of plastic-coated fibres, giving them goodfloating qualities. In more recent times, fly-fishing rodshave benefited from the latest technology in carbonfibre.

Spinning lure techniques have not changed much fromthe early days, except for the tackle, ie rods, lines andreels. The old wooden rods and brass reels with linenlines have been superseded by a big variety offixed-spool and level cross-wind reels. These combin-ations have brought almost every salmon within rangeof the fisherman on nearly every pool of the Dee.Charlie Milne from Glen Tanar told me in 1951 that, inthe old days when they could not cast very far, 'weused to link our lines together with horse hair, andwind the lure back and forth the wide parts, and whenthe salmon took, the hair broke and one was left toland it'. He also told me that pickled gudgeon, loach,sand eel and sprat were the favourite live baits. Spratsdyed with yellow and pickled in a bottle, called 'goldensprats', were acknowledged to be the best. At the

119

present time, the main lures are the wooden 'Devon'and 'Blair' spoon.

5 RecreationOne of the main attractions for tourists on Deeside isthe river, with its spectacle of salmon leaping in thepools such as there is at the Linn of Dee. Salmonfishing is a recreation and is enjoyed by a widespectrum of people, from the Queen to the ordinaryworking man. A new sport on the river has arrived withthe introduction of the fibreglass canoe some 10 yearsago. This has resulted in some disturbance. However,salmon adapt quickly to a floating object, providedthere is no prolonged disturbance. In my experience,canoeists are sensible and respond to requests madeof them. Camping is allowed in some areas near theriver, and some bathing occurs during prolonged spellsof hot weather, but temperatures seldom reach thislevel before the end of June. A favourite pastime ofthe day-tripper is to sit by the river, throw in his rubbishand let it float away. This habit can only be preventedby education.

6 Agriculture and wildlife conservationFast-flowing rivers bring their own problems wherethey run through cultivated land. There is a continualfight to prevent the river cutting into fields and erodingthe banks. During the autumn floods of 1982, treeswere torn out, and stable banks adjoining arableground were ripped away. To counter this on GlenTanar, during the winter thousands of kilograms ofstones were tipped at the vulnerable points. I find thatrocks of 50-100 kg fall into place and lock togetherwell; if filled with some loose earth, they becomecolonized with plants and bushes in a year or so toform a semi-natural river bank profile.

In a natural situation, the river would be unseen by thepasser-by as it would be hidden by trees and bushesalong its banks. Owing to salmon fishing interests, thebanks of the pools have been shorn of their bushyvegetation so that the fisherman can get access to thepools. No more than is absolutely necessary is cut bythe ghillies, as this work is labour-intensive andexpensive.

With a river like the Dee which is virtually free frompollution, great care should be taken in future over anydevelopment, whether industrial or agricultural. Suchitems as large exotic plantations in the catchment areaof the tributaries can alter the water to such an extentthat young salmon cannot live. Large-scale drainage onpeat-covered hills causes flash floods which areunwelcome on any river. Tearing up large areas ofscrub land for re-seeding, or the deep draining ofmarsh land with the consequent loss of insect life, cangreatly affect the diet of young salmon. Many featuresmake a river a good place for fish to live, and to keep itthis way there should be some form of consultation onany development. Bodies like the Nature ConservancyCouncil, the Department of Agriculture and Fisheries

120

for Scotland, and the Institute of Terrestrial Ecologyare readily available, with scientists qualified in everysector of environmental research. Traditional land-owners of the past have had long experience of goodland management, and have cared for the countryside.With the smaller land units coming into being, and newowners, there is a chance of radical new landmanagement, and the new owners should keep thewelfare of the river in mind.

7 Water abstractionWater authorities in Scotland have powers underSection 21 of the Water (Scotland) Act 1946 to acquirewater from any stream or other source by agreementor compulsion. At Common Law, District FisheryBoards may object to the abstraction proposals ofpublic water undertakings and, in some cases, thismay result in a Public Inquiry. Under the existing WaterOrder, the local authority is allowed to extract 'amaximum of 20 M gallons a day' from the Dee, whichgenerally meets the city's daily needs, at least for thepresent. However, in the past 2 dry summers of 1983and 1984, river levels have been very low, and we readin the press that, 'while the number of gallonsextracted from the river above Invercannie will bereduced to 15 M gallons a day, the local authority willbe allowed to take a further 17 M gallons daily(8 500 000 gallons in 2 phases) for the new supplylower downriver at Cults' (Press &Journal, 10 Septem-ber 1984). It is frightening to think that around onethird of the river in a dry spell should be extracted tosupplement other regional supplies. Surely, actionshould be taken to provide adequate storage to fill thisneed in time of drought, as, in my opinion, the level ofabstraction has already reached the maximum that canbe tolerated by fishery interests.

8 FutureWe may have seen the end of good salmon fishing fora long time to come. Gone are the days when we weresure the 'spring runs' would come in. So many factors,very complex and beyond the immediate control of theriver fishing manager, have now come into being. Thetrends (Table 1) show the river catches to be indecline, and the point will come where the impact willbe felt beyond the fishers, and by such people ashoteliers and shop-keepers, coupled with a loss inrates and a loss of employment. The main hope toavoid disaster is the conservation of salmon on thehigh seas. The problems are well enough known; it is amatter of finding the will by international agreement toresolve a complex situation.

9 SummarySalmon fishing on the River Dee is more than a luxuryfor a few sportsmen; it has an important part to play inthe economy of Deeside.

We must look after this asset in every practical waypossible, so that salmon may enjoy the freedom toreturn to the headwaters for many hundreds of yearsto come.

A little over a century ago, our grandfathers took onepositive step in removing all the nets from Banchory toAberdeen, with a subsequent improvement in salmonstocks.

In the last 20 years, we have seen a demolition of thisgood conservation work. Heavy fishing at sea is takinga big toll of salmon on the feeding grounds andhomeward bound routes. Only international agree-ment can resolve this sea fishing problem. At home,increased direct water abstraction from the riverduring spells of drought can only spell disaster forsalmon trying to smell the river out at sea. A largereservoir is needed to supplement the regionaldemands during dry weather, and proposals for furtherdirect abstraction should be resisted.

Land use in the catchment areas should be monitoredcarefully, so that the breeding grounds in the tributar-ies are not destroyed by excessive draining and byplantations of exotic trees. Little is known about thelong-term effects of the policy of blanket planting ofnon-native forests throughout the feeding grounds ofyoung salmon. In many cases, trees of American originare over-hanging the streams for miles on end.

A revival of the anti-net hooks on the river beds wouldallay the worries of fishing proprietors on manystretches that are susceptible to the ravages of thepoachers' nets.

Co-ordination of all the interested parties must prevailif the future of salmon in the river is to be assured. Nosingle group alone can stand up to the pressures of theContemporary world, however well meaning they maybe. Only action on a broad front by people ofknowledge and authority can succeed.

10 ReferencesAssociation of Scottish District Salmon Fishery Boards. 1977.Salmon fisheries of Scotland. Addlestone: Fishing News Books Ltd.Buck,R.J. G. & Hay, D. W. 1984. The relation between stock sizeand progeny of Atlantic salmon (Salmo salad in a Scottish stream. J.Fish. Biol., 24, 1-11.

Calderwood, W. L. 1909. The salmon rivers and lochs of Scotland.London: Arnold.

Fishery Board for Scotland. 1885. Annual report no. 4.

Grimble, A. 1902. The salmon rivers of Scotland. London: KeganPaul, Trench & Trubner.

Victoria, Queen. 1868. Mountain rock and glen: illustrating our lifein the Highlands. London: Bell & D.

Shearer, W. M. & Clarke, D. R. 1983. Long term trends in Scottishsalmon catches. (CM 1983/M:24.1 Copenhagen: International Coun-cil for the Exploration of the Sea, Anadromous and CatadromousFish Committee.

Amenity and recreation on the River DeeR D WATSONNorth East Mountain Trust, 396 King Street, Aberdeen

1 Introduction1.1 Definitions

Amenity commonly includes both services andpleasant surroundings, and both apply here. Informalrecreation, by its very nature, often needs fewservices, but picnic areas and caravan and campingsites are required. Landscape is a term that con-veniently covers much of 'pleasant surroundings', theother aspect of amenity. Countryside users wouldincreasingly include wildlife as part of these pleasantsurroundings. Some points discussed, such as amen-ity and enjoyment, are subjective. Thus, much of whatfollows is personal opinion, albeit shared with manyothers, and obviously cannot be given the status ofscientific conclusions.

There is no precise dividing line between formal andinformal recreation. For example, canoeing encompas-ses formal slalom competitions and informal rivertouring. Similarly, participants in some activities do notrepresent separate 'user groups'. Almost everyonewho is appreciative of countryside is active in morethan one of them. To take an obvious example, thefamily who drive to a favoured place, walk along thebank for pleasure, see interesting flowers and photo-graph them, have participated in a wide range ofactivities. These are problems familiar to anyone whoattempts to categorize the behaviour of people, butthey need not trouble us here, provided we do notattempt to make our classification too rigid.

1.2 Scope and nature of informal recreationThe more popular activities classed as informal recrea-tion on Deeside include touring by road, walking,canoeing, nature study including bird-watching, photo-graphy, swimming, picnicking, camping, and caravan-ning. No overall data exist on the level of participation

Table 1. Proportion of people (%) involved, at least once, in recreational activities in the Scottish countryside in theyear prior to survey (source: Survey Research Associates 1981)

Drives, picnics, Long walks Visitingoutings >3.2 km nature reserves

(Derived from 2905 interviews in 150 polling districts, with sample figures weighted to represent the Scottish adultpopulation)

59 36 12

Table 2. How holiday-makers first heard about their main stop in Scotland (%) (source: as Table 1)

Alwaysknown

about it*

45 29

Recommendedby relative Foundor Men& by chance*

Allmedia

11 3

*The majority of these making repeated visits to their destinationtThe majority of these making first-time visit to their destination

in outdoor recreation by residents in Grampian. It isfairly certain, however, that it has been increasing forsome years, and is still increasing, perhaps rapidly. Theclearest indicator of this increase is for rambling,walking, and canoeing, for which some informationcan be obtained. Both the numbers of clubs and themembership within clubs are increasing fairly rapidly.

These trends may partly reflect national ones, but itwould be difficult to prove that such a national trendhas existed since the late 1970s. A major reason forlocal increases is probably the increase in population inGrampian, which grew by about 50 000 (14%) be-tween 1971 and 1984. This is a large percentageincrease over 13 years. In addition, almost certainly anuntypically high proportion of those attracted to thearea are young, well-paid, middle-class professionalswith considerable disposable income. These are thesocial groups most active in outdoor recreation.

To many local participants in outdoor recreation, thecatchment of the Dee is a huge country park. Peoplewalk, drive, climb, and ski in it, and in later years taketheir children to it as they would to a huge adventureplayground. The social significance of Deeside to theAberdeen urban area as a recreational resource is thusenormous.

2 Informal recreationsThis paper concentrates on the relevant activitiesshown to be the most important in a survey organizedby Survey Research Associates (1981), plus thoselesser activities particularly associated with the river.Considerable information is available on the relativeimportance of the major activities in the Scottishcountryside (Tables 1 & 2).

Visiting stately Pony trekking,homes/museums Fishing riding

10 9 2

School Seen Guidework on book oror trip map gazetteer

2 1 1

TouristOffice

information

121

122

2.1 Road touring and riverside viewsThis survey shows that about 66% of tourists come toScotland by car. It also (Table 1) shows the highimportance of driving in the countryside for localpeople. Driving is also part of most other activities, andfor many people it is the experience of the whole daythat counts, and not just that part of it spent, forexample, canoeing or walking. The importance of the'journey to play', as it has been called, is considerablefor almost all participants in outdoor recreation. Sur-veys show that the dominant factor that attractsvisitors to Scotland, and local people to its countryside,is the scenery. Consequently, the view from the roadis important both socially and economically.

With the car tourer especially, the eye roams at willover the entire landscape, and my comments musttherefore take in a similar sweep of the strath. In thisbroad sweep of the Deeside landscape, the river is animportant unifying feature, giving it a significance fargreater than any quantitative assessment of it as alandscape feature might suggest.

Deeside owes much of its appeal as a tourist route tofrequent and characteristic vistas from the road acrossriver, moors and fields to a backdrop of hills in whichthe varied and subtle hues of bracken, heather, birch,and other local vegetation are the most outstandingfeature. It is this feature of the scenery that isdistinctively Scottish and thus attractive in the eyes offoreign tourists, many of whom are already ac-quainted, in their own countries, with more spectac-ular hills and mountains, and more impressive foreststhan they see here.

Both North and South Deeside Roads give close-upviews of the river itself over short occasionalstretches. Perhaps the finest views are offered at therare places where the road provides a vista up or downthe strath, giving a view of great depth, with the river,fringed with tree and rock, sparkling through themiddle. However, many of these views are beingreduced in quality or being obscured. The main causeis the growth of trees, either by afforestation orthrough natural regeneration following the withdrawalof grazing or burning in some areas. Trees are havingnegative effects on the landscape here.

Often, fringes of regenerating birch add colour andinterest to the view at all seasons. Over quitesubstantial hillsides, such as on parts of Morven,Culblean, and a long series of hills west therefrom,bordering on the north side of the North Deeside Road,mixed regeneration of trees is adding to the variety oftextures and hues. Further, at points, the over-matureamenity plantings around old country houses such asat Durris House, Invercauld, and Mar Lodge contributegreatly to the scene, despite the relatively smallnumbers of trees involved. The impact derives partlyfrom the mature status of the trees, partly from theirparkland setting, and partly from the landscaping skills

of the original planters. It is a matter of concern thatthese impressive but ageing larches and spruce,among other species, are in most instances not beingreplaced. There is surely a good case for a programmeof selective replanting in such circumstances, to becarried out without delay.

On the negative side, large forestry plantations arecovering extensive areas of hill. Large-scale planting ofhill slopes with single species stands of conifers,however well managed, replaces the attractive andvaried shading of the open hill with a more uniform andsombre green, as on the eastern approaches toBanchory and Aboyne and around Bal later. In places,plantations on both sides of the route reduce vision toa conifer-lined tunnel, as on the North Deeside Roadwest of Dinnet. An alternative, under-used, measure isfringing roads with native broadleaves. Among theDeeside estates, only Glen Tanar has done this to areasonable degree.

Where the route rises to well above the river, giving anelevated view of the scene, it provides vistas in depthacross or, along the strath. Most of these have,unfortunately, been obscured by roadside plantationstaken right up to the downhill side of the road margin,along quite extensive stretches, most notably beforeand after the Coilacriech Inn. The same has happenedalong the approach to the Pannanich Wells, till thennoted for superb views of sunset over river and hills.No great restraint would be necessary in managing thewoodland on the downhill side to retain such out-standing views.

Occasionally, mass regeneration of birch is obscuringviews at certain points. In selected places, this mayneed to be restrained, such as below the road atCambus O'May, where birch growth obscures fineviews of the river, and at the old Invercauld bridgewhere it is hiding a famous view of the bridge from theroad. Regrowth of shrubs and trees along the aban-doned Deeside railway line is also producing similareffects due to the siting of the route between river androad.

Given the importance of tourism in the GrampianRegion, and the popularity of driving as a recreationalpastime, the economic and social value of all thesescenes is very great. We cannot afford to allow thequality of our riverside scenery to decline for lack ofthought and management. Table 3 provides some dataon why visitors selected their main stop in Scotland. Itis interesting to see how few were motivated to do soby 'marketing' channels and how many came on thestrength of recommendations by friends or their ownpast experience of the area.

3 Camping, caravanning and picnickingThis group of activities is shared by participants inmany forms of recreation, but it is particularly relatedto the car-borne tourist, and therefore conveniently

Table 3. Holiday-makers in Scotland — activities indulged in at least once during stay (% participants) (source: asTable 1)

Visit loch/ Drives/picnicsriverside in countryside

55 49 47

Visit Long walkssea coast >3 km

considered at this point. A noteworthy characteristic ofthe Dee is that there are extremely few points atwhich the public can gain access officially to the riverbank. Only in the Duthie Park is any sizeable area opento the public for picnicking. That apart, a small area onthe north bank west of Ballater is the only officialaccess for anything except walking. At other points,such as at Aboyne, Park, Potarch and the Linn of Dee,access is 'informal'.

There is some sign that overt attempts by estates torestrict informal access have decreased in recentyears, probably under pressure of numbers of people,and as a response to changes in public attitudes toaccess to the countryside. Nonetheless, in stretchesbelow Aboyne, where the need for such access toserve the local car-borne visitor is greatest, the lack ofthis kind of facility is acute. One questions thewisdom, for example, of promoting the green atPotarch as a picnic area, for which it is now heavilyused, without making any formal access agreementfor the people who then frequent the river in numbers.For the caravanner and camper, there is a similarshortage of sites. The new site at Braemar may helpmatters, but in recent years the traditional 'informal'sites at Potarch, Coilacriech, and Linn of Dee havebeen closed, without a balancing development of newofficial sites. Only at Ballater is there a single,crowded, riverside camp site, and this is too close tothe town. Mainly as a result of this shortage, 'wild'camping and caravanning have become common insome areas such as along the Cluny. There isobviously a need to provide more sites for theseactivities, both for touring holiday-makers and for morelocal visitors.

4 SwimmingClosely associated with outings to the river is thetime-honoured pastime of 'river-dookin'. Traditionalswimming pools are found at various points along theriver, often at bridge points such as at Dinnet, Potarch,and the footbridge at Cambus O'May, althoughperhaps the finest of all is the Chest of Dee above theconfluence of the Dee and the Geldie. Although it canhardly be classed as a mass participation sport, asurprisingly high proportion of the people whofrequent the countryside have enduring memories ofthis pastime, and its significance should not beunder-estimated. Whether swimming people signif-icantly disturb fish for any length of time could be amatter for argument, but if they do the problem doesnot seem to surface as a conflict.

Visit naturereserve

Rowing/sculling/

Fishing canoeing Climbing

33 13 10 5 4

5 WalkingNationally, the importance of walking in the country-side, as opposed to strenuous hill-walking, continuesto be demonstrated as the most important form ofoutdoor recreation by the most recent results fromSurvey Research Associates (1981) (Tables 1 & 2).Locally, Walton (1969) showed its importance onDeeside (Table 4). Other indicators since then aresparse, but significant. The local branch of the Ram-blers' Association, founded in 1983, has reached amembership of over 200, and is still growing. TheNorth East Mountain Trust has gained the impressionthat there remains considerable unsatisfied demandfor this form of recreation. The position for suchpeople in the Aberdeen area is not satisfactory. Thecity is largely surrounded by enclosed farmland, andthe nearby areas of open land such as the Kincardine-shire moors and the Hill of Fare have been afforesteden masse with little regard to the need of people foropen space for such activities as walking. To the north,the south, and in Donside, the countryside is alsoenclosed farmland.

Table 4. Main recreational or leisure pursuits of visitors to a craftcentre in Deeside (source: Walton 1969, Table 5, p 74)

% of visitors

(Sample size is not given in the original)

123

Deeside is the sole large area of opportunity forcountryside walking. However, there are relatively fewstretches of river bank with right of access for walkers,and these tend to be short. Within the city boundary,they occur in the Duthie Park and at Peterculter.Outwith the city, they are found on both banks aboveBanchory, along the old railway line west of theCambus O'May footbridge, and on the north bankwest of Ballater. The lack of riverside walks reflects alack of known public footpaths generally in the strath.Such as there are tend to climb above the valley floor,and the vistas and views of the river from them can bevery fine, as on Craigendarroch.

For the casual walker, in purely physical terms, muchof the river is very accesSible with easy anglers' pathsalong the banks. Except in a few areas, however, littleuse is made of these paths for riverside walking.

124

Casual visitors are probably deterred from using themas it is private land. Modest levels of walking should becompatible with fishing, and it is worth suggesting thatconsideration should be given to access agreementsor other means of facilitating this use.

A potential for conflict exists between the interests ofwildlife and an extension of riverside footpaths.Secluded areas of bank may be disturbed. There is noneed for such walks always to stick to the bank.Discrete diversions, even quite extensive ones, aroundimportant bank areas can leave such sections undis-turbed, and even add variety to the day's walk. Theproblem of disturbance by walkers' dogs is not soeasily solved. Neither, for that matter, are the prob-lems which the same animals create for other walkersby fouling footpaths and picnic areas. It may be that a'Keep Dogs on the Lead' rule would have to beprominently displayed for certain walks.

Most people now agree that the local authoritiesshould have tried to acquire the whole of the disusedDeeside railway line. The assessment that parts of theline did not pass through scenery of outstandingquality on lower Deeside did not allow for the fact that,in areas such as Crathes, the route of the line was theonly way on foot through the countryside, and that itoccurred near a large urban centre. Perhaps we shouldnot give up the idea that the existing bits of the linemight be connected up with old fragments of road, andperhaps the occasional riverside walk, to give a RoyalDeeside Way. It might prove an interesting route, withthe great advantage of frequent, convenient servicesand accommodation centres at reasonable intervals.

Beyond the Linn of Dee, up to the river's source, it ishill-walking that is the dominant form . of informalrecreation, along with back-packing along the import-ant hill passes that converge around the Chest of Deearea. Such conflicts as occur are related to other usesof the hill land, not of the river or its banks. Theirresolution depends on the production of a muchneeded Cairngorms Management Plan, including man-agement of the river.

6 CanoeingPeople are attracted to canoeing by the sheer delightand excitement of moving about on running water andcoping with its variety and challenge, and by the wholeambience of the river: the trees running down to theriver's edge, the quietness, the cowslips in May, andthe sunlight flickering through the branches of thetrees. In both respects, the Aberdeenshire Dee hasmuch to offer, J n d some of its canoeing enthusiastshave called it a 'magic' river.

For the canoeist, the Dee is a world of its own. Itscourse runs mainly away from the road, and there arefew houses along its banks. The banks often risehigher than the canoeist's eye level, restricting thepaddler's view of things beyond. This tendency is

encouraged by the Dee's often steep-sloping banks,and the very extensive wooded fringes. Few humanartefacts can be seen from the river and there is littletraffic noise, with such as there is being drowned bythe noise of the water or filtered out by the trees.There is a sense of isolation and tranquillity, and muchwildlife is seen. The river is now the only way oftravelling along the strath for any distance except ontarmac roads. Between the wooded banks of the Deelies the last lengthy woodland journey one can make inthese islands.

Canoeing is becoming more popular in Grampian.There are approximately 10 canoeing groups or clubsin the Region. Some of them are still small, but 5 havebeen formed within the last year or so, and theAberdeen Kayak Club has more than doubled itsmembership in the last year, from about 35 to nearly80. This is partly due to the publicity that the sport hasreceived on television.

The Dee also offers an excellent range of whitewaters. On its lowest stretches, where it is tidal, arequiet waters suited to the beginner. This stretch alsoharbours a recently formed group of racing paddlers.There are no real conflicts along this stretch. FromBanchory down, there is a stretch of water ofintermediate difficulty, but the stretch from Potarch toBanchory is one of the finest stretches of white waterin Britain. Above Potarch, the river is canoeable only inspate.

Several types of activity are found. There are compet-itions at manoeuvering on white water, slalomscentred at one spot, 'white water races' taking placeover long distances, and perhaps as part of a triathlon,and there is river touring. There is also some simple'mucking about in boats' in pools such as at Dinnet orAboyne. There is some conflict between canoeists andfishing interests, with at least 4 possible kinds ofdisturbance that might be caused by canoeists. Peoplesometimes confuse them, but they are as follows.

i. Disturbance of fish, ie the fish themselves are insome way significantly injured, or their growth orreproduction is harmed.

ii. Disturbance of fishing, ie the business of catch-ing fish is in some way hindered.

iii. Disturbance of fishers. This is a differentphenomenon, in which all that is disturbed is thesense of seclusion or peace on the part of thefisher.

iv. Disturbance of riparian owners. These peoplemay not even be there at the time the canoeistspass, but may feel some anxiety at the activity.

To the best of my knowledge, nobody has eversuggested that canoeing on a river like the Deesignificantly disturbed fish for any length of time. Thequestion of disturbance of fishing seems equallyproblematical. Canoeists report that they almost invar-

iably have cordial relations with fishers on the Dee, andthat the difficulties usually arise with the anxieties ofthe riparian owners. Some disturbance of wildlife onthe river may be caused by canoeing, but probably thishas little long-term effect, because canoeists landinfrequently, and then mainly on open grassy sectionsof bank.

Conflicts stem basically from touring, as slaloms andraces are arranged in co-operation with the riparianowners, with whom the main organizers go toconsiderable lengths to preserve good relations. Con-flict with touring canoeists is minimal because mostcanoeists touring on the Dee are local people, whocanoe mainly on Sundays when there is no fishing, orthey canoe outside the fishing season. Becausefishing on the Dee finishes in late September, themarvellous month of October is free for canoeing.Above Potarch, canoeing largely takes place in spateconditions under which fishing is temporarily sus-pended. Also, canoeing use of the Dee is still at a lowlevel compared with other rivers.

Thus, it may well be that, although the number ofcanoeists has increased and is increasing, conflict willnot become a serious problem, especially if canoeistscontinue to operate the courtesy code of waitingupstream till the fishers signal them to pass, and thenpassing quickly and quietly on the far side of the river.It has been suggested that a formal access agreementcould be arranged, but the Scottish Canoe Associationwould never recognize any agreement between localparties as binding on its membership nationally, andnobody can make an access agreement for the manytouring canoeists unattached to any club or formalgroup. No foreseeable agreement could guaranteeuninterrupted fishing.

7 Natural history study and wildlifeThe millionth visitor to see the ospreys at Boat ofGarten was clear proof, if any were needed, of theextremely widespread interest that the public has inwildlife. Visits to nature reserves figure large incountryside activities (Tables 1 & 2). However, it hasalways seemed to me to be a mistake to think thatwildlife has to be spectacular to hold the public'sinterest. Quite ordinary animals, such as red squirrels,are fascinating to people if a clear view of them isobtainable, and the wildlife does not have to becontinually in view. The memory of short encounterswith wild mammals and birds is often enduring. It isremarkable how a brief sight of an eagle soaring or thediscovery of a rich bank of primroses in flower is a highpoint in the day for many walkers or car-i3orne tourists.The river banks, because of their variety of habitats,and strips of ungrazed, unfertilized floras and frequentscrub, are rich in such memorable encounters.

Deeside, while it is rich in such resources, is very poorin visitor centres where wildlife can be viewed, orwhere general countryside interpretation is offered.

125

Crathes Castle is one of the few places catering forvisitors. The visitor centre at Loch Muick allows deerto be viewed at a distance, but is badly in need ofenhancement of its facilities, and is far off the maintourist routes. There is not a single place where thewildlife of the river can be viewed and where itssignificance is being interpreted to the general public.

8 The use of the river area by childrenThere is one last group of informal recreationists forwhom I would enter a plea. They seldom get men-tioned in surveys or recreational statistics, yet they arevery numerous. I refer to children and young adoles-cents at play. Having spent much of my spare timeover 20 years in youth leadership, I have always beenperplexed at why we are very bad as a society atcatering for youth, and so under-value those earlyadventurous experiences of children. It seems to methat we fail to cater for their informal activities, or moresimply in just leaving them room to get on withclimbing trees, making hideouts in the woods, illicitlyfishing in the river, and so on. These experiences areimportant for the growth of their imaginations, theirsense of adventure, and their love of countryside.

One of the most important parts of the whole riverbank is on the north bank, just above the old Bridge ofDee, in Aberdeen. It is a forgotten corner, but here thechildren from Garthdee and Kincorth come, escapingfrom the carefully laundered grass spreads of thehousing estates, and their boring adventure play-grounds. The children have no right to be there, ofcourse. The landowners have put up fences in the farpast. It's illegal what the children do. They are notsupposed to dangle worms in the water, or gowhooping through the woods disturbing the birds andbreaking the bushes. All up and down the river thereare similar corners sheltering small populations of thisendangered species, the joyful child. I live in fear asthe developers fill in the last bit of 'waste ground', aswe finish planning and planting, and conserving ourwildlife, that these places will have been decentlytidied out of the way, and with them the children.Whatever else this symposium decides to recom-mend, leave room for them.

9 General points on management and recreationAs the question of management always arises wherepublic access is considered, it is worth making a fewgeneral points. First, where 'mass access' is plannedto an area, eg a new picnic or camping site, then it.must be axiomatic that the staff and money tomaintain the area must also be provided.

The second general point is about the introduction ofrules such as bye-laws to regulate informal outdoorrecreation. Spontaneity and freedom are valued char-acteristics of such recreation. Indeed, they are some-where near the heart of the matter to many partici-pants, who are strongly antipathetic to restrictions.Further, such restrictions, if introduced, are often

126

costly and difficult to enforce. In the end, they dependon the voluntary co-operation of participants based onan understanding of the real need for rules, rather thanon simple obedience. It is often better, in fact, to usemore positive and perhaps more subtle managementmethods to achieve ends, such as leading a path off inan interesting diversion, rather than forbidding accessto a sensitive area.

Restrictions should be kept to a minimum, and onlyintroduced when it has been clearly demonstrated thata real conflict exists, and other methods have failed orare very likely to do so. Introduced without publiceducation as to the need for them, they may fail forlack of support or even be counter-productive. Un-necessary rules, or ill-conceived ones can, in fact,become a source of conflict where no real problemexisted before. The bye-laws introduced to the Muir ofDinnet National Nature Reserve seem a good exampleof unnecessary proliferation of such restrictions, creat-ing a potential for conflict.

Is there, in the light of this, a need for a 'river code' onthe Dee at present? I would suggest not. Conflictsbetween canoeists and riparian owners are handled byco-operation between canoeing event organizers andthese owners, and by the canoeists' own code ofconduct when touring. For other forms of recreation,the normal 'country code' seems able to handle suchproblems as might arise.

10 SummaryInformal recreation cannot be distinguished clearlyfrom formal recreation, but includes car-borne touring,walking, canoeing, camping and caravanning, naturestudy, photography, swimming, picnicking, and chil-dren at play. For most people, it is the broad sweep ofscenery that is important, with the river acting as amajor unifying theme in the landscape.

Extensive afforestation is replacing the varied hues ofthe open hill with a more uniform green, and roadsidetree planting, regeneration and afforestation are block-ing economically and socially important views.

For camping, caravanning, and picnicking, the Dee ischaracterized by the extremely limited public access toits banks, especially above Ballater. Very little of theriver bank is officially open to walkers. Most people areuncertain where they may walk. In some areas, deerfences and plantations reduce access.

For canoeists, the attractions of a river lie in the varietyand extent of white water, and the quality of the wholeambience of the river. For both of these, theAberdeenshire Dee is outstanding. Some conflictexists, but the pastime remains almost entirely locallybased, and the problem should not become serious.

From the point of view of all groups, except perhapsthe canoeists, the Dee is a much under-used, high-quality, resource. However, the level of participation inmost activities in the north-east is increasing, perhapssharply. It is suggested that there may be considerableunderlying unsatisfied local demand .for walking, pic-nicking, and some other pursuits.

Given this situation, and the increasing level of publicawareness of such issues as access to the country-side, and interest in wildlife, it is likely that pressure forpicnic sites, river bank walks, and such facilities willincrease. Nonetheless, rules and regulations to controlinformal recreation should be kept to a minimum.Ensuring the wisest use for the interests of everyonerequires the informed co-operation of participantsrather than their obedience to rules. At present, thereappears to be no proven need for restrictions in theuse of the river.

11 ReferencesSurvey Research Associates. 1981. Market survey of holiday-making and countryside recreation in Scotland. Scottish TouristBoard and Countryside Commission for Scotland. (Published private-ly.)

Walton, K. 1969. Royal Grampian country. Aberdeen: Departmentof Geography, University of Aberdeen.

Salmon catch statistics for the River Dee, 1952-83W M SHEARERDepartment of Agriculture and Fisheries for Scotland. Montrose

1 IntroductionThis paper examines the reported catches by each ofthe 2 legal methods of fishing for salmon within theestuarial limits of the Dee (rod and line, and net andcoble) in 1952-83. In Scotland, salmon are harvestedby bag and stake nets along the coast, and by net andcoble (seine nets) and rod and line in estuaries and infreshwater (Strange 1981).

2 HistoryFishing seasons and the number of netting stationshave both changed. In 1852, the season lasted from 1December to 20 September, and from 1902 from 11February to 26 August (nets) or 31 October (rods)(Grimble 1902). Up to 1872, netting extended toBanchory. However, as netting in the river declined(Oswald 1985), coastal fisheries increased. In the Deedistrict, extending 29 km south along the coast, therewere 102 bag and stake nets in 1882 and nearly 200 in1894 (Grimble 1902). A decline began in the late 1950sand has continued.

Published catch records, particularly after the mid-1850s and before 1952, are fragmentary. The growingsecrecy of net fishing proprietors and lessees stem-med from a hostility between netsmen and the riparianowners. The monetary value of leased fishings, whichperiodically came up for renewal, imposed furtherconstraints on the free exchange of information.Because of the many changes in the fisheries,including changes in the start and end of fishingseasons and the inclusion of kelts in earlier catchtotals, the early data can seldom be compared withmore recent information. Nevertheless, the availablehistorical records indicate that periods of low catchesare characteristic of the Scottish salmon fishery. Lowcatches appear to have been associated with changesin the timing of the runs. Following the decline of grilse(fish which have spent one winter in the sea) catchesin the late 1880s, the fishery remained depressed untilthe 1920s when the number of salmon (multi sea-winter fish) caught in the spring began to increase. TheFishery Board for Scotland Report for 1921 stated that'the season of 1921 showed a remarkable recoveryfrom the depressed condition which had been re-ported for some years'. Six years later, the Inspector ofSalmon Fisheries was able to report tHat 'in 1927 thesalmon fisheries of Scotland were even more success-ful than in 1926. The catch sent to the marketamounted to 2910 tons (gross), a figure which has notbeen reached since 1896'. As described, the situationexperienced by the salmon fishermen (rod and net) ofthat time appears to be rather similar to the presentposition.

127

3 Material3.1 Traditional fisheries

Before 1952, there was no statutory obligation tomake data on catch and effort on salmon catchesavailable to the Department of Agriculture and Fisher-ies for Scotland (DAFS). However, since then, thisinformation has been provided annually, in confidence,to the Department as a statutory requirement underthe Salmon and Freshwater Fisheries (Protection)(Scotland) Act 1951. Catches, divided between sal-mon, grilse and sea trout, are recorded monthly bynumber, by weight and by method of capture (fixed-engine, net and coble, and rod and line). In addition,net fisheries are required to supply the minimum andmaximum number of netting crews fishing eachmonth, and fixed-engine (ie fixed nets, mainly in thesea) fisheries the minimum and maximum number oftraps operated. No effort data are requested from rodfisheries. As a supplement to these data, the annualcatches of salmon and grilse taken during the period1872-1984 in the net and coble fishery operated byAberdeen Harbour Board were readily made available.

The catch figures examined are those supplied, inconfidence, to DAFS by the proprietors and lessees offishings within the estuarial limits of the Dee. Catchestaken outwith the river but still within the area of theDee District Salmon Fishery Board have been ex-cluded from the analysis mainly because they couldcontain a substantial proportion of fish not native to theDee. To maintain confidentiality, it has not beenpossible, except in one instance, to quote actual catchfigures. The Aberdeen Harbour Board, who suppliedthe catch data from its net and coble fishery datingback to 1872, has given permission to use thisinformation without restriction. Otherwise, all catcheshave had to be expressed as percentages, which isunfortunate because it obscures the relative mag-nitude of the catches taken by the different gears.Fishing effort has been omitted from the analysismainly because it is not available for the rod fishery,and the units which have been used to collect nettingeffort have been found to be insensitive to observedchanges in the fishing intensity.

3.2 Interception fisheries

Since the early 1960s, a number of high-sea fisherieshave caught fish of Dee origin, among others. Salmoncatches in excess of 100 tonnes (t) were first recordedat west Greenland in 1961, and over 2600 t werereported caught in 1971. Recently, agreed quotas forthis fishery were increased from 1190 t (1976) to1270 t (1981), primarily to take account of thedecrease in the number of fish caught per tonne as aresult of the later opening date. Subsequently, the

128

quota was decreased to 870 t (1984) in order to affordsome protection to multi sea-winter salmon stocks.Fishing normally takes place between August andOctober. For comparison, 1010 t were caught by allmethods in Scotland in 1976. However, the agreedquota for Greenland has not been caught in the last 4years, and in 1983 the shortfall amounted to some980 t. Apart from fish of Greenland origin, no fishwhich would have returned to 'home waters' as grilseare taken in this fishery, and the average stockcomposition is of 40% North American and 60%European origin, as determined by scale discriminantanalysis. Presently, it is not possible to break down theEuropean component into countries of origin.

Since 1968, a long-line fishery has developed about112 km north of the Faroes, with catches up to 40 t in1977. From 1977-82, Danish vessels took part in thisfishery. Total landings rose to about 1000 t in 1981,partly because the season was lengthened in 1979 torun from October to June, and because the number ofbOats increased from 9 in 1977 to 44 in 1981. Inaddition, the fishery moved further north (up to576 km) in 1977-83, and at this time the averageweight of the fish increased. Until the late 1970s, mostfish caught in this fishery were grilse. The recentextension in fishing areas resulted in the exploitationof salmon with a higher average sea age. In March1981, and in 1982/83 and 1983/84, about 99% of thecatch landed was of 2 or more sea-winter fish and onlyabout 1% was one sea-winter fish. In addition, about5% of the catch was discarded and consisted pre-dominantly of one sea-winter fish. In 1982, theFaroese government agreed to a total catch limit of750 t and to 625 t (25 boats allowed a quota of 25 teach) in 1983 and 1984. All boats participating do sounder licence, and from 1983 the fishery was confinedwithin the Faroese 320 km limit under the terms of theConvention for the Conservation of Salmon in theNorth Atlantic Ocean. At present, it is not possible toapportion the catch to countries of origin.

Two other fisheries are also known, from the captureof adults tagged as smolts, to exploit salmon ofScottish, including Dee, origin. These are drift netfisheries off Ireland and in the North Sea off North-umberland and Yorkshire. Both these fisheries tend tocatch a preponderance of grilse. In 1982, these 2fisheries caught about 710 and 177 t respectively.Although the rivers of origin of these fish are notprecisely known, present data suggest that thosesituated between the Ugie and the Tweed make thebiggest contribution to the fishery located off theNorthumberland and Yorkshire coasts. In 1961 and1962, a legal drift net fishery operated off the Scottishcoast, but this was banned following the introductionof legislation in 1962. However, an illegal fishery usingsimilar techniques continues to operate off the Scot-tish coast, in spite of a considerable enforcementeffort.

3.3 Fishing season

Fishing seasons on the Dee now extend from11 February to 26 August for netting (usually stoppingearlier) and for angling between 1 February and30 September. In addition to the Annual Close Time,there is a Weekly Close Time for nets from noonSaturday to 0600 h on Monday and for rods on Sunday.

Since 1983, DAFS staff have sampled the Dee net andcoble catch weekly to obtain information on theweight, length, and the river and sea ages of the fishbeing exploited.

3.4 Biases in data

3.4.1 Relationship between catch and stock

Salmon runs do not necessarily conform to fishingseason. In 1981, 1982 and 1983 (the years for whichdata are available), approximately 40% of the NorthEsk spawning stock moved upstream over the Logiefish counter (3 km upstream from the sea) after theend of the net fishing season. Fish which enter riversoutwith the fishing season do not contribute tocatches taken in rivers. Thus, changes in the timing ofthe runs could bias estimates of annual stock abund-ance based solely on catch data in the fishing season.Futhermore, it is unlikely that the amount of fishingeffort or the rate at which the stock is exploited or thecatchability of fish remains constant throughout afishing season or from one season to another.

Biological differences between catch and stock canalso exist, for a number of reasons, including gearselectivity and differences in the level at which thevarious components of the stock are harvested. Also, arelationship is known to exist between river and seaage and the calendar date when fish belonging toparticular river and sea age groups return to freshwater (Shearer 1984). The age composition of thecatch will be biased towards those age groups whichreturn in greatest proportion during the fishing season.

The rod catch could be particularly sensitive to physicalchanges in the river system, such as the gravelling ofpools and changes in discharge rates and temperature.Furthermore, any change in the sea age compositionof stock could be important because it would alter thetiming of the runs and, therefore, the availability ofcatchable fish.

In addition, losses attributable to non-catch fishingmortality include the non-reporting of catches by someowners and lessees of fishings and the number of fishcaught annually by illegal methods. These figures arenot included in the official returns and no reliableestimate is available.

3.4.2 Limitations of the data

i. Annual fluctuation in fish numbers

All the data discussed refer to reported catch figures.Fluctuations occur in annual catches for a number ofreasons. A bad year in a single river or fishery must not

be assumed to be evidence of a general deteriorationof the stocks of salmon, although it is reasonable toassume that increased catches are usually related toan increase in the total number of salmon available forexploitation during the fishing season. The spawningescapement, ie the proportion of returning salmon thatis not taken by nets or rods, cannot be estimateddirectly from numbers of fish caught, because a largenumber could migrate after the end of the fishingseason, having made no contribution to catches, andexploitation rates could vary.

'Grilse error'In the catch returns submitted to DAFS, lessees andowners of salmon fisheries, with few exceptions, haveseparated their catches into salmon and grilse on thebasis of weight. Fish weighing less than 3.6 kg havebeen classed as grilse and the rest as salmon whenboth sea age groups were present. In 1952-83, theproportions of fish classed as grilse by this method inthe annual catches have varied as a result of changesin the growth rate of grilse in the sea. Furthermore,since grilse generally increase in weight as the seasonadvances, the magnitude of this reporting error doesnot remain constant throughout the fishing season. InAugust, 'salmon' catches will contain relatively moreover-weight grilse than in June or July. In general,those years when grilse were most abundant werealso characterized by above average proportions ofover-sized grilse.

4 Analyses of available dataSalmon catches can conveniently be broken down into3 major components. These are (i) grilse, (ii) spring fish(multi sea-winter fish caught up to and including 30April), and (rii) summer fish (multi sea-winter fishcaught after 30 April). Typically, grilse and summersalmon come into rivers from May to December andspring fish from November to April. Some fish runninginto freshwater in November 1984 can remain in theriver and not spawn until October 1985. If these fishmove upstream of the netting zone before 11 February,they will not be available to the nets but they will beavailable to angling for all of the fishing season,although not necessarily catchable.

Total catches for the Dee of salmon plus grilse,salmon, and grilse caught by the 2 methods of fishingare expressed as percentages of the 1952-83 means inFigure 1. Figure 2 shows these data broken down intotheir angling and netting components. Figure 3 showstotal numbers caught for all Scotland, and Figure 4illustrates the total, rod and line, and net and coblespring catches as percentages of the respectivecombined catches. Figure 5 gives the catch data forthe net and coble fishery operated in the estuary of theDee by the Aberdeen Harbour Board from 1872 to1984 expressed as 5-year means, and Figure 6 showsthe proportions of grilse in these catches. Figures 7-9compare data from the Dee with those from othernearby rivers.

129

The river and sea age composition and the averageweight of the various sea age groups of fish caught in1983 by the nets are compared with similar datarelating to the 1920s (Menzies 1922; Menzies &Macfarlane 1924, 1926, 1927, 1932).

5 Results5.1 Catch data

The total catches of salmon, grilse (uncorrected forgrilse error) and salmon plus grilse fluctuated widelybetween years (Figure 1). Because grilse catchesthroughout the period 1952-83 were relatively small inrelation to total catches, their influence when added tothe salmon catch was negligible. Salmon and grilsecatches combined, and salmon catches alone fell into2 distinct groups, 1952-66 and 1967-83. There was abig decrease between these 2 periods, and the groupswere significantly different from each other at the0.1% level. Whereas most annual catches prior to1966 were above the long-term mean, all catches but2 after 1966 were below it. There was no obvioustrend in 1952-66, but, following the big decrease, anupward trend was apparent in 1967-83. Grilse catchesfluctuated widely but showed an upward trend latterly.The smallest and largest catches of grilse were in 1964and 1973 respectively, and the second largest in 1982.

Catches of salmon plus grilse and salmon taken by the2 methods of fishing fell into the same 2 groups but,whereas the decrease in net catches began about1963, it was delayed until 1966 in the case of rodcatches (Figure 2). Subsequently, the rod catchesfluctuated around the new lower level, but the netcatches tended to show a slight increase. Nonethe-less, most annual catches in the second periodremained below the long-term mean. The differencebetween the results from the 2 catching methodsoccurred because relatively few (5-21% of the total)grilse were caught by rod compared with by net,especially since the mid-1960s. However, since theearly 1960s, the increase in the average weight ofgrilse, and presumably also in the number of grilse of3.6 kg and over, led to a large but variable over-statement of the number of salmon reported caught.These data need further analysis.

For the whole of Scotland, while the catches by the 2netting methods varied together, the trend shown bythe rod and line catches was rather different. Catchesin the net fisheries increased in 1962 and this newcatch level was maintained until about 1975. Since1976, catches by both types of netting have declinedand recent figures are among the lowest recordedsince 1952. On the other hand, annual catches ofsalmon by rod and line have not decreased signif-icantly in recent years (Figure 3). Catches of springsalmon taken by all methods showed a gradual butsteady downward trend which began during the early1960s. However, when the rod and line catches wereexamined separately, no trend was evident. Neverthe-less, it is perhaps worth noting that, while spring

130

Salmon + grilse Salmon

240

180

60

Grilse

240

180

120

01950 1960 1970 1980 1990

Years

60

01950 1960 1970 1980 1990

Years

Figure 1. Total catches of salmon, salmon plus grilse,and grilse on the Dee, expressed as proportions of themeans for 1952-83 (at 100%, the horizontal line is thelong-term mean)

catches in the Dee in 1981, 1982 and 1983 wereproportionally the smallest in the time series, the valuefor 1980 was the highest since 1963 (Figure 4). Theproportion of the total salmon and grilse catch in theDee taken by rods increased between 1952 and 1983(40% in 1952, and 78% in 1971).

In 1872-1984, annual net and coble catches of salmonincreased (Figure 5), mostly due to more salmon butwith more grilse after 1955. However, there were bigfluctuations and the mean catch of salmon in 5-yearperiods fell to its lowest level during the last 50 years

240

180

60

01950 1960 1970 1980 1990

Years

in the period 1970-74; since then, it has steadilyincreased. On the other hand, catches of grilseremained remarkably stable until 1960-64, since whenthey too have shown a steady increase.

Expressed as percentages, the grilse proportion of thetotal salmon plus grilse catch varied widely (Figure 6).From 1875, a decline continued until the second half ofthe 1920s. Subsequently, the proportion of grilse inthe net catches increased nearly to the same level asin the late 1800s. Because the catchability of grilse byrod appears to be considerably less than that ofsalmon, any factor which increased the ratio of grilseto salmon in a stock returning to spawn withoutchanging its overall numerical strength will depressthe rod catch without having a similar effect on netcatches.

5.2 Biological data

In 1983, grilse comprised 49% of the total net catchbut in 1921-82 they did not exceed 32% in any year.Nevertheless, grilse were relatively more abundant inMay in the earlier period, contributing 8% to the totalcatch in that month compared with only 2% in 1983.The position was reversed later in the season and byAugust the corresponding values were 32% and 62%,suggesting that the peak of the grilse run into freshwater had shifted and was now nearer the end of thenet fishing season than formerly.

Table 1 shows that, in months with comparable data,the proportions of grilse and 2 sea-winter salmon weresimilar in February-July in the 1920s and in 1983.However, in August 1983, grilse were dominant in thecatch. The proportion of 3 sea-winter fish in the overallcatch did not change, but there were small differencesin the manner in which this age group was distributedthroughout the season. Some fish which had spent 4

Salmon + grilse by rod Salmon + grilse by net300 300

225 225

150

-

- - -

a)coco

a)a

75 75

0 01950 1960 1970 1980 1990 1950 1960 1970 1980 1990

Years Years

Salmon by rod Salmon by net300 300

225 225

a) a)co c)I2 coc ta) 150 w 1502 00 iiio_ a

75 75

0 01950 1960 1970 1980 1990 1950 1960 1970 1980 1990

Years Years

Grilse by rod Grilse by net

300 300

225 225

75 75

- -

131

0 01950 1960 1970 1980 1990 1950 1960 1970 1980 1990

Years Years

Figure 2. Total catches on the Dee according to method of capture, expressed as proportions of means for 1952-83

132

Table 1. Percentage of fish of different sea age caught in each month, February-August, in 1921-25 and in 1983(sources (data for 1921-25): Menzies 1922; Menzies & Macfarlane 1924, 1926, 1927, 1932)

Percentage monthly sea age composition of net and coble catch

1950 1960 1970 1980Years

Figure 3. Total catches of salmon plus grilse in Scotland in 1952-83

winters in the sea were present in the catches taken inthe 1920s (<1% of the total catch) but not in 1983.

Two-year-old smolts formed the most numeroussingle age group in the catches in the 1920s and 1983.However, approximately 78% of the catch in the1920s had migrated to sea after spending 2 years infreshwater, compared with 56% in 1983. The increasein the average smolt age was not limited to oneparticular sea age group, but it was most marked in the2 sea-winter age group (Table 2).

Table 2. Smolt age (%) of fish caught in the Dee by net and coble(sources (data for 1921-25): Menzies 1922; Menzies &

1 2

Sea age (years)

3 4

Month 1921-25 1983 1921-25 1983 1921-25 1983 1921-25 1983

February 90.0 9.9 0.1

March 89.3 82.3 10.5 17.7 0.2

April 0.2 92.4_ 70.6 7.4 29.4 0.1

May 8.1 1.9 86.4 89.4 5.5 8.7 0.1

June 79.3 75.6 19.4 22.8 1.3 1.6 <0.1

July 83.3 85.5 15.4 14.1 1.2 0.4 <0.1

August 31.8 61.9 59.7 38.1 8.5

640 000

480 000

a)ca

All methods

Tow320 000

-c?

160 000 Fixed engineNet & coble

Rod & line

Macfarlane 1924, 1926, 1927, 1932)

River age(years) 1921-25 1983

1 0.8 2.22 78.4 56.03 20.4 36.44 0.4 5.4

All methods Rod & linetoo too

75 75

a 0a) a)aZ V,a) 50 cco 500ir3 2a_ 0la_ •

25 25

100

0 01950 1960 1970 1980 1990 1950 1960 1970 1980 1990

Years Years

Figure 4. Catches of spring salmon on the Dee,

Net & coble expressed as proportions of the mean for 1952-83

The other biological parameter examined was averageweight (Table 3). Grilse caught each month in May-August 1983 were on average heavier than compar

75

-able fish taken in the 1920s. Over the season, thedifference between the 2 mean values was 0.4 kg orapproximately 20%. Although there was no difference

0.) between the annual average weight of the 2 sea-t:»winter fish caught in the 1920s and 1983, those caught

.2)50 in 5 of the 6 months for which comparable data wereavailable were lighter on average in 1983 than in the1920s. If data had been available for February 1983, itis a reasonable assumption that the overall figure for

25 that month would also have been less. The differencesobserved were greater in July and August than inMarch and April. Three sea-winter fish caught in 1983were on average lighter both at the monthly andannual levels, compared with the same age group

1950 1960 1970 1980 1990 caught in the 1920s. In most months, these differ-Years ences amounted to more than 1 kg.

Table 3. Average weight (kg) of fish caught in the Dee by net and coble (sources (data for 1921-251: Menzies 1922;Menzies & Macfarlane 1924, 1926, 1927, 19321

133

134

Salmon6000

4500

_oE 3000

1500

Grilse

0

6000

4500

J=IE 3000z

1500

0

Period

Period

Figure 5. Catches of salmon and grilse by net and coble on the Dee in 1872-1984, expressedas 5-year means

Salmon + grilse

6000

4500

en

-a 3000

1500

6 Discussion6.1 Main features of these data

Period

Figure 5 (continued). Catches of salmon and grilse by net and coble on the Dee in 1872-1984,expressed as 5-year means

5.3 River characteristics

Not every river and tributary contains spring salmon,summer salmon and grilse. In those that do, theproportions can vary both between rivers and betweenyears. Although all 3 groups are found in the Dee, thefishery is mainly in the spring. Figures 7, 8 and 9compare data from the Dee with those from neigh-bouring rivers which support both net and coble androd and line fisheries. The data show that catches in allrivers have fluctuated, without following the samepattern between years. For example, catches in theTay fisheries in the 1960s and 1970s reached a peak ata time when other rivers yielded low catches. Therelative importance of the contribution of one sea-winter fish (grilse) to total catches has varied bothbetween rivers and between years. Grilse have beenimportant in the Findhorn fishery and relatively unim-portant in the Dee fishery.

The main features of these data are the following.i. There was a decrease in reported catches in the

Dee in the 1960s which has not been made up.

ii. This decrease was shown by both net and rodfisheries.

135

iii. However, in the case of data for Scotland as awhole, the main decrease occurred in the early1970s, and was mainly confined to the netfishery. There has been little change overall inthe numbers caught by rods.

iv. Catches of salmon, but not of grilse, taken in thenet and coble fishery operated by the AberdeenHarbour Board have shown an increase over thelast 100 years, despite large annual fluctuations.Grilse catches only show an increase from the1950s.

v. The proportion of grilse caught by the net andcoble fishery operated by the Aberdeen HarbourBoard has fluctuated, with peaks in the 1870sand the 1970s and a big trough in the 1920-30s.

vi. Decreases in catches by all methods haveoccurred in 4 out of 5 east coast rivers, but not inthe fifth, where there was an increase around1970 in contrast to the others.

vii. The proportion of grilse caught by all methodshas fluctuated in all rivers but is increasing in theDee, a trend sustained since the 1950s.

viii. However, grilse are relatively unimportant in theDee catches and the main feature characterizingthe Dee fishery is a decline over the last 20 yearsin the spring catch of multi sea-winter salmontaken by the net and coble fishery.

136

100

75

25

1920 1930 1940 1950 1960 1970 19801924 1934 1944 1954 1964 1974 1984

Period

Figure 6. Proportion of grilse caught by net and coble, expressed as 5-year means, in 1872-1984

6.2 Stock genetics

Salmon biologists tend to describe individual riverstocks in terms of salmon and grilse. The salmonfraction can be broken down into spring and summercomponents, depending on the calendar date whenthe fish return to fresh water. At present, it is not at allcertain whether these groups are genetically distinctor whether the distinction is based on environmentaland other factors to which the fish were exposedeither during their early freshwater life or their periodin the sea. Nevertheless, at spawning time, there issome degree of isolation because spring salmon arenormally found in those tributaries which join the mainriver nearer its source than those frequented bysummer salmon and grilse. In addition, spring salmontend to spawn earlier.

6.3 Effort

It was not possible to ascertain whether any of theobserved catch trends were related to changes ineffort or to the availability of salmon. Neither, with the

data available, is it possible to quantify the effectswhich observed reductions in net fishing effort havehad on total catches. Stations which stop fishing firsttend to have the lowest catch rates. Therefore,reductions in total catch due to the closing of fishingstations may not be directly proportional to the numberof stations which have closed. No data describing theeffort for rod fisheries are available, because it cannotbe assumed that all anglers and tackle are equallyefficient. Nonetheless, variation in effort possiblypresents a major difficulty in interpreting the data.

6.4 Effects of Faroese and Greenland fisheries

Although it is by no means certain that all salmon feedat some time during their life cycle in either of these 2areas, estimating total losses to European home waterstocks for each tonne landed from the Faroese andGreenland fisheries presents no great difficulty.However, the biological data currently available areinsufficient to calculate these losses at the individualriver level. Nevertheless, as both Faroese and Green-

Findhorn240

60

180

co

120ca

60

01950

Tay240

•• ••

1960 1970 1980 1990Years

Spey240

180 180

60

o 01950 1960 1970 1980 1990 1950 1960 1970 1980 1990

Years Years

Dee N. Esk240 240

180

60

OM=

137

1950 1960 1970 1980 1990Years

Figure 7. Total annual catches of salmon for 5 eastcoast rivers, expressed as proportions of the mean for1952-83

180 land fisheries exploit salmon rather than grilse, anyeffect on rivers possessing only a grilse stock is likelyto be marginal. The results of smolt tagging experi-

o) ments in home rivers and adult tagging experiments at120(.) Greenland suggest that most salmon of Scottish origin

which feed off the Greenland coast have come fromtributaries normally favoured by spring, rather thansummer, fish. Furthermore, the majority of fish tagged

60 at Greenland and recaptured in home waters werecaught in the spring rather than the summer of theyear following that in which they were tagged (Jensen1980). Because the Dee fishery is largely dependent

0on multi sea-winter fish returning to the home waters

1950 1960 1970 1980 1990 in the spring, the effects of the Faroese and GreenlandYears fisheries on Dee stocks are likely to be greater than the

138

Findhorn240

180 180

0 01950 1960 1970 1980 1990 1950 1960 1970 1980 1990

Years YearsDee N. Esk

240 240

180

a) a)a) a)LS co

tc 120 120a) co

å)a)a_ a_

70

0

1950 1960 1970 1980 1990Years

Figure 8. Total annual catches of grilse for 5 east coastrivers, expressed as proportions of the mean for1952-83

60 60

1950 1960 1970 1980 1990Years

Spey

240

180

60

average figure (mostly for summer fish only) derivedfrom the estimated total European loss.

6.5 Sea temperature

After studying Aberdeen Harbour Board catch datarecorded between 1872 and 1960 and the sea surfacetemperatures at Grimsey, north Iceland, over roughlythe same time period, Martin and Mitchell (1985)suggested that the sub-arctiC sea temperature,although not influencing the total number of fishreturning from particular smolt year classes, has amajor influence on whether a fish returns as a grilse ora salmon. During periods when sub-arctic temper-

Findhorn Spey300

225

Co

150

75

225

139

0 o1950 1960 1970 1980 1990 1950 1960 1970 1980 1990

Years Years

Tay Figure 9. Total annual catches of salmon plus grilse for300 5 east coast rivers, expressed as proportions of the

mean for 7952-83

atures were below average, the percentage of fishreturning as grilse increased. Since 1960, the grilseproportion of the catch returning after one sea-winterhas risen as the surface sea temperature in thesub-arctic has fallen steeply.

75

1%/\1)

6.6 Mortality

For most years in 1961-76, the estimated natural seamortality of North Esk salmon varied between 54%and 79% (Shearer 1984). In 1980-82, it increased to

o 86% and did not fall below 83% in any year. During the

1650 1960 1970 1980 1990 same period, the smolt production of the North Esk didYears not show any marked downward trend.

140

6.7 Age and weight

The increase in the average smolt age noted in theDee net catch between the 1920s and 1983 did notoccur in the Spey. Both at the Spey, and at the NorthEsk in 1964-84, the average smolt age decreased. Theincrease in the contribution which grilse made to theDee August catch, the later arrival of the grilse run atthe net and coble fishery, and the increase in theirmonthly average weight in 1983 compared with the1920s are similar to data from catches in the Spey andNorth Esk nets over similar periods. The generaldecrease in the average weight of the Dee multisea-winter fish did not occur at either of the other 2 netand coble fisheries.

6.8 Juvenile migrant production

Investigations carried out in the Girnock burn, atributary in the headwaters of the Dee, in 1966-76have shown that, although the number of female adultsalmon spawning in that burn each year varied widely(28-127), the annual variation in the number of juvenilemigrants produced was much smaller (2900-5600). Asa result, it was suggested that there may be noadvantage in allowing ova deposition to exceed a levelaround 200000. This number would give rise to anaverage of about 4000 juvenile migrants per season.The number of juvenile migrants leaving the Girnockburn has not been limited by the number of spawnersin recent years. In most years, the Girnock burnappears to have excess spawning. Therefore, extracropping of adults would seem possible in some yearsas the burn appears to have more than sufficient adultfish returning to spawn (Buck & Hay 1984).

6.9 Conclusions

Climate and ocean conditions interact with density-dependent growth and smolt production to produceshort-term and long-term fluctuations in abundanceand yield of salmon and the proportion of different agegroups of fish occurring in individual years. It remainsto be seen whether the marked reduction in theGreenland fishery in 1983 and 1984 will have anyeffect on Dee salmon catches. However, the only firmdata on fish numbers are those in Figure 3 for thewhole of Scotland and in Figure 5 for the Dee nets.They show for all Scotland that the big increase insalmon caught in the late 1960s and early 1970s wasnot sustained and that numbers are now back to thelevel of the 1950s. However, for the net and coblefishery operated by the Aberdeen Harbour Board, thetrend over the last 100 years is for a steady increase incatches. Firm conclusions about the rod fishery cannotbe drawn without more precise data on catch andeffort. Although the sea ages of salmon caught off thewest coast of Greenland and north of the Faroes aresimilar to those caught on the Dee, it has not beenpossible to relate the decline in the total Dee springcatch directly to either of these 2 fisheries.

7 Summary7.1 This paper analyses catch statistics from the

Dee, mainly for 1952-83, but includes data from

the net and coble fishery operated by theAberdeen Harbour Board from 1872. The dataare analysed in terms of percentage variationfrom annual or longer period means. The relativenumbers of fish caught cannot be given. It ispossible to analyse the catch data qualitatively interms of relative changes in the different com-ponents of the population and to establishwhether trends occur in some or all of thesegroups.

7.2 Catches on the Dee of salmon, grilse, andsalmon plus grilse taken by all methods fluc-

. tuated widely between years. Catches of sal-mon, and salmon plus grilse, prior to 1966,except for one year, were above the long-termmean, while catches after that date, except for 2years, were below it. No marked downwardtrend was noted in 1952-66 but the trend hasbeen upward since 1970.

7.3 Over the period 1952-83, grilse made a relativelysmall contribution to both the rod and line andnet and coble catches.

7.4 Since 1960, spring salmon net and coble catcheshave gradually declined, but no trend was evid-ent in the corresponding rod and line catch.

7.5 Each year, the rods caught between 40% and78% of the total catch of salmon and grilse.

7.6 A marked upward trend was observed in the netand coble catches in 1872-1984. The ratio ofgrilse to salmon in the catches varied widely,principally due to variations in the salmon ratherthan the grilse component (Figures 5 and 7).

7.7 Although the grilse caught by net and coble in1983 were on average heavier than in the 1920s,the multi sea-winter sea age groups of salmon innet catches were on average lighter.

7.8 Fish which had spent 4 winters in the sea weremissing from the net and coble catch in 1983,and the peak of the grilse run may now be laterthan in the early 1920s.

7.9 The average smolt ages of the one and multisea-winter fish caught in 1983 were greater thanthe corresponding figures for captures in theearlier period.

7.10 The data available from the Girnock Burn suggestthat juvenile migrant production is not presently alimiting factor.

7.11 Historical records suggest that not only havesalmon catches fluctuated widely but also thetiming of the return migration to fresh water.

7.12 Present data are insufficient to relate the allegeddecline in the spring run of salmon in the Dee tothe operation of the Faroese and Greenlandfisheries which are known to catch fish originat-ing in eastern Scotland.

7.13 The main requirements now for further analysisare informationon the effort put into catchsalmon by each method, including rod and line,and for consistent catch reporting.

8 AcknowledgementsI am most grateful to the Aberdeen Harbour Board, forpermission to use the catch data from the net andcoble fishery operated by the Board in the estuary ofthe Dee.

9 ReferencesAnon. 1825. Minutes of 1825 Salmon Commission. Edinburgh.

Buck, R. J. G. & Hay, D. W. 1984. The relation between stock sizeand progeny of Atlantic salmon, Salmo salar L., in a Scottish stream.J. Fish Biol., 24, 1-11.

Fishery Board for Scotland. 1922. Annual report, 1921.

Fishery Board for Scotland. 1928. Annual report, 1927.

Grimble A. 1902. The salmon rivers of Scotland. London: KeganPaul, Trench & Trubner.

Jensen, J. M. 1980. Recaptures from international tagging experi-ment at west Greenland. In: ICES/ICNAF joint investigation on NorthAtlantic salmon, edited by B. B. Parrish & SV. AA. Horsted, 114-135.Copenhagen: International Council for the Exploration of the Sea.

141

Martin, J. H. A. & Mitchell, K. A. 1985. Influence of seatemperature upon the numbers of grilse and multi-sea-winterAtlantic salmon (Salmo safer) caught in the vicinity of the River Dee(Aberdeenshire). Can. J. Fish. Aguat. Sci., 42, 1513-1521.

Menzies, W. J. M. 1922. Salmon investigations in Scotland, 1921.1.Salmon of the River Dee (Aberdeenshire). Salm. Fish., Edinb.,3-48.

Menzies, W. J. M. & Macfarlane, P. R. C. 1924. Salmoninvestigations in Scotland, 1922. II. Salmon of the River Dee(Aberdeenshire). Salm. Fish., Edinb., (3), 5-52.

Menzies, W. J. M. & Macfarlane, P. R. C. 1926. Salmoninvestigations in Scotland, 1923. I. Salmon of the River Dee(Aberdeenshire). Salm. Fish., Edinb., (4) 4-46.

Menzies, W. J. M. & Macfarlane, P. R. C. 1927. Salmon of the RiverDee (Aberdeen) 1924. Salm. Fish., Edinb., (3), 3-30.

Menzies, W. J. M. & Macfarlane, P. R. C. 1932. Salmon of the RiverDee (Aberdeen) in 1925. Salm. Fish., Edinb., 1931, (4), 3-22.

Oswald, J. 1985. Fishery management. In: The biology andmanagement of the River Dee, edited by D. Jenkins, 117-126. (ITEsymposium no. 14.) Abbots Ripton: Institute of Terrestrial Ecology.

Shearer, W. M. 1984. Research activities within the Spey SalmonFishery District. Annu. Rep. Spey District Salmon Fishery Board,1-44.

Shearer, W. M. 1984. The natural mortality at sea for North Esksalmon. (ICES CM 1984/M:23.) Copenhagen: International Councilfor the Exploration of the Sea.

Shearer, W. M. 1984. The relationship between both river and seaage and return to home waters in Atlantic salmon. (ICES CM1984/M:24.) Copenhagen: International Council for the Explorationof the Sea.

Strange, E. S. 1981. An introduction to commercial fishing gear andmethods used in Scotland. Scott. Fish. Inf. Pam., no. 1.

142

The status of the River Dee in a national andinternational contextP S MAITLANDInstitute of Terrestrial Ecology, Edinburgh

1 IntroductionThe classification of river systems has receivedconsiderable attention from freshwater scientists,and a wide variety of parameters has been proposedas the basis of different schemes of typology. Earlyworkers used physical attributes such as flow andtemperature (Ricker 1934), geological origin and sub-strate (Carpenter 1927; Butcher 1933) and the erosionand deposition of sediments (Moon 1938). Otherworkers have used biological features, such as fish(Huet 1946), invertebrates (lilies 1953) and plants(Has lam 1982), and there is often a reasonable cross-correlation among the different schemes (Pennak1971). Many of the different schemes have beenreviewed by Hawkes (1975) and recent proposals(Maitland 1979; Wright et al. 1984) have movedtowards sophisticated multivariate analyses usingboth physico-chemical and biological data. Constantproblems are that the characteristics of rivers change(sometimes dramatically) from source to mouth(Maitland 1966; Cummins et al. 1966), and that manydifferences in the biota (eg the fish: Maitland 1985)may be due to historical and not ecological factors.

The conservation of fresh waters has received moreattention in recent years, but the attention given torunning waters has lagged far behind that given tostanding waters. Thus, in Project Aqua (Luther &Rzoska 1971), 'a source book of inland waters pro-posed for conservation', a total of 108 running watersis listed (Table 1), compared with a standing water listof 526. Within the United Kingdom, about 52 standingwaters and one running water are listed for England,about 14 standing waters and no running waters forWales, and about 15 standing waters and no runningwaters for Scotland. The single running water listed forthe United Kingdom (Bere Stream) is actually not ofparticular conservation importance.

Table 1. Fresh waters listed for their conservation value

Source Running waters Standing waters

Luther and Rzoska (1971)International 108 526Great Britain 1 81

Ratcliffe (1977)Great Britain 19 104

In Great Britain, the most recent and authoritativereview of sites of national importance is that ofRatcliffe (1977). It gives less attention to runningwaters than to standing waters, but the coverage isnevertheless much better than in most comparable

documents. In the index of open water sites (grades 1and 2), there are 104 standing waters and 19 runningwaters. Eighteen of the standing waters and 2 of therunning waters are regarded as internationally impor-tant. The Dee (Aberdeenshire/Kincardineshire) is listedas an upland running water of grade 1 national (but notinternational) importance.

2 The River DeeThe Dee rises as a large number of streams drainingthe east central highlands of Scotland, some of whichform the outflows of high-altitude lochs such as LochEtchachan and Loch Muick. The traditional source isthe Springs of Dee, but the highest sources are foundon the slopes of Ben Macdui within the CairngormsNational Nature Reserve (Ratcliffe 1977). Compared tothe source streams of the Spey, within the Cairn-gorms, those of the Dee are more precipitous, andstreams, such as the Allt a' Choire Mhoir, after risingon a gently sloping boulder field at 1220 m above ODthen plunge 600 m down the 40° slopes of the LairigGhru. In its upper reaches, the substrate of this streamis granite gravel and stones, but on the steep slopesthe stream cascades over bedrock and boulders. Inboth sections, most of the rock surfaces are covered indense growths of bryophytes such as Scapaniaundulata, Jungermannia cordifolia and Marsupellaemarginata. The maximum width here is about 3 mand the greatest depth only 20 cm. The water is low inpH (5.1) and extremely deficient in dissolved salts, theconductivity being low (12-19 .1,mhos), but nitratelevels are relatively high (0.24-0.25 mg I-1 NO3-N).For most of the year, the flow consists largely ofsnowmelt and even in late summer the temperaturerarely rises much above 5°C. The invertebrate faunaincludes species of Oligochaeta, Plecoptera, Tricladi-da, Trichoptera and Diptera typical of arctic-alpinestreams, with the larvae of orthoclad chironomidsdominating numerically. There are no fish or aquaticangiosperms.

Lower down, these small alpine streams converge toform a small fast-flowing river, prone to heavy spating.It is characterized by clear waters and unstable shinglebeds, interrupted at some points by areas of smoothflat bedrock. The water is extremely poor in nutrientsbut is less acid than higher up. Higher aquaticvegetation is virtually absent, and the stones arealmost free of epilithic algae. The invertebrate fauna issparse, and large areas of the bed can be virtuallydevoid of animals after floods. The fauna still containsan important arctic-alpine element, but also speciescharacteristic of lower elevations, such as the caddis

Polycentropus flavomaculatus, the alderfly Sialis fulig-inosa, the stonefly Taeniopteryx nebulosa, the waterbug Gerris costai and Hydracarina.

Below the Linn of Dee, the river becomes broader andis moderate to fast flowing, with a more stable bedconsisting of stones and gravel, a few boulders and alittle sand. It remains similar throughout most of itslength, gradually increasing in size so that at Cults it isup to 60 m wide. The dissolved nutrient content of thewater also increases downstream as the river entersfarmland, and at Peterculter it is mesotrophic (alkalinity16 mg .1 as CaCO3) but nitrate levels are stillrelatively low (average 0.5 mg NO3—N). In its lowerreaches, the water is less clear and sometimes slightlydiscoloured by plankton from the Loch of Skene,which drains into the river. Macrophytic vegetation isvirtually absent throughout the river, only a few clumpsof bryophytes being found.

Associated with the increasing nutrient content, thereis an increase in the biomass and variety of benthicinvertebrates as one moves downstream. A number ofmontane species such as Simuliurn monticola,Protonemura montana, Diura bicaudata and Crenobiaalpina disappear, and in the lower mesotrophic sec-tions the fauna is augmented by Polycelis tenuis/nigra,Gammarus pulex, Simuliurn reptans, Baetis pumilus,Ephemerella ignita, a number of caddis species and afew gastropods, all of which appear in increasingnumbers downstream.

3 International aspectsConservation assessment criteria are notoriously sub-jective, and there have been far too few seriousattempts to outline objective criteria which are quan-tifiable in some way, and which would serve asguidelines both for comparing one system withanother and also for assessing any changes in timewithin one system. The initial selection of aquaticnature reserves in Great Britain was by ornithologists,and only recently have other (more fundamental)criteria been used (Ratcliffe 1977). A wide variety ofsuch criteria must be considered in any generalscheme of site selection.

Although the Dee is listed as a grade 1 site of nationalimportance, there has been no suggestion so far that itcould be important internationally. Examination of theavailable data seems to confirm that it is not. Allphysical comparisons with important internationalrunning water systems would seem to place the Deelow down, regardless of the parameter or ratingsystem used (Table 2). The same seems to be true ofbiological factors, including important features likeplant, invertebrate, fish and bird diversity, as well asvarious facets of organic production and communitystructure.

The importance of the Dee must be viewed in anational, rather than an international, context. Does it

Table 2_ Comparative data for world and British rivers (source:Lewin 1981)

143

justify its rating (Ratcliffe 1977) as a grade 1 nationalsite? If so, can the assessment criteria used be mademore objective and quantifiable than has been thecase in the past?

4 National comparisons4.1 Physical characteristics

Ward (in Lewin 1981) has ranked major British rivers interms of length, area and mean annual discharge. TheDee is included in his list of rivers and falls 10th inlength, 21st in catchment area, and 18th in terms offlow, eventually being ranked 18th in terms of these 3characters. However, there are many other relevantriver characteristics, some of them probably muchmore important in biological terms, which should beconsidered when comparing the attributes of differentrunning water systems. Two of the most important ofthese are the altitude of the upper reaches and theextent of the lowland (and estuary) sections of eachriver.

Table 3 indicates that, according to the Water DataUnit (1982), the Dee rises higher than any other majorBritish river (1310 m)*, closely followed by the Spey(1309 m), the Tay (1215 m) and the Ness. No otherlarge river rises above 1000 m. As discussed below,this high-altitude section can be of considerableecological importance. At the other end of the river,the Dee (unlike most of the other large rivers) is almostdevoid of an estuary and has a relatively short lowlandsection.

In physical terms, therefore, it would be true to saythat the Dee is one of the most highland in character ofall the large British rivers. This statement is confirmedby an analysis of its flow regime which Warren (1985)describes as 'alpine' in character.

4.2 Chemical characteristics

It is difficult to obtain comparable chemical data forrivers in different areas. However, Table 4 gives ageneral chemical classification of series of stations*but see Maizels (1985) who states that the Dee rises at 1220 m(Ed).

144

Table 3 Major British rivers ranked in terms of flow, length,catchment areas and maximum altitude

Table 4 Chemical classification of various stations on the large Scottish rivers (source: SDD 1976). The classes are:1. unpolluted; 2. fairly good quality; 3. poor quality; 4. grossly polluted. The sampling stations on each riverare listed numerically from mouth towards source, but the distance between each varies

from near the mouth towards the source of 12 of thelarger Scottish rivers. The chemical classification isrelated to pollution (Scottish Development Depart-ment 1976) and the classes are defined as follows:1. unpolluted; 2. fairly good quality; 3. poor quality;4. grossly polluted.

It is clear that, although several of the rivers concernedhave sections where the chemical quality of the wateris poor, this is not true of the Dee, which is one of theleast contaminated of the larger Scottish rivers. Itshould follow, therefore, that its biota include mainlynatural communities, relatively little influenced byman.

Pugh (1985) describes the Dee as being in pristinecondition, emphasizing that its size and chemicalnature set it apart from many other rivers. It is a fineexample of a river which is oligotrophic from source tomouth.

4.3 Botanical characteristics

Morgan and Britton (in Ratcliffe 1977) give a briefdescription of the macroflora of the Dee, emphasizingthe importance of the bryophytes in the upperreaches. In a more detailed survey, Holmes (1985)showed that several zones of vegetation could bedistinguished on passing downstream and that, whencompared to other large rivers in Great Britain, the Deehas a unique succession of communities, exemplifyinga large oligotrophic highland river. Due to this repres-entativeness and its naturalness, Holmes confirmsthe Dee as being of prime nature conservationinterest.

4.4 Invertebrate characteristics

The invertebrate fauna of many Scottish rivers wasstudied at the same time as their water chemistry(Table 3) and comparable data are available for 1974and 1980 (Scottish Development Department 1976,1983). Both sets of samples were taken in the spring,

but the 1974 data (Table 5) are expressed in the formof Trent indices, whereas the 1980 data are repres-ented as biological scores (Scottish DevelopmentDepartment 1983). As with the chemical data, it isclear that, unlike several of the other large Scottishrivers, the Dee gives every indication of healthy,uncontaminated invertebrate communities along mostof its length.

Other, more detailed, studies have been made of theinvertebrate fauna of the Dee (Ratcliffe 1977; Wrightet al. 1984; Davidson et al. 1985). Unfortunately, onlyMorgan and Britton (in Ratcliffe 1977) covered sites inthe upper reaches of the river. Wright et al. (1984)compared the invertebrate fauna of the Dee withthose of 41 other river systems in Great Britain, usingan analysis of data at species level by multivariatestatistical techniques. The Dee was characterized inthis study as being 'upland' for most of its length,confirming the physical data discussed above. How-

Table 5. Invertebrate classification of various stations on the larger Scottish rivers (source: SDD 1976). The samplingstations on each river are listed numerically from mouth towards source, but the distance between eachvaries

NessSpeyDonDeeTayForthTyneTweedAnnanStincharClydeLeven

STATIONS: MOUTH SOURCE

ever, as none of the stations sampled in this studywere above 375 m and as the river rises above1000 m, it is likely that the really interesting andcharacteristic part of this river has been missed by thisand several other studies.

Davidson et al. (1985) found through an associationanalysis of invertebrate samples that sites on the mainriver were similar to each other and characteristic of anunpolluted highland river system.

4.5 Fish

A comparison of the species composition of the fishfauna of the larger Scottish rivers (Table 6) shows that,although the Dee is certainly truly highland in char-acter, there is nothing especially remarkable about itsfish community. Application of the criteria for theselection of important systems for freshwater fishrecommended by Maitland (1985) confirms that theDee is not especially noteworthy, as far as its fish areconcerned.

Table 6. The fish fauna of the larger Scottish rivers(updated from Maitland 1972)

1 2 3 4 5 6 7 8 9 10 11 12 13

X X X X XX IX X X X X X IX X VIIIVI IX X IX X X X X X XVIII IX X X VII VIII IX X X IX X VIIIX X X X X X X X X X X X XVI X IX X X IX IXVI VIII IV III IV VII VIII VIII VIII VII VIII IX XIX IX IX X X X X X X X X X XX VIII X X X

VIII IX VIII VIII VIII IXII III VII VI VIII IX VIII VIII VIII VIII X IX VIIX VIII VII X X

145

5 DiscussionThe problem of reaching an objective decision aboutthe status of any ecosystem is a difficult one and hasbeen discussed by a number of authors. Ratcliffe(1977) used the following criteria in his assessment ofBritish ecosystems: extent, diversity, naturalness,rarity, fragility, representativeness, recorded history,position in an eco-geographical unit and potentialvalue. In estimating the importance of the Dee, theauthor developed the procedure shown in Table 7 forthe selection of running waters of conservation value.It follows the pattern of the successful procedurepreviously used for selecting waters important for fishconservation (Maitland 1985).

It is clear that the Dee is of relatively little importanceinternationally, except perhaps as a prime example of awest European highland river. It is of national impor-tance in this category, and is probably the bestexample in the British Isles of a relatively naturalhighland river. Its particular characteristics include a

-0.2 -6EL co(r)a') a'(1)

_C

_c2

0 01.1J F— cc) Q Li_ 1—

+ + — + 9+ + + + 11+ + + + 13+ + + + 11+ + + + 15+ + + + 13+ + + + 13+ + + + 16+ + + + 15+ + — + 9+ + — 12+ + + + 14

146

greater altitudinal range than any other large Britishriver, virtual absence of a meandering lowland sectionor estuary, and very little impact from man in terms ofpollution, abstraction and hydro-electric or other ab-stractions. Its flora and fauna are typical of anuncontaminated highland system typically rather lowin diversity, especially compared to lowland systems.

There were 2 main problems in making this review.The first is that there is still insufficient ecological

1500

1000

500

Table 7. Procedure for the selection of running waters of conservation value

1. Is sufficient information available about the system to characterize it from sourceto mouth?

2. Are there any physico-chemical or biological features sufficiently outstandingto warrant international status?

3. Are any of the physico-chemical or biological characteristics sufficiently importantto warrant national recognition?

4. Are any man-made developments likely to be having a significant effect onthe river?

A. More research is required.B. The system should be notified internationally.C. The system should be notified nationally.D. The conservation value of the system is reduced.E. The conservation value of the system is increased.

YESNO

YESNO

YESNO

YESNO

2A

3

4

information concerning the Dee. This is particularlytrue of most aspects of the extremely important (andpotentially unique) upper reaches but also of somebasic aspects such as the distribution and status of itsfish populations (Jenkins & Bell 1985). Figure 1emphasizes the upland character of the Dee comparedto another, even larger, British river, the Trent. Thedistribution of stations for most types of samplingseems to be limited by accessibility from road bridges,and more research is needed in the mountainous

— River Dee. . .. River Trent/ Sampling point• Road bridge

....•••••• ...........................................................................................

0 50 100Length (%)

Figure 1. A profile of the River Dee (140 km) from source to mouth contrasting with the River Trent (256 km) drawnto the same percentage scale. Recent sampling points on the Dee by various organizations are shown just abovethe profile in relation to the main road bridges, below

upper reaches; these are, after all, one of the mostimportant features of this river.

The second problem is that comparisons with othersystems are difficult because of the absence of dataand also because data are rarely available in aconsistent form. For this reason, much of the infor-mation produced by the Scottish River PurificationBoards and published in their Annual Reports cannotbe used directly. The same is true for many datapresented by local authorities and other regionalorganizations.

This review confirms the grade 1 national rating givento the Dee during the Nature Conservation Review(Ratcliffe 1977). However, in spite of this rating,relatively little specific action, in nature conservationterms, has been taken. The North East River Purif-ication Board is well aware of the high quality of theriver and continues its efforts to maintain this, butthere remains a need for more integrated conserv-ation plan for the whole river basin which takes into

Table 8. Why is the River Dee important? A summary of features of importance in a national context

1. An excellent example of a highland eroding river.2. Its headwaters are probably the best alpine streams in Great Britain.3. It has virtually no lowland depositing section or estuary.4. Its waters are nutrient-poor and water quality is always high.5. Man-made impacts are low, so the system is highly natural.6. The plant communities show a unique succession, and are intact and representative.7. The invertebrate communities are characteristically highland throughout most of the river.8. The fish and other vertebrates are poor in diversity, but the salmonid fishery is important.9. The catchment area is scenically attractive.

10. The recreational value of the river system is high.

account all the various demands on the resource whilekeeping as its top priority the conservation of waterquality and the flora and fauna of the main river. Itsmajor features of national importance are summarizedin Table 8.

6 SummaryThe general ecological status of the Dee is reviewed.In international terms, for virtually all assessmentcriteria, the Dee is unimportant compared to the largerivers of the world. Its most important role is as thebest example of a large natural highland river inScotland. Its most characteristic and important feat-ures are as follows: its headwaters are among thehighest of any river in the British Isles; there is virtuallyno lowland section (or estuary) and the substrate iseroding virtually to the river mouth; the aquatic floraand fauna are highly characteristic of this type ofsystem and are relatively unaffected by man. In view ofits national importance, strong conservation measuresshould be adopted to preserve its status.

7 ReferencesButcher, R. W. 1933. Studies on the ecology of rivers: 1. On thedistribution of the macrophytic vegetation in the rivers of Britain. J.Ecol., 21, 58-91.

147

Carpenter, K. E. 1927. Faunistic ecology of some Cardiganshirestreams. J. EcoL, 15, 33-54.

Cummins, K. W., Coffman, W. P. & Roff, P. A. 1966. Trophicrelationships in a small woodland stream. Verh. int. Verein. theor.angew. LimnoL, 16, 627-638.

Davidson, M. B., Owen, R. P. & Young, M. R. 1985. Invertebratesof the River Dee. In: The biology and management of the River Dee,edited by D. Jenkins, 64-82. (ITE symposium volume no. 14.) AbbotsRipton: Institute of Terrestrial Ecology.

Has lam, S. M. 1982. Vegetation in British rivers. Huntingdon:Nature Conservancy Council.

Hawkes, H. A. 1975. River zonation and classification. In: Riverecology, edited by B. A. Whitton, 312-374. Oxford: BlackwellScientific.

Holmes, N. T. H. 1985. Vegetation of the River Dee. In: The biologyand management of the River Dee, edited by D. Jenkins, 42-55. (ITEsymposium no. 14.) Abbots Ripton: Institute of Terrestrial Ecology.

Huet, M. 1946. Note préliminaire sur les relations entre la pente etles populations piscicoles des eaux courantes. Biol. Jaarb., 13,232-243.

lilies, J. 1953. Die Besiedlung der Fulda (insbesondere das Benthosder Salmonidregion) nach dem jetzigen Stand der Untersuchungen.Ber. limnol. Flussstn Freudenthal, 5, 1-28.

Jenkins, D. & Bell, M. V. 1985. Vertebrates, except salmon andtrout, of the River Dee. In: The biology and management of the RiverDee, edited by D. Jenkins, 83-93. (ITE symposium no. 14.1 AbbotsRipton: Institute of Terrestrial Ecology.

Lewin, J., ed. 1981. British rivers. London: Allen & Unwin.

Luther, H. & Rzoska, J. 1971. Project Aqua: a source book of inlandwaters proposed for conservation. Oxford: Blackwell Scientific.

Maitland, P. S. 1966. The fauna of the River Endrick. Glasgow:Blackie.

Maitland, P. S. 1972. A key to the freshwater fishea of the BritishIsles, with notes on their distribution and ecology. (Scientificpublications no. 27.) Windermere: Freshwater Biological Associa-tion.

Maitland, P. S. 1979. Synoptic limnology. The analysis of Britishfreshwater ecosystems. Cambridge: Institute of Terrestrial Ecology.

Maitland, P. S. 1985. Criteria for the selection of important sites forfreshwater fish in the British Isles. Biol. Conserv, 31, 335-353.

Maizels, J. K. 1985. The physical background of the River Dee. In:The biology and management of the River Dee, edited by D. Jenkins,7-22. (ITE symposium no. 14.) Abbots Ripton: Institute of TerrestrialEcology.

Moon, H. P. 1938. Importance of the substratum to the invertebrateanimals on which fish feed. Ann. Rep. Avon. biol. Res., 6th, 36-40.

Pennak, R. W. 1971. Towards a classification of lotic habitats.Hydrobiologia, 38, 321-324. (ITE symposium no. 14.1 Abbots Ripton:Institute of Terrestrial Ecology.

Pugh, K. B. 1985. The chemistry of the River Dee. In: The biologyand management of the River Dee, edited by D. Jenkins, 34-41.

148

Ratcliffe, D. A., ed. 1977. A nature conservation review: theselection of biological sites of national importance to natureconservation in Britain. 2v. Cambridge: Cambridge University Press.

Ricker, W. E. 1934. An ecological classification of certain Ontariostreams. Univ. Toronto Stud. biol. Ser., 37, 1-114.

Scottish Development Department. 1976. Towards cleaner water1975. Edinburgh: HMSO.

Scottish Development Department. 1983. Water pollution controlin Scotland: recent developments. Edinburgh: SDD.

Warren, J. S. 1985. The hydrology of the River Dee. In: The biologyand management of the River Dee, edited by D. Jenkins, 23-28. CITEsymposium no. 14.1 Abbots Ripton: Institute of Terrestrial Ecology.

Water Data Unit. 1982. Surface water: United Kingdom 7974-76.London: HMSO.

Wright, J. F., Moss, D., Armitage, P. D. & Furse, M. T. 1984. Apreliminary classification of running-water sites in Great Britainbased on macro-invertebrate species and the prediction of commun-ity type using environmental data. Freshwater Biol., 14, 221-256.

Concluding commentsC H GIMINGHAMDepartment of Plant Science, University of Aberdeen

1 IntroductionIt will be evident from the collected papers that theconference effectively combined presentation ofscientific studies of the River Dee and its surroundingswith discussions of practical problems relating tomanagement. The assembled information rangesacross a wide spectrum of disciplines, and the viewsexpressed represent a considerable variety of in-terests. Much of the success of the symposiumderived from the way in which scientists andhistorians, landowners and engineers, managers andsportsmen, conservationists and 'recreationists', for-esters and agriculturists were all drawn into a discus-sion on the past, present and future of what, bycommon agreement, was regarded as a resource ofunique value.

There is little point in rehearSing here the contents ofthe various contributions; however, it may be useful inthese concluding comments to touch on the extent towhich the symposium achieved its objectives, and topresent some of the views expressed in the finaldiscussions, particularly as regards future action. Thestated aims were 'To assess existing knowledge ofthe river system and to identify gaps in knowledge, toassess the interests of people in the river system andpossible conflicts between these interests, and toconsider the current management of the system'.

2 The scientific assessmentAn impressive array of scientific observations and datawas presented, extending from the geomorphologicalevolution of the river system and its bed, to features ofthe flow and chemical composition of the water, theaquatic and riverside vegetation, and the invertebrateand vertebrate fauna. While in the course of timechanges of land use in the catchment, modifications ofthe course of the river, management of its banks andextraction of its water have all had their effects, thefact remains that the Dee has been much lessinfluenced by such activities than most other Britishrivers of equivalent size. The water quality remainsvirtually undiminished throughout its length, pollutionis minimal, quickly diluted and dispersed, and the riveris oligotrophic from source to sea. These features,together with the 'alpine' type of flow regime, conferupon it genuine distinction in scientific terms. Fur-thermore, although the pure acidic and nutrient-poorwater does not encourage a highly diverse or speciesrich fauna and flora, there are nonetheless numerouscomponents of special interest, and in several casesrarity. The landscape through which the river flows isof spectacular beauty and charm, and in terms ofecological and natural history interest ranks highly.During the symposium, a very positive, affirmative

149

answer was given to the question 'Is the River Dee ofexceptional value scientifically, and is it sufficientlyimportant to merit efforts to ensure its conservation?'

Detailed attention to this question led to the concl-usion that the river and its catchment are, indeed, ofnational importance in conservation terms.

The scientific assessment also involved considerationof the history of the river system and the nature of theprocesses at work at the present time. Past changes inthe course of the river, with the creation of terraces,abandoned channels, islands and other features, havegreatly enhanced the diversity of habitats and con-sequently of the fauna and flora. However, increasinguse of the floodplain in the lower reaches of the valleyfor agriculture and other purposes has led to a varietyof measures designed to control or regulate naturalprocesses such as flooding or bank erosion. Theimplications of these measures are often difficult topredict, and their ecological consequences may beunknown.

This highlighted the incompleteness of the scientificassessment, which, though extensive, revealed manygaps. For example, while flow rates have beenmonitored at several stations over a considerableperiod of years and correlated with climatic fluctua-tions, little is known of whether the changes representlong-term trends, or may be cyclic in nature. In view ofthe apparent increased incidence of periods of lowflow in recent years, this kind of information is neededto answer the question as to whether these arecaused by changes in the hydrology of the river, or bychanges in climate.

There are, also, few data on current rates of bankerosion, or of the wider effects of control measures.The sources and higher reaches of the river havereceived relatively little attention, and there are con-spicuous gaps in the documentation of the flora andfauna. In particular, there is much detail still to beadded to the description of the often rich and variedplant communities of the river banks, while theinvertebrate fauna would also repay more intensivesampling. Of the larger animals, little is known of thesmall mammals (eg water voles and water shrews) orof the sorts of places where otters have their young,while the status and distribution of the coarse fish areinsufficiently investigated. Apart from the straightfor-ward documentation, more attention to the ecology ofthe plant and animal populations of the river wouldhelp to establish some important points which remainobscure, for example the effects upon them of pastlosses of deciduous woodland which formerly fringed

150

the river throughout much of its length, and thechanges which might be expected if it were to berestored in selected locations.

3 Assessment of interests and demands upon the riverand its catchmentThe symposium drew attention to the fact that theriver and its surroundings are of social and economicimportance in a widely ranging variety of ways. Pastpatterns of agriculture and forestry in the area have,perhaps, had rather little direct effect, except wherethey have brought about replacement of deciduoustrees by conifers, or by fields close to the river banks.Eutrophication from agricultural fertilizers, however,has evidently not been a major problem. The 2 maindemands upon the river itself concern, first, waterabstraction, chiefly to supply the needs of Aberdeencity, and, second, fishing. Demands for water haverisen over the years at an increasing pace, to a point atwhich they may conflict with the interests of fishing attimes when the river is low. The importance of theDee as a salmon river was undisputed, and wasregarded as conferring upon it national, if not inter-national, significance. Understandably, this is anaspect of major concern to the landowners throughwhose estates the river flows.

People appreciate the river and its surroundings for avariety of other reasons. Its varied scenery has a wideappeal and positive economic value as a touristattraction, while the banks provide opportunities forinformal recreation and the river itself for canoeing.Certain conflicts were identified among these de-mands, but in general these were regarded as of minorsignificance and capable of solution by consultationbetween the parties concerned. However, there is nodoubt that the pressures are increasing and thedangers inherent in uncontrolled access to this amen-ity are real.

4 Rates of change — cause for concern?Most of the papers in the symposium were concernedwith establishing the scientific importance of the riverand its surroundings, and identifying the values placedupon it by the community. Having recognized it as aunique asset, the question arose as to whether it canconfidently be expected that its special qualities willsurvive substantially undiminished, in the absence ofmeasures designed to conserve them. While therewas a fair degree of optimism, it was also felt that inview of the rapidity of environmental change atpresent, even in sparsely populated areas, it would bea mistake to be complacent. What was important wasto be aware of, and in a position to monitor, changesthat may be taking place.

In certain respects, valuable monitoring is alreadybeing carried out. Rainfall is measured at a number ofstations, and both flow and chemical composition arerecorded regularly at selected sites along the wholelength of the river. Biological monitoring is alsoconducted on a regular basis. Statistics are available on

the salmon catches over a period of at least 30 years.However, in the course of discussion, it becameevident that in a number of respects the research basefor detecting potentially damaging changes is in-adequate or lacking. Reference was made, for exam-ple, to the need to follow the effects on soil or water ofchanges of land use in the catchment affecting thebalance between livestock and arable farming, orbetween agriculture and forestry, which may beinitiated by alterations in EEC policies or other socialchanges. The possibility that further extensiveafforestation in the catchment might affect itshydrology was evidently very much a matter of debate.Emphasis was therefore placed on the need to obtainactual measurements of the effects of plantation onevapotranspiration in this area. This would require asubstantial research input. Similarly, the extent towhich fertilizer applications during the establishmentof plantations might lead to elevated concentrations ofcertain nutrients (notably phosphorus) in the riverwater is relatively unexplored. This factor was thoughtunlikely to be of significance because it would applyonly where the water drains entirely through peat.Care is normally taken to ensure that drainage watermust pass through mineral material, which extractsthe excess phosphorus. However, this is a topic whichmerits further investigation with a view to checkingwhether or not recommended practice is alwaysfollowed.

Another urgent problem may be the effects of acidrain. The North East River Purification Board, recogniz-ing the possibility of damaging consequences, hasinitiated monthly monitoring at a number of sites in theupper reaches of the river system, and has identified asmall number of locations which may be especially atrisk in respect of the ability of the river water to bufferchanges in acidity. However, the Board has neither thecapacity nor the funds to undertake the intensiveresearch necessary for continual monitoring, in orderto detect what might be infrequent but heavy pollution'events' (eg acid snow). It was argued that here thereis a major research programme waiting to be under-taken if resources could be obtained, and a frameworkset up to promote collaboration between differentorganizations. Furthermore, apart from the question ofacid rain, there was evidently a need for much moreinformation, both chemical and biological, from theupper sections of the river system.

With the greatly increased abstraction of water fromthe river and new 'bed intakes', the point was madethat, while flow is recorded at various stations alongthe river, management requires detailed knowledge ofconditions at the point of intervention. Attempts toreconcile measurements from different gaugingstations show that significant losses and gains mayoccur unexpectedly, possibly as a result of exchangeswith the gravels of the floodplain. This again repre-sents a potential cause of change in the system aboutwhich there is insufficient information.

The conditions of flow in the river, especially at timesof low flow, affect the movement of salmon. Whilethere has been effective co-operation between theWater Services of the Grampian Regional Council andvarious salmon fishing interests, knowledge of varia-tion in the fish populations of the river is limited to thestatistics of the catches. The Dee salmon constitutessuch an important resource, which is susceptible tochange induced by a variety of causes, that it was feltto be a matter of urgency to rectify this situation andseek means to conduct a much more thoroughexamination of the ecology of salmon in the Dee, alonglines which have been successfully developed in otherrivers.

5 A proposal for actionThe river is a changing system; it has altered a greatdeal in the past and will continue to alter in the future.It was recognized that there is no merit in trying topreserve the river exactly as it is at present. However,while most other rivers have lost much or all of theiroriginal character, the Dee has not suffered to thesame extent. It is therefore a strong contender forconservation as an example of an oligotrophic river. Agood deal has already been achieved by informalconsultation between interested groups. However,members of the symposium felt strongly that thereshould be moves towards a more integrated type of

151

management, not imposing arbitrary rules but promot-ing communication and understanding of the varyingattributes of the river which people appreciate, and theseveral demands made upon it. Landowners, man-agers, naturalists, and representatives of water auth-orities, fishing and recreational interests all expressedtheir sympathy with conservation objectives and theirwillingness to collaborate to promote these and toremove any sources of conflict. (An encouragingexample of such collaboration was the successfulestablishment of an 'otter haven' on the river.) Theneed for public understanding and support, and thedesirability of education to that end, were alsostressed.

It became evident that there was a considerable bodyof opinion in favour of setting up an advisory group,which would provide the opportunity for all interests tointeract and work towardS the integration of manage-ment activities. Such a group could also focus atten-tion on the gaps in knowledge and understanding ofthe processes at work in the river and its catchment,and seek resources to encourage the necessaryresearch. It was felt that it would be most appropriatefor a group of this kind to have a formal link with thelocal authority and its planning department, and it wasdecided to explore this proposal further.

152

Appendix ILIST OF PARTICIPANTS

Mr C 0 Badenoch, Nature Conservancy Council, Galahill House, BarrRoad, Galashiels, TD1 3HX.

Mr R Balharry, Nature Conservancy Council, Wynne-EdwardsHouse, 17 Rubislaw Terrace, Aberdeen, AB1 1XE.

Dr N G Bayfield, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Dr M V Bell, Institute of Marine Biochemistry, St Fittick's Road,Torry, Aberdeen, AB1 3RA.

Mr H L Black, North of Scotland College of Agriculture, AgricultureCollege Office, 21 Evan Street, Stonehaven, AB3 2E0.

Mrs M Borthwick, Department of Physical Planning, GrampianRegional Council, Woodhill House, Ashgrove- Road West, Aberdeen,AB9 2LU.

Mr I D Brown, Director of Water Services, Grampian RegionalCouncil, Woodhill House, Ashgrove Road West, Aberdeen, AB92LU.

Mr R N Campbell, Tigh-Ur, Killiecrankie, by Pitlochry, Perthshire,PH16 5NP.

Dr R N B Campbell, Institute of Terrestrial Ecology, Bush Estate,Penicuik, Midlothian, EH 26 OQB.

Mr J W H Conroy, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Miss S M M Cooper, Rector, Aboyne Academy, Bridgeview Road,Aboyne, Aberdeenshire, AB3 5JN.

Dr G Dalton, North of Scotland College of Agriculture, School ofAgriculture, 581 King Street, Aberdeen, AB9 IUD.

Mr M B Davidson, North East River Purification Board, WoodsideHouse, Persley, Aberdeen, AB2 2UQ.

Mr R Dennis, Royal Society for the Protection of Birds, RSPBHighland Office, Munlochy, Ross & Cromarty, IV8 8ND.

Mr D Dunk ley, Department of Agriculture and Fisheries for Scotland,6 California Street, Montrose, Angus, DD10 8DN.

Dr E Farr, North of Scotland College of Agriculture, School ofAgriculture, 581 King Street, Aberdeen, AB9 IUD.

Mr J A Forster, Careers Advisory Office, University of Aberdeen,University Office, Regent Walk, Aberdeen, AB9 1FX.

Major H F B Foster, Park House, Drumoak, Banchory, Kincardine-shire, AB3 3AD.

Mr I Fozzard, Forth River Purification Board, Causewayhead OldSchool, Alloa Road, Stirling, FK9 5NP.

Mr W Gardner, 3 Mill of Dinnet, Dinnet, Aboyne, Aberdeenshire,AB3 5LA.

Professor C H Gimingham, Department of Botany, University ofAberdeen, St Machar Drive, Aberdeen, AB9 2UD.

Miss J Green, Nature Conservancy Council, Wynne-Edwards House,17 Rubislaw Terrace, Aberdeen, AB1 1XE.

Dr D Groman, North East River Purification Board, Woodside House,Persley, Aberdeen, AB2 2UQ.

Mr H A Hawkes, 154 Orphanage Road, Erdington, Birmingham,B24 OAA.

Mr & Mrs J Hepburn, Badcall Bay, Scourie, by Lairg, Sutherland,IV27 4TH.

Mr A Holden, Ach-na-Sithe, Manse Road, Moulin, Pitlochry,Perthshire.

Dr N T H Holmes, Nature Conservancy Council, NorthminsterHouse, Peterborough, PE1 1UA.

Mr J M M Humphrey, Estate Office, Dinnet, Aboyne, Aberdeen-shire, AB3 5LL.

Dr D Jenkins, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Mr W Kellas, 138 Gray Street, Aberdeen, AB1 6JU.

Dr J B Kenworthy, Department of Botany, University of Aberdeen,St Machar Drive, Aberdeen, AB9 2UD.

Dr H Kruuk, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Mr J A Lewis, 80 Beaconsfield Place, Aberdeen, AB2 4AJ.

Mr D Litten, North East River Purification Board, Woodside House,Persley, Aberdeen, AB2 2U0.

Mr F Little, North East River Purification Board, Woodside House,•Persley, Aberdeen, AB2 2U0.

Mr C I Lumsden, Laramie, Muiresk Drive, Turriff, Aberdeenshire,AB5 7BB.

Mr G Macdonald, Environmental Education Resource Base, Bal laterOutdoor Centre, School Lane, Ballater, Aberdeenshire, AB3 5RJ.

Mr T D Macdonald, Scottish Development Department, PentlandHouse, 47 Robbs Loan, Edinburgh, EH14 1TY.

Professor A D McIntyre, Department of Agriculture and Fisheries forScotland, Marine Laboratory, PO Box 101, Victoria Road, Torry,Aberdeen, AB9 8DB.

Dr P S Maitland, Institute of Terrestrial Ecology, Bush Estate,Penicuik, Midlothian, EH26 OQB.

Dr Judith K Maizels, University of Aberdeen, Department ofGeography, St Mary's, High Street, Old Aberdeen, AB9 2UF.

Mr P R Marren, Nature Conservancy Council, Foxhold House,Thornford Road, Crookham Common, Newbury, Berkshire, RG158EL.Mr N G Marr, Kincardine & Deeside District Council, Carlton House,Arduthie Road, Stonehaven, AB3 2DP.Dr M Marquiss, Institute of Terrestrial Ecology, Bush Estate,Penicuik, Midlothian, EH26 008.

Mr E M Matthew, Nature Conservancy Council, Wynne-EdwardsHouse, 17 Rubislaw Terrace, Aberdeen, AB1 1XE.Mr J F Mellanby, Grampian Regional Council, Woodhill House,Ashgrove Road West, Aberdeen, AB9 2LU.Dr J Miles, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Dr G R Miller, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Professor H G Miller, Department of Forestry, University ofAberdeen, St Machar Drive, Aberdeen, AB9 2UU.

Miss K Miller, North East River Purification Board, Woodside House,Persley, Aberdeen, AB2 2UQ.

Mr R C Milligan, Shell UK Exploration and Production, 1 Altens FarmRoad, Nigg, Aberdeen, AB9 2HY.

Mr James M Mitchell, Messrs A J Tulloch, Commercial Quay,Aberdeen, AB1 2PH.

Dr E A Needham, Hayes McCubbin Macfarlane, 11 RubislawTerrace, Aberdeen, AB-1 1XE.

Dr C Newbold, Nature Conservancy Council, Northminster House,Peterborough, PE1 1UA.

Dr I Newton, Institute of Terrestrial Ecology, Monks Wood Experi-mental Station, Abbots Ripton, Huntingdon, Cambs, PE17 2LS.

Mr I A Nicholson, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Mr A H Nicol, Estates Office, Ballogie, Aboyne, Aberdeenshire, AB35DT.

Mr R M G O'Brien, Education Department, Grampian RegionalCouncil, Moray Divisional Education Offices, Academy Street, Elgin,IV30 1LL.Mr J Oswald, 3 Bush Cottages, Glen Tanar, Aboyne, Aberdeenshire,AB3 5EU.

Dr R P Owen. North East River Purification Board. Woodside House.Persley. Aberdeen, AB2 2UO.

Mr J Parkin, Nature Conservancy Council, New Kinord House.Dinnet, Aboyne, Aberdeenshire, AB3 5LO.

Mr I S Paterson, Institute of Terrestrial Ecology, Hill of Brathens,Banchory, Kincardineshire, AB3 48Y.

Mr A M Pelham-Bum, Knappach, Banchoty, Kincardineshire, AB33JS.

Mr N Picozzi, Institute of Terrestrial Ecology. Hill of Brathens,Banchory, Kincardineshire, AB3 4BY.

Dr J D Pirie, Department of Natural Philosophy, University ofAberdeen, Aberdeen, AB9 2UE.

Dr K B Pugh, North East River Purification Board, Woodside House,Persley, Aberdeen, AB2 2U0.

Professor P A Racey, Department of Zoology. University ofAberdeen, Tillydrone Avenue, Aberdeen, AB9 2TN.

Mr D Rice, Countryside Commission for Scotland, Battleby, Red-gorton, Perth, PH1 3EW.

Mr K Robertson, Laird's Cast, Inchmarlo Road, Banchory, Kin-cardineshire, AB3 3RR.

Mr P I Rothwell, Nature Conservancy Council, Wynne-EdwardsHouse, 17 Rubislaw Terrace, Aberdeen, AB1 1XE.

Mr J A Ruxton, Department of Water Services, Grampian RegionalCouncil, Woodhill House, Ashgrove Road West, Aberdeen, AB92LU.

Mr S H Ali Al-Thanoon, Department of Botany, University ofAberdeen, St Machar Drive, Aberdeen, AB9 2UD.

Mr J R Scott, Secretary, Aberdeen Harbour Board, 16 Regent Quay,Aberdeen, AB1 2AE.

Mr W M Shearer, Department of Agriculture and Fisheries forScotland, Freshwater Fisheries Laboratory, 6 California Street,Montrose, Angus, DD10 8DW.

Dr R G J Shelton, Department of Agriculture and Fisheries forScotland, Freshwater Fisheries Laboratory, Faskally, PitloChry,Perthshire, PH16 5LB.

Mr I R Smith, Institute of Terrestrial Ecology, Bush Estate, Penicuik,Midlothian, EH26 008.

153

Dr J S Smith. Department of Geography. University of Aberdeen,St Mary's, High Street. Old Aberdeen, AB9 2UF.

Dr J Speakman, Department of Zoology. University of Aberdeen,Tillydrone Avenue, Aberdeen, AB9 2TN.

Dr B W Staines, Institute of Terrestrial Ecology, Hill of Brathens.Banchoty, Kincardineshire, AB3 48Y.

Mr M Stephenson. North East River Purification Board. WoodsideHouse, Persley, Aberdeen, AB2 2U0.

Mr B L Strachan, Mill of Strachan. Strachan, by Bancholy,Kincardineshire, AB3 3NS.

Mr G G M Taylor, Forestry Commission, East (Scotland) Conser-vancy, 6 Queen's Gate, Aberdeen, AB9 2NO.

Dr J Treasurer, Hospital Services Administrator, Woodend GeneralHospital, Eday Road, Aberdeen, AB9 2YS.

Mr G Tuley, Forestry Commission. Kirkton of Durris, Banchory,Kincardineshire, AB3 3BP.

Mr J S Warren, North East River Purification Board, WoodsideHouse, Persley, Aberdeen, AB2 2UQ.

Mr R D Watson, North East Mountain Trust, 396 King Street,Aberdeen, A82 3BY.

The Hon D N Weir, c/o Highland Wildlife Park. Kincraig, Kingussie,Inverness-shire, PH21 1ND.

Dr B B Willetts, Department of Engineering. University of Aberdeen,Marischal College, Aberdeen, AB9 1 AS.

Dr W Williams, Nature Conservancy Council, Wynne-EdwardsHouse, 17 Rubislaw Terrace, Aberdeen, AB1 1XE.

Miss E Woodward, Department of Physical Planning, GrampianRegional Council, Woodhill House, Ashgrove Road West, Aberdeen,AB9 2LU.

Mrs R Worsman, North East River Purification Board, WoodsideHouse, Persley, Aberdeen, AB2 2UQ.

Dr R A D Wright, Shell UK Exploration and Production, 1 Altens FarmRoad, Nigg, Aberdeen, AB9 2HY.

Dr M R Young, Department of Zoology, University of Aberdeen,Tillydrone Avenue, Aberdeen, AB9 2TN.

154

Appendix IILIST OF PLANT AND ANIMAL SPECIES REFERRED TO IN THE TEXT WITH BOTH THEIRCOMMON ENGLISH NAME AND THEIR LATIN BINOMIAL

1 Plants (common English names are rarely used either for algae, lichens, liverworts andmosses, or for most invertebrates. These species have therefore only been included in thelists of Latin names)1.1 Listed by common English name

Alder Alnus glutinosa Cottongrass Eriophorurn spp.Alternate water-milfoil Myriophyllurn alterniflorurn Creeping bent Agrostis stoloniferaAmphibious bistort Polygonurn amphibium Creeping buttercup Ranunculus repensAnnual meadow-grass Poa annua Creeping yellow-cress Rorippa sylvestrisAsh Fraxinus excelsior Crested dog's-tail Cynosurus cristatusAspen Populus trernula Cuckooflower Cardarnine pratensisBay willow Salix pentandra Curled pondweed Potamogeton crispusBeech Fagus sylvatica Deergrass Trichophorurn cespitosumBell heather Erica cinerea Devil's-bit scabious Succisa pratensisBilberry Vaccinium myrtillus Dog's mercury Mercurialis perennisBird cherry Prunus padus Douglas fir Pseudotsuga rnenziesiiBistort Polygonurn crispus Downy birch Betula pubescensBistort Polygonurn natans Eared willow Salix auritaBistort Polygonurn pectinatus European larch Larix deciduaBistort Polygonum perfoliatus Eyebright Euphrasia officinalisBitter vetch Lathyrus rnontanus False oat-grass Arrhenatherum elatiusBittersweet Solanurn dulcamara Fescue Festuca spp.Blinks Montia fontana Field forget-me-not Myosotis arvensisBlue water-speedwell Veronica anagallis-aquatica Field horsetail Equisetum arvenseBog pondweed Potarnogeton polygonifolius Fine-leaved sheep's-fescue Festuca tenuifoliaBog stitchwort Stellaria alsine Floating bur-reed Sparganiurn angustifoliurnBottle sedge Carex rostrata Floating club-rush Scirpus fluitansBracken Pteridiurn aquilinurn Floating sweet-grass Glyceria fluitansBramble Rubus fruticosus Fool's water-cress Apium nodiflorurnBranched bur-reed Sparganium erecturn Forget-me-not Myosotis palustrisBroad-leaved pondweed Potarnogeton natans Foxglove Digitalis purpureaBrook lime Veronica beccabunga Germander speedwell Veronica charnaedrysBroom Sarothamnus scoparius Glaucous sedge Carex flaccaBrown bent Agrostis canina Globeflower Trollius europaeusBugle Ajuga reptans Goat willow SalrX capreaBulbous rush Juncus bulbosus Goldenrod Solidago virgaureaBurnet-saxifrage Pimpinella saxifraga Goldilocks buttercup Ranunculus auricornusBur-reed Sparganiurn spp. Gorse Ulex europaeusBush vetch Vicia sepium Grand fir Abies grandisButterbur Petasites hybridus Great willowherb Epilobiurn hirsutumButtercup Ranunculus calcareus Great wood-rush Luzula sylvaticaCanadian waterweed Elodea canadensis Great yellow-cress Rorippa arnphibiaCanary-grass Phalaris canariensis Greater bird's-foot-trefoil Lotus uliginosusCarnation sedge Carex panicea Greater pond-sedge Carex ripariaCat's-ear Hypochoeris radicata Green water-cress Nasturtium officinaleCelery-leaved buttercup Ranunculus scleratus Green-ribbed sedge Carex binervisClover Trifoliurn spp. Grey willow Salix cinereaCock's-foot Dactylis glornerata Hairy sedge Carex hirtaColt's-foot Tussilago farfara Hairy St John's-wort Hypericurn hirsuturnCommon bent Agrostis tenuis Hard rush Juncus inflexusCommon bistort Polygonurn bistorta Harebell Carnpanula rotundifoliaCommon comfrey Syrnphyturn officinale Hazel Corylus avellanaCommon dog-violet Viola riviniana Heath bedstraw Galium saxatileCommon duckweek Lernna minor Heath rush Juncus squarrosusCommon figwort Scrophularia nodosa Heath speedwell Veronica officinalisCommon knapweed Centaurea nigra Heather Calluna vulgarisCommon marsh-bedstraw Galiurn palustre Hedge woundwort Stachys sylvaticaCommon nettle Urtica dioica Hemlock water-dropwort Oenanthe crocataCommon ragwort Senecio jacobaea Herb-Robert Geraniurn robertianumCommon rock-rose Helianthernum charnaecistus Hogweed Heracleum sphondyliurnCommon sedge Carex nigra Horned pondweed Zannichellia palustrisCommon spike-rush Eleocharis palustris Imperforate St John's-wort Hypericurn maculatumCommon valerian Valeriana officinalis Indian balsam lrnpatiens glanduliferaCommon water-crowfoot Ranunculus aquatilis Japanese larch Larix leptolepisCommon water-starwort Callitriche stagnalis Jointed rush Juncus articulatusCommon wintergreen Pyrola rninor Lady's bedstraw Galiurn verumCommon yellow-sedge Carex dernissa Lady's mantle Alchernilla glabra

155

156

Betula pendula Silver birch Equisetum hyemale Rough horsetailBetula pubescens Downy birch Equisetum palustre Marsh horsetailBlasia pusilla (Liverwort) Equisetum sylvaticum Wood horsetailBlindia acuta (Moss) Equisetum x trachyodon Mackay's horsetailBrachythecium mildeanum (Moss) Erica cinerea Bell heatherBrachythecium plumosum (Moss) Eriophorum spp. CottongrassBrachythecium rivulare (Moss) Euphrasia officinalis EyebrightBrachythecium rutabulum (Moss) Fagus sylvatica BeechBriza media Quaking-grass Festuca arundinacea Tall fescueBryum alpinum (Moss) Festuca ovina Sheep's-fescueBryum argenteum (Moss) Festuca tenuifolia Fine-leaved sheep's-fescueBryum pallens (Moss) Filipendula ulmaria MeadowsweetBryum pseudotriquetrum (Moss) Fissidens vindulus (Moss)Calliergon cuspidatum (Moss) Fontinalis antipyretica (Moss)Calliergon giganteum (Moss) Fontinalis squamosa (Moss)Callitriche hamulata Starwort Fragaria vesca Wild strawberryCallitriche stagnalis Common water-starwort Fraxinus excelsior AshCalluna vulgaris Heather Galium boreale Northern bedstrawCaltha palustris Marsh-marigold Galium palustre Common marsh-bedstrawCampanula rotundifolia Harebell Galium saxatile Heath bedstrawCardamine amara Large bitter-cress Galium verum Lady's bedstrawCardamine pratensis Cuckooflower Geranium robertianum Herb-RobertCardaminopsis petraea Northern rock-cress Geranium sylvaticum Wood crane's-billCarex aquatilis Water sedge Geum rivale Water avensCarex binervis Green-ribbed sedge Geum urbanum Wood avensCarex demissa Common yellow-sedge Glyceria fluitans Floating sweet-grassCarex echinata Star sedge Glyceria maxima Reed sweet-grassCarex flacca Glaucous sedge Gymnostomum aeruginosum (Moss)Carex hirta Hairy sedge Helianthemum chamaecistus Common rock-roseCarex lepidocarpa Long-stalked yellow-sedge Helictotrichon pratense Meadow oat-grassCarex nigra Common sedge Heracleum sphondylium HogweedCarex ovalis Oval sedge Hieracium pilosella Mouse-ear hawkweedCarex panicea Carnation sedge Hildenbrandia rivularis (Alga)Carex remota Remote sedge Hygrobiella laxifolia (Liverwort)Carex riparia Greater pond-sedge Hygrohypnum luridum (Moss)Carex rostrata Bottle sedge Hygrohypnum ochraceum (Moss)Centaurea nigra Common knapweed Hygrohypnum (Moss)Cephalozia bicuspidata (Liverwort) Hylocomium splendens (Moss)Chamaesiphon spp. (Algae) Hyocomium armoricum (Moss)Chamerion angustifolium Rosebay willowherb Hypericum hirsutum Hairy St John's-wortChiloscyphus polyanthos (Liverwort) Hypericum maculatum Imperforate St John's-wortCinclidotus fontinaloides (Moss) Hypericum perforatum Perforate St John's-wortCirsium heterophyllum Melancholy thistle Hypericum pulchrum Slender St John's-wortCladophora aegagropila (Alga) Hypochoeris radicata Cat's-earCladophora glomerata (Alga) Hypnum cupressiforme (Moss)Climacium dendroides (Moss) Impatiens glandulifera Indian balsamClinopodium vulgare Wild basil Iris pseudacorus Yellow irisCochlearia spp. Scurvygrass Juncus acutiflorus Sharp-flowered rushCollema fluviatile (Lichen) Juncus articulatus Jointed rushConocephalum conicum (Liverwort) Juncus bufonius Toad rushConopodium majus Pignut Juncus bulbosus Bulbous rushCorylus avellana Hazel Juncus effusus Soft rushCynosurus cristatus Crested dog's-tail Juncus inflexus Hard rushDactylis glomerata Cock's-foot Juncus squarrosus Heath rushDactylorhiza purpurella Northern marsh-orchid Jungermannia cordifolia (Liverwort)Dermatocarpon fluviatile (Lichen) Larix decidua European larchDeschampsia cespitosa Tufted hair-grass Larix leptolepis Japanese larchDianthus deltoides Maiden pink Lathyrus montanus Bitter vetchDichodontium flavescens (Moss) Lathyrus pratensis Meadow vetchlingDichodontium pellucidum (Moss) Lemanea fluviatilis (Alga)Dicranella palustris (Moss) Lemna minor Common duckweedDicranella rufescens (Moss) Lepidium heterophyllum Smith's pepperwortDidymosphenia geminata (Alga) Leucanthemum vulgare Oxeye daisyDigitalis purpurea Foxglove Littorella uniflora ShoreweedDoronicum pardalianches Leopard's-bane Lofium perenne Perennial rye-grassDrepanocladus revolvens (Moss) Lotus uliginosus Greater bird's-foot-trefoilEleocharis palustris Common spike-rush Lunularia cruciata (Liverwort)Elodea canadensis Canadian waterweed Lupinus nootkatensis Nootka lupinEnteromorpha spp. (Algae) Luzula sylvatica Great wood-rush[Ephydatia fluviatilis (Sponge)] Lychnis flos-cuculi Ragged-robinEpilobium hirsutum Great willowherb Lysimachia. nemorum Yellow pimpernelEquisetum arvense Field horsetail Marchantia polymorpha (Liverwort)Equisetum fluviatile Water horsetail Marsupella emarginata (Liverwort)

157

158

Veronica officinalis Heath speedwellVerrucaria praetermissa (Lichen)Verrucaria spp. (Lichen)Vicia cracca Tufted vetchVicia sepium Bush vetchViola palustris Marsh violetViola riviniana Common dog-violetViola tricolor Wild pansyZannichellia palustris Horned pondweed

2 Animals2.1 Vertebrates

2.1.1 Listed by common English name

Adder Opera berusBarnacle goose Branta leucopsisBean goose Anser fabalisBlack grouse Lyrurus tetnXBlack-headed gull Larus ridibundusBrook lamprey Lampetra planeriBrown bear Ursus arctosBrown hare Lepus capensisBrown long-eared bat Plecotus auritusBrown rat Rattus norvegicusBrown trout Salmo truttaB'uzzard Buteo buteoCanada goose Branta canadensisCapercaillie Tetrao urogallusCharr Salve linus spp.Chub Leuciscus cephalusCommon frog Rana temporariaCommon gull Larus canusCommon sandpiper Tringa hypoleucosCommon scoter Melanitta nigraCommon seal Phoca vitulinaCommon tern Sterna hirundoCommon toad Bufo bufoCoot • Fulica atraCurlew Numenius arquataDace Leuciscus leuciscusDaubenton's bat Myotis daubentoniDipper Cinclus cinclusDun lin Calidris alpinaEel Anguilla anguillaEider Somateria mollissimaGreat crested newt Triturus cristatusFlounder Platichthys flesusGadwall Anas streperaGoldeneye Bucephala clangulaGoosander Mergus merganserGrasshopper warbler Locustella naeviaGrayling Thymallus thymallusGreat black-backed gull Larus marinusGreat crested grebe Podiceps cristatusGreenshank Tringa nebulariaGrey wagtail Motacilla cinereaGrey lag goose Anser anserGudgeon Gobio gobioHen harrier Circus cyaneusHeron Ardea cinereaHerring gull Larus argentatusKingfisher Alcedo atthisLapwing Vanellus vanellusLong-tailed duck Clangula hyemalisMallard Anas platyrhynchosMarsh harrier Circus aeruginosusMink Mustela visonMinnow Phoxinus phoxinusMoorhen Gallinula chloropusMountain hare Lepus timidusMute swan Cygnus olorNatterer's bat Myotis nattereriNoctule Nyctalus noctula

OspreyOtterOystercatcherPalmate newtParti-coloured batPartridgePerchPheasantPied wagtailPikePink-footed goosePintailPipistrelle batPochardRabbitRainbow troutRed deerRed grouseRed-breasted merganserRedshankRiver lampreyRoachRoe deerRuffeSalmonSand eelSand martinScaupSea lampreySpratSea troutSedge warblerShelduckShovelerSlow wormSmeltSmewSmooth newtSnow gooseStarlingStone loachSwallowTealThree-spined sticklebackTufted duckViviparous lizardWater railWater shrewWater voleWhite-fronted gooseWhooper swanWigeonWolf

2.1.2 Listed by Latin binomial

Acrocephalus schoenobaenusAlcedo atthisAmmodytes marinusAnas acutaAnas clypeataAnas creccaAnas penelopeAnas platyrhynchosAnas streperaAnguilla anguillaAnguis fragilisAnser albifronsAnser anserAnser brachyrhynchusAnser caerulescensAnser fabalisArdea cinereaArvicola terrestris

Pandion haliaetusLutra lutraHaematopus ostralegusTriturus helveticusVespertilio murinusPerdix perdixPerca fluviatilisPhasianus colchicusMotacilla albaEsox luciusAnser brachyrhynchusAnas acutaPipistrellus pipistrellusAythya ferinaOryctolagus cuniculusSalmo gairdneriCervus elaphusLagopus lagopus scoticusMergus serratorTringa totanusLampetra fluviatilisRutilus rutilusCapreolus capreolusGymnocephalus cernuaSalmo salarAmmodytes marinusRiparia ripariaAythya marilaPetromyzon marinusSprattus sprattusSalmo truttaAcrocephalus schoenobaenusTadorna tadornaAnas clypeataAnguis fragilisOsmerus eperlanusMergus albellusTriturus vulgarisAnser caerulescensSturnus vulgarisNoemacheilus barbatulusHirundo rusticaAnas creccaGasterosteus aculeatusAythya fuligulaLacerta viviparaRallus aquaticusNeomys fodiensArvicola terrestrisAnser albifronsCygnus cygnusAnas penelopeCanis lupus

Sedge warblerKingfisherSand eelPintailShovelerTealWigeonMallardGadwallEelSlow wormWhite-fronted gooseGrey lag goosePink-footed gooseSnow gooseBean gooseHeronWater vole

159

160

Gammarus pulex (Water shrimp) Perla bipunctata (Stonefly)Gammarus zaddachi (Water shrimp) Per lodes microcephala (Stonefly)Gerris costai (Water bug) Philopotamus montanus (Caddis fly)Glossiphonia complanata (Leech) Physa fontinalis (Water snail)Glossosoma spp. (Caddis flies) Pisidiurn spp. (Water pea mussels)Goera pilosa (Caddis fly) Planorbis albus (Water snail)Halesus radiatus/digitatus (Caddis fly) Planorbis contortus (Water snail)Haliplus spp. (Beetles) Platarnbus maculatus (Beetle)Helobdella stagnalis (Leech) Plectrocnemia conspersa (Caddis fly)Helophorus spp. . (Beetles) Polycelis felina (Flatworm)Hemerodromia spp. (Two-winged flies) Polycelis tenuis/nigra (Flatworm)Heptagenia sulphurea (Mayfly) Polycentropus flavomaculatus (Caddis fly)Hexatoma spp. (Two-winged flies) Potamonectes depressus/Hydraena gracilis (Beetle) elegans (Beetle)Hydropsyche angustipennis (Caddis fly) Potamophylax cingulatus (Caddis fly)Hydropsyche cohtubernalis (Caddis fly) Potarnophylax latipennis (Caddis fly)Hydropsyche instabilis (Caddis fly) Potamopyrgus jenkinsi (Water snail)Hydropsyche pellucidula (Caddis fly) Protonernura meyeri (Stonefly)Hydropsyche siltalai (Caddis fly) Protonemura montana .(Stonefly)Hydroptilidae (Caddis flies) Protonemura praecox (Stonefly)Isoperla grammatica (Stonefly) Psychodidae (Two-winged flies)Lasiocephala basalis (Caddis fly) Rhithrogena semicolorata (Mayfly)Lepidostoma hirtum (Caddis fly) Rhyacophila dorsalis (Caddis fly)Leptophlebia marginata (Mayfly) Rhyacophila obliterata (Caddis fly)Leuctra fusca (Stonefly) Sericostoma personaturn (Caddis fly)Leuctra geniculata (Stonefly) Sialis fuliginosa (Alderfly)Leuctra hippopus (Stonefly) Sialis lutaria (Alderfly)

Leuctra inermis (Stonefly) Sigara concinna (Water bug)Limnephilus spp. (Caddis fly) pallipes (Caddis fly)Limnius volckmari (Beetle) Simulidae (Two-winged flies)Limnophora spp. (Two-winged flies Simuliurn monticola (Two-winged fly)Lumbricidae (Segmented worms) Simuliurn reptans (Two-winged fly)Lumbriculidae (Segmented worms) Sphaeriurn corneum (Water pea mussel)Lymnaea peregra (Water snail) Sphaerium spp. (Water pea mussels)Naididae (Segmented worms) Stenophylax lateralis/sequax (Caddis fly)Nemoura avicularis (Stonefly) Taeniopteryx nebulosa (Stonefly)Nemoura cambrica (Stonefly) Taphrophila spp. (Two-winged flies)Nemoura cinerea (Stonefly) Therornyzon tessulatum (Leech)

Odontocerum albicome (Caddis fly) Tipula spp. - (Two-winged flies)Oreodytes sanmarki (Beetle) Trocheta bykowskii (Leech)Oulimnius tuberculatus (Beetle) Tubificidae (Segmented worms)Paraleptophlebia submarginata (Mayfly) Wormaldia spp. (Caddisflies)

á

ISBN 0904282 88 0ISSN 0263-8614