tGJ
!~,
\~
.UnitedDepartmentStates of
Agriculture
Forest
Service
Miscellaneous
Publication 1465
Cover-The semi-arid mountainsof the Wallowa National Forest, Oregoncontain
numerousexamplesof the landform influenceson ecosystempatterns. (USDA
Forest Servicephoto).
..\
~~~
United
DepartmentStates
tW
Agriculture
G
of
Forest
Service
Miscellaneous
Publication
1465
Robert G. Bailey1, Geographer
Land Management Planning Staff
USDA Forest Service
Washington, DC
November
1988
ICorrespondence
address: Land Management Planning Systems,
USDA Forest Service, 3825 East Mulberry Street, Fort Collins, CO
80524 (USA)
Bailey, Robert G. 1988. Ecogeographicanalysis:a guide
to the ecological division of land for resourcemanagement. Misc. Publ. 1465. Washington,DC: U. S.
Departmentof Agriculture. 18 p.
Ecological units of different sizes for predictive modeling
of resourceproductivity and ecologicalresponseto
managementneedto be identified and mapped. A set of
criteria for subdividing a landscapeinto ecosystemunits
of different sizesis presented,basedon differencesin factors important in differentiating ecosystemsat varying
scalesin a hierarchy. Practical applicationsof suchunits
are discussed.
Keywords: ecologicalland classification,landscapeecology, scale, ecosystemmapping,resourceplanning model
ii
Contents
Page
Introduction.
Role
Scale
Differentiating
Mesoscale:
Macroscale:
Microscale:
of of
Climate
Ecosystem
Landform
Edaphic-topoclimatic
Macroclimatic
Factors in
Ecosystem
Units
and Differentiation.Scale.
Differentiation.
Differentiation.
Differentiation.
..
1
2
2
3
3
6
8
LiteratureCited
Discussion.
Practical
Applications
13
14
16
iii
Introduction
Wildland planning efforts today focus largely on trying to
predict or model the behaviorof the ecosystemunder
differing kinds and intensitiesof management.In order to
do this, the land must be divided into component
ecosystemsthat reflect significant differencesin response
to managementand resourceproductioncapability.
The questionnow arises: How much detail is adequateto
model the area under analysis?
The answerto this questiondependson the complexity of
the area and the level of precisionneededby the manager. Ecosystemsoften exist naturally in very different
sizes,and can be identified at severalgeogtaphicalscales
and levels of detail in a hierarchical manner, ranging
from site-specificecosystemsto groups of spatially related
systemsknow as regions. One of the essentialframeworks
for predictive modeling, therefore, is an identification of
ecosystemsat different levels of detail that can be linked
to different levels of analysis.Ecogeographicanalysis is
the subdivisionof a landscapefor this purpose.
This approachstarts with mappingthe ecosystemsof various sizesunderlying the area for analysis.This requiresa
union of ecologyand geography.In thesefields there are
numeroustextbooksthat are lengthy, complete, definitive
works. Ecosystemclassificationhas beendescribedby
USDA ForestService (1982) and Driscoll and others
(1984). So far ecosystemboundarieshave receivedlittle
systematictreatment. Most textbooksdeal with boundary
problems superficially, and few specialessaysare available. One could argue, therefore, that a short accessible
monographthat treats the establishmentof ecosystem
boundariesis neededas a guide for thosewhose professions involve the analysisand managementof land. This
monographis written in an attemptto fulfill this need.
The basic conceptsaboutscaleand ecosystemsare discussedin textbooks on landscapeecologyand geography
(Isachenko1973; Leser 1976; Formanand Godron 1986).
A synthesisof theseconceptshas beenpresentedelsewhere by Bailey (1985). In a brief followup article,
Bailey (1987) reviewedthe literature to suggestpossible
criteria to make the conceptsoperationalthrough mapping. This monographexpandsupon thesearticles to
elaborateand illustrate the conceptsand criteria. It also
provides a discussionof practical applicationsin planning
and management.
Scale of EcosystemUnits
Role of Climate in EcosystemDifferentiation
Scaleimplies a certain level of perceiveddetail. Suppose,
for example, that an area of intermixed grasslandand
pine forest is examinedcarefully. At one scale, the grassland and the stand of pine are each spatiallyhomogeneous
and look uniform. Yet linkages of energyand material
exist betweenthesesystems.Having determinedthese
linkages,the locationally separatesystemsare intellectually combinedinto a new entity of higher order and
greatersize. These larger systemsrepresentpatternsor
associationsof linked, smallerecosystems.
Ecosystemsof different climatic areasdiffer sig~ificantly.
Thesedifferencesare the result of factors that control climatic regime defined as the diurnal and seasonalfluxes of
energyand moisture. The basic pattern of .fluctuationof
solar energyis responsiblefor the largestshareof periodicity and spatial variation of the earth's environment.
The changing intensity and duration of solar radiation
brings aboutchangesin the troposphere,stratosphere,
earth temperatureand seatemperature;influencesmigratory and hibernationpatternsof animals; and controls the
life cycles of much of the biota. Thus, an energy flow
that varies systematicallythrough time and spaceresults
in an environmentthat also varies through time and
space.The samecan be said of the moisture flow. Energy
and moisture regimes, in combination,are the dominant
controls of all biophysicalprocesses.
Schemesfor recognizing suchlinkageshave beenproposedand implementedin a number of countries (for
example, Zonneveld 1972). The nomenclatureand number
of levels in theseschemesvary. One scheme,proposedby
Miller (1978), recognizeslinkages at three scalesof perception. While not definitive, it illustratesthe nature of
theseschemes.The smallest, or local, ecosystems
(microecosystems)
are the homogeneoussites commonly
recognizedby forestersand range scientists.They are of
the size of hectares.
Linked sites create a landscapemosaic (mesoecosystem)
that looks like a patchwork. A landscapemosaicis made
up of spatially contiguoussites distinguishedby material
and energyexchange.They range in size from 10 km2 to
severalthousandkm2.
A classicexample of a landscapemosaicis a mountain
landscape.Betweenthe componentsystemsof a mountain
range there is lively exchangeof materials: water and
products of erosion move down the mountains;updrafts
carry dust and pieces of organic matter upward, and
downdrafts carry them downward; animals can move from
one systeminto the next; seedsare easily scatteredby the
wind or propagatedby birds.
At broader scales, landscapemosaicsare connectedto
form larger units (macroecosystems).
Mountainsand
plains are a case in point. For example,the lowland
plains of the Western United Statesas a mosaiccontrasts
with steeplandscapesin adjacentmountainranges.As
water from .themountainsflows to the valley, and as the
mountainsaffect the climate of the valley through sheltering, two large-scalelinkages are evident. Suchlinkages
createreal economicand ecologic units. This unit with
connectedmosaicsis called a region. Regionsare in
many scales(Bailey 1983). Like landscapes,they stand in
contrast with one anotherand also are connectedthrough
long-distancelinkages. Finally this progressionreaches
the scale of the planet.
2
Climatic regime, in turn, is channeled,shaped,and transformed by the structural characteristicsof ecosystems,
that is, by the nature of the earth's surface. In this sense,
all ecosystems,macro and micro, are respondingto climatic influencesat different scales.The primary controls
over the climatic effects changewith the scale of observation. Latitude, continentality,and mountainsall differentiate regional climate, while landforms and local vegetation
on them differentiatelocal climate. Setting ecosystem
boundariesinvolves the understandingof thesefactors on
a scale-relatedbasis.
Differentiating Factors And Scale
The factors that are thoughtto differentiate eco-climatic
units, and the scale at which they operate,are described
as follows:
Macroscale: Macroclimatic Differentiation
To make comparisonsof climates on a macroscaleor
global level, it is necessaryto considerclimatic conditions
that prevail over large areas. Unfortunately, climate
changeswithin short distancesdue to variations in local
landform featuresand the vegetationthat developson
them. It is necessary,therefore, to postulatea climate that
lies just beyond the local modifying irregularities of landform and vegetation.To this climate, the term' 'macroclimate" is applied. Variations in macroclimate(as
determinedby the observationsof meteorologicalstations)
are related to severalfactors.
Latitude- The primary control of climate at the global
level is latitude, resulting in irregular solar energy at
different latitudes. Solar radiation experiencesa generally
latitudinal decreasefrom equatorto pole due to increases
in the angle of incidence of the sun's rays and to the
thicknessof the atmosphere.The resulting generallyeastwest belts or zones correspondto life zones, plant formations, and biomes commonlyrecognizedby ecologistsand
biogeographers(Whittaker 1975). The boundariesof
zonesare determinedby thermal and moisture limits for
plant growth.
Three major thermally defined zonescan be defined: (1) a
winterless climatic zone of low latitude, (2) temperateclimates of mid latitudes with both a summerand winter,
and (3) a summerlessclimate of high latitude. A winterless climate is commonly defined as one in which no
month of the year has a mean monthly temperaturelower
than 64 of (18°C). This isothermapproximatesthe position of the boundary of the poleward limit of plants
characteristicof the humid tropics. A summerlessclimate
is one in which no month has a mean monthly temperature higher than 50°F (10°C). The 50°F isotherm closely
coincides with the northernmostlimit of tree growth;
hence, it separatesthe regions of boreal forest from the
treelesstundra.
The relative amplitudesof the periodicities of annualand
diurnal energycycles vary in eachzone. Within the
tropics, the diurnal range is of greatermagnitudethan the
annual cycle. Within temperatezones,the annualrange
exceedsthe diurnal range, althoughthe diurnal can be
very large. Within the polar zones,the annual range is far
greaterthan the diurnal range.
Precipitationalso follows a zonal pattern. The dry zones
are controlled by the subtropicalhigh-pressurecells centered on the tropics of Cancerand Capricorn(231h0 N
and g). Thesezonesare too dry for tree growth.
Continental position-At any given latitude the summers
are hotter and the winters colder over the land than over
the oceans,giving rise to the distinction betweenmarine
and continentalclimates.
The distribution of land and seaalso complicatesprecipitation patterns. Evaporationis rapid over warm water,
and therefore precipitationis generally greater over the
margins of the continentsbathed by warm water.
By combiningthe thermally defined zones with the moisture zones, it is possibleto delineatefour eco-climatic
zones:hunrid tropics, humid temperate,polar, and dry.
Within eachof these zones,one or severalclimatic gradients may affect the potential distribution of the dominant
vegetation.Within the Humid Tropical zone, for example,
rainforeststhat have year-round precipitationcan be distinguished from savannasthat receive seasonalprecipitation.
The analysis of eachzone results in the identification of a
number of climatic subzones.Theseclimatic subzonesare
correlated with actual climatic types, using the systemof
climatic classificationdevelopedby Koppen (1931) and
modified by Trewartha (1968). Koppen's systemis simple, is basedon quantitativecriteria, and correlateswell
with the distribution of many natural phenomena,suchas
vegetationand soil.
Other bioclimatic methodsfor mappingzones at global
levels exist (for example,Holdridge 1947, Troll 1964,
Walter and others.1975). All use selectedclimatic characteristics that outline zones within which certain general
level vegetationhomogeneityshould be found. They also
suggesta strong similarity of vegetationin equivalent
bioclimatic zones in different parts of the globe. All of
the methodsappearto work better in someareasthan in
others and have gained their own following. Koppen's
systemhas becomethe most widely used climatic classification for geographicalpurposesand is, therefore,
presentedhere to illustrate the basis for zone delineation.
3
~
1~
0'1
.&
By applying Koppen's system,the following thirteenbasic
climates result:
Do
Ar
E
Temperateoceanic: 8 months over 50 of, warmest
month below 72 of (22°C).
Temperatecontinental: 4 to 8 months over 50of.
Boreal or subarctic: one warmestmonth 50 of or
above.
Tundra: all months below 50of.
Polar ice cap: all months below 32of.
Dc
Tropical wet: all months above64 of (18 °C) and no
dry season.
Aw Tropical wet-dry: all months above 64 of and 2
months dry in the winter.
BSh Tropical/subtropicalsemi-arid: evaporationexceeds
precipitationand all months over 32 of (0 °C).
BWh Tropical/subtropicalarid: one-half the precipitation
of the semi-arid and all months over 32of.
BSk Temperatesemi-arid: sameas BSh but with at least
1 month colder than 32°F.
BWk Temperatearid: sameas BWh but with at least 1
month colder than 32 OF.
Cs Subtropical dry summer(Mediterranean):8 months
50°F (10°C) or more, summerdry.
Cf Subtropicalhumid: 8 months over 50of.
Ft
Fi
The distribution of theseclimatesis shown in figure 1.
Eachclimatic subzoneis clearly defined by a particular
type of climatic regime, and, with a few exceptions,the
subzoneslargely correspondto zonal soil types and zonal
vegetation.Table 1 showsthe relations betweenzonal
types and climatesas classified by Koppen.
Zonal soil types and vegetationoccur on sites supporting
climatic climax vegetation. Suchsites are uplands, that is,
sites with a well-drained surface, moderatesurfaceslope,
~
~-~~L.
e'
~(
'\
'
~ ~l
~\\
J~,
~
'\;II
r
'OJ
'.
;tJ
,
I
Ar Tropical
Aw Tropical
Wet
Wet-Dry
BSh Tropical/Subtropical
'31
Semi-arid
BWh Tropical/Subtropical
BSk Temperate Semi-arid
~
BWk Temperate Arid
Cs Subtropical Dry Summer
Figure l-Eco-climatic zones of the world (after Trewartha1968).
4
Arid
U" ~
I//~r I
BS
Boreal
Table I-Zonal relationshipsbetweenclimate, soil, and vegetation
Eco-climatic
Zonal soil type!
Zonal vegetation
zone
AT
Latisols (Oxisols)
Evergreentropical rain forest (selva)
Aw
Latisois (OxisoIs)
Tropical deciduousforest or savannas
Chestnut,Brown soils and Sierozems(Mollisols
Aridisols)
Short grass
Desert (Aridisols)
Shrubsor sparsegrasses
Cs
Mediterraneanbrown earths
Sclerophyllouswoodlands
Cf
Red and Yellow Podzolics (illtisols)
Coniferousand mixed coniferous-deciduousforest
Brown Forestand Gray-brownPodzolic (Alfisols)
Coniferousforest
Dc
Gray-brown Podzolic (Alfisols)
Deciduousand mixed coniferous-deciduousforest
E
Podzolic (Spodosolsand associatedHistosols)
Ft
Tundra humus soils with solifluction (Entisols,
Inceptisolsand associatedHistosols)
BW
Do
coniferous forest (taiga)
Tundra vegetation(treeless)
Fi
1Names in parenthesesare Soil Taxonomy soil orders (USDA Soil ConservationService 1975).
and well-developedsoils. The climax vegetationcorrespondsto the major plant formation (for example,
deciduousforest) that is the presumedresult of succession, given enoughtime.
It is possibleto subdividezonesinto finer ecological
units. For example,the vegetationcover of the savanna
zone is highly differentiated. It has heavy forest near its
boundary with the equatorialzone and sparseshrubsand
grassesnear its arid border. Variation in the length and
intensity of the rainy seasonrelate to both the variety of
vegetationand to soil and hydrologic conditions.
Elevation-The zonal correspondenceof climate with latitude and continentalposition is broken by featuresthat
are dependenton differencesin elevation, or relief.
Becausethey can occur in any zone they are referred to
as azonal.
Becauseof elevation, high mountainsare differentiated
into vertical zones. Every mountainwithin a zone has a
typical sequenceof altitudinal belts that differs according
to the zone in which it is located (seefig. 2). Two series
of eco-climaticunits can, therefore, be established:
lowlands and highlands.
The effects of latitude, continentalposition, and elevation,
togetherwith other climatic factors, combineto form the
eco-climaticzonesof the world. Figure I showsthe climatic zoneswithin which distinct ecosystemassemblages
might be expectedto occur. This map showsclimatic
units that appearto be important to the climatologistand
can be usedto help determineecosystemboundariesat the
macroscale.
Sincemeteorologicalstationsare too sparsein many
areas,dataare simply not availableto map more precisely
5
1
D
2
~
3
~
4
~
5
m
6
~
7
[:3
8
~
Figure 2- Vertical zonation in different eco-c1imatic zonesalong the easternslopes ofthe RockyMountains (after
Schmithiisen1976). 1, ice region; 2, mountain vegetationabovetree line; 3, boreal and subpolar open
coniferouswoodland; 4, boreal evergreenconiferousforest; 5, boreal evergreenmountainconiferous
forest; 6, coniferousdry forest; 7, short grass dry steppe;8, boreal evergreenconiferousforest with
cold-deciduousbroodleavedtrees. A, ecotone;B, Boreal; C, ecotone;D, TemperateSemi-arid.
the distribution of theseecologicalclimates. Thus, we
generallysubstituteother distributions. The composition
and distribution of vegetationwas used by Koppen in
his searchfor significant climatic boundaries,and vegetation is a major criterion in the ecosystemregion maps
of Bailey (1983) and Walter and Box (1976).
Climatic differencesuseful in recognizingunits at this
level can be reflected in the vegetationin severalways
(Damman 1979): (1) changesin forest stand structure,
dominantlife forms, and topographyof organic deposits;
(2) changesin dominantspeciesand in the toposequence
of plant communities;and (3) displacementof plant communities, changesin the chronosequenceof a habitat, and
minor changesin the speciescompositionof comparable
plant communities. Other differencesare given by
Kuchler (1974)and van der Maarel (1976).
Traditionally, the principal source of such information has
been vegetationmappingby-ground survey. If large areas
are to be surveyedthis approachis not very practical, and
satellite remote-sensingdata with its synoptic overview is
used to look for zoneswhere vegetationcover is relatively uniform. Thesezonesare especiallyapparentin
low-resolutionremote sensingimagery (Tucker and others
1985).
In someareas,pr.oblemsresulting from disturbanceand
the occurrenceof an-intricate patternof secondarysucces6
sional stagesmake regionalboundary placementdifficult.
In suchareas,theseproblems can be overcomeby consideringthe patternsdisplayedon soil maps of broad
regions, suchas ttle FAO/UNESCO World Soil Map
(FAO/UNESCO 1971-78). Becausesoils tend to be more
stable than the vegetation,they provide supplementalbasis
for recognizingecosystemsregardlessof presentland use
or existing vegetation.
Mesoscale:Landform Differentiation
Macroclimateaccountsfor the largestshare of systematic
environmentalvariation at the macroscaleor regional
level. At the mesoscale,the broad patternsare broken up
by geologyand topography(landform). For example,
solar energy will be receivedand processoodifferently by
a field of sanddunes,lacrustrineplain, or an upland
hummockymoraine.
Landforms (with their geologic substrate,surface shape,
and relief) influence place-to-placevariation in ecological
factors suchas water availability and exposureto radiant
solar energy. Through varying height and degreeof inclination of the ground surface,landforms interact with climate and directly influence hydrologic and soil-forming
processes.
In short, the best correlation of vegetationand soil patterns at mesoand inicroscalesis landform becauseit controls the intensitiesof key factors importantto plants and
to the soils that developwith them (Hack and Goodlet
1960; Swansonand others 1988). Realizationof the
importanceof landform is apparentin a number of
approachesto classificationof forest land (for example,
Barnesand others 1982).
three major characteristics:relative amountof gently sloping land (less than 8 percent), local relief, and generalized
profIle (that is, where and how much of the gently sloping land is located in the valley bottoms or in the
Landforms come in all scalesand in a great array of
shapes.On a continentalscale within the samemacroclimate there commonly exist severalbroad-scalelandform
patterns that break up the zonal patterns(figure 3). The
landform classificationof Hammond (1954, 1964), who
classifiedland-surfaceforms in terms of existing surface
geometry, is useful in determiningthe limits of various
mesoecosystems
or landscapemosaics.In the Hammond
system, summarizedin table 2, landforms are identified
on the basis of similarity and differenceswith respectto
On the basis of thesecharacteristicsalone, it is possible
to distinguishamong (1) plains having a predominanceof
gently sloping land, coupled with low relief, (2) plains
with somefeature of considerablerelief, (3) hills with
gently sloping land and low to moderaterelief, and
(4) mountainsthat have little gently slopingland and high
local relief.
uplands).
The secondgroup may be subdividedon the basis of
where the gently sloping land occurs in the profile into
Plain
Figure 3-Land-surface form typesand their effect on zonal climate.
,.,
Table 2-Hammond's schemeof landform classification
Symbol Definition
Slope
A.
B .
C.
D.
~,
More than 80 percentof area gently sloping.
50-80 percentof area gently sloping.
20-50 percentof area gently sloping.
Less than 20 percentof area gently sloping.
Local Relief
2.. .
3.. .
4. ..
5.. .
6.. .
0-100 feet.
100-300 feet.
300-500 feet.
500-1000 feet.
1000-3000feet.
Over 3000 feet.
Profile type
a..
bo
Coo
do.
0
More than 75 percentof gentle slope is in lowland.
50-75 percentof gentle slopeis in lowland.
50-75 percentof gentle slopeis on upland.
More than 75 percentof gentle slopeis on upland.
plains with hills, mountains,or tablelands.Approximate
definitions of the grouping or generalizedterrain types are
as follows:
.Nearly flat plains: AI; any profile
.Rolling and irregular plains: A2 , Bl, B2; any profile.
.Plains with widely-spacedhills or mountains:A3a or
b, B3a or b to B6a or b.
.Partially dissectedtablelands:B3c or ~ to B6c or d.
.Hills:
D3, 04; any profile.
.Low mountains:D5; any profile.
.High mountains:D6; any profile.
Of course, within theseclassesthere exists much variety.
Someplains, for instance,are flat and swampy,others
rolling and well drained, and still others are simply broad
expansesof smoothice. Similarly, somemountainsare
low, smoothed-sloped,and arrangedin parallel ridges,
while others are exceedinglyhigh, with rugged, rocky
slopesand glaciers and snowfields.
Figure 4 shows how someof theseclasses(landscape
mosaics)are distributed in Koppen's Mediterranean(Cs),
or SubtropicalDry Summer,zone.
According to its physiographicnature, a landform unit
consistsof a certain set of sites. A delta has differing
types of ecosystemsfrom those of a moraine landscape
next to it. Within a landscapemosaic, the sites are
arrangedin a specific pattern. The tablelandsof the westcentral part of the North American continentare a case in
point (seefig. 5). For example,the Colorado Plateauis
made up of various site-specificecosystems,including
valleys of various sizes, smoothuplands,streamchannels
(mostly dry), individual slopes,terraces,sandbarsin the
streamchannels,and several small and shallowdepressions in the uplands.
Units at this level can be most accuratelydelineatedby
considering the toposequence(Major 1951), or catenaof
site types, throughoutthe unit.
To account for some of this variability, two additional
classes are identified in the plains areas. The two added
classes are the following:
.Ice
cap: More than 50 percent of the area is covered
by permanent ice.
Microscale:Edaphic-topoclimatic
Differentiation
.Poorly
drained lands: More than 10 percent of the
area is covered by lake or swamp.
latter is the edaphic (related to soil) factor.
8
Although the distribution of ecological zones is controlled
by macroclimate and broad-scale landform patterns, local
differences are controlled chiefly by microclimate and
ground conditions, especially moisture availability.
The
30
40
50
Figure 4-Landscapes of the SubtropicalDry Summerzone (from Thrower and Bradbury 1973; redrawn with permissionfrom Springer-Verlag,Heidelberg).
Within a landform there exist slight differencesin slope
and aspectthat modify the macroclimateto topoclimate
(Thomthwaite 1953). There are three classesof topoclimate: normal, hotter than normal, and colder than normal
(figure 6). The units derived from theseclassesare
referred to as site classes(Hills, 1952).
Deviations from normal topoclimateand mesic soil moisture occur in various combinationswithin a region, and
are referred to as site types (Hills 1952). As a result,
every regional system-regardlessof size or rank-is
characterizedby the associationof three types of local
ecosystemsor site types:
In differentiating local sites within topoclimates,soil
moisture regimeshave beenfound to be the feature that
provide the most significant segregationof the plant communities. A toposequenceof a drainagecatenaillustrates
this phenomenon(figure 7). A commondivision of the
soil moisture gradientis: very dry, dry, fresh, moist and
Zonal site types- Thesesites are characterizedby normal
topoclimateand fresh and moist soil moisture.
wet.
Azonal site types-These sites are zonal in a neighboring
zone but are confined to an extra-zonalenvironmentin a
given zone. For instance,in the northern hemisphere,
south-facingslopesreceive more solar radiation than
9
10
Figure S-A variety of ecosystems
form this mosaic of riparian, grazing, and woodlandsitesin UnaweepCanyon
and vicinity in western Colorado, a well-developedtableland in the TemperateSemi-aridzone. (USDA
Forest $ervice photo.)
north,-facingslopes,and thus south-facingslopestend to
be warmer, drier, less thickly vegetated,and covered by
thinner soils than north-facing slopes. In arid mountains,
the south-facingslopes.arecommonlycovered by grass,
while steepernorth-facing slopesare forested.Azonal
sites are hotter, colder, wetter, and drier than zonal sites.
Intrazonal site types-These sites occur in exceptional
situationswithin a zone. They are presentedby small
areas with extremetypes of soil and intrazonalvegetation.
Vegetationis influencedto a greaterextent by soil than
by climate, and thus the samevegetationforms may occur
on similar soil in a nUmberof zones. They are differentiated into four groups:
First, there are those are that are unbalanced chemically.
Someexamplesfrom the United Statesare the specialized
plant standson serpentine(magnesiumrich) soils in the
California CoastRanges.Other examplesare the belts of
grasslandon the lime-rich black belts of Alabama,Mississippi, and Texas and the low mat saltbush(Atriplex corrugata) on shaledesertsof the Utah desert, which
contrastswith upright shrubson adjacentsandyground.
The kind and amountof dissolved matter in groundwater
also affect plant distribution. It is especiallyobvious along
coastsand along edgesof desertbasinswhere the water is
brackish or saline.
11
N
---$-
Second,very wet sites are where intrazonal plant distributions are controlled by ground-watertable. The plants
of thesesites are phreatophytesthat sendroots to the
water table. Examplesare provided by riparian zones in
the desertsof the southwesternUnited States,suchas cottonwood (Populusdeltoides)floodplain forest, and the
cypress (Taxodiumdistichum)and tupelo (Nyssaaquatica)
forests of the Southeast.
Third, very dry sites with sandysoils, becauseof limited
moisture-holdingcapacity,are drier thnn the generalclimate. At the extreme, sanddunes fail to supportany
The use of the word' 'potential" is critical becauseit
allows a single site to include different kinds of vegetation
as long as they representdifferent stagesof biotic succession from weedypioneersto "climax" forest or grasslands. It is possibleto identify anotherlevel (provisionally
called the site phase)to allow the classificationto communicatethe agesand speciescompositionof existing
vegetation.Thesecorrespondto forest and range cover
types that are commonlymappedby the use of remote
sensingimagery.
vegetation.
Fourth, there are very shallow sites. Soil depth, as a factor in plant distribution, may be controlled by depthto a
water table or depthto bedrock. Vegetationgrowing
along a stream or pond differs from that growing some
distanceaway where the depthto the water table is
greater. Examplesof the influence of depthto bedrock on
plant distribution can be seenin mountainousareaswhere
bare rock surfacesthat support only lichens are surrounded by distinctive flowering plants growing where
thin soil overlapsthe rock, and is, in turn, surroundedby
forest where the soil deepens.
Figure 8, in a very simplified way, illustrates how topography, even in areasof uniform macroclimate,leadsto
differences in local climates and soil conditions. The climatic climax theoretically would occur over the entire
region but for topographyleadingto different local climates, which partially determineedaphicor soil conditions. On theseareas different edaphicclimaxesoccur;
climatic climaxes occur only on mesic soils.
The units at this scalecorrespondto units with similar
soil particle size, mineralogicalclasses,soil moisture, and
temperatureregimes. Theseare generallythe same
differentiating criteria used to define families of soils in
the Systemof Soil Taxonomy of the National Cooperative
Soil Survey (USDA Soil ConservationService 1975).
The potential, or climax, vegetationof theseunits is the
plant community with the rank of association,which is
the basic unit of phytocenology.Associations(also called
habitattypes in the WesternUnited Statesby Pfister and
Arno (1980) are named after the dominantspeciesof the
overstory and of the understory(Daubenmire1968).
12
It
..-:::
I
I11I1
::::
HabltatMicroclimate and sol/
-
---Normal
IiffiIiiiiIffij
===
microclimate
over moist soil
Normal microclimate
over wet soil
---Normal
microclimate
ffifjj
over dry soil
It t t!
-
Warmer microclimate
Climax biotic
community
Maple-beech
}
CLIMATIC
CLIMAX
Oak-ash
Oak-hickory
Tulip-walnut
DnnIDmID over moist soil
Ittlt
t t t II
C:::::J
III
II
Warmer microclimate
over wet soil
Sycamore-tulip
Warmer microclimate
over dry soil
Oak-chestnut
Colder microclimate
Elm-ash-oak
EDAPHIC
CLIMAXES
IJnIDII!DID over moist soil
IIIII
Colder microclimate
over wet soil
White sprucebalsam fir
III II
",
Colder microclimate
over dry soil
Hemlockyellow birch
PracticalApplications
Ecogeographicanalysis may be carried out at different
levels (at different scales).The particular rank of the ecosystemin the hierarchy of ecologicalunits dependson the
purposeand scaleof analysis.We can thus define relationships betweenthe levels of analysisand the rank of
the ecological unit subjectto investigation:
Regions delimit large areaswithin which local ecosystems
recur more or less throughoutthe region in a predictable
fashion. By observingthe behavior of the different kinds
of systemswithin a region, it is possibleto predict the
behavior of an unvisited one.
Regionshave two important functions for management.
First, a map of suchregions suggestsover what area the
knowledge about ecosystembehaviorderived from experiments and experiencecan be applied without too much
adjustment,for example, silvicultural practices,estimates
of forest yields, and seeduse. Experienceconcerningland
use, suchas terrain sensitivityto acid rain, suitability for
agriculture, and effectivenessof best managementpractices in protecting fisheries, can be applied to similar sites
within each region. Second,regions provide a geographical framework in which similar responsesmay be
expectedwithin similarly defined systems.It is thus possible to formulate managementpolicy and apply it on a
regionwide basis rather than on a site-by-sitebasis. This
increasesthe utility of site-specificinformation and cuts
down on the cost of environmentalinventoriesand
monitoring.
Landscape mosaics delimit areasthat representdifferent
patterns or combinationsof sites within a regional ecosystem. The interaction betweensites causesprocessesto
emerge that were not presentor evident at the site level.
The processesof a landscapemosaicare more than those
of its separateecosystemsbecausethe mosaicinternalizes
exchangesamong componentparts. For example,a snowforest landscapeincludes dark pines that convert solar
radiationinto sensibleheat that movesto the snow cover
and melts it fasterthan would happenin either a wholly
snow-coveredor a wholly forestedbasin. The pines are
the intermediariesthat speedup the processand affect the
timing of the water runoff. Watershedmanagersattempt
to producethe sameeffects by strip cutting extensive
forests. It is the understandingof landscapeprocessthat
makespossiblethe analysisof the effects of management
of a site on surroundingsite, and therebythe assessment
of the cumulativeeffects that may occur from a proposed
managementactivity.
Furthermore, a map of suchmosaicsrevealsthe relative
diversity of the landscape.Planningand managementof
diverse and complexlandscapes(for example, mountains)
is problematicaland difficult, while uniform ones (for
example, plains) presentproblems of relative simplicity.
Solving problemsrelated to land use suchas erosionand
revegetationdependon understandingthe complexity of
the landscape.By knowing the characterof the mosaic
and the landscapeprocessesit is possibleto analyzeand
mitigate the problemsassociatedwith management
activities.
Sites are, for practical purposes,relatively homogeneous
with respectto all the biophysicalcomponents.Suchareal
site units are the base for productivity assessments,
silvicultural prescription, and forest management.
Site maps are general purpose ecosystem maps. Applied
ecosystem maps can be developed by interpreting and
grouping the basic ecosystem units shown on the general
purpose map. For example, a general purpose map can be
interpreted to show units with high arboreal productivity
and low potential for slope failure. A further interpretation can be made to place units with such a combination
into a category of high suitability for forestry. Different
types of applied ecosystem maps will differ only in the
interpretation and grouping of the basic ecosystem units.
11
Highland
Macro
Discussion
All natural ecosystemsare recognizedby differencesin
climatic regime. The basic idea here is that climate, as a
source of energyand moisture, acts as the primary control for the ecosystem.As this componentchanges,the
other componentschangein response.The primary controls over the climatic effects changewith scale. Regional
ecosystemsare areasof essentiallyhomogeneousmacroclimate. Landform is an important criterion for recognizing smallerdivisions within macroclimaticunits.
Landform modifies climatic regime at all scaleswithin
macroclimaticzones. It is the causeof the modification of
macroclimateto local climate. Thus, landform provides
the bestmeansof identifying local ecosystems.At the
mesooscale,the landform and landform pattern form a
natural ecologicalunit. At the microscale, suchpatterns
can be divided topographicallyinto slopeand aspectunits
that are relatively consistentas to soil moisture regime,
soil temperatureregime, and plant association;that is, the
homogeneous"site."
Therefore, the answerto the questionof boundarycriteria
is that climate, as modified by landform, offers the logical
basis for delineating both large and small ecosystems.
Basedon the foregoing analysis,criteria indicative of climatic changesof different magnitudeare presentedin
table 3. Figure 9 illustrates the use of thesecriteria. They
are offered as suggestionsto guide the mappingof
ecosystemsof different size. In broad outline, the criteria
for delineationare quite different at eachof three scales
of analysis.The results of this review are not meantto be
definitive, but are an attemptto illustrate criteria that
appearto be importantand that can be usedto establish
ecosystemboundaries.
With referenceto the generalprinciples involved in
assigningprime importanceat the different scalelevels to
different criteria, it shouldbe noted that Rowe (1980) has
raised the need for a caveat. Although the levels can be
mappedby referenceto single physical and biological features, they must always be checkedto assurethat the
boundarieshave ecologicalsignificance. A climatic map
showing such key factors as temperatureand precipitation
is not necessarilyan ecologicalmap, until its boundaries
are shownto correspondto significantbiologic boundaries. Likewise, maps of landform, vegetation,and soils
are not necessarilyecologicalmaps until it has been
shownthat the types have the samedistribution. Before
any map is used, it should be thoroughly testedand modified if necessary(Bailey 1984).
It is importantto link the ecosystemwith management
hierarchies. It is not suggestedin the foregoing that three
levels of ecologicalpartitioning are everywheredesirable;
there could be two or nine dependingon the kind of que~tion being askedand the scale of the study. However, if
is advantageous
to have a basic framework consistingof a
relatively few units to which all ecologicalland mappers
can relate and betweenwhich other units can be defined
as required.
Table 3-Mapping criteria for ecosystemunits at different scaleswith examples
Name
Scale
14
of
unit
Examplesof units
Criteria
Lowland series
series
Regionor
zone
BcD-climaticzone
(Koppen 1931)
Temperatesemi-arid (BSk)
Temperatesemi-arid
regime highlands (H)
Landscape
mosaic
Land-surfaceform class
(Hammond 1954)
Nearly flat plains (AI)
High mountains(D6)
Site
Microclimate and soil
Normal microclimate over
moist soil
Normal microclimate over
moist soil
2 mi
Figure 9-Ecosystem maps of different scales.
15
Literature Cited
Bailey, R.G. 1983. Delineation of ecosystemregions. Environmental
Management.7: 365-373.
Bailey, R.G. 1984. Testing an ecosystemregionalization.Journal of
EnvironmentalManagement.19: 239-248.
Bailey, R.G. 1985. The factor of scale in ecosystemmapping. EnvironmentalManagement.9: 271-276.
Bailey, R.G. 1987. Suggestedhierarchy of criteria for multi-scale ecosystemmapping. Landscapeand Urban Planning. 14: 313-319.
Barnes,B.V.; Pregitzer,K.S.; Spies,T.A.; SpoonerV.H. 1982. Ecological forest site classification. Journal of Forestry. 80: 493-498.
Damman,A. W.H. 1979. The role of vegetationanalysisin land classification. ForestryChronicle. 55: 175-182.
Daubenrnire,R. 1968. Plant communities:a text book on plant synecology. New York: Harper and Row. 300 p.
Driscoll,R.S,; Merkel, D.L.; Radloff, D.L.; Snyder,D.E.; Hagihara,
J.S. 1984. An ecological land classificationframework for the United
States.Misc. Publ. 1439. Washington,DC.: U. S. Departmentof
Agriculture. 56 p.
FAO/UNESCO. 1971-1978. FAO/UNESCO soil map of the world 1:5
million. North America, SouthAmerica, Mexico and Central
America, Europe, Africa, South Asia, North and Central Asia, Australia. Paris: UNESCO.
Forman R.T.T.; Godron, M. 1986. Landscapeecology. New York:
JohnWiley. 619 p.
Hack, J.T.; Goodlet, J.C. 1960. Geomorphologyand forest ecology of a
mountain region in the central Appalachians.Prof. Pap. 347.
Washington,DC: U.S. GeologicalSurvey. 66 p.
Hammond,E.H. 1954. Small-scalecontinentallandform maps. Annals
AssociationAmerican Geographers.44: 33-42.
Hammond,E.H. 1964. Classesof land-surfaceform in the forty eight
States.U.S.A. Annals AssociationAmerican Geographers54. Map
supplementno. 4, scale 1:5 million.
Hills, A. 1952. The classificationand evaluationof site for forestry.
Res. Rep. 24. Toronto: Ontario Departmentof Lands and Forest.
41 p.
Holdridge, L.R. 1947. Determinationof world plant formations from
simple climatic data. Science.105: 367-368.
Isachenko,A.G. 1973. Principles of landscapescienceand physicalgeographicregionalization(trans. from Russianby J.S. Massey).
Carlton, Victoria, Australia: Melbourne University Press.311 p.
Koppen,W. 1931. Grondriss der klimakunde. Berlin: Walter de
Gruyter, 388 p.
Kuchler, A.W. 1974. Boundarieson vegetationmaps. In: Tuxen, R.,
editor. Tatsachenund probleme der grenzenin der vegetation. Lehre,
FR.of Germany: Verlag von J. Cramer. p. 415-427.
Leser, H. 1976. Landscaftsokologie.Stuttgart: Ulmer. 432 p.
Major, J. 1951. A functional, factorial approachto plant ecology. Ecology. 32: 392-412.
16
Miller, D.H. 1978. The factor of scale: ecosystem,landscapemosaic,
and region. In: K.A. Hammond,editor. Sourcebookon the environment. Chicago: University of ChicagoPress:p. 63-88.
Pfister, R.D.; Arno, S.F. 1980. Classifying forest habitat types basedon
potential climax vegetation.ForestScience.26: 52-70.
Rowe, J.S. 1980. The commondenominatorin land classificationin
Canada:an ecological approachto mapping. Forestry Chronicle.
56: 19-20.
Schmithiisen,J. 1976. Atlas zur biogeographie.Mannheim-Wien-Zurich:
BibliographischesInstitut. 88 p.
Swanson,F.J.; Kratz, T.K.; Caine, N.; Woodmansee,R.G. 1988.
Landform effects on ecosystempatternsand processes.Bioscience38:
92-98.
Thornthwaite,C. W. 1953. Topoclimatology.In: Proceedingsof the
Toronto Meteorological Conference,September9-15, 1953. Toronto:
Royal MeterologicalSociety. p. 227-232.
Thrower, N. J.W.; Bradbury, D.E. 1973. The physiographyof the
Mediterraneanlandscapewith special emphasison California and
Chile. In: F. di Castri and H.A. Mooney, editors. Mediterranean
Type Ecosystems:Origin and Structure. Ecological Studies: Analysis
and Synthesis,vol. 7. Heidelberg: Springer-Verlag.p. 37-52.
Trewartha,G.T. 1968. An introduction to climate, 4th ed. New York:
McGraw-Hill. 408 p.
Troll, C. 1964. l<arte der jahrzeiten-klimateder erde. Erdkunde.
17: 5-28.
Tucker, C.J,; Townshend,J.R.G.; Goff, T.E. 1985. African land-cover
classificationusing satellitedata. Science.227: 369-375.
USDA ForestService. 1982. Ecosystemclassification,interpretation,
andapplication. ForestService Manual 2061, 2062, 2063. Washington, D.C.
USDA Soil ConservationService. 1975. Soil taxonomy: a basic system
for making and interpreting soil surveys.Agric. Handb. 436.
WashingtonDC: U.S. Departmentof Agriculture. 754 p.
van der Maarel, E. 1976. On the establishmentof plant community
boundaries.Bericht der DeutschenbotanischenGesellschaft.89:
415-433.
Walter, H.; Harnickell, E.; Mueller-Dombois,D. 1975. Climatediagram mapsof the individual continentsand the ecological climatic
regions of the earth (mapsand manual). Heidelberg: Springer-Verlag.
36p.
Walter, H.; Box, E. 1976. Global classificationof natural terrestrial
ecosystems.Vegetatio. 32: 75-81.
Whittaker, R.H. 1975. Communitiesandecosystems.2d ed. New York:
MacMillan. 387 p.
Zonneveld,I.S. 1972. Land evaluationand land(scape)science. Use of
aerial photographsin geographyand geomorphology.ITC Textbook of
PhotointerpretationVol. Vll. Enschede,The Netherlands:International
Training Centre for Aerial Survey. 106p.