int j remote sensing 2002 vol 23 no 21 4551ndash4570

The Circumpolar Arctic Vegetation Map AVHRR-derived base mapsenvironmental controls and integrated mapping procedures

D A WALKER W A GOULD H A MAIER andM K RAYNOLDS

Institute of Arctic Biology University of Alaska Fairbanks FairbanksAK 99775ndash7000 USA e-mail Vdawuafedu

Abstract A new false-colour-infrared image derived from biweekly 1993 and1995 Advanced Very High Resolution Radiometer (AVHRR) data provides asnow-free and cloud-free base image for the interpretation of vegetation as partof a 175 M-scale Circumpolar Arctic Vegetation Map (CAVM) A maximum-NDVI (Normalized DiVerence Vegetation Index) image prepared from the samedata provides a circumpolar view of vegetation green-biomass density across theArctic This paper describes the remote sensing products the environmentalfactors that control the principal vegetation patterns at this small scale and theintegrated geographic information-system (GIS) methods used in making theCAVM

1 IntroductionRemote sensing products from Earth-orbiting satellites provide the image base

to make the rst detailed vegetation map of an entire global biome the arctic tundraA new vegetation map of the Arctic is needed for numerous international eVortsincluding global-change and conservation studies land-use planning large-scaleresource development and education The Circumpolar Arctic Vegetation Map(CAVM) will be based on current knowledge of arctic plant communities and theirenvironmental controls During the past six years the CAVM participants havede ned the project organization and methods in a series of workshops (Walker 1995D A Walker et al 1995 Walker and Markon 1996 Walker and Lillie 1997 Markonand Walker 1999 Walker 2000 Raynolds and Markon 2002) Six groups of collabor-ators are now working on regional maps of Alaska Canada Greenland IcelandSvalbard and Russia This paper presents two new AVHRR-derived images of thecircumpolar Arctic the key environmental variables used for interpreting thevegetation and the GIS framework for the mapping methods

2 Remote-sensing products21 AVHRR-derived base map

A high-resolution false colour-infrared (CIR) image of the circumpolar regionrepresenting the vegetation at maximum greenness is an essential element of the

This paper was presented at the 6th Circumpolar Symposium on Remote Sensing of PolarEnvironments held in Yellowknife Northwest Territories Canada from 12ndash14 June 2000

International Journal of Remote SensingISSN 0143-1161 printISSN 1366-5901 online copy 2002 Taylor amp Francis Ltd

httpwwwtandfcoukjournalsDOI 10108001431160110113854

D A Walker et al4552

mapping method ( gure 1) This image is derived from Advanced Very-HighResolution Radiometer (AVHRR) data that were obtained from the US GeologicalSurvey (USGS) Alaska Geographical Science OYce It is composed of 1 kmtimes1 kmpicture elements (pixels) The image is a composite of pixels where each pixel wasselected by using the date with highest Normalized DiVerence Vegetation Index(NDVI) for that pixel among biweekly images taken during the periods from 1 Aprilto 31 October in 1993 and 1995 These periods cover the vegetation green-up-to-senescence period during two relatively warm years when summer-snow cover wasat minimum in the Arctic This permitted the construction of an image with minimumsnow and cloud cover The white areas are glaciers that were masked out usinginformation from the Digital Chart of the World The ocean ice was masked outusing the coastlines from the Digital Chart of the World The southern border ofthe CIR image is clipped to tree line which was derived from the best available

Figure 1 AVHRR-derived false colour-infrared image of the circumpolar Arctic The imageis derived from the pixels with highest NDVI among biweekly images from 1993 and1995 The southern border of the CIR image is clipped to treeline which was derivedfrom the best available vegetation maps

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4553

vegetation maps and the most recent information (Raynolds and Markon 2002)The image is used as a base for drawing vegetation map polygons Most boundarieson the vegetation map correspond to features that can be seen on the AVHRR image(see sect4)

22 Maximum-NDVI imageAn image portraying maximum NDVI ( gure 2) is used to delineate areas of

high and low vegetation biomass This image was created from the same data as thefalse CIR image ( gure 1) NDVI=(NIRplusmn IR)(NIR+IR) where IR is the spectralre ectance in the AVHRR near-infrared channel (0725ndash11 mm) where light-re ectance from the plant canopy is dominant and R is the re ectance in the redchannel (05ndash068 mm) the portion of the spectrum where chlorophyll absorbs maxim-ally The NDVI values were grouped into eight classes that meaningfully separatethe vegetation according to biomass Red and orange areas in gure 2 are areas ofshrubby vegetation with high biomass and blue and purple areas are areas with lowbiomass

Figure 2 Maximum NDVI for the circumpolar Arctic derived from the same data as gure 5

D A Walker et al4554

3 Environmental controls of vegetation at 175 M scaleThe approach used for making the CAVM is based on manual lsquophoto-

interpretationrsquo of the AVHRR satellite image A map based purely on automatedremote sensing procedures could not portray the details of plant communities forthe entire circumpolar region because there is large variability of tundra vegetationwith similar spectral properties Due to the small scale of the map and lack ofprevious vegetation maps for much of the Arctic vegetation information must beinferred from expert knowledge of the plant communities in relation to principallandforms and other terrain features that are visible on small-scale satellite imageryIn the Arctic vegetation of a given landscape can be predicted on the basis ofsummer temperature regime (bioclimatic subzones) available plants in the regional ora ( oristic sectors) soil chemistry and prevailing drainage conditions (Walker2000)

31 Bioclimatic zonationA fundamental problem for the CAVM is how to characterize the transitions in

vegetation that occur across the roughly 10degC mean July temperature gradient fromthe tree line to the coldest parts of the Arctic The important role of summertemperature has been noted with respect to a wide variety of ecological phenomenaincluding phenology (Sorensen 1941) species diversity (Young 1971 Rannie 1986)plant community composition (Matveyeva 1998) biomass (Bliss and Matveyeva1992 Bazilevich et al 1997) and invertebrate and vertebrate diversity (Chernov andMatveyeva 1997) Various authors working with diVerent geobotanical traditionshave divided the Arctic into bioclimatic regions using a variety of terminologies(table 1) The origins of these diVerent terms and approaches have been reviewedfor the Panarctic Flora (PAF) initiative (Elvebakk 1999) The PAF and CAVM haveaccepted a ve-subzone version of the Russian zonal approach ( gure 3) The subzoneboundaries are somewhat modi ed from the phytogeographic subzones of Yurtsev(1994) based on recent information from a variety of sources (Raynolds and Markon2002) A brief description of each subzone follows

311 Subzone A Herb subzoneSubzone A includes mostly fog-shrouded islands within the permanent arctic ice

pack where July mean temperatures are less than about 2ndash3degC such as Ellef RingnesAmund Ringnes King Christian northern Prince Patrick and nearby islands in thenorthwest corner of the Canadian Archipelago It also includes the coastal fringe ofnorthernmost Greenland and northern Ellesmere and northern Axel Heiberg islandsthe northeastern portion of Svalbard Franz Josef Land Severnaya Zemlya thenorthern tips of the Taimyr Peninsula and northern tip of Novaya Zemlya Thesummer temperatures in these areas are near freezing all summer due to a combina-tion of nearly continuous cloud and fog cover which limits solar radiation and theclose proximity to the ice-covered ocean (Bay 1997 Razzhivin 1999) More contin-ental inland areas of the larger islands are often considerably warmer Permanentice covers large areas of the land Major parts of the nonglacial land surfaces arelargely barren often with lt5 cover of vascular plants however meadow-likeplant communities are not uncommon on mesic ne-grained soils where there issuYcient moisture provided by the cold humid oceanic climate

Woody plants are absent on zonal sites Zonal sites are at or gently slopingmoderately drained sites with ne-grained soils that are not in uenced by extremes

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4555

of soil moisture snow soil chemistry or disturbance and which fully express thein uence of the prevailing regional climate Lichens bryophytes cyanobacteria andscattered forbs (eg Papaver Draba Saxifraga and Stellaria) are the dominant plantsMany of the forbs lichens and mosses have a compact cushion growth form Inmidsummer the arctic poppy Papaver radicatum is the most conspicuous plant overlarge portions of this subzone Other important low-growing cushion-forb generainclude Minuartia and Cerastium Soil lichens mosses and liverworts can cover ahigh percentage of the surface particularly in more maritime areas such as NovayaZemlya (Alexandrova 1980) Rushes (L uzula and Juncus) and grasses (AlopecurusPuccinellia Phippsia and Dupontia) are the main graminoid groups Sedges(Cyperaceae) are rare and wetlands lack organic peat layers There is little contrastin the composition of vegetation on mesic sites streamside sites and snowbeds Thevascular plant ora is extremely depauperate consisting of only about 50ndash60 species(Young 1971) On ne grained soils the extremely cold temperatures and the thinsparse plant canopy induce intense frost activity which forms networks of small(lt50 cm diameter) nonsorted hummocks and polygons and plants are con nedmainly to the depressions between the hummocks (Chernov and Matveyeva 1997)

312 Subzone B Prostrate dwarf-shrub subzoneSubzone B includes parts of the southern Queen Elizabeth Islands the eastern

fringe of Ellesmere Island much of Peary Land in Greenland the eastern portionof Svalbard central Novaya Zemlya the northern coast of the Taimyr Peninsulaand the New Siberian Islands Because of its small size and weakly de ned distin-guishing characteristics Subzone B could be characterized as a transition zone toSubzone C Several treatments of arctic zonation include subzones B and C in asingle subzone (Matveyeva 1998 Yurtsev 1994) The mean July temperature at thesouthern boundary of Subzone B is approximately 5degC This subzone has scatteredprostrate (creeping) dwarf shrubs on zonal soils Erect shrubby vegetation is lackingMesic low-elevation surfaces with ne-grained soils generally have open patchyplant cover generally with 5ndash25 cover of vascular plants Well-diVerentiatedsnowbed and riparian plant communities are lacking Nonsorted circles stripes andice-wedge polygons are common Vascular plant vegetation is often con ned tocracks and depressions in the polygonal network and areas irrigated by runoV fromsnow patches The dominant growth forms on mesic sites are prostrate dwarf shrubs

(eg Dryas Salix arctica and S polaris) forbs (eg Draba Saxifraga MinuartiaCerastium and Papaver) graminoids (eg Carex stans Carex rupestris Alopecurusalpinus Deschampsia borealis and L uzula confusa) mosses and lichens Rushes(L uzula) are also an important component of many mesic vegetation types Sedges

(Carex Eriophorum) often are dominant in wet areas

313 Subzone C Hemiprostrate dwarf-shrub subzone

Subzone C includes northern Banks Island northern Victoria Island DevonIsland Prince of Wales Island Somerset Island northern BaYn Island most ofEllesmere Island Axel Heiberg Islands the west coast of Greenland and inner ordregions of northeast Greenland southern Novaya Zemlya some coastal areas of

Yakutia and Chukotka and the northernmost coast of Alaska (Yurtsev 1994) Themean July temperature at the southern boundary of Subzone C is about 7degC Mesiczonal surfaces in this subzone have more luxuriant plant growth than that in SubzoneB Zonal sites are generally well vegetated but vascular-plant cover is still open and

D A Walker et al4556

Table

1

App

roxim

ate

equ

ivale

nt

sub

div

isio

ns

of

the

Arc

tic

Zo

ne

inR

uss

ia

No

rth

Am

eric

aand

Fen

nosc

and

ia

Ru

ssia

Nort

hA

mer

ica

Fen

nosc

andia

Edlu

nd

and

Yurt

sev

Ale

xan

dro

vaM

atv

eyev

aW

alk

erP

olu

nin

Alt

(198

9)

Tuhk

anen

Elv

ebakk

Su

bzo

ne

(199

4)

(198

0)

(1998

)(2

000

)(1

951

19

60)

Edlu

nd

(199

6)

Bliss

(1997

)(1

986

)(1

999

)

AH

igh

arct

icN

ort

her

nP

ola

rdes

ert

Cush

ion

-forb

Hig

harc

tic3

Her

bac

eou

sand

Hig

harc

tic

Inner

pola

rA

rcti

cpola

rtu

ndra

pola

rdes

ert

subzo

ne

crypto

gam

zone

des

ert

zone

(C=

00

deg)5

South

ern

Oute

rpola

rp

ola

rdes

ert

(1ndash3deg

C

zone

(15

ndash2deg

C)1

(2ndash3degC

)250

ndash15

0T

DD

)4(C

=0

5deg)

BA

rcti

ctu

nd

randash

No

rth

ern

Arc

tic

tun

dra

Pro

stra

teH

erb

-pro

stra

tesh

rub

Nort

her

nN

ort

her

nnort

her

nvari

ant

arc

tic

tundra

dw

arf

-shru

btr

ansi

tion

arc

tic

zone

arc

tic

tundra

subzo

ne

(3ndash4deg

C

zone

150ndash250

TD

D)

Pro

stra

tesh

rub (4

ndash6deg

C

250ndash350

TD

D)

(C=

10deg)

CA

rcti

ctu

nd

randash

South

ern

Mid

dle

arc

tic

Dw

arf

and

pro

stra

teM

idd

lear

ctic

Mid

dle

arc

tic

south

ern

vari

ant

arc

tic

tundra

shru

bzo

ne

tundra

zone

(5ndash7deg

C

(5ndash6degC

)(7

degC)

gt350

TD

D)

(C=

17

5deg)

DN

ort

her

nN

ort

her

nT

ypic

altu

ndra

Ere

ctdw

arf-

Low

arct

icL

ow

erec

tsh

rub

Low

arct

icSo

uth

ern

South

ern

hyp

oar

ctic

subarc

tic

shru

bsu

bzo

ne

arc

tic

zone

arc

tic

tundra

tundra

tundra

zone

Mid

dle

subarc

tic

tundra

(8ndash10

degC)

(9degC

)

ESouth

ern

South

ern

South

ern

Low

-shru

bA

rcti

csh

rub

hyp

oar

ctic

subarc

tic

tundra

subzo

ne

tundra

zone

tundra

tundra

(10ndash12

degC)

(10ndash12deg

C)

(7ndash10deg

C)

(C=

25deg)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4557

interrupted by frost scars and other periglacial features The main features distin-guishing Subzone C from Subzone B are the presence of the hemiprostrate shrubCassiope tetragona and well-diVerentiated plant communities in mires snowbeds andcreek sides Cassiope covers large areas on mesic sites in areas with acidic parentmaterial such as on Svalbard BaYn Island and interior portions of GreenlandHowever Cassiope is lacking on mesic alkaline surfaces in the western CanadianIslands In these areas Cassiope is generally con ned to snow accumulation areasSubzone C has much greater species diversity than Subzone B Sedges such asKobresia myosuroides Carex bigelowii and Carex rupestris are common on uplandsurfaces and numerous forbs such as members of the Fabaceae (eg OxytropisAstragalus) are important in nonacidic regions Communities in intrazonal habitats(snowbeds and creek sides) are well developed Wetland communities are much betterdeveloped than in Subzone B Creek sides have distinctive Epilobium latifoliumcommunities

314 Subzone D Erect dwarf-shrub subzoneSubzone D covers the southern parts of Banks Island and Victoria Island much

of Keewatin southern BaYn Island most of southern Greenland and a broad bandacross Siberia and Chukotka Climatically Subzone D periodically receives relativelytemperate air from the south during the summer while Subzone C receives predomin-ately arctic air masses The mean July temperature at the southern boundary ofSubzone D is about 9degC The boundary between subzones C and D is considered ofhighest rank because it separates the northern drier tundras on mineral soils fromthe southern relatively moist tundras with moss carpets and peaty soils (Alexandrova1980) This is approximately equivalent to the boundary between Blissrsquo High andLow Arctic (Bliss 1997) The major diVerence in pedology causes dramatic changesto the vegetation The plants in Subzone D have strong hypoarctic (boreal forest)aYnities (Yurtsev et al 1978) Important hypoarctic species such as birch (eg Betulanana) alder (Alnus) willow (Salix) and heath plants (Ericaceae and Empetraceae)extend their ranges from the lower layer of subarctic woodlands but are not dominantas they are in Subzone E Low shrubs (gt40 cm tall ) occur along streams Overallthe role of shrublands is much less prominent than in Subzone E The plant canopyis usually interrupted by patches of bare soil caused by nonsorted circles stripesand a variety of other periglacial features (lsquospotty tundrarsquo in the Russian literature)Vascular plants generally cover about 50ndash80 of the surface Zonal vegetation ongently sloping upland surfaces consists of sedges (eg Carex bigelowii Carex mem-branacea Eriophorum triste E vaginatum and Kobresia myosuroides) prostrate anderect dwarf (lt40 cm tall ) shrubs (eg Salix planifolia S lanata ssp richardsonii

Footnote to Table 1

1Mateveyeva (1998) Range of mean July temperature at southern boundary of subzone2Walker (2000) Approximate mean July temperature at the southern boundary of the

subzone3Polunin (1951) Zones based roughly on openness of ground cover High arctic very

sparsely vegetated middle arctic open vegetation carpet low arctic closed vegetation carpet4Edlund and Alt (1989) Zones de ned on the bases of mean July temperature and sum

of mean temperature of days exceeding 0degC (thawing degree days TDD)5Tuhkanen (1986) Southern boundary of zones de ned by Holdridge biotemperature as

the sum of mean monthly temperatures gt0degC divided by 12

D A Walker et al4558

Figure 3 Bioclimatic subzones in the Arctic Modi ed from Elvebakk et al (1999)

S reticulata S arctica Betula exilis and Dryas integrifolia) and mosses Prostrate-dwarf-shrub communities which were common on zonal surfaces in Subzone C arecon ned mainly to wind-swept sites The moss layer consisting primarily of

T omentypnum Hylocomium Aulacomnium and Sphagnum contributes to the develop-ment of organic soil horizons on ne-grained soils Soils in Subzone D have thinpeaty horizons and ne-grained soils are often nonacidic whereas soils in SubzoneE are usually acidic regardless of texture

There is more regional variation in the zonal vegetation than in subzones A Band C Tussock tundra consisting of cottongrass tussocks (Eriophorum vaginatum)and dwarf shrubs is common on ne-grained acidic soils over much of northeasternSiberia and northern Alaska (Walker et al 1994) particularly in areas that wereunglaciated during the last part of the Pleistocene In transitional areas to SubzoneC and on nonacidic loess Dryas spp and Cassiope tetragona are important (Walkerand Everett 1991) Some continental areas of Russia have dry steppe tundras thatare relicts of cold dry Pleistocene vegetation (Yurtsev 1982)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4553

vegetation maps and the most recent information (Raynolds and Markon 2002)The image is used as a base for drawing vegetation map polygons Most boundarieson the vegetation map correspond to features that can be seen on the AVHRR image(see sect4)

22 Maximum-NDVI imageAn image portraying maximum NDVI ( gure 2) is used to delineate areas of

high and low vegetation biomass This image was created from the same data as thefalse CIR image ( gure 1) NDVI=(NIRplusmn IR)(NIR+IR) where IR is the spectralre ectance in the AVHRR near-infrared channel (0725ndash11 mm) where light-re ectance from the plant canopy is dominant and R is the re ectance in the redchannel (05ndash068 mm) the portion of the spectrum where chlorophyll absorbs maxim-ally The NDVI values were grouped into eight classes that meaningfully separatethe vegetation according to biomass Red and orange areas in gure 2 are areas ofshrubby vegetation with high biomass and blue and purple areas are areas with lowbiomass

Figure 2 Maximum NDVI for the circumpolar Arctic derived from the same data as gure 5

D A Walker et al4554

3 Environmental controls of vegetation at 175 M scaleThe approach used for making the CAVM is based on manual lsquophoto-

interpretationrsquo of the AVHRR satellite image A map based purely on automatedremote sensing procedures could not portray the details of plant communities forthe entire circumpolar region because there is large variability of tundra vegetationwith similar spectral properties Due to the small scale of the map and lack ofprevious vegetation maps for much of the Arctic vegetation information must beinferred from expert knowledge of the plant communities in relation to principallandforms and other terrain features that are visible on small-scale satellite imageryIn the Arctic vegetation of a given landscape can be predicted on the basis ofsummer temperature regime (bioclimatic subzones) available plants in the regional ora ( oristic sectors) soil chemistry and prevailing drainage conditions (Walker2000)

31 Bioclimatic zonationA fundamental problem for the CAVM is how to characterize the transitions in

vegetation that occur across the roughly 10degC mean July temperature gradient fromthe tree line to the coldest parts of the Arctic The important role of summertemperature has been noted with respect to a wide variety of ecological phenomenaincluding phenology (Sorensen 1941) species diversity (Young 1971 Rannie 1986)plant community composition (Matveyeva 1998) biomass (Bliss and Matveyeva1992 Bazilevich et al 1997) and invertebrate and vertebrate diversity (Chernov andMatveyeva 1997) Various authors working with diVerent geobotanical traditionshave divided the Arctic into bioclimatic regions using a variety of terminologies(table 1) The origins of these diVerent terms and approaches have been reviewedfor the Panarctic Flora (PAF) initiative (Elvebakk 1999) The PAF and CAVM haveaccepted a ve-subzone version of the Russian zonal approach ( gure 3) The subzoneboundaries are somewhat modi ed from the phytogeographic subzones of Yurtsev(1994) based on recent information from a variety of sources (Raynolds and Markon2002) A brief description of each subzone follows

311 Subzone A Herb subzoneSubzone A includes mostly fog-shrouded islands within the permanent arctic ice

pack where July mean temperatures are less than about 2ndash3degC such as Ellef RingnesAmund Ringnes King Christian northern Prince Patrick and nearby islands in thenorthwest corner of the Canadian Archipelago It also includes the coastal fringe ofnorthernmost Greenland and northern Ellesmere and northern Axel Heiberg islandsthe northeastern portion of Svalbard Franz Josef Land Severnaya Zemlya thenorthern tips of the Taimyr Peninsula and northern tip of Novaya Zemlya Thesummer temperatures in these areas are near freezing all summer due to a combina-tion of nearly continuous cloud and fog cover which limits solar radiation and theclose proximity to the ice-covered ocean (Bay 1997 Razzhivin 1999) More contin-ental inland areas of the larger islands are often considerably warmer Permanentice covers large areas of the land Major parts of the nonglacial land surfaces arelargely barren often with lt5 cover of vascular plants however meadow-likeplant communities are not uncommon on mesic ne-grained soils where there issuYcient moisture provided by the cold humid oceanic climate

Woody plants are absent on zonal sites Zonal sites are at or gently slopingmoderately drained sites with ne-grained soils that are not in uenced by extremes

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4555

of soil moisture snow soil chemistry or disturbance and which fully express thein uence of the prevailing regional climate Lichens bryophytes cyanobacteria andscattered forbs (eg Papaver Draba Saxifraga and Stellaria) are the dominant plantsMany of the forbs lichens and mosses have a compact cushion growth form Inmidsummer the arctic poppy Papaver radicatum is the most conspicuous plant overlarge portions of this subzone Other important low-growing cushion-forb generainclude Minuartia and Cerastium Soil lichens mosses and liverworts can cover ahigh percentage of the surface particularly in more maritime areas such as NovayaZemlya (Alexandrova 1980) Rushes (L uzula and Juncus) and grasses (AlopecurusPuccinellia Phippsia and Dupontia) are the main graminoid groups Sedges(Cyperaceae) are rare and wetlands lack organic peat layers There is little contrastin the composition of vegetation on mesic sites streamside sites and snowbeds Thevascular plant ora is extremely depauperate consisting of only about 50ndash60 species(Young 1971) On ne grained soils the extremely cold temperatures and the thinsparse plant canopy induce intense frost activity which forms networks of small(lt50 cm diameter) nonsorted hummocks and polygons and plants are con nedmainly to the depressions between the hummocks (Chernov and Matveyeva 1997)

312 Subzone B Prostrate dwarf-shrub subzoneSubzone B includes parts of the southern Queen Elizabeth Islands the eastern

fringe of Ellesmere Island much of Peary Land in Greenland the eastern portionof Svalbard central Novaya Zemlya the northern coast of the Taimyr Peninsulaand the New Siberian Islands Because of its small size and weakly de ned distin-guishing characteristics Subzone B could be characterized as a transition zone toSubzone C Several treatments of arctic zonation include subzones B and C in asingle subzone (Matveyeva 1998 Yurtsev 1994) The mean July temperature at thesouthern boundary of Subzone B is approximately 5degC This subzone has scatteredprostrate (creeping) dwarf shrubs on zonal soils Erect shrubby vegetation is lackingMesic low-elevation surfaces with ne-grained soils generally have open patchyplant cover generally with 5ndash25 cover of vascular plants Well-diVerentiatedsnowbed and riparian plant communities are lacking Nonsorted circles stripes andice-wedge polygons are common Vascular plant vegetation is often con ned tocracks and depressions in the polygonal network and areas irrigated by runoV fromsnow patches The dominant growth forms on mesic sites are prostrate dwarf shrubs

(eg Dryas Salix arctica and S polaris) forbs (eg Draba Saxifraga MinuartiaCerastium and Papaver) graminoids (eg Carex stans Carex rupestris Alopecurusalpinus Deschampsia borealis and L uzula confusa) mosses and lichens Rushes(L uzula) are also an important component of many mesic vegetation types Sedges

(Carex Eriophorum) often are dominant in wet areas

313 Subzone C Hemiprostrate dwarf-shrub subzone

Subzone C includes northern Banks Island northern Victoria Island DevonIsland Prince of Wales Island Somerset Island northern BaYn Island most ofEllesmere Island Axel Heiberg Islands the west coast of Greenland and inner ordregions of northeast Greenland southern Novaya Zemlya some coastal areas of

Yakutia and Chukotka and the northernmost coast of Alaska (Yurtsev 1994) Themean July temperature at the southern boundary of Subzone C is about 7degC Mesiczonal surfaces in this subzone have more luxuriant plant growth than that in SubzoneB Zonal sites are generally well vegetated but vascular-plant cover is still open and

D A Walker et al4556

Table

1

App

roxim

ate

equ

ivale

nt

sub

div

isio

ns

of

the

Arc

tic

Zo

ne

inR

uss

ia

No

rth

Am

eric

aand

Fen

nosc

and

ia

Ru

ssia

Nort

hA

mer

ica

Fen

nosc

andia

Edlu

nd

and

Yurt

sev

Ale

xan

dro

vaM

atv

eyev

aW

alk

erP

olu

nin

Alt

(198

9)

Tuhk

anen

Elv

ebakk

Su

bzo

ne

(199

4)

(198

0)

(1998

)(2

000

)(1

951

19

60)

Edlu

nd

(199

6)

Bliss

(1997

)(1

986

)(1

999

)

AH

igh

arct

icN

ort

her

nP

ola

rdes

ert

Cush

ion

-forb

Hig

harc

tic3

Her

bac

eou

sand

Hig

harc

tic

Inner

pola

rA

rcti

cpola

rtu

ndra

pola

rdes

ert

subzo

ne

crypto

gam

zone

des

ert

zone

(C=

00

deg)5

South

ern

Oute

rpola

rp

ola

rdes

ert

(1ndash3deg

C

zone

(15

ndash2deg

C)1

(2ndash3degC

)250

ndash15

0T

DD

)4(C

=0

5deg)

BA

rcti

ctu

nd

randash

No

rth

ern

Arc

tic

tun

dra

Pro

stra

teH

erb

-pro

stra

tesh

rub

Nort

her

nN

ort

her

nnort

her

nvari

ant

arc

tic

tundra

dw

arf

-shru

btr

ansi

tion

arc

tic

zone

arc

tic

tundra

subzo

ne

(3ndash4deg

C

zone

150ndash250

TD

D)

Pro

stra

tesh

rub (4

ndash6deg

C

250ndash350

TD

D)

(C=

10deg)

CA

rcti

ctu

nd

randash

South

ern

Mid

dle

arc

tic

Dw

arf

and

pro

stra

teM

idd

lear

ctic

Mid

dle

arc

tic

south

ern

vari

ant

arc

tic

tundra

shru

bzo

ne

tundra

zone

(5ndash7deg

C

(5ndash6degC

)(7

degC)

gt350

TD

D)

(C=

17

5deg)

DN

ort

her

nN

ort

her

nT

ypic

altu

ndra

Ere

ctdw

arf-

Low

arct

icL

ow

erec

tsh

rub

Low

arct

icSo

uth

ern

South

ern

hyp

oar

ctic

subarc

tic

shru

bsu

bzo

ne

arc

tic

zone

arc

tic

tundra

tundra

tundra

zone

Mid

dle

subarc

tic

tundra

(8ndash10

degC)

(9degC

)

ESouth

ern

South

ern

South

ern

Low

-shru

bA

rcti

csh

rub

hyp

oar

ctic

subarc

tic

tundra

subzo

ne

tundra

zone

tundra

tundra

(10ndash12

degC)

(10ndash12deg

C)

(7ndash10deg

C)

(C=

25deg)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4557

interrupted by frost scars and other periglacial features The main features distin-guishing Subzone C from Subzone B are the presence of the hemiprostrate shrubCassiope tetragona and well-diVerentiated plant communities in mires snowbeds andcreek sides Cassiope covers large areas on mesic sites in areas with acidic parentmaterial such as on Svalbard BaYn Island and interior portions of GreenlandHowever Cassiope is lacking on mesic alkaline surfaces in the western CanadianIslands In these areas Cassiope is generally con ned to snow accumulation areasSubzone C has much greater species diversity than Subzone B Sedges such asKobresia myosuroides Carex bigelowii and Carex rupestris are common on uplandsurfaces and numerous forbs such as members of the Fabaceae (eg OxytropisAstragalus) are important in nonacidic regions Communities in intrazonal habitats(snowbeds and creek sides) are well developed Wetland communities are much betterdeveloped than in Subzone B Creek sides have distinctive Epilobium latifoliumcommunities

314 Subzone D Erect dwarf-shrub subzoneSubzone D covers the southern parts of Banks Island and Victoria Island much

of Keewatin southern BaYn Island most of southern Greenland and a broad bandacross Siberia and Chukotka Climatically Subzone D periodically receives relativelytemperate air from the south during the summer while Subzone C receives predomin-ately arctic air masses The mean July temperature at the southern boundary ofSubzone D is about 9degC The boundary between subzones C and D is considered ofhighest rank because it separates the northern drier tundras on mineral soils fromthe southern relatively moist tundras with moss carpets and peaty soils (Alexandrova1980) This is approximately equivalent to the boundary between Blissrsquo High andLow Arctic (Bliss 1997) The major diVerence in pedology causes dramatic changesto the vegetation The plants in Subzone D have strong hypoarctic (boreal forest)aYnities (Yurtsev et al 1978) Important hypoarctic species such as birch (eg Betulanana) alder (Alnus) willow (Salix) and heath plants (Ericaceae and Empetraceae)extend their ranges from the lower layer of subarctic woodlands but are not dominantas they are in Subzone E Low shrubs (gt40 cm tall ) occur along streams Overallthe role of shrublands is much less prominent than in Subzone E The plant canopyis usually interrupted by patches of bare soil caused by nonsorted circles stripesand a variety of other periglacial features (lsquospotty tundrarsquo in the Russian literature)Vascular plants generally cover about 50ndash80 of the surface Zonal vegetation ongently sloping upland surfaces consists of sedges (eg Carex bigelowii Carex mem-branacea Eriophorum triste E vaginatum and Kobresia myosuroides) prostrate anderect dwarf (lt40 cm tall ) shrubs (eg Salix planifolia S lanata ssp richardsonii

Footnote to Table 1

1Mateveyeva (1998) Range of mean July temperature at southern boundary of subzone2Walker (2000) Approximate mean July temperature at the southern boundary of the

subzone3Polunin (1951) Zones based roughly on openness of ground cover High arctic very

sparsely vegetated middle arctic open vegetation carpet low arctic closed vegetation carpet4Edlund and Alt (1989) Zones de ned on the bases of mean July temperature and sum

of mean temperature of days exceeding 0degC (thawing degree days TDD)5Tuhkanen (1986) Southern boundary of zones de ned by Holdridge biotemperature as

the sum of mean monthly temperatures gt0degC divided by 12

D A Walker et al4558

Figure 3 Bioclimatic subzones in the Arctic Modi ed from Elvebakk et al (1999)

S reticulata S arctica Betula exilis and Dryas integrifolia) and mosses Prostrate-dwarf-shrub communities which were common on zonal surfaces in Subzone C arecon ned mainly to wind-swept sites The moss layer consisting primarily of

T omentypnum Hylocomium Aulacomnium and Sphagnum contributes to the develop-ment of organic soil horizons on ne-grained soils Soils in Subzone D have thinpeaty horizons and ne-grained soils are often nonacidic whereas soils in SubzoneE are usually acidic regardless of texture

There is more regional variation in the zonal vegetation than in subzones A Band C Tussock tundra consisting of cottongrass tussocks (Eriophorum vaginatum)and dwarf shrubs is common on ne-grained acidic soils over much of northeasternSiberia and northern Alaska (Walker et al 1994) particularly in areas that wereunglaciated during the last part of the Pleistocene In transitional areas to SubzoneC and on nonacidic loess Dryas spp and Cassiope tetragona are important (Walkerand Everett 1991) Some continental areas of Russia have dry steppe tundras thatare relicts of cold dry Pleistocene vegetation (Yurtsev 1982)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4554

3 Environmental controls of vegetation at 175 M scaleThe approach used for making the CAVM is based on manual lsquophoto-

interpretationrsquo of the AVHRR satellite image A map based purely on automatedremote sensing procedures could not portray the details of plant communities forthe entire circumpolar region because there is large variability of tundra vegetationwith similar spectral properties Due to the small scale of the map and lack ofprevious vegetation maps for much of the Arctic vegetation information must beinferred from expert knowledge of the plant communities in relation to principallandforms and other terrain features that are visible on small-scale satellite imageryIn the Arctic vegetation of a given landscape can be predicted on the basis ofsummer temperature regime (bioclimatic subzones) available plants in the regional ora ( oristic sectors) soil chemistry and prevailing drainage conditions (Walker2000)

31 Bioclimatic zonationA fundamental problem for the CAVM is how to characterize the transitions in

vegetation that occur across the roughly 10degC mean July temperature gradient fromthe tree line to the coldest parts of the Arctic The important role of summertemperature has been noted with respect to a wide variety of ecological phenomenaincluding phenology (Sorensen 1941) species diversity (Young 1971 Rannie 1986)plant community composition (Matveyeva 1998) biomass (Bliss and Matveyeva1992 Bazilevich et al 1997) and invertebrate and vertebrate diversity (Chernov andMatveyeva 1997) Various authors working with diVerent geobotanical traditionshave divided the Arctic into bioclimatic regions using a variety of terminologies(table 1) The origins of these diVerent terms and approaches have been reviewedfor the Panarctic Flora (PAF) initiative (Elvebakk 1999) The PAF and CAVM haveaccepted a ve-subzone version of the Russian zonal approach ( gure 3) The subzoneboundaries are somewhat modi ed from the phytogeographic subzones of Yurtsev(1994) based on recent information from a variety of sources (Raynolds and Markon2002) A brief description of each subzone follows

311 Subzone A Herb subzoneSubzone A includes mostly fog-shrouded islands within the permanent arctic ice

pack where July mean temperatures are less than about 2ndash3degC such as Ellef RingnesAmund Ringnes King Christian northern Prince Patrick and nearby islands in thenorthwest corner of the Canadian Archipelago It also includes the coastal fringe ofnorthernmost Greenland and northern Ellesmere and northern Axel Heiberg islandsthe northeastern portion of Svalbard Franz Josef Land Severnaya Zemlya thenorthern tips of the Taimyr Peninsula and northern tip of Novaya Zemlya Thesummer temperatures in these areas are near freezing all summer due to a combina-tion of nearly continuous cloud and fog cover which limits solar radiation and theclose proximity to the ice-covered ocean (Bay 1997 Razzhivin 1999) More contin-ental inland areas of the larger islands are often considerably warmer Permanentice covers large areas of the land Major parts of the nonglacial land surfaces arelargely barren often with lt5 cover of vascular plants however meadow-likeplant communities are not uncommon on mesic ne-grained soils where there issuYcient moisture provided by the cold humid oceanic climate

Woody plants are absent on zonal sites Zonal sites are at or gently slopingmoderately drained sites with ne-grained soils that are not in uenced by extremes

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4555

of soil moisture snow soil chemistry or disturbance and which fully express thein uence of the prevailing regional climate Lichens bryophytes cyanobacteria andscattered forbs (eg Papaver Draba Saxifraga and Stellaria) are the dominant plantsMany of the forbs lichens and mosses have a compact cushion growth form Inmidsummer the arctic poppy Papaver radicatum is the most conspicuous plant overlarge portions of this subzone Other important low-growing cushion-forb generainclude Minuartia and Cerastium Soil lichens mosses and liverworts can cover ahigh percentage of the surface particularly in more maritime areas such as NovayaZemlya (Alexandrova 1980) Rushes (L uzula and Juncus) and grasses (AlopecurusPuccinellia Phippsia and Dupontia) are the main graminoid groups Sedges(Cyperaceae) are rare and wetlands lack organic peat layers There is little contrastin the composition of vegetation on mesic sites streamside sites and snowbeds Thevascular plant ora is extremely depauperate consisting of only about 50ndash60 species(Young 1971) On ne grained soils the extremely cold temperatures and the thinsparse plant canopy induce intense frost activity which forms networks of small(lt50 cm diameter) nonsorted hummocks and polygons and plants are con nedmainly to the depressions between the hummocks (Chernov and Matveyeva 1997)

312 Subzone B Prostrate dwarf-shrub subzoneSubzone B includes parts of the southern Queen Elizabeth Islands the eastern

fringe of Ellesmere Island much of Peary Land in Greenland the eastern portionof Svalbard central Novaya Zemlya the northern coast of the Taimyr Peninsulaand the New Siberian Islands Because of its small size and weakly de ned distin-guishing characteristics Subzone B could be characterized as a transition zone toSubzone C Several treatments of arctic zonation include subzones B and C in asingle subzone (Matveyeva 1998 Yurtsev 1994) The mean July temperature at thesouthern boundary of Subzone B is approximately 5degC This subzone has scatteredprostrate (creeping) dwarf shrubs on zonal soils Erect shrubby vegetation is lackingMesic low-elevation surfaces with ne-grained soils generally have open patchyplant cover generally with 5ndash25 cover of vascular plants Well-diVerentiatedsnowbed and riparian plant communities are lacking Nonsorted circles stripes andice-wedge polygons are common Vascular plant vegetation is often con ned tocracks and depressions in the polygonal network and areas irrigated by runoV fromsnow patches The dominant growth forms on mesic sites are prostrate dwarf shrubs

(eg Dryas Salix arctica and S polaris) forbs (eg Draba Saxifraga MinuartiaCerastium and Papaver) graminoids (eg Carex stans Carex rupestris Alopecurusalpinus Deschampsia borealis and L uzula confusa) mosses and lichens Rushes(L uzula) are also an important component of many mesic vegetation types Sedges

(Carex Eriophorum) often are dominant in wet areas

313 Subzone C Hemiprostrate dwarf-shrub subzone

Subzone C includes northern Banks Island northern Victoria Island DevonIsland Prince of Wales Island Somerset Island northern BaYn Island most ofEllesmere Island Axel Heiberg Islands the west coast of Greenland and inner ordregions of northeast Greenland southern Novaya Zemlya some coastal areas of

Yakutia and Chukotka and the northernmost coast of Alaska (Yurtsev 1994) Themean July temperature at the southern boundary of Subzone C is about 7degC Mesiczonal surfaces in this subzone have more luxuriant plant growth than that in SubzoneB Zonal sites are generally well vegetated but vascular-plant cover is still open and

D A Walker et al4556

Table

1

App

roxim

ate

equ

ivale

nt

sub

div

isio

ns

of

the

Arc

tic

Zo

ne

inR

uss

ia

No

rth

Am

eric

aand

Fen

nosc

and

ia

Ru

ssia

Nort

hA

mer

ica

Fen

nosc

andia

Edlu

nd

and

Yurt

sev

Ale

xan

dro

vaM

atv

eyev

aW

alk

erP

olu

nin

Alt

(198

9)

Tuhk

anen

Elv

ebakk

Su

bzo

ne

(199

4)

(198

0)

(1998

)(2

000

)(1

951

19

60)

Edlu

nd

(199

6)

Bliss

(1997

)(1

986

)(1

999

)

AH

igh

arct

icN

ort

her

nP

ola

rdes

ert

Cush

ion

-forb

Hig

harc

tic3

Her

bac

eou

sand

Hig

harc

tic

Inner

pola

rA

rcti

cpola

rtu

ndra

pola

rdes

ert

subzo

ne

crypto

gam

zone

des

ert

zone

(C=

00

deg)5

South

ern

Oute

rpola

rp

ola

rdes

ert

(1ndash3deg

C

zone

(15

ndash2deg

C)1

(2ndash3degC

)250

ndash15

0T

DD

)4(C

=0

5deg)

BA

rcti

ctu

nd

randash

No

rth

ern

Arc

tic

tun

dra

Pro

stra

teH

erb

-pro

stra

tesh

rub

Nort

her

nN

ort

her

nnort

her

nvari

ant

arc

tic

tundra

dw

arf

-shru

btr

ansi

tion

arc

tic

zone

arc

tic

tundra

subzo

ne

(3ndash4deg

C

zone

150ndash250

TD

D)

Pro

stra

tesh

rub (4

ndash6deg

C

250ndash350

TD

D)

(C=

10deg)

CA

rcti

ctu

nd

randash

South

ern

Mid

dle

arc

tic

Dw

arf

and

pro

stra

teM

idd

lear

ctic

Mid

dle

arc

tic

south

ern

vari

ant

arc

tic

tundra

shru

bzo

ne

tundra

zone

(5ndash7deg

C

(5ndash6degC

)(7

degC)

gt350

TD

D)

(C=

17

5deg)

DN

ort

her

nN

ort

her

nT

ypic

altu

ndra

Ere

ctdw

arf-

Low

arct

icL

ow

erec

tsh

rub

Low

arct

icSo

uth

ern

South

ern

hyp

oar

ctic

subarc

tic

shru

bsu

bzo

ne

arc

tic

zone

arc

tic

tundra

tundra

tundra

zone

Mid

dle

subarc

tic

tundra

(8ndash10

degC)

(9degC

)

ESouth

ern

South

ern

South

ern

Low

-shru

bA

rcti

csh

rub

hyp

oar

ctic

subarc

tic

tundra

subzo

ne

tundra

zone

tundra

tundra

(10ndash12

degC)

(10ndash12deg

C)

(7ndash10deg

C)

(C=

25deg)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4557

interrupted by frost scars and other periglacial features The main features distin-guishing Subzone C from Subzone B are the presence of the hemiprostrate shrubCassiope tetragona and well-diVerentiated plant communities in mires snowbeds andcreek sides Cassiope covers large areas on mesic sites in areas with acidic parentmaterial such as on Svalbard BaYn Island and interior portions of GreenlandHowever Cassiope is lacking on mesic alkaline surfaces in the western CanadianIslands In these areas Cassiope is generally con ned to snow accumulation areasSubzone C has much greater species diversity than Subzone B Sedges such asKobresia myosuroides Carex bigelowii and Carex rupestris are common on uplandsurfaces and numerous forbs such as members of the Fabaceae (eg OxytropisAstragalus) are important in nonacidic regions Communities in intrazonal habitats(snowbeds and creek sides) are well developed Wetland communities are much betterdeveloped than in Subzone B Creek sides have distinctive Epilobium latifoliumcommunities

314 Subzone D Erect dwarf-shrub subzoneSubzone D covers the southern parts of Banks Island and Victoria Island much

of Keewatin southern BaYn Island most of southern Greenland and a broad bandacross Siberia and Chukotka Climatically Subzone D periodically receives relativelytemperate air from the south during the summer while Subzone C receives predomin-ately arctic air masses The mean July temperature at the southern boundary ofSubzone D is about 9degC The boundary between subzones C and D is considered ofhighest rank because it separates the northern drier tundras on mineral soils fromthe southern relatively moist tundras with moss carpets and peaty soils (Alexandrova1980) This is approximately equivalent to the boundary between Blissrsquo High andLow Arctic (Bliss 1997) The major diVerence in pedology causes dramatic changesto the vegetation The plants in Subzone D have strong hypoarctic (boreal forest)aYnities (Yurtsev et al 1978) Important hypoarctic species such as birch (eg Betulanana) alder (Alnus) willow (Salix) and heath plants (Ericaceae and Empetraceae)extend their ranges from the lower layer of subarctic woodlands but are not dominantas they are in Subzone E Low shrubs (gt40 cm tall ) occur along streams Overallthe role of shrublands is much less prominent than in Subzone E The plant canopyis usually interrupted by patches of bare soil caused by nonsorted circles stripesand a variety of other periglacial features (lsquospotty tundrarsquo in the Russian literature)Vascular plants generally cover about 50ndash80 of the surface Zonal vegetation ongently sloping upland surfaces consists of sedges (eg Carex bigelowii Carex mem-branacea Eriophorum triste E vaginatum and Kobresia myosuroides) prostrate anderect dwarf (lt40 cm tall ) shrubs (eg Salix planifolia S lanata ssp richardsonii

Footnote to Table 1

1Mateveyeva (1998) Range of mean July temperature at southern boundary of subzone2Walker (2000) Approximate mean July temperature at the southern boundary of the

subzone3Polunin (1951) Zones based roughly on openness of ground cover High arctic very

sparsely vegetated middle arctic open vegetation carpet low arctic closed vegetation carpet4Edlund and Alt (1989) Zones de ned on the bases of mean July temperature and sum

of mean temperature of days exceeding 0degC (thawing degree days TDD)5Tuhkanen (1986) Southern boundary of zones de ned by Holdridge biotemperature as

the sum of mean monthly temperatures gt0degC divided by 12

D A Walker et al4558

Figure 3 Bioclimatic subzones in the Arctic Modi ed from Elvebakk et al (1999)

S reticulata S arctica Betula exilis and Dryas integrifolia) and mosses Prostrate-dwarf-shrub communities which were common on zonal surfaces in Subzone C arecon ned mainly to wind-swept sites The moss layer consisting primarily of

T omentypnum Hylocomium Aulacomnium and Sphagnum contributes to the develop-ment of organic soil horizons on ne-grained soils Soils in Subzone D have thinpeaty horizons and ne-grained soils are often nonacidic whereas soils in SubzoneE are usually acidic regardless of texture

There is more regional variation in the zonal vegetation than in subzones A Band C Tussock tundra consisting of cottongrass tussocks (Eriophorum vaginatum)and dwarf shrubs is common on ne-grained acidic soils over much of northeasternSiberia and northern Alaska (Walker et al 1994) particularly in areas that wereunglaciated during the last part of the Pleistocene In transitional areas to SubzoneC and on nonacidic loess Dryas spp and Cassiope tetragona are important (Walkerand Everett 1991) Some continental areas of Russia have dry steppe tundras thatare relicts of cold dry Pleistocene vegetation (Yurtsev 1982)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4555

of soil moisture snow soil chemistry or disturbance and which fully express thein uence of the prevailing regional climate Lichens bryophytes cyanobacteria andscattered forbs (eg Papaver Draba Saxifraga and Stellaria) are the dominant plantsMany of the forbs lichens and mosses have a compact cushion growth form Inmidsummer the arctic poppy Papaver radicatum is the most conspicuous plant overlarge portions of this subzone Other important low-growing cushion-forb generainclude Minuartia and Cerastium Soil lichens mosses and liverworts can cover ahigh percentage of the surface particularly in more maritime areas such as NovayaZemlya (Alexandrova 1980) Rushes (L uzula and Juncus) and grasses (AlopecurusPuccinellia Phippsia and Dupontia) are the main graminoid groups Sedges(Cyperaceae) are rare and wetlands lack organic peat layers There is little contrastin the composition of vegetation on mesic sites streamside sites and snowbeds Thevascular plant ora is extremely depauperate consisting of only about 50ndash60 species(Young 1971) On ne grained soils the extremely cold temperatures and the thinsparse plant canopy induce intense frost activity which forms networks of small(lt50 cm diameter) nonsorted hummocks and polygons and plants are con nedmainly to the depressions between the hummocks (Chernov and Matveyeva 1997)

312 Subzone B Prostrate dwarf-shrub subzoneSubzone B includes parts of the southern Queen Elizabeth Islands the eastern

fringe of Ellesmere Island much of Peary Land in Greenland the eastern portionof Svalbard central Novaya Zemlya the northern coast of the Taimyr Peninsulaand the New Siberian Islands Because of its small size and weakly de ned distin-guishing characteristics Subzone B could be characterized as a transition zone toSubzone C Several treatments of arctic zonation include subzones B and C in asingle subzone (Matveyeva 1998 Yurtsev 1994) The mean July temperature at thesouthern boundary of Subzone B is approximately 5degC This subzone has scatteredprostrate (creeping) dwarf shrubs on zonal soils Erect shrubby vegetation is lackingMesic low-elevation surfaces with ne-grained soils generally have open patchyplant cover generally with 5ndash25 cover of vascular plants Well-diVerentiatedsnowbed and riparian plant communities are lacking Nonsorted circles stripes andice-wedge polygons are common Vascular plant vegetation is often con ned tocracks and depressions in the polygonal network and areas irrigated by runoV fromsnow patches The dominant growth forms on mesic sites are prostrate dwarf shrubs

(eg Dryas Salix arctica and S polaris) forbs (eg Draba Saxifraga MinuartiaCerastium and Papaver) graminoids (eg Carex stans Carex rupestris Alopecurusalpinus Deschampsia borealis and L uzula confusa) mosses and lichens Rushes(L uzula) are also an important component of many mesic vegetation types Sedges

(Carex Eriophorum) often are dominant in wet areas

313 Subzone C Hemiprostrate dwarf-shrub subzone

Subzone C includes northern Banks Island northern Victoria Island DevonIsland Prince of Wales Island Somerset Island northern BaYn Island most ofEllesmere Island Axel Heiberg Islands the west coast of Greenland and inner ordregions of northeast Greenland southern Novaya Zemlya some coastal areas of

Yakutia and Chukotka and the northernmost coast of Alaska (Yurtsev 1994) Themean July temperature at the southern boundary of Subzone C is about 7degC Mesiczonal surfaces in this subzone have more luxuriant plant growth than that in SubzoneB Zonal sites are generally well vegetated but vascular-plant cover is still open and

D A Walker et al4556

Table

1

App

roxim

ate

equ

ivale

nt

sub

div

isio

ns

of

the

Arc

tic

Zo

ne

inR

uss

ia

No

rth

Am

eric

aand

Fen

nosc

and

ia

Ru

ssia

Nort

hA

mer

ica

Fen

nosc

andia

Edlu

nd

and

Yurt

sev

Ale

xan

dro

vaM

atv

eyev

aW

alk

erP

olu

nin

Alt

(198

9)

Tuhk

anen

Elv

ebakk

Su

bzo

ne

(199

4)

(198

0)

(1998

)(2

000

)(1

951

19

60)

Edlu

nd

(199

6)

Bliss

(1997

)(1

986

)(1

999

)

AH

igh

arct

icN

ort

her

nP

ola

rdes

ert

Cush

ion

-forb

Hig

harc

tic3

Her

bac

eou

sand

Hig

harc

tic

Inner

pola

rA

rcti

cpola

rtu

ndra

pola

rdes

ert

subzo

ne

crypto

gam

zone

des

ert

zone

(C=

00

deg)5

South

ern

Oute

rpola

rp

ola

rdes

ert

(1ndash3deg

C

zone

(15

ndash2deg

C)1

(2ndash3degC

)250

ndash15

0T

DD

)4(C

=0

5deg)

BA

rcti

ctu

nd

randash

No

rth

ern

Arc

tic

tun

dra

Pro

stra

teH

erb

-pro

stra

tesh

rub

Nort

her

nN

ort

her

nnort

her

nvari

ant

arc

tic

tundra

dw

arf

-shru

btr

ansi

tion

arc

tic

zone

arc

tic

tundra

subzo

ne

(3ndash4deg

C

zone

150ndash250

TD

D)

Pro

stra

tesh

rub (4

ndash6deg

C

250ndash350

TD

D)

(C=

10deg)

CA

rcti

ctu

nd

randash

South

ern

Mid

dle

arc

tic

Dw

arf

and

pro

stra

teM

idd

lear

ctic

Mid

dle

arc

tic

south

ern

vari

ant

arc

tic

tundra

shru

bzo

ne

tundra

zone

(5ndash7deg

C

(5ndash6degC

)(7

degC)

gt350

TD

D)

(C=

17

5deg)

DN

ort

her

nN

ort

her

nT

ypic

altu

ndra

Ere

ctdw

arf-

Low

arct

icL

ow

erec

tsh

rub

Low

arct

icSo

uth

ern

South

ern

hyp

oar

ctic

subarc

tic

shru

bsu

bzo

ne

arc

tic

zone

arc

tic

tundra

tundra

tundra

zone

Mid

dle

subarc

tic

tundra

(8ndash10

degC)

(9degC

)

ESouth

ern

South

ern

South

ern

Low

-shru

bA

rcti

csh

rub

hyp

oar

ctic

subarc

tic

tundra

subzo

ne

tundra

zone

tundra

tundra

(10ndash12

degC)

(10ndash12deg

C)

(7ndash10deg

C)

(C=

25deg)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4557

interrupted by frost scars and other periglacial features The main features distin-guishing Subzone C from Subzone B are the presence of the hemiprostrate shrubCassiope tetragona and well-diVerentiated plant communities in mires snowbeds andcreek sides Cassiope covers large areas on mesic sites in areas with acidic parentmaterial such as on Svalbard BaYn Island and interior portions of GreenlandHowever Cassiope is lacking on mesic alkaline surfaces in the western CanadianIslands In these areas Cassiope is generally con ned to snow accumulation areasSubzone C has much greater species diversity than Subzone B Sedges such asKobresia myosuroides Carex bigelowii and Carex rupestris are common on uplandsurfaces and numerous forbs such as members of the Fabaceae (eg OxytropisAstragalus) are important in nonacidic regions Communities in intrazonal habitats(snowbeds and creek sides) are well developed Wetland communities are much betterdeveloped than in Subzone B Creek sides have distinctive Epilobium latifoliumcommunities

314 Subzone D Erect dwarf-shrub subzoneSubzone D covers the southern parts of Banks Island and Victoria Island much

of Keewatin southern BaYn Island most of southern Greenland and a broad bandacross Siberia and Chukotka Climatically Subzone D periodically receives relativelytemperate air from the south during the summer while Subzone C receives predomin-ately arctic air masses The mean July temperature at the southern boundary ofSubzone D is about 9degC The boundary between subzones C and D is considered ofhighest rank because it separates the northern drier tundras on mineral soils fromthe southern relatively moist tundras with moss carpets and peaty soils (Alexandrova1980) This is approximately equivalent to the boundary between Blissrsquo High andLow Arctic (Bliss 1997) The major diVerence in pedology causes dramatic changesto the vegetation The plants in Subzone D have strong hypoarctic (boreal forest)aYnities (Yurtsev et al 1978) Important hypoarctic species such as birch (eg Betulanana) alder (Alnus) willow (Salix) and heath plants (Ericaceae and Empetraceae)extend their ranges from the lower layer of subarctic woodlands but are not dominantas they are in Subzone E Low shrubs (gt40 cm tall ) occur along streams Overallthe role of shrublands is much less prominent than in Subzone E The plant canopyis usually interrupted by patches of bare soil caused by nonsorted circles stripesand a variety of other periglacial features (lsquospotty tundrarsquo in the Russian literature)Vascular plants generally cover about 50ndash80 of the surface Zonal vegetation ongently sloping upland surfaces consists of sedges (eg Carex bigelowii Carex mem-branacea Eriophorum triste E vaginatum and Kobresia myosuroides) prostrate anderect dwarf (lt40 cm tall ) shrubs (eg Salix planifolia S lanata ssp richardsonii

Footnote to Table 1

1Mateveyeva (1998) Range of mean July temperature at southern boundary of subzone2Walker (2000) Approximate mean July temperature at the southern boundary of the

subzone3Polunin (1951) Zones based roughly on openness of ground cover High arctic very

sparsely vegetated middle arctic open vegetation carpet low arctic closed vegetation carpet4Edlund and Alt (1989) Zones de ned on the bases of mean July temperature and sum

of mean temperature of days exceeding 0degC (thawing degree days TDD)5Tuhkanen (1986) Southern boundary of zones de ned by Holdridge biotemperature as

the sum of mean monthly temperatures gt0degC divided by 12

D A Walker et al4558

Figure 3 Bioclimatic subzones in the Arctic Modi ed from Elvebakk et al (1999)

S reticulata S arctica Betula exilis and Dryas integrifolia) and mosses Prostrate-dwarf-shrub communities which were common on zonal surfaces in Subzone C arecon ned mainly to wind-swept sites The moss layer consisting primarily of

T omentypnum Hylocomium Aulacomnium and Sphagnum contributes to the develop-ment of organic soil horizons on ne-grained soils Soils in Subzone D have thinpeaty horizons and ne-grained soils are often nonacidic whereas soils in SubzoneE are usually acidic regardless of texture

There is more regional variation in the zonal vegetation than in subzones A Band C Tussock tundra consisting of cottongrass tussocks (Eriophorum vaginatum)and dwarf shrubs is common on ne-grained acidic soils over much of northeasternSiberia and northern Alaska (Walker et al 1994) particularly in areas that wereunglaciated during the last part of the Pleistocene In transitional areas to SubzoneC and on nonacidic loess Dryas spp and Cassiope tetragona are important (Walkerand Everett 1991) Some continental areas of Russia have dry steppe tundras thatare relicts of cold dry Pleistocene vegetation (Yurtsev 1982)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4556

Table

1

App

roxim

ate

equ

ivale

nt

sub

div

isio

ns

of

the

Arc

tic

Zo

ne

inR

uss

ia

No

rth

Am

eric

aand

Fen

nosc

and

ia

Ru

ssia

Nort

hA

mer

ica

Fen

nosc

andia

Edlu

nd

and

Yurt

sev

Ale

xan

dro

vaM

atv

eyev

aW

alk

erP

olu

nin

Alt

(198

9)

Tuhk

anen

Elv

ebakk

Su

bzo

ne

(199

4)

(198

0)

(1998

)(2

000

)(1

951

19

60)

Edlu

nd

(199

6)

Bliss

(1997

)(1

986

)(1

999

)

AH

igh

arct

icN

ort

her

nP

ola

rdes

ert

Cush

ion

-forb

Hig

harc

tic3

Her

bac

eou

sand

Hig

harc

tic

Inner

pola

rA

rcti

cpola

rtu

ndra

pola

rdes

ert

subzo

ne

crypto

gam

zone

des

ert

zone

(C=

00

deg)5

South

ern

Oute

rpola

rp

ola

rdes

ert

(1ndash3deg

C

zone

(15

ndash2deg

C)1

(2ndash3degC

)250

ndash15

0T

DD

)4(C

=0

5deg)

BA

rcti

ctu

nd

randash

No

rth

ern

Arc

tic

tun

dra

Pro

stra

teH

erb

-pro

stra

tesh

rub

Nort

her

nN

ort

her

nnort

her

nvari

ant

arc

tic

tundra

dw

arf

-shru

btr

ansi

tion

arc

tic

zone

arc

tic

tundra

subzo

ne

(3ndash4deg

C

zone

150ndash250

TD

D)

Pro

stra

tesh

rub (4

ndash6deg

C

250ndash350

TD

D)

(C=

10deg)

CA

rcti

ctu

nd

randash

South

ern

Mid

dle

arc

tic

Dw

arf

and

pro

stra

teM

idd

lear

ctic

Mid

dle

arc

tic

south

ern

vari

ant

arc

tic

tundra

shru

bzo

ne

tundra

zone

(5ndash7deg

C

(5ndash6degC

)(7

degC)

gt350

TD

D)

(C=

17

5deg)

DN

ort

her

nN

ort

her

nT

ypic

altu

ndra

Ere

ctdw

arf-

Low

arct

icL

ow

erec

tsh

rub

Low

arct

icSo

uth

ern

South

ern

hyp

oar

ctic

subarc

tic

shru

bsu

bzo

ne

arc

tic

zone

arc

tic

tundra

tundra

tundra

zone

Mid

dle

subarc

tic

tundra

(8ndash10

degC)

(9degC

)

ESouth

ern

South

ern

South

ern

Low

-shru

bA

rcti

csh

rub

hyp

oar

ctic

subarc

tic

tundra

subzo

ne

tundra

zone

tundra

tundra

(10ndash12

degC)

(10ndash12deg

C)

(7ndash10deg

C)

(C=

25deg)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4557

interrupted by frost scars and other periglacial features The main features distin-guishing Subzone C from Subzone B are the presence of the hemiprostrate shrubCassiope tetragona and well-diVerentiated plant communities in mires snowbeds andcreek sides Cassiope covers large areas on mesic sites in areas with acidic parentmaterial such as on Svalbard BaYn Island and interior portions of GreenlandHowever Cassiope is lacking on mesic alkaline surfaces in the western CanadianIslands In these areas Cassiope is generally con ned to snow accumulation areasSubzone C has much greater species diversity than Subzone B Sedges such asKobresia myosuroides Carex bigelowii and Carex rupestris are common on uplandsurfaces and numerous forbs such as members of the Fabaceae (eg OxytropisAstragalus) are important in nonacidic regions Communities in intrazonal habitats(snowbeds and creek sides) are well developed Wetland communities are much betterdeveloped than in Subzone B Creek sides have distinctive Epilobium latifoliumcommunities

314 Subzone D Erect dwarf-shrub subzoneSubzone D covers the southern parts of Banks Island and Victoria Island much

of Keewatin southern BaYn Island most of southern Greenland and a broad bandacross Siberia and Chukotka Climatically Subzone D periodically receives relativelytemperate air from the south during the summer while Subzone C receives predomin-ately arctic air masses The mean July temperature at the southern boundary ofSubzone D is about 9degC The boundary between subzones C and D is considered ofhighest rank because it separates the northern drier tundras on mineral soils fromthe southern relatively moist tundras with moss carpets and peaty soils (Alexandrova1980) This is approximately equivalent to the boundary between Blissrsquo High andLow Arctic (Bliss 1997) The major diVerence in pedology causes dramatic changesto the vegetation The plants in Subzone D have strong hypoarctic (boreal forest)aYnities (Yurtsev et al 1978) Important hypoarctic species such as birch (eg Betulanana) alder (Alnus) willow (Salix) and heath plants (Ericaceae and Empetraceae)extend their ranges from the lower layer of subarctic woodlands but are not dominantas they are in Subzone E Low shrubs (gt40 cm tall ) occur along streams Overallthe role of shrublands is much less prominent than in Subzone E The plant canopyis usually interrupted by patches of bare soil caused by nonsorted circles stripesand a variety of other periglacial features (lsquospotty tundrarsquo in the Russian literature)Vascular plants generally cover about 50ndash80 of the surface Zonal vegetation ongently sloping upland surfaces consists of sedges (eg Carex bigelowii Carex mem-branacea Eriophorum triste E vaginatum and Kobresia myosuroides) prostrate anderect dwarf (lt40 cm tall ) shrubs (eg Salix planifolia S lanata ssp richardsonii

Footnote to Table 1

1Mateveyeva (1998) Range of mean July temperature at southern boundary of subzone2Walker (2000) Approximate mean July temperature at the southern boundary of the

subzone3Polunin (1951) Zones based roughly on openness of ground cover High arctic very

sparsely vegetated middle arctic open vegetation carpet low arctic closed vegetation carpet4Edlund and Alt (1989) Zones de ned on the bases of mean July temperature and sum

of mean temperature of days exceeding 0degC (thawing degree days TDD)5Tuhkanen (1986) Southern boundary of zones de ned by Holdridge biotemperature as

the sum of mean monthly temperatures gt0degC divided by 12

D A Walker et al4558

Figure 3 Bioclimatic subzones in the Arctic Modi ed from Elvebakk et al (1999)

S reticulata S arctica Betula exilis and Dryas integrifolia) and mosses Prostrate-dwarf-shrub communities which were common on zonal surfaces in Subzone C arecon ned mainly to wind-swept sites The moss layer consisting primarily of

T omentypnum Hylocomium Aulacomnium and Sphagnum contributes to the develop-ment of organic soil horizons on ne-grained soils Soils in Subzone D have thinpeaty horizons and ne-grained soils are often nonacidic whereas soils in SubzoneE are usually acidic regardless of texture

There is more regional variation in the zonal vegetation than in subzones A Band C Tussock tundra consisting of cottongrass tussocks (Eriophorum vaginatum)and dwarf shrubs is common on ne-grained acidic soils over much of northeasternSiberia and northern Alaska (Walker et al 1994) particularly in areas that wereunglaciated during the last part of the Pleistocene In transitional areas to SubzoneC and on nonacidic loess Dryas spp and Cassiope tetragona are important (Walkerand Everett 1991) Some continental areas of Russia have dry steppe tundras thatare relicts of cold dry Pleistocene vegetation (Yurtsev 1982)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4557

interrupted by frost scars and other periglacial features The main features distin-guishing Subzone C from Subzone B are the presence of the hemiprostrate shrubCassiope tetragona and well-diVerentiated plant communities in mires snowbeds andcreek sides Cassiope covers large areas on mesic sites in areas with acidic parentmaterial such as on Svalbard BaYn Island and interior portions of GreenlandHowever Cassiope is lacking on mesic alkaline surfaces in the western CanadianIslands In these areas Cassiope is generally con ned to snow accumulation areasSubzone C has much greater species diversity than Subzone B Sedges such asKobresia myosuroides Carex bigelowii and Carex rupestris are common on uplandsurfaces and numerous forbs such as members of the Fabaceae (eg OxytropisAstragalus) are important in nonacidic regions Communities in intrazonal habitats(snowbeds and creek sides) are well developed Wetland communities are much betterdeveloped than in Subzone B Creek sides have distinctive Epilobium latifoliumcommunities

314 Subzone D Erect dwarf-shrub subzoneSubzone D covers the southern parts of Banks Island and Victoria Island much

of Keewatin southern BaYn Island most of southern Greenland and a broad bandacross Siberia and Chukotka Climatically Subzone D periodically receives relativelytemperate air from the south during the summer while Subzone C receives predomin-ately arctic air masses The mean July temperature at the southern boundary ofSubzone D is about 9degC The boundary between subzones C and D is considered ofhighest rank because it separates the northern drier tundras on mineral soils fromthe southern relatively moist tundras with moss carpets and peaty soils (Alexandrova1980) This is approximately equivalent to the boundary between Blissrsquo High andLow Arctic (Bliss 1997) The major diVerence in pedology causes dramatic changesto the vegetation The plants in Subzone D have strong hypoarctic (boreal forest)aYnities (Yurtsev et al 1978) Important hypoarctic species such as birch (eg Betulanana) alder (Alnus) willow (Salix) and heath plants (Ericaceae and Empetraceae)extend their ranges from the lower layer of subarctic woodlands but are not dominantas they are in Subzone E Low shrubs (gt40 cm tall ) occur along streams Overallthe role of shrublands is much less prominent than in Subzone E The plant canopyis usually interrupted by patches of bare soil caused by nonsorted circles stripesand a variety of other periglacial features (lsquospotty tundrarsquo in the Russian literature)Vascular plants generally cover about 50ndash80 of the surface Zonal vegetation ongently sloping upland surfaces consists of sedges (eg Carex bigelowii Carex mem-branacea Eriophorum triste E vaginatum and Kobresia myosuroides) prostrate anderect dwarf (lt40 cm tall ) shrubs (eg Salix planifolia S lanata ssp richardsonii

Footnote to Table 1

1Mateveyeva (1998) Range of mean July temperature at southern boundary of subzone2Walker (2000) Approximate mean July temperature at the southern boundary of the

subzone3Polunin (1951) Zones based roughly on openness of ground cover High arctic very

sparsely vegetated middle arctic open vegetation carpet low arctic closed vegetation carpet4Edlund and Alt (1989) Zones de ned on the bases of mean July temperature and sum

of mean temperature of days exceeding 0degC (thawing degree days TDD)5Tuhkanen (1986) Southern boundary of zones de ned by Holdridge biotemperature as

the sum of mean monthly temperatures gt0degC divided by 12

D A Walker et al4558

Figure 3 Bioclimatic subzones in the Arctic Modi ed from Elvebakk et al (1999)

S reticulata S arctica Betula exilis and Dryas integrifolia) and mosses Prostrate-dwarf-shrub communities which were common on zonal surfaces in Subzone C arecon ned mainly to wind-swept sites The moss layer consisting primarily of

T omentypnum Hylocomium Aulacomnium and Sphagnum contributes to the develop-ment of organic soil horizons on ne-grained soils Soils in Subzone D have thinpeaty horizons and ne-grained soils are often nonacidic whereas soils in SubzoneE are usually acidic regardless of texture

There is more regional variation in the zonal vegetation than in subzones A Band C Tussock tundra consisting of cottongrass tussocks (Eriophorum vaginatum)and dwarf shrubs is common on ne-grained acidic soils over much of northeasternSiberia and northern Alaska (Walker et al 1994) particularly in areas that wereunglaciated during the last part of the Pleistocene In transitional areas to SubzoneC and on nonacidic loess Dryas spp and Cassiope tetragona are important (Walkerand Everett 1991) Some continental areas of Russia have dry steppe tundras thatare relicts of cold dry Pleistocene vegetation (Yurtsev 1982)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4558

Figure 3 Bioclimatic subzones in the Arctic Modi ed from Elvebakk et al (1999)

S reticulata S arctica Betula exilis and Dryas integrifolia) and mosses Prostrate-dwarf-shrub communities which were common on zonal surfaces in Subzone C arecon ned mainly to wind-swept sites The moss layer consisting primarily of

T omentypnum Hylocomium Aulacomnium and Sphagnum contributes to the develop-ment of organic soil horizons on ne-grained soils Soils in Subzone D have thinpeaty horizons and ne-grained soils are often nonacidic whereas soils in SubzoneE are usually acidic regardless of texture

There is more regional variation in the zonal vegetation than in subzones A Band C Tussock tundra consisting of cottongrass tussocks (Eriophorum vaginatum)and dwarf shrubs is common on ne-grained acidic soils over much of northeasternSiberia and northern Alaska (Walker et al 1994) particularly in areas that wereunglaciated during the last part of the Pleistocene In transitional areas to SubzoneC and on nonacidic loess Dryas spp and Cassiope tetragona are important (Walkerand Everett 1991) Some continental areas of Russia have dry steppe tundras thatare relicts of cold dry Pleistocene vegetation (Yurtsev 1982)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4559

315 Subzone E L ow-shrub subzoneSubzone E is the warmest part of the Arctic Tundra Zone with mean July

temperatures of 10ndash12degC In Subzone E the zonal vegetation is dominated by hypo-arctic low shrubs that are often greater than 40 cm tall (eg Betula nana B exilisB glandulosa Salix glauca S phylicifolia S planifolia S richardsonii and Alnusspp) True shrub tundra with dense canopies of birch willows and sometimes alder(Alnus) occur in many areas Birch or willow thickets 80 to 200cm tall occur onzonal sites in some moister areas such as west Siberia and northwest Alaska Inmore continental areas and areas with less snow cover the shrubs are shorter andform a more open canopy Tussock tundra is common in northern Alaska andeastern Siberia and contains more shrubs than in Subzone D (Alexandrova 1980)Low and tall (gt2 m) shrubs are abundant in most watercourses Toward the southernpart of Subzone E in at areas that are continuous with the boreal forest patchesof open forest penetrate into this area along riparian corridors These woodlandsconsist of a variety of species of spruce (Picea) pine (Pinus) cottonwood (Populus)and larch (L arix) and tree birches (Betula) Peat plateaus (palsas) up to 15 m talloccur in lowland areas

316 Altitudinal zonationAdiabatic cooling of air masses at higher elevations causes altitudinal zonation

For continental areas the environmental adiabatic lapse rate is about 6degC per 1000 m(Barry 1981) This is equivalent to a shift in bioclimatic subzone for about every333-m elevation gain A topographic map was made with elevation isolines at 333667 1000 1333 1667 and 2000m ( gure 4) These show roughly where altitudinalzonation shifts can be expected

32 Floristic variation within the zonesRussian geobotanists have described longitudinal subdivisions within the sub-

zones which are based primarily on oristic diVerences (Alexandrova 1980 Yurtsev1994) These divisions are useful for characterizing the considerable eastndashwest oristicvariation within the subzones particularly in subzones C D and E In the morenorthern two subzones the Arctic has a relatively consistent core of circumpolararctic plant species that occur around the circumpolar region Further south localeast-west variation is related to a variety of factors including diVerent paleo-historiesand the greater climatic heterogeneity Large north-south trending mountain rangesprimarily in Asia have also restricted the exchange of species between parts of theArctic (Alexandrova 1980) Yurtsev (1994) delineated six oristic provinces and 22subprovinces and has discussed their characteristics These have recently been revisedby the Panarctic Flora Project ( gure 5) (Elvebakk et al 1999) The Northern Alaskasubprovince is used below in an example for a framework for a circumpolar arcticvegetation map (see table 2) This area covers the region north of the Brooks Rangefrom the Mackenzie River westward to about Point Lay

33 T he role of parent materialVegetation patterns related to parent material diVerences are extensive and there-

fore important to global and regional scale modelling eVorts There is a rich literaturedescribing the peculiarities of oras and vegetation on carbonate and ultrama crocks saline soils and ne vs coarse textured soils in the Arctic (Edlund 1982Elvebakk 1982 Cooper 1986 Walker et al 1998) These diVerences in parent material

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4560

NW

C

a n a d a

Ka ni n - Pechor a

J a n May en

S va l bard -

F J L an d

N ov a ya Zem ly a

Pola r U ral -

Ya mal - G ydan

T a i m y r

Kh a rau la k hYana - Kolym a

W C huk otka

S Chu kotka

E C h uk otka

NAlaska - Yukon

Be rin gianAlas ka

Cen tr al Canada E llesm ere - Pe ary Land

Hudso n

-Labrador

WG reen land

E G re en landN Ice land -

N Fennosc and ia

W ra ng el Isl

Anabar - O le nye k

Figure 4 Floristic sectors within the Arctic From Elvebakk et al (1999)

have important eVects on ecosystem patterns and processes Some of the mostimportant vegetation eVects are related to substrate pH Sylvia Edlundrsquos diagram in gure 6 illustrates well diVerences in plant communities on alkaline and acidicsubstrates across zonal boundaries in the Canadian High Arctic Similar patternshave been described on Svalbard (Elvebakk 1985) and northern Alaska (Walkeret al 1994) Vegetation on nonacidic and acidic parent materials can be determinedfor most regions of the Arctic using available soil and sur cial-geology maps in aGIS context Extensive saline areas occur in parts of the Russian arctic Other parentmaterial subdivisions could be made for global mapping if they were found to beregionally extensive and important to ecosystem processes

34 T he role of topographyAt landscape scales (1ndash100 km2 ) topography strongly in uences soil moisture

which is an important control of patterns of tundra plant communities and tundraecosystem processes (Webber 1978) The eVect of topography can be portrayed alonghill-slopes as a mesotopographic gradient ( gure 7 and Billings 1973) Generally

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4561

Figure 5 Topography of the circumpolar Arctic showing colour breaks corresponding toapproximate altitudinal zonation boundaries The environmental adiabatic lapse rateof 6degC per 1000m creates a shift in equivalent bioclimatic subzone about every 333-melevation gain

only extensive wetlands and dry mountainous areas are large enough to be displayedat the 175M scale Plant communities in snowbeds and along streams are usuallytoo small to map but are important components of regional vegetation mosaics andare used to help diVerentiate the vegetation in diVerent landscape units and complexesof plant communities in diVerent bioclimate subzones

35 Hierarchical tables summarizing regional plant communitiesThe plant community summary table and associated look-up tables are the

foundation of vegetation knowledge for the CAVM Separate tables have beencompiled for each oristic province An example from northern Alaska is shown intable 2 The major columns of the tables are the subzones and the subcolumns arethe various parent material types The rows are the major habitats along the mesoto-pographic gradient Plant communities are listed with a unique identi cation (ID)

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4562T

able

2

Sum

mary

tab

lefo

rp

lan

tco

mm

un

itie

sin

the

No

rth

Ala

ska

Sub

pro

vin

ce(p

art

of

Ala

ska

o

rist

icre

gio

n)

Do

min

ant

pla

nt

com

mun

itie

socc

urr

ing

inm

ajo

rhabit

ats

alo

ng

the

mes

oto

pogra

ph

icgra

die

nt

on

aci

dic

and

no

naci

dic

subst

rate

s1in

Sub

zon

esC

D

an

dE

S

ub

zon

esA

and

Bare

mis

sing

inN

ort

her

nA

lask

a

The

bre

ak

bet

wee

naci

dic

and

non

acid

icso

ils

isapp

roxi

mate

lyp

H55

D

ata

are

mis

sing

for

the

Ber

ingia

nA

lask

aS

ub

pro

vin

ceand

are

assu

med

tobe

sim

ilar

ton

ort

her

nA

lask

a

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Dry

exp

ose

dsi

tes

IS

paer

ophoru

sglo

bosu

sndash8

Ox

ytr

opis

bry

ophil

a-

13

Dia

pen

sia

lapponic

a-

18

Ox

ytr

opis

bry

ophil

a-

26

Sel

agin

ello

sibir

icaendash

38

Sel

agin

ello

sibir

icaendash

Luzu

laco

nfu

sasu

bty

pe

Dry

as

inte

gri

foli

aco

mm

S

ali

xphle

bophyll

aco

mm

D

ryas

inte

gri

foli

aco

mm

D

ryadet

um

oct

opet

ala

eD

ryadet

um

oct

opet

ala

eS

ali

xro

tundif

oli

aco

mm

(=

Walk

er(1

985

)T

yp

e(=

Ko

mark

ova

an

d(=

Walk

er(1

985

)T

yp

e(W

alk

eret

al

(1994

))(W

alk

eret

al

1994

)sa

me

(Eli

as

etal

(1996

)=B

12

=U

nit

18

this

tab

le)

Web

ber

(1980

)M

ap

2

B1

=U

nit

8th

ista

ble

(D

rygra

vel

lyso

ils)

as

Un

it26

this

tab

le(D

ryN

od

um

II

Web

ber

(Dry

no

naci

dic

gra

vel

lyU

nit

6)

(Dry

san

dy

site

s)als

oM

D

W

alk

ergra

vel

lyso

ils)

27

Sali

ciphle

bophyll

aendash

(1978

))(D

ryb

each

and

site

s)(1

990

))(D

ryn

on

aci

dic

Arc

toet

um

alp

inae

river

terr

ace

s)gra

vel

lysi

tes)

(Walk

eret

al

(1994

))(D

ryaci

dic

org

an

icso

ils)

Mes

iczo

nal

site

s2

Sphaer

ophoru

s9

Dry

as

inte

gri

foli

andash

14

Led

um

dec

um

ben

sndash19

Dry

as

inte

gri

foli

andash

28

Sphagno

ndash39

Dry

ado

inte

gri

folia

glo

bosu

sndashca

rex

aquati

lis

com

m

Eri

ophoru

mva

gin

atu

mE

riophoru

mtr

iste

Eri

ophore

tum

vagin

ati

Cari

cum

big

elow

iL

usu

laco

nfu

sasu

bty

pe

[=T

ype

U12

Walk

erco

mm

(=

Sta

nd

Typ

eU

3

(Walk

eret

al

(1994

))(W

alk

eret

al

(1994

)(m

ois

tS

ax

ifra

ga

foli

olo

saco

mm

1985

](M

esic

calc

are

ou

s(=

Map

2

Un

it8

Walk

er(1

985

))(M

esic

(Tu

sso

cktu

nd

rao

nca

lcare

ou

s(t

un

dra

)(E

lias

etal

(1996

)=co

ast

al

mea

do

ws)

Ko

mark

ova

an

dW

ebb

ern

on

aci

dic

loes

s)aci

dic

n

egra

ined

soils)

No

du

m1

Web

ber

(1978

)(1

980

))(M

esic

stab

iliz

ied

29

Sphagno

ndashT

yp

e5

Walk

er(1

977

))sa

nd

s)E

riophore

tum

vagin

ati

(Mes

ich

igh

-cen

tere

dbet

ulo

sum

nanae

po

lygo

ns)

(Walk

eret

al

(1994

))3

Sax

ifra

ga

cern

uandash

(Sh

rub

tun

dra

inw

arm

erC

are

xaquati

lis

com

m

mes

ocl

imati

co

ase

so

fth

e(E

lias

etal

(1996

)=fo

oth

ills

als

oalo

ng

wate

rN

od

um

IV

Web

ber

track

san

din

basi

ns)

(1978

)T

yp

e6

an

d7

30

Clim

ace

um

Walk

er(1

977

))(M

ois

tden

dro

ides

ndashaci

dic

n

e-gra

ined

soil

s)A

lnus

viri

dis

com

m

(Walk

eret

al

(1997

)sa

me

as

Un

it45

this

tab

le)

(Op

enald

ersh

rub

savann

as

inw

arm

erm

eso

clim

ati

co

ase

so

fth

efo

oth

ills

)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4563

Table

2

(Con

tinue

d)

Sub

zon

eC

Su

bzo

ne

DS

ub

zon

eE

Hab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(B

arr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(P

rudoe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Wet

site

s4

Callie

rgon

10

Dre

panocl

adus

15

Eri

ophoru

m20

Dre

panocl

adus

31

Sphagnum

ori

enta

lendash

40

Eri

ophoru

msa

rmen

tosu

mndash

Care

xbre

vifo

lius-

Care

xaquati

lis

angust

ifoli

um

ndashbre

vifo

liusndash

Care

xE

riophoru

msc

heu

chze

riangust

ifoli

um

ndashC

are

xaquati

lis

com

m

com

m

Care

xaquati

lis

com

m

aquata

lis

com

m

com

m

aquati

lis

com

m

(=N

od

aV

an

dV

I(=

Typ

eM

10

Walk

er(=

Map

2

Un

it16

J=st

an

dT

yp

eM

2

(Walk

erand

Walk

er(M

D

W

alk

eret

al

1996

)W

ebb

er(1

978

)T

yp

e9

(1985

))(W

etca

lcare

ou

sK

om

ark

ova

an

dW

ebb

erW

alk

er(1

985

)(W

et1996

)(W

etm

icro

site

sin

(No

naci

dic

mars

hes

)10

12

an

d13

Walk

erco

ast

al

mea

do

ws)

(1980

))(W

etm

ars

hes

)ca

lcare

ou

sm

ead

ow

sic

e-w

etaci

dic

tun

dra

in(1

977

)E

riophoru

mw

edge

po

lygo

nce

nte

rs

foo

thills

)angust

ifolium

ndashla

ke

marg

ins)

32

Spagnum

lenen

sendash

Care

xaquati

lis

com

m

Salix

fusc

ensc

ens

com

m

Eli

as

etal

(1996

))(W

et(W

alk

erand

Walk

eraci

dic

site

sw

ith

ou

t1996

)(R

ais

edm

icro

site

sst

an

din

gw

ate

r)in

aci

dic

wet

lan

ds)

Sn

ow

bed

s5

Salix

rotu

ndifoli

andash

11

No

data

P

rob

ab

ly16

Boykin

iari

chard

sonii

ndash21

Dry

as

inte

gri

foli

andash

33

Cari

cim

icro

chaet

aendash

41

Dry

as

inte

gri

foli

andash

Cet

rari

adel

esii

com

m

sim

ilar

toU

nit

21

this

Cass

iope

tetr

agona

com

m

Cass

iope

teta

gona

com

m

Cass

iope

tetr

agona

com

m

Cass

iope

tetr

agona

com

m

(Eli

as

etal

(1996

))ta

ble

(=M

ap

2

Un

it11

(=S

tan

dT

yp

eU

6

(Walk

eret

al

1994

)(M

D

W

alk

eret

al

(1994

))(E

arl

y-m

elti

ng

sno

wb

eds

Ko

mark

ova

an

dW

ebb

erW

alk

er(1

985

)S

tan

d(E

arl

y-m

elti

ng

aci

dic

(Earl

y-m

elti

ng

no

naci

dic

nea

rth

eco

ast

)(1

980

))(E

arl

y-m

elti

ng

typ

eC

ass

iope

tetr

adona-

sno

wb

eds)

sno

wb

eds)

sno

wbed

ssi

mil

ar

toD

ryas

inte

gri

foli

a

M

D

6

Phip

psi

aalg

idandash

34

Saxif

raga

niv

alisndash

42

Salix

rotu

ndif

olia

ndashU

nit

33

this

tab

le)

Walk

er(1

990

))(E

arl

yS

ax

ifra

ga

rivu

lari

sco

mm

S

alix

rotu

ndif

olia

com

m

Sax

rifa

ga

riva

lis

com

m

mel

tin

gn

on

aci

dic

(=N

od

um

VII

IW

ebb

er(=

M

D

Walk

eret

al

(=M

D

W

alk

eret

al

sno

wb

eds)

(1978

)T

yp

e15

Walk

er(1

989

)si

mil

ar

toU

nit

22

(1989

)(s

imilar

toU

nit

22

(1977

))(L

ate

-mel

tin

g22

Eri

ophoru

mtr

iste

ndashth

ista

ble

))(L

ate

-mel

tin

gth

ista

ble

)(L

ate

mel

tin

gsn

ow

bed

sn

ear

the

coast

)S

ali

xro

tundif

olia

com

m

sno

wb

eds)

sno

wb

eds)

(=S

tan

dT

yp

eU

7

Walk

er(1

985

)S

alix

rotu

ndifoli

andashE

riophoru

mtr

iste

M

D

W

alk

er(1

990

))(L

ate

mel

tin

gn

on

aci

dic

sno

wb

eds)

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4564

Table

2

(Conti

nue

d)

Su

bzo

ne

CS

ub

zon

eD

Su

bzo

ne

EH

ab

itat

alo

ng

mes

oto

po

gra

ph

icA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sA

cidic

subst

rate

sN

onaci

dic

subst

rate

sgra

die

nt

(Barr

ow

)(P

rudhoe

Bay

coast

)(A

tqasu

k)

(Pru

doe

Bay

inla

nd)

(Im

nava

itC

reek

)(T

ooli

kL

ake)

Str

eam

sid

es7

Phip

psi

aalg

idandash

12

Epil

obiu

mla

tifo

lium

ndash17

Sali

xla

nata

ndashS

alix

23

Epilobio

lati

foli

indash35

Vale

riano

capit

ata

endash

43

Epilobio

lati

folii-

Coch

leari

aoY

cinali

sA

rtem

sia

arc

tica

ala

xen

sis

Sali

cetu

mala

xen

sis

ass

S

ali

cetu

mpla

nif

oliae

(Ass

S

alice

tum

ala

xen

sis

(Ass

(=

No

du

mV

III

Web

ber

(Po

ssib

ly=

Epil

obio

(=M

ap

2

Un

it17

pro

v

[S

chik

oV

inp

ress

]p

rov

Sch

iko

V

Inp

rov

Sch

iko

Vin

pre

ss)

(1978

)T

yp

e15

Walk

erla

tifo

liandashS

ali

cetu

mK

om

ark

ova

an

dW

ebb

er(a

ctiv

en

on

aci

dic

pre

ss=

Eri

ophoru

m(A

ctiv

e

oo

dp

lain

s)(1

977

)p

oss

ibly

equ

al

toala

xen

sis

ass

p

rov

(1980

))(s

trea

mb

an

k

oo

dp

lain

s)angust

ifoli

um

mdashS

ali

x44

Tall

shru

bver

sio

no

fU

nit

7

this

tab

le)

Sch

iko

Vin

pre

p

(Act

ive

shru

bla

nd

sp

rob

ab

lypurl

chra

com

m

24

Lo

wsh

rub

ver

sio

no

fS

alice

tum

gla

uco

ndash(U

nst

ab

lest

ream

marg

ins

no

naci

dic

o

od

pla

ins

equ

al

toU

nit

23

this

(Walk

eret

al

(1994

))S

ali

cetu

mgla

uco

ndashri

chard

sonii

at

coast

)n

ear

coast

)ta

ble

)(A

cid

icto

circ

um

neu

tral

rich

ard

sonii

(Ass

p

rov

Sch

iko

Vin

pre

ss)

wate

rtr

ack

sst

ream

sid

es)

(Ass

p

rov

Sch

iko

Vin

(Wil

low

shru

bla

nd

so

np

ress

=

Sta

nd

Typ

eU

8)

36

Lo

wsh

rub

ver

sio

no

fst

ab

le

oo

dp

lain

s)W

alk

er(1

985

)(W

illo

wS

ali

cetu

mgla

uco

ndash45

Cll

imace

um

den

dro

ides

ndashsh

rub

lan

ds

on

stab

leri

chard

sonii

Aln

us

viri

dis

com

m

o

od

pla

ins)

(Ass

p

rov

Sch

iko

Vin

(Walk

eret

al

(1997

))p

ress

)(W

illo

wsh

rub

lan

ds

25

Dry

as

inte

gri

foli

andash

(Ald

ersh

rub

lan

ds

on

stab

le

oo

dp

lain

s)L

upin

isa

rcti

caco

mm

oo

dp

lain

s)(W

alk

eret

al

(1997

))37

Clim

ace

um

46

Dry

as

inte

gri

foli

andash

(Dry

river

terr

ace

s)den

dro

ides

ndashAlm

us

viri

dis

Lupin

us

arc

tica

com

m

com

m

(Walk

eret

al

(1997

))(D

ry(W

alk

eret

al

1997

)ri

ver

terr

ace

s)(A

lder

shru

bla

nd

s

oo

dp

lain

s)

1Su

rface

dep

osi

tsat

the

rep

rese

nta

tive

site

sB

arro

wmdash

acid

icm

ari

ne

sand

san

dgra

vels

P

red

ho

eB

ayco

astmdash

calc

are

ou

sgl

aci

al

ou

twash

A

tqas

uk

mdashaci

dic

colian

sand

sP

rud

ho

eB

ayis

lan

dmdash

calc

are

ou

slo

ess

Imm

ava

itC

reek

mdashaci

dic

mid

-Ple

isto

cene

gla

cial

tilt

T

oo

lik

Lakemdash

calc

areo

us

lake-

Ple

isto

cen

egla

cial

tilt

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4565

Figure 6 Vegetation on acidic and alkaline substrates in subzones A and B Modi ed fromEdlund and Alt (1989) Species shown are 1 L uzula confusa 1b L arctica 2 Papaverradicatum 3 Potentilla hyparctica 4 Alopecurus alpinus 5 Phippsia algida 6 Saxifragaoppositifolia 7 Poa abbreviata 8 Draba sp 11 Carex aquatilis var stans 12Pleuropogon sabinei 14 Eriophorum triste 15 E scheuchzeri 17 Dryas integrifolia 18Cassiope tetragona 19 Arctophila fulva 20 Hippuris vulgaris 21 Oxytropis arctobia22 Hierochloe alpina

Figure 7 Conceptual mesotopographic gradient for arctic landscapes

number a two-species plant community name and the publication where the com-munity is described Plant communities that are only described locally are givenlsquoplant community typersquo (comm) designations Those that have lsquoassociationrsquo (suYx-etum) names have been compared globally to types described from other areas andhave been incorporated into the BraunndashBlanquet syntaxonomic nomenclature system(WesthoV and van der Maarel 1978) This European phytosociological approachhas many advantages as an international classi cation system at the plant communitylevel because of its well-established procedures long history and wide applicationthroughout the Arctic (Daniels 1994 Elvebakk 1994 Dierssen 1996 M D Walkeret al 1995 Matveyeva 1998) An example table is shown for the Alaska oristicprovince (table 1)

More detailed information for the plant communities is contained in separate

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4566

look-up tables that are linked to this table via the plant community identi cation(ID) number The linked tables contain information such as dominant plant growthforms plant species horizontal structure vertical structure primary production andbiomass (see look-up table 2 in Walker 1999)

4 Integrated mapping methodsVegetation is interpreted on the basis of lsquointegrated vegetation complexesrsquo (IVC)

These IVCs are derived from a combination of information on satellite-derivedimages and information from existing source maps such as bedrock geology sur cialgeology soils hydrology and previous vegetation maps The procedures forde ning these complexes are summarized in gure 7 and described brie y belowSupplementary information can be found in Walker (1999)

First level geological features (mountains hills plains tablelands) are rst inter-preted directly from the AVHRR image (see step 1 and layer 1 in gure 8)

(a) (b)

Figure 8 Six-step procedure for making the CAVM See text for brief synopsis of each stepSummarized from Walker (1999)

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4567

Information from other simpli ed source maps such as bedrock geology sur cialgeology per cent water cover and soils are used to create additional map boundaries(steps 2 and 3 in gure 8) These procedures are based on the principle that geo-morphic features are the primary controls of vegetation boundaries This processconsiders water and landform boundaries as unalterable (hard) boundariesBoundaries controlling variables such as soils and vegetation which change acrossecotones or gradients are considered soft boundaries which are made to conformto landscape boundaries wherever possible The number of additional polygons isminimized through the integrated mapping procedures The result is the lsquointegratedland-unit maprsquo (ILUM) This map has all the essential terrain information

At the next step vegetation information from a variety of sources is used to createadditional vegetation boundaries on the integrated map (step 4 gure 8) The NDVImap ( gure 2) and other vegetation maps are used to help delineate other featuresthat may be visible on the AVHRR imagery Where possible the polygon boundariesare made to conform to those of the ILUM This new map is called the lsquointegratedvegetation complexrsquo map (IVCM) The names of the complexes generally correspondto landscape features such as lsquoacidic mountain complexrsquo lsquoacidic mire complex withless than 25 lakesrsquo lsquononacidic plateau basin or plain complexrsquo lsquonunatak complexrsquolsquoisland complexrsquo or lsquolarge river oodplain complexrsquo The IVCM is the only map thatneeds to be digitized in the mapping procedures Each polygon on the map is givena unique ID number An attribute table contains each polygon ID number followedby a series of codes that describe the polygonrsquos attributes (vegetation complex

bedrock geology sur cial geology per cent water cover etc)Step 5 of the mapping procedures has already been described in the creation of

the plant community summary table (see table 2) which contains the basic plant-community information for a given oristic region A look-up table contains detailedinformation for each community (plant species growth forms production biomassetc) This table is not shown here but an example can be found in table 3 of Walker(1999) The creation of the nal maps assumes that similar vegetation complexeswithin a given combination of oristic region and bioclimatic subzone will all havethe same suite of plant communities If there are known exceptions that do notconform to this logic these map polygons can be selected out and recoded Primarysecondary and tertiary (dominant and subdominant ) plant communities are listedfor each combination of subzone oristic region and vegetation complex At step6 a wide variety of vegetation-related maps can be created by reference to the linked

tables that list growth forms production and dominant species (Walker 1999)

5 ConclusionCircumpolar AVHRR-derived false (CIR) and maximum NDVI images are the

rst products from the Circumpolar Arctic Vegetation Mapping project They provide

the key means of making an image-interpreted vegetation map of the arctic tundrabiome The CAVM is essentially a GIS database that uses expert knowledge of therelationship of plant communities to climate parent material and topographic factorsto predict vegetation within landscape units that can be delineated on 175M scale

AVHRR images The database will be a source for creating a wide variety of mapproducts that will be useful for many aspects of arctic research and education The rst draft of the CAVM is expected in 2002 The methods outlined in this papercould be applied to vegetation maps of other biomes The images in this paper are

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

D A Walker et al4568

available through the UCARJoint OYce for Science Support Boulder CO USAe-mail jmooreucaredu

AcknowledgmentsAll the present and past participants in the CAVM project have contributed a

great deal to the ideas presented here Listed alphabetically they are Galina AnanievaNancy Auerbach Christian Bay Larry Bliss Udo Bohn Fred Daniels NickolaiDenisov Dmitri Drozdov Sylvia Edlund Eythor Einarsson Arve Elvebakk HelmutEpp Mike Fleming Gudmundur Gudjonsson Elena Golubeva Irina Ilyina BerntJohansen Seppo Kaitala Andrei Kapitsa Adrian Katenin Sergei Kholod YuriKorostelev Carl Markon Nadya Matveyeva Evgeny Melnikov Lia Meltzer DaveMurray Nataly Moskalenko Valentina Per lieva Alexei Polezhaev RaisaSchelkunova Christopher Smith Martha Raynolds Irina Safronova Steve TalbotSvein Tveitdal Maike Wilhelm Steve Young Konstantin Volotovskyi TatyanaYurkovskaya Boris Yurtsev and Steve Zoltai Special thanks to Kent Lethcoe andSteve Muller for help in preparation of the AVHRR images and to Fred DanielsArve Elvebakk Dave Murray and Boris Yurtsev for their thoughts on vegetationzonation Funding for much of the CAVM work and this paper came from theNational Science Foundation OPP-9908829

References

Alexandrova V D 1980 T he Arctic and Antarctic T heir Division into Geobotanical Areas(Cambridge Cambridge University Press)

Barry R G 1981 Mountain Weather and Climate (London Methuen)Bay C 1997 Floristical and ecological characterization of the polar desert zone of Greenland

Journal of Vegetation Science 8 685ndash696Bazilevich N I Tishkov A A and Vilchek G E 1997 Live and dead reserves and

primary production in polar desert tundra and forest tundra of the former SovietUnion In Polar and Alpine T undra edited by F E Wielgolaski (Amsterdam Elsevier)pp 509ndash539

Billings W D 1973 Arctic and alpine vegetation similarities diVerences and susceptibilityto disturbances BioScience 23 697ndash704

Bliss L C 1997 Arctic ecosystems of North America In Polar and Alpine T undra editedby F E Wielgolaski (Amsterdam Elsevier) pp 551ndash683

Bliss L C and Matveyeva N V 1992 Circumpolar arctic vegetation In Arctic Ecosystemsin a Changing Climate An Ecophysiological Perspective edited by F S Chapin IIIR L JeVeries J F Reynolds G R Shaver J Svoboda and E W Chu (San DiegoCA Academic Press Inc) pp 59ndash89

Chernov Y I and Matveyeva N V 1997 Arctic ecosystems in Russia In Polar andAlpine T undra edited by F E Wielgolaski (Amsterdam Elvesier) pp 361ndash507

Cooper D J 1986 Arctic-alpine tundra vegetation of the Arrigetch Creek Valley BrooksRange Alaska Phytocoenologia 14 467ndash555

Daniels F J A 1994 Vegetation classi cations in Greenland Journal of Vegetation Science5 781

Dierssen K 1996 Vegetation Nordeuropas (Stuttgart Ulmer)Edlund S A 1982 Vegetation of eastern Melville Island Open File no 852 Geological

Survey of CanadaEdlund S A and Alt B T 1989 Regional congruence of vegetation and summer climate

patterns in the Queen Elizabeth Islands Northwest Territories Canada Arctic 423ndash23

Elias S A Short S K Walker D A and Auerbach N A 1996 Final ReportHistorical Biodiversity at Remote Air Force Sites in Alaska Institute of Arctic andAlpine Research Boulder CO Final Report to the Department of Defense Air ForceLegacy Resource Management Program Project 0742

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments 4569

Elvebakk A 1982 Geological preferences among Svalbard plants Inter-Nord 16 11ndash31Elvebakk A 1985 Higher phytosociologicalsyntaxa on Svalbard and their use in subdivision

of the Arctic Nordic Journal of Botany 5 273ndash284Elvebakk A 1994 A survey of plant associations and alliances from Svalbard Journal of

Vegetation Science 5 791Elvebakk A 1999 Bioclimatic delimitation and subdivision of the Arctic In T he Species

Concept in the High NorthmdashA Panarctic Flora Initiative edited by I Nordal andV Y Razzhivin (Oslo The Norwegian Academy of Science and Letters) pp 81ndash112

Elvebakk A Elven R and Razzhivin V Y 1999 Delimitation zonal and sectorialsubdivision of the Arctic for the Panarctic Flora Project In T he Species Concept inthe High NorthmdashA Panarctic Flora Initiative edited by I Nordal and V Y Razzhivin(Oslo The Norwegian Academy of Science and Letters) pp 375ndash386

Komarsquorkovarsquo V and Webber P J 1980 Two Low Arctic vegetation maps near AtkasookAlaska Arctic and Alpine Research 12 447ndash472

Markon C J 1999 Characteristics of the Alaskan 1-km Advanced Very High ResolutionRadiometer data sets used for analysis of vegetation biophysical properties USGSOpen File Report 99ndash088 US Geological Survey

Markon C J and Walker D A 1999 Proceedings of the T hird International CircumpolarArctic Vegetation Mapping Workshop Open File Report 99ndash551 US Geological Survey

Matveyeva N V 1998 Zonation in Plant Cover of the Arctic Proceedings of the KomarovBotanical Institute No 21 (Russian Academy of Sciences) (in Russian)

Rannie W F 1986 Summer air temperature and number of vascular species in arcticCanada Arctic 39 133ndash137

Raynolds M K and Markon C J 2002 Fourth International Circumpolar ArcticVegetation Mapping Workshop Proceedings US Geological Survey Open File Report02-181

Razzhivin V Y 1999 Zonation of vegetation in the Russian Arctic In T he Species Conceptin the High NorthmdashA Panarctic Flora Initiative edited by I Nordal and V Y Razzhivin(Oslo The Norwegian Academy of Science and Letters) pp 113ndash130

Schickhoff U Walker M D and Walker D A 2002 Riparian willow communitieson the arctic Slope of Alaska and their environmental relationships a classi cationand ordination analysis Phytocoenologia in press

Sorensen T 1941 Temperature relations and phenology of the northeast Greenland oweringplants Meddelelser om Gronland 125 1ndash315

Walker D A 1977 The analysis of the eVectiveness of a television scanning densitometerfor indicating geobotanical features in an ice-wedge polygon complex at BarrowAlaska MA thesis University of Colorado Boulder

Walker D A 1985 Vegetation and environmental gradients of the Prudhoe Bay regionAlaska Hanover NH US Army Cold Regions Research and Engineering LaboratoryReport 85-14

Walker D A 1995 Toward a new circumpolar arctic vegetation map Arctic and AlpineResearch 31 169ndash178

Walker D A 1999 An integrated vegetation mapping approach for northern Alaska(1 4 M scale) International Journal of Remote Sensing 20 2895ndash2920

Walker D A 2000 Hierarchical subdivision of arctic tundra based on vegetation responseto climate parent material and topography Global Change Biology 6 19ndash34

Walker D A and Everett K R 1991 Loess ecosystems of northern Alaska regionalgradient and toposequence at Prudhoe Bay Ecological Monographs 61 437ndash464

Walker D A and Lillie A C editors 1997 Proceedings of the Second Circumpolar ArcticVegetation Mapping Workshop Arendal Norway 19ndash24 May 1996 and the CAVM-North America Workshop Anchorage Alaska USA 14ndash16 January 1997 (BoulderCO Institute of Arctic and Alpine Research) Occasional Paper No 52

Walker D A and Markon C J 1996 Circumpolar arctic vegetation mapping workshopAbstracts and short papers (Reston Virginia US Geological Survey) Open FileReport 96ndash251

Walker D A and Walker M D 1996 Terrain and vegetation of the Imnavait CreekWatershed In L andscape Function Implications for Ecosystem Disturbance a CaseStudy in Arctic T undra edited by J F Reynolds and J D Tenhunen (New YorkSpringer-Verlag) pp 73ndash108

Sixth Circumpolar Symposium on Remote Sensing of Polar Environments4570

Walker D A Bay C Daniels F J A Einarsson E Elvebakk A Johansen B EKapitsa A Kholod S S Murray D F Talbot S S Yurtsev B A andZoltai S C 1995 Toward a new arctic vegetation map a review of existing mapsJournal of Vegetation Science 6 427ndash436

Walker D A Auerbach N A Nettleton T K Gallant A and Murphy S M1997 Happy Valley Permanent Vegetation Plots Study Data Report Arctic SystemScience Flux Boulder CO Tundra Ecosystem Analysis and Mapping LaboratoryInstitute of Arctic and Alpine Research University of Colorado

Walker D A Auerbach N A Bockheim J G Chapin F S I Eugster W KingJ Y McFadden J P Michaelson G J Nelson F E Oechel W C PingC L Reeburg W S Regli S Shiklomanov N I and Vourlitis G L 1998Energy and trace-gas uxes across a soil pH boundary in the Arctic Nature 394469ndash472

Walker M D 1990 Vegetation and oristics of pingos Central Arctic Coastal Plain AlaskaDissertationes Botanicae 149 (Stuttgart J Cramer)

Walker M D Walker D A and Everett K R 1989 Wetland soils and vegetationArctic Foothills Alaska US Fish and Wildlife Service Biological Report 89(7)

Walker M D Walker D A and Auerbach N A 1994 Plant communities of a tussocktundra landscape in the Brooks Range Foothills Alaska Journal of Vegetation Science5 843ndash866

Walker M D Daniels F J A and van der Maarel E 1995 Circumpolar arcticvegetation Special Features in Vegetation Science 7 176

Webber P J 1978 Spatial and temporal variation of the vegetation and its productivityBarrow Alaska In Vegetation and Production Ecology of an Alaskan Arctic T undraedited by L L Tieszen (New York Springer-Verlag) pp 37ndash112

Westhoff V and van der Maarel E 1978 The BraunmdashBlanquet approach InClassi cation of plant communities edited by R H Whittaker (Den Haag Dr W Junk)pp 287ndash399

Young S B 1971 The vascular ora of St Lawrence Island with special reference to oristiczonation in the arctic regions Contributions from the Gray Herbarium 201 11ndash115

Yurtsev B A 1982 Relicts of the xerophyte vegetation of Beringia in northeastern Asia InPaleoecology of Beringia edited by D M Hopkins J V Matthews Jr C E Schwegerand S B Young (New York Academic Press) pp 157ndash177

Yurtsev B A 1994 Floristic division of the Arctic Journal of Vegetation Science 5 765ndash776Yurtsev B A Tolmachev A I and Rebristaya O V 1978 The oristic delimitation

and subdivision of the Arctic In T he Arctic Floristic Region edited by B A Yurtsev(Leningrad Nauka) pp 9ndash104