PART III: DISCUSSION

95

Chapter 8: Landscape and environment

The sediment sequences and environments ofdeposition (Fig. 8.1)Several recurring sedimentary units and facies typeswere identified throughout the route sections duringthe initial fieldwork stages, correlated throughdetailed lithological description as well as associatedfeatures and artefactual assemblages (Appendix 2).Interpretation of the sequences was enhancedduring the assessment and analysis stages with aprogramme of radiocarbon dating (Appendix 1),together with detail provided by the examination ofpalaeoecological indicators (Appendix 3).

Superficially the Holocene sequences can becorrelated with the typical Thames tri-partitesequence of clay-silt, peat and clay-silt as describedby Long et al. (2000) and in general there are similar-ities with profiles recorded in other areas of theLower Thames floodplain (for example Devoy 1979;1982). For the purposes of this discussion generalcorrelations have been made within and betweensites and with the Cultural Landscape Model (CLM)of Bates and Whittaker (2004) referred to in Chapter2 (Table 1.3), in order to contextualise the associatedarchaeological remains (Table 8.1). There are, how -

ever, differences, as the sequences from the A13 sitescan be traced into edge marginal sequences thatabut the rising gravel terraces. The investigationshave demonstrated that significant complexityexists within the sediment sequences at terrace edgelocations and in the vicinity of palaeochannels,particularly at Movers Lane. Here, sequences are areflection of complex local as well as regionalfactors. Figure 8.1 provides a summary of the keysample sequences examined from each site andFigure 8.2 shows the site locations.

The following section briefly describes the keycharacteristics of each CLM Stage as presented inBates and Whittaker (2004), followed by asummary of the evidence recorded during the A13investigations.

Late Pleistocene (Fig. 8.3)The sequence of Pleistocene river terrace formationin the Thames is relatively well-documented (egBridgland 1994; Gibbard 1977; 1985; 1994). Thedeposition of sand and gravel sequences associatedwith the East Tilbury Marshes Gravel, which forms

Plate 19 Braided river, Denali National Park, Alaska (photo by Nick McPhee)

Landscape and Prehistory of the East London Wetlands

96

Fig. 8.1

Summ

ary of the route-w

ide samp

led sequences

Chapter 8

97

the higher terrace in the vicinity of the A13, occurredduring MIS 6 through 5e (Ipswichian Interglacial) toMIS 3. This terrace includes both temperate and coldclimate deposits which accumulated in a range ofenvironments, from braided to meandering riverchannels and estuarine situations. From c 30,000 BP(CLM Stage 1a, Bates and Whittaker 2004; Fig. 8.3),prior to and during downcutting under the sea-level

lowstand associated with the Last Glacial Maximum(LGM: 18,000-19,000 BP), reworking of the surface ofthe East Tilbury Marshes Gravel occurred by cold-climate solifluction processes. The SheppertonGravel deposited following the LGM (CLM Stage1b, Bates and Whittaker 2004, Fig. 8.3) forms theterrace that lies beneath Holocene sequences.Aggradation of the Shepperton Gravel is likely to

Fig. 8.2 Location of sites referred to in Chapter 8

Landscape and Prehistory of the East London Wetlands

98

Table 8.1 Sum

mary of route-w

ide sedim

ent sequences

Lower T

hames

General

Environments of deposition Inferred

date CL

M

Sedim

ent units

Associated

archaeology

tri-part sequence

description

Uppe

r allu

vium

Blue

-gre

y cla

y sil

t mos

tly

Low

ener

gy w

etlan

d en

viro

nmen

t Po

st 3k

a B.P.

5-6

CT5

PRL5

WM

W5

ML5

RB

5Se

als fe

ature

s of N

eolit

hic t

o str

uctu

reles

s. Oc

casio

nally

in

itiall

y fre

sh or

brac

kish

wate

r Ro

man

date

at h

ighe

r lam

inate

d w

ith oc

casio

nal

cond

ition

s, gi

ving

way

to sa

lt m

arsh

/ ele

vatio

ns.

coar

se fl

int g

rave

l hor

izons

inter

tida

l mud

flats.

Pos

sible

post-

de

posit

iona

l ped

ogen

ic ac

tivity

Orga

nic c

ompl

exH

ighl

y va

riabl

e org

anic

silts

Alde

r car

r wetl

and

with

pha

ses

3-

6ka B

.P.4

CT4

PRL4

WM

W4

ML4

RB

4BA

trac

kway

s, str

uctu

res

and

peats

exten

ding

acro

ss

of m

iner

ogen

ic in

put.

Up-p

rofil

e an

d as

socia

ted ar

tefac

t sca

tters

low

-lyin

g ar

ea.

poss

ibly

beco

min

g re

edsw

amp

or

brac

kish

mar

shlan

d. R

apid

lat

eral

chan

gesin

sedi

men

t and

ve

getat

ion/

spec

ies co

mpo

sitio

n.

Low

er al

luvi

umBl

ue g

rey

or li

ght b

row

n Sa

nd ba

rs in

mea

nder

ing

chan

nel

? Pre

6ka

2/3

CT3

PRL2

WM

W3

ML3

RB

2Re

wor

ked

andin situ ar

tefac

tcla

ys si

lts or

sand

y sil

ts

syste

ms o

r san

d sh

eets

rew

orke

d

B.P.

WM

W2

ML2

scatt

ers.

Mes

olith

ic, N

eolit

hic

and

sand

s, so

meti

mes

well

by su

rface

run-

off,

follo

wed

by

and

Bron

ze A

ge lo

cated

on u

pper

sorte

d, w

eakl

y lam

inate

dse

a-lev

el ris

e res

ultin

g in

sedi

men

ts w

eath

ered

surfa

cesa

nds g

radi

ng u

pwar

dsfro

m ba

cked

up

fresh

wate

r sys

tems,

A co

ncen

tratio

n of

early

Neo

lithi

cin

to si

lts, w

ith oc

casio

nal

grad

ing

upw

ards

to sa

lt m

arsh

N

eolit

hic f

lint,

cera

mics

and

grav

el cla

stssu

rface

/int

er ti

dal c

hann

el sy

stem

s ch

arre

d pl

ant r

emain

s, lo

cated

onFr

eque

ntly

exhi

bit e

vide

nce

(ML

and

WM

W)

a pos

sible

sand

-bar

(WM

W)

of ro

otin

g on

upp

erPo

st de

posit

iona

l ped

ogen

ic ac

tivity

/Fe

ature

s cut

into

upp

er w

eath

ered

wea

ther

ed su

rface

uppe

r sur

face

wea

ther

edw

eath

ered

surfa

ce. P

artic

ular

ly at

high

er el

evati

ons a

t edg

e of t

erra

ceIso

lated

pos

sible

Neoli

thic

featu

res

Bron

ze A

ge ac

tivity

indi

cativ

e of

dom

estic

and

agric

ultu

ral a

ctivit

ies.

Rom

an fi

eld sy

stem

? (PR

L)

Grav

elsSa

nd an

d gr

avel

horiz

ons,

Seve

re co

ld w

ith fr

ozen

gro

und.

? P

re-1

5ka

1

a-1b

-

PRL1

a-

ML1

a-

clear

bedd

ing

of sa

nds,

Perio

dic w

ashi

ng of

eros

ion

of

B.P.

sand

y gr

avels

and

silts.

high

er sa

nds,

solif

lucti

on an

d

Occa

siona

lly g

radi

ng

eros

ion

proc

esse

s alte

rnati

ng

upw

ards

into

fine

silty

sand

w

ith m

ore s

table

surfa

ces

Soft

brow

n cla

y, po

ckets

of

Low

ener

gy ch

anne

l inf

ill fe

ature

s, -

CT1b

--

--

fibro

us p

eat o

r dar

k br

own

prob

ably

Late

Glac

ialor

gani

c clay

Poor

ly so

rted

coar

se fl

int

Hig

h en

ergy

fluv

ial en

viro

nmen

t, >3

0ka t

o 1b

CT1

PRL1

WM

W1

ML1

RB1

grav

elspr

obab

ly br

aided

stre

am ch

anne

ls10

-15k

a B.P.

Ironbridge-CanningTownPrinceRegentLane

WoolwichManorWay

OldRodingBridge

MoversLane

have occurred under cold-climate conditions inhigh-energy fast flowing braided river channelsaccompanied by continued down slope erosion ofthe higher terrace. The braided river probablycomprised a network of transient channels withsand and gravel bars, similar to those that flow inareas of Alaska today (Plate 19).

Pleistocene gravel deposits were exposedthroughout the A13 route, although due to problemswith ground water and trench collapse they wererarely exposed to any great extent. The depositslargely comprised coarse flint sandy gravels andsands of fluvial origin (CT1, PRL1, WMW1, RB1 andML1), although sand, silt and gravel horizons,possibly deposited by sub-aerial erosion, wererecorded at Movers Lane and Prince Regent Lane

(ML1a and PRL1a). Although there was an absenceof associated palaeoenvironmental material withinthis latter group of sediments, their charactersuggests that they probably represent slopewash,colluviation and solifluction during intervals ofsevere cold when frozen ground dominated.Periodic washing of the surface and erosion of sandsfrom higher gravel terraces is likely to havecontributed to the sediment stack and slope erosionmay have alternated with periods of relative surfacestability. These sediments would have accumulatedon a slope southwards towards the contemporaryThames channel (a braided system across the valleyfloor at this time) and active downward erosionwould characterise this phase of fluvial activityaway from the immediate site vicinity. At MoversLane these deposits demonstrated considerablecomplexity grading upwards into fine silty sand andwere dated by luminescence techniques from15,800±850 BP to 23,900±1300 BP, clearly indicatingaccumulation immediately prior to and followingthe LGM. At Prince Regent Lane extensive depositsof bedded sand, sandy gravels and silts overlyingthe fluvial gravels were dated from 15,800±840 BP to16,300±820 BP (see Appendix 1).

At Canning Town, a sediment unit identified atthe base of gravels, or associated with the upper partof the Shepperton gravels, comprised soft brownclay, pockets of fibrous peat or dark brown organicclay (CT1b). Although these deposits were identifiedin geotechnical boreholes, unfortunately they werenot sampled by the purposive geoarchaeologicalboreholes and no sediment was available foranalysis. Organic deposits of similar descriptionoccur at the base of the lowermost gravels in the LeaValley and are well known from sites in TempleMills and Edmonton (the Lea Valley Arctic Bedsdescribed by Warren 1912; 1916; 1938). They repre-sent severely cold climate deposits dated 21,530 to28,000 BP, comprising broken rafts of peaty material,frequently rich in plant macrofossils and molluscs.Sediments of this type are rarely penetrated in theLea Valley and are of considerable interest as theyhave the potential to provide regional palaeoenvi-ronmental records and age estimates for this period(see Corcoran et al. 2011, 145-9 for the most recentassessment). The elevation of the deposits identifiedduring this study, however, would perhaps suggestchannel infill features of Late Glacial date is a moreappropriate interpretation. Similar channel depositsoccurring within the valley floor gravel were mostrecently identified during borehole work at theOlympics Park in Stratford (Fig. 8.2), spanning boththe Windermere Interstadial (warm phase) and LochLomand Stadial (cold phase) at c 11,000 -13,000 BP(Corcoran et al. 2011, 150).

Early–mid Holocene (Fig. 8.4)The surface of the Pleistocene deposits describedabove would have defined the topography of theearly Holocene landscape, which in turn would

Chapter 8

99

Fig. 8.3 Cultural Landscape Model (CLM) Stages 1aand 1b

have influenced patterns of later sediment accumu-lation. During CLM Stage 2 (Bates and Whittaker2004; Fig. 8.4) following climatic amelioration butprior to sea-level rise attaining near present daylevels, the area is likely to have been characterisedby relict Late Glacial features, but with a stablechannel within the old Late Glacial main channel.The floodplain of the river adjacent to the mainchannel would have stabilised with the develop-ment of the Holocene vegetation. Local pockets ofsediment accumulation are likely to have accruedduring this time in channels and hollows on thegravel surface, for example at Bramcote Green(Thomas and Rackham 1996). A key ecotonal areaprobably existed adjacent to the main Thameschannel and floodplain tributaries. Higher groundwould have provided additional landscaperesources within different environments.

During CLM Stage 3 (Bates and Whittaker 2004;Fig. 8.4) sea-level rise resulted in inundation of theformer dry land surface and began to influencesedimentation and fluvial dynamics within the

valley floor area. As the sea-level rose, channelstability will have decreased causing the start offlooding of low-lying areas. The floodplain surfaceis likely to have become unstable due to widespreadflooding and rapid sedimentation. Minerogenicsedimentation probably characterises this phase.While sediment accumulation during this stage willhave begun under freshwater conditions, it wouldhave been subsequently transformed by the onset ofestuarine conditions as marine inundationoccurred. During this period the ecotonal zonebetween wet and dry ground will have migratedinland and risen in datum across the flood surface.Thus wetland environments began to expand at theexpense of the dry ground areas. Temporarylandsurfaces may have existed within the flood areabut these are likely to have been ephemeral and oflocal significance only. Activity would probablyhave remained focused on channel marginal situa-tions and areas of the floodplain not inundated.Later more extensive inundation would eventuallyfocus activity on the floodplain margins and any

Landscape and Prehistory of the East London Wetlands

100

Fig. 8.4 Cultural Landcape Model (CLM) Stages 2 and 3

remnant islands of sand and gravel within thefloodplain.

Deposits relating to these landscape stages wereidentified throughout the A13 route corridor. Theycan be variously described as blue grey or lightbrown clays, silts or sandy silts and sands,sometimes well sorted, weakly laminated sandsgrading upwards into silts, with occasional gravelclasts. In general they lie between the basal fluvialgravels and the overlying organic complex.

The presence of a weathered horizon and associ-ated archaeological remains at the upper contact ofthe sandy facies at Prince Regent Lane, MoversLane and Woolwich Manor Way indicates that thesands at these locations were exposed as a land-surface prior to inundation. The landscape associ-ated with this horizon, after climatic ameliorationbut prior to the onset of sedimentation onto thetopographic template, is assumed to have been oneof relative stability. There would have beenminimal sediment deposition and little chance ofstratigraphic development, resulting in theconcatenated assemblages of Mesolithic andNeolithic artefacts in low lying areas, with theaddition of artefacts from later periods at higherelevations, where burial by sedimentation occurredlater (Fig. 8.5).

At Canning Town a series of blue grey clay siltsor sandy silts (CT3) were recorded at the base ofTP29. These pre-date c 3600 cal BC and appear tohave accumulated in freshwater conditions, in oradjacent to an active channel. The depositsincluded opercula of the freshwater molluscBithynia as well as the remains of bank vole,amphibians and cyprinid (the carp family). AtPrince Regent Lane in the western part of the site,blue grey clay-silts or sandy-silts (PRL2) beneaththe organic complex also accumulated under fresh-water conditions, evidenced by the presence ofoogonia (calcified fruiting bodies) of stonewort andwater flea eggs. Investigation of the top of thesequence in the ‘sump’ at Freemasons Roadindicated that an environment of damp alder carr

and sedge had developed prior to around 2000-2300 cal BC.

A possible early phase of marine incursion wasrecorded at both Woolwich Manor Way and MoversLane. At Woolwich Manor Way deposits beneaththe main peat body contained pollen of the goose-foot family, which includes species found insaltmarsh environments (Plate 20). This, togetherwith dinoflagellate cysts, sponge spicule fragmentsand a single example of the brackish diatomCyclotella striata, hint at the proximity of intertidalconditions. Radio carbon age determinations fromthe overlying peat suggest that the depositsaccumulated during the late Mesolithic period,prior to about 4600 cal BC. In the easternpalaeochannel at Movers Lane (TP39) a gradualchange from sand to organic silt accumulationsuggests a shift to slow moving conditions duringthe Mesolithic, prior to about 4000 cal BC. Althoughpollen was poorly preserved, the presence of goose-foots, together with dinoflagellates, perhapssuggests some marine influence with localsaltmarsh and some coastal woodland. Estuarineconditions were not recorded at the Roding Bridgesite, located between Woolwich Manor Way andMovers Lane, probably due to the much higherelevations recorded from the surface of thePleistocene gravels.

Mid Holocene (Fig. 8.6)CLM Stage 4 (Bates and Whittaker 2004; Fig. 8.6) ischaracterised by apparent fluctuating sea-levels inwhich alterations between organic and inorganicsedimentation dominate the area. A major expan-sion of freshwater wetlands is associated with thisphase and the temporary emergence of surfaces toor above flooding level will have stimulated thegrowth of organic sediments and led to peatgrowth. The ecotonal zone between wetland anddryland continued to move inland, causing loss oftopographic variation. During times of peataccumulation complex boundaries between peat

Chapter 8

101

Fig. 8.5 A model for temporal separation of artefact assemblages below the floodplain surface, after URN and URS (1999)

and non-peat wetland ecosystems will haveemerged within the wetland. Wetland would nowdominate in the floodplain area as dry groundzones shrank rapidly.

Inter-bedded peat or organic deposits wererecorded extensively along the A13 route. Thesedeposits in general represent alder carr wetland(Plate 21) with phases of minerogenic input, up-profile becoming reedswamp and brackish marsh-land. Radio carbon dates for the accumulation of theseorganic deposits broadly place them within the Neo -lithic and Bronze Age. This compares well with themodel for estuary contraction at around 4900-1250 calBC proposed by Long et al. (2000) and data fromnearby Crossness, where the channel was estimatedto have contracted from 4700m to 670m between3600-2000 BC (Devoy 1979). The sequences alsobroadly fit within the time range for Devoy’s TilburyIII (3550-2050 BC) and IV (1450 BC-AD 200) peat.

Considerable variation occurs, however, withinand between route sections within this organiccomplex, with peat formation occurring withindifferent timeframes at differing elevations. Thismay be a response to local factors. It is likely thatpools of water existed on the floodplain betweenstands of alder, willow and hazel. Peat will havebuilt up non-uniformly as pools were graduallyfilled in and new pools formed. Seasonal flooding,and probably an increase in river levels generally,

may have deposited sediment, particularly in thoselocations adjacent to tributaries.

At Canning Town, on the floodplain of the RiverLea, deep organic silt and peat units (CT4) accumu-lated over a period of about 2500 years between c3650 to 1210 cal BC, from the early Neolithic tomiddle Bronze Age period. At Prince Regent Lane,located at a higher elevation at the interfacebetween the terrace edge and the Thames flood-plain, the peat/organic silts (PRL4) were muchreduced in thickness. Radiocarbon dates suggestaccumulation here occurred over a significantlyshorter period of time of about 1500 years. Theearliest organic sedimentation in the vicinity ofFreemasons Road appears to have occurred in theslightly lower lying hollows, followed by morewidespread peat formation. Radiocarbon dating ofthe base of the peat in the ‘sump’ sequence suggeststhat accumulation occurred between about 2300and 800 cal BC. At Woolwich Manor Way peat andorganic sediments (WMW4) varied considerablyalong the route corridor. A lower woody peat andan upper amorphous peat were noted towards thewest, replaced by more minerogenic sediments tothe east where this group of sediments wedged outagainst the rising gravel surface. Radiocarbon datesfrom the thickest sequence in TP1, at the westernextent of the route section, indicate accumulationbetween about 4600 and 1450 cal BC, from the late

Landscape and Prehistory of the East London Wetlands

102

Plate 20 Saltmarsh, Fambridge, Essex (photo by Andy Roberts)

Mesolithic to middle Bronze Age. The peat appearsto have accumulated throughout the Neolithic andBronze Age, gradually encroaching onto the higherground to the east. At Movers Lane radiocarbondates from peat deposits (ML4) in TP39 indicate thataccumulation at this location occurred between c4000 to 900 cal BC; from the early Neolithic to thelate Bronze Age.

Late Holocene (Fig. 8.6)CLM Stage 5 (Bates and Whittaker 2004; Fig. 8.6) ischaracterised by the final submergence of the formerfloodplain topography and the loss of much of thefloodplain diversity. Typically organic sediment

growth appears to cease after topographicallyelevated areas became buried, and brackish waterconditions dominated. Sporadic occupation foreconomic use of the floodplain will have continuedthrough hunting/shell fish gathering as well as useof the river and tributaries for transport purposes.Eventual land reclamation and drainage allowedoccupation of the land for farming and habitation.

Sediments associated with this phase oflandscape development were recorded at all of theA13 sites, evident by the accumulation of predomi-nantly minerogenic silt-clays from the latter half ofthe 2nd millennium BC. These deposits occurred inevery profile examined and represent an ingress oftidal waters as a result of rising sea levels.

Chapter 8

103

Plate 21 Flooded alder carr, Brownsea Island, Poole Harbour (photo by David J Glaves)

Landscape and Prehistory of the East London Wetlands

104

The onset of minerogenic sedimentation atFreemasons Road (PRL5) dates from the lateBronze Age to early Iron Age. The pollen from the‘sump’ sequence at Freemasons Road, immediatelyabove the peat interface, shows little change apartfrom a slight rise in pollen of the goosefoot family,which includes species that grow on saltmarshes.However, several erosion channels cutting into thetop of the peats in the main excavation areacontained common sponge spicule fragments, aswell as marine and brackish diatoms, whichsuggests encroaching estuarine conditions. Theupper part of the alluvium from the excavationswas not analysed in detail but pollen work from theevaluation stage in T23 indicated the gradualdevelopment of open saltmarsh vegetation, withfreshwater marshes on the inland edge.Foraminifera and ostracod evidence suggestsmixed brackish and freshwater conditions, inwhich the introduction of freshwater speciesprobably derives from influxes of freshwater from

streams draining the inland marshes. On the higherground at Prince Regent Lane this sedimentarycomplex directly overlay Pleistocene colluvialdeposits, but also sealed ditches of Roman date.This implies that the higher ground remainedrelatively dry ground well into the historicalperiod.

At Woolwich Manor Way, minerogenic sedimen-tation (WMW5) occurred slightly earlier during themiddle to late Bronze Age. Accumulation appearsto have continued throughout the later prehistoricand into the historic period with evidence of Romanoccupation occurring within the alluvium in T15and TP13 in the east of the site. The environmentalevidence indicates that these deposits similarlyformed in a predominantly open estuarine environ-ment, with tidal creeks and saltmarsh, althoughhigh in the tidal frame with freshwater environ-ments co-existing nearby. A similar change is inevidence at Movers Lane with the deposition of claysilt (ML5) across the whole site, sealing the earlier

Fig. 8.6 Cultural Landsape Model (CLM) Stages 4 and 5

palaeochannels and the middle and late Bronze Agefeatures on the higher terrace.

The nature and speed of landscape change (Fig. 8.7)Despite the apparent variation in the date andelevations for the main period of peat accumulation,Figure 8.7 (updated from Bates and Whittaker 2004)illustrates that the onset of accumulation at thethree key A13 sites generally compares well withother sites in the vicinity (the location and details ofeach site is presented in Table 8.2 and Fig. 8.2). For amore comprehensive review of data derived fromunpublished grey literature reports the reader isreferred to Batchelor (2009) and Chapter 10. Theplot (Fig. 8.7) demonstrates a steep trend associatedwith rising sea-level prior to 6000 BP, with thegradient reducing following sea-level attainingclose to modern elevations. Calculation of the slopeof regression lines allows a time-depth model to beproduced which can be used to estimate the speedat which dry ground areas were lost to theexpanding wetland front during the mid Holocene.This model was applied by Bates to the BarkingReach area, immediately to the east of Movers Lane(Fig. 8.2), demonstrating 75% of dry ground waslost in low lying areas of about -6m to -3m OD overaround 700 years during the later Mesolithic period(4700-4000 cal BC, approximately 4.3mm/year

datum) (Bates 1998; 1999; Bates and Whittaker2004). With reference sites at higher elevations,including A13 sites, the plot broadly shows theslower rate of expansion of the wetlands ontohigher terrace edge locations from approximately -3m to 0m OD over a period of around 2500 yearsfrom the beginning of the Neolithic to middleBronze Age (1.2mm/year datum). These figures,although approximations, suggest the loss of dryground areas to the encroaching alder carr andassociated changes in flora and fauna may havebeen perceptible to local communities within a fewgenerations.

Evident at all of the A13 sites was a rising watertable within the upper parts of the peat profilesfollowed by a change to minerogenic sedimentationduring the later prehistoric period. This develop-ment has been recorded at many other East andCentral London sites and represents a gradualchange from a predominantly freshwater regime toa tidal river (see Sidell et al. 2000). The primaryfactors initiating this change may be attributed toincreases in the rate of sea-level rise combined withisostatic downwarp (Devoy 1979; Fairbridge 1983;Shennan 1987 cited in Sidell et al. 2000, 109).However, the precise timing of this change atindividual sites is most likely to have been influ-enced by more local factors such as the location ofpalaeochannels, elevation and position relative tothe main Thames channel.

Chapter 8

105

Table 8.2 Radiocarbon age estimates for selected sites in the Lower Thames where age estimates are available forcontexts overlying non-compressible sediments (after Bates and Whittaker 2004)

Site Elevation mOD 14C yr BP Lab code Reference

Tilbury -13.32 8170±110 Q1426 Devoy 1982Ebbsfleet Valley -0.7 4540±40 Beta-108114 Wenban-Smith et al. forthcoming Ebbsfleet Valley -2.32 4926±35 NZA-29080 Wenban-Smith et al. forthcoming Broadness -8.57 6620±90 Q1339 Devoy 1982West Thurrock -8.45 6450±120 IGS-C14/153 Devoy 1982Stone -8.82 6970±90 Q1334 Devoy 1982Slade Green Relief Road -2 4390±70 Beta-726204 Bates and Williamson 1995Crossness -5.3 5850±70 Beta-76991 Pine et al. 1994Dagenham Vale (HS1) -3.88 5751±40 NZA-16264 Bates and Stafford forthcomingRipple Lane Portal (HS1) -3.75 5773±40 NZA-28794 Bates and Stafford forthcomingMovers Lane (A3) -0.6 3950±35 SUERC-25572 (GU-19431) This volumeMovers Lane (TP39) -1.9 4680±35 SUERC-25567 (GU-19426) This volumeWoolwich Manor Way (A2) -1.3 4265±35 SUERC-25563 (GU-19425) This volumeWoolwich Manor Way (T16) -3.13 5460±80 Beta-152740 This volumeWoolwich Manor Way (T17) -3.19 5510±70 Beta-152741 This volumeWoolwich Manor Way (TP1) -3.65 5630±60 Beta-147956 This volumeFreemasons Road (Sump) -0.6 3745±35 SUERC-24600 (GU-18961) This volumeFort Street, Silvertown -2.52 4750±70 Beta-93683 Wilkinson et al. 2000Fort Street, Silvertown -3.3 5660±100 Beta-93689 Wilkinson et al. 2000Westmoor Street -0.3 3280±80 Beta-81970 Bates unpublishedBorax Works -8 6850±70 Beta-76200 Bates unpublishedCanning Town -1.5 4030±60 Beta-70248 Bates unpublishedBellot Street -0.6 3600±70 CIB-325 Unpublished archiveWest Ferry Road -3.2 5460±80 Beta-84317 Pine et al. 1994Phoenix Wharf 0.3 3110±40 BM-2766 Sidell et al. 2002

Landscape and Prehistory of the East London Wetlands

106

Fig. 8.7 Conventional radiocarbon age estimates plotted against depth for organic onto gravel situations in theLower Thames (mid estuary) area (after Bates and Whittaker 2004)

Plate 22 Thames Barrier at Woolwich (photo by Herry Lawford)

Chapter 8

107

At Crossness, located well into the floodplain onthe opposite side of the Thames to the A13 sites,estuarine conditions are evident much earlier; fromthe late Neolithic at 2800-2400 cal BC (Devoy 1979).Immediately adjacent to the current Thameschannel a sequence from the Isle of Dogs was datedto the early Bronze Age at 2200 cal BC (Wilkinson1995). The middle to late Bronze Age dates from theA13 sites probably relate to their position at higherelevations, close to the margins with the terraceedge. Similarly late dates were apparent during theinvestigations associated with the construction ofHigh Speed 1 to the east (formerly known as theChannel Tunnel Rail Link); at Ripple Lane Portal,Dagenham and East of Ferry Lane, Rainham theonset of brackish water conditions was apparentfrom the middle Bronze Age to middle Iron Age(Bates and Stafford forthcoming). In central Londonevidence from sites along the Jubilee Line Extensionindicates the development of full tidal conditionsfrom the late Bronze Age (c 1000 BC), leading Sidellet al. to suggest an average movement of the tidalhead upstream from Crossness of 5.4m a year (Sidellet al. 2000, 109; 2004, 41).

Deposition of estuarine sediments, however,which may represent earlier episodic inundations,was noted during the late Neolithic (c 2800-2550 calBC) period at both the Union Street and Joan Streetsites in Southwark (ibid). Similar evidence forestuarine flood events was noted within the parts ofthe peat sequences at Woolwich Manor Way, for

example in TP1, which contained rare valves of themarine and brackish water diatoms (Pseudo stelligerawestii, Cyclotella striata, Diploneis didyma andActinoptychus senarius) as well as sponge spiculefragments. The building of the trackways duringthe latter half of the Bronze Age, such as thoserecorded along the A13 at Woolwich Manor Wayand Movers Lane, may be seen as a response bylocal communities to increasingly wet conditionsand seasonal flooding to maintain access routesthough the marshes. However, the evidence atMovers Lane suggests repeated episodes of highenergy flooding, erosion and sand deposition inclose proximity to active channels during thisperiod. It is possible that some of these flood eventsmay be related to higher magnitude tidal surges, towhich historically the Thames estuary is particu-larly vulnerable. It should be noted, however, thatprior to extensive floodplain reclamation the river islikely to have been much wider and shallower thanit is today. During the 1st century AD the tidal rangeis likely to have been approximately 2m and thewidth of the river at high water was close to akilometre (Graham 1978, Milne et al. 1983 cited inSidell et al. 2000, 17). Extensive areas of flankingmarsh and creek systems along the estuary arelikely have facilitated the dissipation of floodwaters. Drainage and reclamation through the con -struction of embankments from the medievalperiod onwards has, however, substantiallynarrowed the width of river. This together with an

Plate 23 Flood defences at Barking Creek on the River Roding (photo by Lars Plougmann)

Landscape and Prehistory of the East London Wetlands

108

increased depth due to dredging activities hasresulted in an increase in tidal flow and funnellingeffect, producing a modern tidal range of up to 7min central London (Haughey 2008). During themedieval period there are numerous historicalrecords of flood events affecting the Thamesestuary. A period of increased storminess during the13th to 15th centuries caused repeated breaches ofthe river walls. Recurrent flooding affected thelands of the Abbess of Barking, and other breachesoccurred near Rotherhithe and in the stretch of theThames between Woolwich and Greenwich (Gallo -way 2009). The devastating flood of 1953, caused bya tidal surge in the North Sea, affected extensiveareas of the Thames Estuary and was one of thefactors that led to the construction of the Thamesbarrier at Woolwich (opened in 1982) and mainte-nance of around 300km of associated sea defences.As recently as September 2000, however, the RiverRoding overflowed its defences and flooded 320properties in the Wanstead and Woodford areas ofEast London. Since the 1980s, growing awareness ofglobal warming and sea-level rise has led to areassessment of coastal defences along the estuarywith alternative strategies focused on managedrealignment schemes in Kent and Essex andincreased flood storage capacity.

Vegetation patterns and human influences (Fig. 8.8)Information on vegetation patterns along the routeis largely gleaned from the pollen assemblages

recovered from the A13 sites (see Haggart, Peglarand Druce in Appendix 3), although more site-specific data is provided by other categories ofmaterial such as macroscopic plant remains(Pelling in Appendix 3) and insects (Smith inAppendix 3). The earliest deposits that produceduseful information date from the late Mesolithic toearly Neolithic, from sequences at WoolwichManor Way and Movers Lane. The sequence atFreemasons Road is a little later, dating from theearly Bronze Age.

Mesolithic woodlandsEvidence for the vegetation of this area of theThames Valley in the earlier part of the Holocenehas been reviewed in Sidell et al. 2000 and morerecently by Batchelor 2009; suffice to say sites withgood pollen preservation are sparse. Overall itappears that an initial phase of birch and pinewoodland was superceded by the mid Holocenewith lime, oak, elm, hazel and alder, which isevident at sites such as Hampstead Heath andRunneymede Bridge (Scaife in Sidell et al. 2000,111). Prior to its decline, associated with laterprehistoric clearance, lime was of great impor-tance and there is increasing evidence to show thatas well as being a major constituent of thewoodland on better drained sandy soils of theterraces, lime may also have been growing indamp woodland on peaty substrates (ibid) andthis may have been the case in the early phases ofpeat development at Woolwich Manor Way where

Fig. 8.8 Generic model of wetland zones

high levels of lime pollen have been recorded(Haggart in Appendix 3).

Late Mesolithic to early Bronze Age alder carr Along the A13 alder carr wetland environmentsappear to have developed in the low-lying areasfrom very early in the sampled sequences and areassociated with deposition of both the lowerminerogenic alluvium and the main period of peatformation. At Woolwich Manor Way the sequencebegins in the late Mesolithic period, prior to 4580-4350 cal BC (Beta-147956: 5630±60 BP), which is thedate from the base of the main peat unit in TP1 inthe western part of the site. The pollen assemblagesfrom the lower alluvial deposits beneath the peat,although poorly preserved, suggest a wetlandenvironment was present in the western part of thesite (Haggart in Appendix 3). The eastern part of thesite, which lay at higher elevations, remainedrelatively dry ground and this is reflected in thepoorly preserved waterlogged plant remains recov-ered from the sandy palaeosol and the large assem-blage of charred emmer wheat (Pelling in Appendix3). Locally the environment appears to have beeninitially rather open with the total herb pollenreaching 73%. The vegetation was domin ated bysedges and grasses, with some bracken and ferns.As the alluvium became more organic approachingthe interface with the over lying peat, alder began todominate the assemblages (60% TLP). This, alongwith willow and hazel, suggests a carr environmentwith perhaps oak growing at higher elevations.However, the occurrence of pollen of the goosefootfamily, together with dinoflagellate cysts, spongespicule fragments and a single example of thebrackish diatom Cyclotella striata, hint at theproximity of intertidal conditions. Increasingwetness and the formation of alder car is alsosuggested by the improved preservation of water-logged plant remains, with the remains of alderand seeds from plants typical of wet grassy groundsuch as rushes, sedges, celery-leaved crowfoot andgypsywort (Pelling in Appendix 3). The very baseof the peat contained unusually high percentages oflime pollen (> 30% TLP). The pollen grains of lime,however, are very resistant to decay which tends tomean it is well-represented in mineral soils. Thistogether with the presence of fungal spores indica-tive of dry mesotrophic conditions suggestsinitially a rather dry alder carr environment. Themain body of the peat was dominated by alder, oak,and hazel, with lime at low but constant frequen-cies; sedges and grasses were few. This suggeststhat the alder carr became very dense. It is duringthis period that early Neolithic activity was occur-ring on the higher ground in the vicinity of T15.Sporadic occurrences of yew, buckthorn andviburnum (probably Viburnum lantana, wayfaringtree) pollen occur, a finding mirroring that of Scaife,who comments on the species richness of fen carrpeats during the Neolithic and early to middle

Bronze Age at the Union Street and Joan Street sitesin Southwark (Scaife in Sidell et al. 2000). AtWoolwich Manor Way the pollen assemblages fromthe late Neolithic to early Bronze Age peat, fromthe excavation areas further to the east, producedsimilar results to TP1 but also a diverse range ofseeds of wetlands plants: sedges, crowfoots, waterdropwort and water pepper/mite. Disturbedground species were also present and includedchickweeds, black nightshade and brambles(Haggart and Pelling in Appendix 3).

The pollen sequence analysed from TP39 atMovers Lane begins in the early Neolithic period, inan organic silt at the base of the main peat bed,dated to 3960-3770 cal BC (SUERC-25568: 5055±35BP). Similar to Woolwich Manor Way, tree andshrub pollen dominated, with alder again being themost abundant. Other important trees includedlime, oak and hazel. Ferns were also present andcharcoal particles were quite high. Overall thissuggests that during the early Neolithic, freshwateralder carr was growing locally in wet places andalong riverbanks, with deciduous woodland on thedrier ground, but as at Woolwich Manor Way thealder carr eventually became quite dense (Peglar inAppendix 3).

The presence of yew (Taxus baccata) in the pollensequences is of particular note. Yew was alsoutilised in some of the Bronze Age timber trackwaysat Woolwich Manor Way as well as the platformstructure at the adjacent Golf Driving Range site(Carew 2010, Goodburn 2003b). A small scatter ofyew wood chips was recovered from Bronze Agedeposits at Movers Lane and an axe trimmedsection of yew at Freemasons Road (see Chapter 10).Yew was apparently a major component of coastalwoodland at this time and timber has been identi-fied at several floodplain sites within GreaterLondon, for example at Wennington (Sidell 1996),Dagenham (Divers 1994) and Beckton (Meddensand Sidell 1995; Scaife 1997). Sidell (1996) hassuggested that yew was an important woodlandtaxon, and that the low pollen representation wasdue to taphonomic factors and Scaife (2000) goes onto suggest suggests that this species rich aldercarr/coastal woodland may have no modernanalogue. Yew is also frequently found withinlowland wetland coastal and estuarine peat inBelgium, Germany and The Netherlands (Deforceand Bastiaens 2004). Across Europe there is a recog-nised shift in yew from lowland wetlands to uplanddryland during the Holocene, which may be attrib-uted to a change in ecological preference (Pelling inAppendix 3).

There is little evidence from the A13 sites forhuman impact on vegetation of the area during theNeolithic period. The analysed pollen profiles donot show a marked elm decline despite the twoprofiles from TP1 at Woolwich Manor Way andTP39 at Movers Lane potentially containingsediments spanning the event. An early Neolithicelm decline has been recorded at other sites in the

Chapter 8

109

region and constitutes a broadly synchronous eventdated to around 4300-3800 BC (see discussion ofScaife in Sidell et al. 2000, 112). It is a now widelyaccepted view, first proposed by Girling forHampstead Heath, that the demise of elm wasmainly due to a disease carried by the elm barkbeetle Scolytus scolytus, although the spread of thedisease was probably aided by Neolithic woodlandclearance (Girling and Grieg 1977; 1985; Grieg 1989;1992). The absence of evidence on the A13 sites maybe due to the floodplain location where local siteconditions such as increased waterlogging andalder carr formation diluted the influence of nearbydryland tree taxa including elm. However, this ishard to reconcile with a date for an elm decline fromnearby Silvertown which was placed at 5070 ± 70 BP(Wilkinson et al. 2000).

There was some tentative evidence in the pollenrecord of cereal cultivation from the earlyNeolithic onwards. However, the similarity ofcereal pollen with some wild grasses such asGlyceria (sweet-grasses) means that the evidenceremains equivocal (Haggart in Appendix 3). AtMovers Lane cereal-type pollen was identified inthe basal part of the peat in TP39 dated to the earlyNeolithic (c 4000-3600 cal BC) and here there wassome also evidence for open areas with grassland,possibly used for pasture prior to the alder carrbecoming quite dense (Peglar in Appendix 3). Anearly decline in lime pollen, along with other treeswas also noted at the junction of the silty peat andwood peat, dated to 3600-3360 cal BC, althoughthis could be attributed to human activity it couldalso have been related to increased water levels.The latter interpretation was also suggested for anlime decline observed in early Neolithic peatdeposits from the Ebbsfleet Valley (Huckerby et al.in Wenban-Smith et al. forthcoming). Alternativelythe thick alder carr may have prevented otherpollen from reaching the site. The lime decline isnormally associated with the Neo lithic/BronzeAge activity and this is thought to have led to theabsence of lime in the Thames valley at this time(Devoy 1979; 1980; 2000; Scaife in Sidell et al. 2000;Scaife in MoLAS 2001; Scaife 2006; Wilkinson et al.2000; Druce in Appendix 3). In the Thames Valleythere is an increasing body of evidence at anumber of sites suggesting an earlier temporarydecline in lime during the early Neolithic(Huckerby et al. in Wenban-Smith et al. forth-coming). In TP1 at Woolwich Manor Way cereal-type pollen was also present in low frequencies inthe upper levels of the peat dated to between theearly Neolithic and early Bronze Age (3400-1500cal BC), and the recovery of an assemblage ofcharred emmer wheat from T15 at WoolwichManor Way provides clear evidence for cultivationin the vicinity. At Freemasons Road, although thesedi ment sequence is a little later, the basal alluvialdeposits pre-dating 2300-2000 cal BC alsoproduced cereal type-pollen. An increase in fernand bracken spores and pollen of grassland plants

accompanied by a temporary decline in limepollen may suggest a clearance episode (Druce inAppendix 3).

Reedswamp, sedge fen and marsh environments ofthe 2nd millennium BC There is evidence towards the top of the peat ofincreasingly wet conditions and more openenvironments developing in many of the sequencesexamined during the first half of the second millen-nium BC. This is evident at Woolwich Manor Wayfrom increases in pollen of plantains (including seaplantain) and pondweeds. This period is alsocharacterised by a gradual decline in arborealpollen, most notably alder and oak, and an increasein pollen of grasses and goosefoots as well as seedsof meadow buttercup and rushes, suggesting moreopen areas of damp grassland (Haggart inAppendix 3).

At Movers Lane the upper part of the peat alsoshows a rapid increase in herbs and fern spores anda concomitant decrease in tree and shrub pollen,mainly due to a large drop in alder, and increases ingrasses and sedges (Peglar in Appendix 3). The plantassemblage provided evidence for alder (seeds andcones); wet ground species such as water-dropwort,fool’s water-cress, gypsywort, watermint, commonspike-rush, branched bur-reed, crowfoots, water-worts, water-plantain, red-shank, meadowsweet,sedges and rushes suggest marshy grassy condi-tions, and a small number of seeds of duckweedsand caddisfly larval cases may point to some openbodies of water. The drier ground species wererelatively limited but indicate a background ofscrubby vegetation, ruderal and disturbed habitatswith plants such as bramble, fat hen, stingingnettles, orache, knotgrass, cinquefoils, chickweed/stitchworts, black nightshade and thistle growing inthe vicinity (Pelling in Appendix 3).

At Freemasons Road alder carr gave way todamp species-rich sedge fen and grassland devel-oped immediately at the site. Freshwater pools andstreams were prevalent with green algae, aquatics,and bulrushes (Druce in Appendix 3). The macro-scopic remains produced broadly similar results tothe pollen: seeds and cones of alder and fruit ofbranched bur-reed indicate carr or fen conditions.Aquatic species included crowfoots, water-plantain,water-pepper and occasional seeds of duckweed, aswell as oogonia (calcified fruiting bodies) ofstonewort and the larval cases of caddisfly. Seeds ofwaterside vegetation which might have includedspecies growing within the shallow muddy waterinclude branched bur-reed, club-rushes, waterdropwort, fool’s water-cress, gypsywort and water-mint. Seeds of elder and bramble may indicateshrubby disturbed ground. Certainly disturbedhabitats and nitrogen-rich soil are suggested by fathen, stinging nettle, black nightshade, hairybuttercup and docks. Wet or damp grassland isindicated by meadow species, including possible

Landscape and Prehistory of the East London Wetlands

110

meadow rue, ragged robin, and buttercups (Pellingin Appendix 3).

The changing hydrological conditions at the sitesduring the 2nd millennium would undoubtedlyhave affected the local vegetation in terms of thereduction of alder, encroachment of wetter condi-tions onto drier ground and more open conditionsdeveloping as a result of a reduction in the localalder canopy. The more open environment is likelyto have increased the pollen catchment by allowingpollen from plants growing on the dryland, whichwas previously masked by the abundant alderpollen, to be better represented. Further increases incereal-type pollen during this period, along with anincrease of microscopic charcoal and seeds ofdisturbed ground species such as nettle andbramble, may point to increased clearance and culti-vation on the nearby higher terraces, although asmentioned above the distinction between cereal-type pollen and the pollen from some wild grassesfound growing in freshwater and coastal situationsis problematic. The evidence from the plant remainsat all three sites is equally enigmatic as seeds fromdisturbed, open and wet ground with some aquaticplants have all been recorded. At Freemasons Road,along with cereal type pollen, a significant declinein lime woodland on the surrounding slopes wasevident during this period, accompanied by adecline in fern spores and a slight increase in thosefrom bracken. The latter may be indicative ofincreased grazing (Behre 1986) or alternativelyrecent burning episodes, as bracken is known torapidly invade areas cleared by fire (Druce inAppendix 3). The presence of small numbers of‘dung beetles’ (eg Plate 24) and ‘chafers’ may alsosuggest that some of the landscape in the vicinitywas cleared of woodland or used as pasture (Smithin Appendix 3). At Movers Lane there was a smallbackground assemblage of charred plant remains, asignificant quantity of charcoal in some samples,and waterlogged plant remains suggestive ofscrubby vegetation and disturbed ground.

The interpretation of environmental proxiesidentified in estuarine and fluvial situations, partic-ularly pollen, is not easy. The sources of the pollenare likely to be widespread and grains will havearrived on site in a variety of ways. Some pollen islikely to have come from the local plant communi-ties, while wind dispersed pollen will representplants from further afield and grains may also betransported by water from both river and sea. Thelocal human and animal populations may also haveintroduced pollen and other plant and invertebrateremains into the sediments. At Woolwich ManorWay, Movers Lane and Freemasons Road theevidence is perhaps best summed up in the insectreport which states that ‘one major difficulty withthese faunas and their archaeological circumstancesis the degree to which they can be used reliably toreconstruct woodland surrounding the site’ (Smithin Appendix 3). The same could be said of both thepollen and other plant remains when used to distin-guish individual plant communities in isolation, butwhen considered together with the insects theysuggest the nature of both the very local wetlandand the surrounding dryland vegetation.

Chapter 8

111

Plate 24 Dung beetle (Aphodius granarius) modernspecimen and elytra (photo by Professor Mark Robinson)