Annu. Rev. Ecol. Syst. 2001. 32:127-57
URBAN ECOLOGICAL
SYSTEMS:Linking Terrestrial
Ecological,Physical,and Socioeconomic
Components of MetropolitanAreas*
S. T. A. Pickett,1M. L. Cadenasso,1J.M. Grove,2
C. H. Nilon,3 R. V. Pouyat,4W. C. Zipperer,4
and R. Costanza5
'Instituteof EcosystemStudies,Millbrook,New York;e-mail:picketts@ecostudies.org,
cadenassom@ecostudies.org
2USDAForestService,NortheasternResearchStation,Burlington,Vermont;
e-mail:jmgrove@att.net
3Fisheriesand Wildlife,Universityof Missouri, Columbia,Missouri;
e-mail: nilonc@missouri.edu
4USDAForestService,NortheasternResearchStation,Syracuse,New York;
e-mail: ipouyat@aol.com,wzipperer@fsfed.us
5Centerfor Environmentaland Estuar-ineStudies, Universityof Maryland,Solomans,
Maiyland; e-mail: costza@cbl.cees.edu
Key Words
city, hierarchy theory, integration, patch dynamics, urban ecology
* Abstract
Ecological studies of terrestrial urban systems have been approached
along several kinds of contrasts: ecology in as opposed to ecology of cities; biogeochemical compared to organismal perspectives, land use planning versus biological,
and disciplinary versus interdisciplinary. In order to point out how urban ecological
studies are poised for significant integration, we review key aspects of these disparate
literatures. We emphasize an open definition of urban systems that accounts for the
exchanges of material and influence between cities and surrounding landscapes. Research on ecology in urban systems highlights the nature of the physical environment,
including urban climate, hydrology, and soils. Biotic research has studied flora, fauna,
and vegetation, including trophic effects of wildlife and pets. Unexpected interactions
among soil chemistry, leaf litter quality, and exotic invertebrates exemplify the novel
kinds of interactions that can occur in urban systems. Vegetation and faunal responses
suggest that the configuration of spatial heterogeneity is especially important in urban
systems. This insight parallels the concern in the literature on the ecological dimensions
of land use planning. The contrasting approach of ecology of cities has used a strategy
of biogeochemical budgets, ecological footprints, and summaries of citywide species
richness. Contemporary ecosystem approaches have begun to integrate organismal,
*TheU.S. Governmenthas the rightto retaina nonexclusive,royalty-freelicense in and to
any copyrightcoveringthis paper.
127
128
PICKETTET AL.
the social
nutrient,andenergeticapproaches,andto showthe need for understanding
dimensionsof urbanecology.Social structureandthe social allocationof naturaland
institutionalresourcesaresubjectsthatarewell understoodwithinsocialsciences,and
thatcan be readilyaccommodatedin ecosystemmodelsof metropolitanareas.Likeof spatialdimensionsof social differentiation
wise, the sophisticatedunderstanding
has parallelswithconceptsanddataon patchdynamicsin ecology andsets the stage
of urbanecosystems.The linkagesare capturedin
for comprehensiveunderstanding
the humanecosystemframework.
FORURBAN
INTRODUCTION:JUSTIFICATION
STUDIES
ECOLOGICAL
Urbanizationis a dominantdemographictrend and an importantcomponent of
global land transformation.Slightly less than half of the world's populationnow
resides in cities, but this is projectedto rise to nearly 60% in the next 30 years
(United Nations 1993). The developednationshave more urbanizedpopulations;
for example, close to 80% of the US populationis urban.Urbanizationhas also
resultedin a dramaticrise in the size of cities: over 300 cities have more than 106
inhabitantsand 14 megacities exceed 107. The increasingpopulationand spatial
prominenceof urbanareasis reasonenoughto studythem,butecologists mustalso
inform decision makers involved in regional planning and conservation.Proper
managementof cities will ensure that they are reasonableplaces to live in the
future.
In additionto its global reach, urbanizationhas importanteffects in regional
landscapes.For example, in industrializednations, the conversion of land from
wild and agriculturaluses to urbanand suburbanoccupancyis growing at a faster
rate than the population in urban areas. Cities are no longer compact, isodiametric aggregations;rather,they sprawl in fractal or spider-likeconfigurations
(Makse et al. 1995). Consequently,urban areas increasingly abut and interdigitate with wild lands. Indeed, even for many rapidly growing metropolitanareas, the suburbanzones are growing faster than other zones (Katz & Bradley
1999). The resulting new forms of urban development, including edge cities
(Garreau1991) and housing interspersedin forest, shrubland,and desert habitats, bring people possessing equity generatedin urban systems, expressing urban habits, and drawingupon urbanexperiences,into daily contact with habitats
formerly controlled by agriculturalists,foresters, and conservationists(Bradley
1995).
Urbanhabitatsconstitutean open frontierfor ecological research.Ecologists
have come to recognize that few ecosystems are totally devoid of direct or subtle
humaninfluence(McDonnell& Pickett 1993). Yeturbansystems arerelativelyneglected as anendmemberwith whichto comparetherole of humansin ecosystems.
Notably,many classic geographicstudies of cities, which offer valuableinsights
to ecologists, arebased on outmodedecological theorysuch as deterministicmodels of succession and assumptionsof equilibriumdynamicsof ecosystems. Hence
TERRESTRIAL
URBANECOLOGY
129
classical ecological approachesandthe geographicstudiesthathavereliedon them
have not been as useful as they would be otherwise(Zimmerer1994).
Althoughthe ecology of urbanareashas long elicited the academicattentionof
ecologists, physical and social scientists, and regionalplanners,thereis much opportunityto extend andintegrateknowledge of the metropolisusing an ecological
lens. The purposeof this paperis to review the statusof ecological knowledge of
the terrestrialcomponentsof urbanareasandto presenta frameworkfor continued
ecological researchandintegrationwith social and economic understanding.This
papercomplementsthe review of aquaticcomponentsof urbansystems by Paul &
Meyer (2001).
Definition and Roots of Urban Ecology
Urban ecosystems are those in which people live at high densities, or where the
builtinfrastructurecovers a largeproportionof the land surface.The US Bureauof
the Census definesurbanareasas those in which the humanpopulationreachesor
exceeds densitiesof 186 people perkm2.However,an ecological understandingof
urbansystemsalso mustincludeless denselypopulatedareasbecauseof reciprocal
flows and influences between densely and sparsely settled areas. Comparisons
along gradientsof urbanizationcan capturethe full range of urbaneffects as well
as the existence of thresholds.Therefore,in the broadestsense, urbanecosystems
comprisesuburbanareas,exurbs,sparselysettledvillages connectedby commuting
corridorsor by utilities,andhinterlandsdirectlymanagedor affectedby the energy
and materialfrom the urbancore and suburbanlands.
The boundariesof urban ecosystems are often set by watersheds, airsheds,
commutingradii,or convenience.In otherwords,boundariesof urbanecosystems
are set in the same ways and for the same reasons as are the boundariesin any
otherecosystem study.In the case of urbanecosystems, it is clearthatmanyfluxes
and interactionsextend well beyond the urban boundariesdefined by political,
research,or biophysical reasons. Urban ecology, as an integrativesubdiscipline
of the science of ecology, focuses on urbansystems as broadlyconceived above.
Thereis little to be gainedfrom seeking distinctionsbetween"urban"andabutting
"wild"lands, as a comprehensive,spatially extensive, systems approachis most
valuablefor science (Pickettet al. 1997) and management(Rowntree1995).
There are two distinct meanings of urban ecology in the literature(Sukopp
1998). One is a scientific definition,and the otheremerges from urbanplanning.
In ecology, the term urbanecology refers to studies of the distributionand abundance of organismsin and aroundcities, and on the biogeochemical budgets of
urban areas. In planning, urbanecology has focused on designing the environmental amenities of cities for people, and on reducingenvironmentalimpacts of
urbanregions (Deelstra 1998). The planningperspectiveis normativeand claims
ecological justificationfor specific planningapproachesand goals. We review key
aspects of these complementaryapproachesand then frame a social-ecological
approachto integratethese two approaches.
130
PICKETTET AL.
BIOGEOPHYSICAL
APPROACHES
There aretwo aspectsto the biogeophysicalapproachto urbanecological studies.
One, the pioneeringand most common approach,examines ecological structure
andfunctionof habitatsor organismswithincities. This approachis called ecology
in cities. The second, more recent and still emerging approach,examines entire
cities or metropolitanareasfrom an ecological perspective.The second approach
is labeled ecology of cities (Grimmet al. 2000). Although the differences in the
prepositionsin the phrasesidentifyingthe two contrastingapproachesmay appear
subtle, the understandingachieved by identifying them as poles between which
urbanecological studies sort out, is crucial to understandingthe history of urban
ecology, andthe integrationit is now poised to make.We review literaturethathas
takenthese two contrastingapproachesin turn.
Ecologyin the City
The study of ecology in the city has focused on the physical environment,soils,
plants and vegetation,and animalsand wildlife. These studies are the foundation
for understandingurbanecosystems. The literaturein this area has taken a case
studyapproach,andunifyingthemesarestill to emerge.Wehighlightkey examples
from amongthe many cases.
URBAN PHYSICALENVIRONMENT The urbanheat island constitutesclimate modification directly relatedto urbanland cover and humanenergy use (Oke 1995).
The urbanheat island describes the difference between urbanand ruraltemperatures.Such differences often are negligible in the daytime but develop rapidly
aftersunset,peaking 2-3 h later.Ambient air temperaturesmay reach maximaof
5-10?C warmerthan hinterlands(Zippereret al. 1997). For example, New York
City is, on average,2-3?C warmerthanany otherlocationalong a 130-kmtransect
into surroundingruralareas(McDonnellet al. 1993). The durationandmagnitude
of the temperaturedifferentialdepend on the spatial heterogeneityof the urban
landscape.As the percentageof artificialor human-madesurfacesincreases, the
temperaturedifferentialincreases. Hence, the urbancore is warmerthan neighboringresidentialareas,which are warmerthanneighboringfarmlandsor forests.
The differences also change seasonally. For example, cities in mid-lattitudesof
the United States aretypically 1-2?C warmerthanthe surroundingsin winter,and
0.5-1.0?C warmerin summer(Botkin & Beveridge 1997).
The heat island effect varies by region, as seen in a comparison between
Baltimore,Maryland,andPhoenix, Arizona(Brazelet al. 2000). Duringthe summer, mean maximum temperaturesin Baltimore were greaterthan in the rural
landscape.Phoenix, on the otherhand,became an oasis, with cooler temperatures
than the surroundingdesert. The cooling of Phoenix is due to the watering of
mesic plantings in the city. In contrast,the mean minimum temperatureswere
warmerin both cities thanin the respectiveneighboringrurallandscape,although
the differentialin Phoenix was greater.
TERRESTRIAL
URBANECOLOGY
131
The heat island intensity also is related to city size and population density
(Oke 1973, Brazel et al. 2000). For example, Baltimore'smean minimum temperaturedifferentialincreaseduntil the 1970s when the city experienceda decline
in population.Since 1970, the mean minimaltemperaturedifferentialhas leveled
off. Phoenix also showed an increase in mean minimumtemperaturedifferential
with an increase in population.However,Phoenix is the second fastest growing
metropolitanareain the UnitedStates,so the differentialhas continuedto increase
with populationgrowth.In general,a nonlinearrelationshipexists between mean
minimumtemperaturedifferentialand populationdensity (Brazel et al. 2000).
The differencesin climatebetween city and countrysidehave biological implications. For example, as a resultof climatic modificationin temperatezone cities,
leaf emergence and floweringtimes are earlier,and leaf drop is later than in the
surroundingcountryside(Sukopp 1998). Increasedtemperaturesin and around
cities enhance ozone formation,and increasethe numberof officially recognized
pollutiondays andtracegas emissions (Sukopp 1998). Ozone concentrationstend
to be highest in and aroundurbanareas. As urbanareas have expandedthrough
processes of suburbansprawl,the spatialinfluenceof urbanizationhas increased.
Within regions of ozone pollution, agriculturalcrops may be adversely affected
and yields decrease 5-10% (Chameideset al. 1994). Crop type and stage of development,andthe degree, spatialextent, and durationof ozone exposuremay all
influencethe decreasein production(Chameideset al. 1994).
Precipitationis enhanced in and downwind of cities as a result of the higher
concentrationsof particulatecondensationnuclei in urbanatmospheres.Precipitation can be up to 5-10% higherin cities, which can experiencegreatercloudiness
and fog (Botkin & Beveridge 1997). The probabilityof precipitationincreases
towardthe end of the work week and on weekendsdue to a buildupof particulates
resultingfrom manufacturingand transportation(Collins et al. 2000).
Urban hydrology is drastically modified compared to agriculturaland wild
lands.Thistopicis coveredmorefully in a companionreview(Paul& Meyer2001).
Relativizinga waterbudgetto 100 units of precipitation,and comparingurbanto
nonurbanareas,evapotranspiration
decreasesfrom a value of 40%to 25%, surface
runoffincreasesfrom 10%to 30%, and groundwaterdecreasesfrom 50% to 32%
(Hough 1995). Forty-threepercentof precipitationexits the urbanareavia storm
sewers, with 13%of that having firstfallen on buildings.The role of impervious
surfacesis crucialto the functioningof urbanwatersheds(Dow & DeWalle2000).
The hydrologyin urbanareascan be furthermodifiedby ecological structures.For
example,reducedtreecoverin urbanareasincreasestherateof runoffanddecreases
the time lag between initiationof storms and initiationof runoff (Hough 1995).
The increasedrunoff in urban areas changes the morphology of urbanstreams,
which become deeply incised in their floodplains. Remnantriparianvegetation
may suffer as a result of isolation from the watertable.
Soils in urbanlandscapesretain and supply nutrients,serve as a
growthmediumand substratefor soil faunaand flora,and absorband store water.
URBAN SOILS
132
PICKETT
ETAL.
Soils also interceptcontaminantssuch as pesticides and other toxics generated
throughhumanactivities(Pouyat& McDonnell 1991). However,in urbansettings
soils are modified by human activity and, consequently,are functionally altered
(Effland& Pouyat 1997). In addition,completely new substratesare createdby
depositionof debris, soil, and rock in urbansites. Such new substratesare called
made land.
As land is convertedto urbanuse, both direct and indirectfactors can affect
the functioning of soils. Direct effects include physical disturbances,burial of
soil by fill material,coverage by impervioussurfaces,and additionsof chemicals
and water (e.g., fertilization and irrigation).Direct effects often lead to highly
modified substratesin which soil developmentthen proceeds (Effland& Pouyat
1997). Indirecteffects change the abiotic and biotic environment,which in turn
can influence soil developmentand ecological processes in intact soils. Indirect
effects include the urbanheat island (Oke 1995), soil hydrophobicity(White &
McDonnell 1988), introductionsof nonnativeplant and animal species (Airola &
Buchholz 1984, Steinberget al. 1997), and atmosphericdeposition of pollutants
(Lovett et al. 2000). Moreover, toxic, sublethal, or stress effects of the urban
environmenton soil decomposersand primaryproducerscan significantlyaffect
the qualityof organicmatterand subsequentsoil processes (Pouyatet al. 1997).
The results from a transectalong an urban-ruralland-use gradientin the New
YorkCity metropolitanarea show the influence of urbanenvironmentson intact
forest soils (McDonnell et al. 1997). Along this transect,humanpopulationdensity, percentimpervioussurface,and automobiletrafficvolume were significantly
higher at the urbanthan the ruralend of the gradient(Medley et al. 1995). Soil
chemical and physical properties,soil organism abundances,and C and N processes were investigatedalong this gradientto assess the sometimescomplex and
sometimescontradictoryinteractionsbetweenurbansoil chemistry,local leaf litter
quality,and exotic species.
Soil chemistrysignificantlycorrelatedwith measuresof urbanization.Higher
concentrationsof heavy metals (Pb, Cu, Ni), organicmatter,salts, and soil acidity
were found in the surface 10 cm of forest soils at the urbanend of the transect
(Pouyatet al. 1995). The most probablefactor is metropolitan-wideatmospheric
deposition.
Litter decomposition studies in both the field and the laboratorydetermined
thaturban-derivedoak litterdecomposedmore slowly thanrural-derivedoak litter
underconstantconditions(Carreiroet al. 1999), with lignin concentrationexplaining 50% of the variation.Moreover,a site effect was measuredfor reciprocally
transplantedlitter,as decompositionwas fasterin the urbanthanin the ruralsites,
regardlessof litter origin. This result is surprisingbecause the lower litter fungal
biomass and microinvertebrateabundancesfound in urban stands compared to
rural(Pouyat et al. 1994) would lead to an expectationof slower litter turnover
rates in urbanforests. However,abundantearthworms(Steinberget al. 1997) and
highersoil temperaturesin the urbanstandsmay compensatefor lowerlitterquality,
lower fungalbiomass,andlower microinvertebrate
abundancesin the urbanstands
TERRESTRIAL
URBANECOLOGY
133
(R.V. Pouyat & M.M. Carreiro,submittedfor publication).Such compensation
likely explains the fasterdecompositionrates in the urbanstands.
Litterquality,site environment,and soil organismdifferencesbetween the urban and ruralsites also affected C and N pools and processes in the soil. Urban
litter was found to be of lower quality (higher C:N ratio) than rurallitter.Typically, poor qualitylittereither decreasesthe rate at which labile C mineralizesN
or increasesthe amountof organicmattertransferredto recalcitrantpools, or both.
Hence, urbanlitter is expected to be more recalcitrantthan rural litter. Indeed,
measurementsof soil C pools along the urban-ruraltransectsuggest that recalcitrant C pools are higher and passive pools lower in urbanforest soils relative
to ruralforest soils (Groffmanet al. 1995). Consequently,N mineralizationrates
were expected to be lower in urbanstands.However,the opposite was found. Net
potentialN-mineralizationrates in the A-horizonwere higher in the urbanstands
than in the rural stands (Pouyat et al. 1997). Soil cores taken from urbanareas
accumulatedmore NH$ and NOI than soil from ruralareas. In addition,from a
reciprocaltransplantexperiment,soil cores incubatedat urbansites accumulated
more inorganicN than cores incubatedat the ruralsites, regardlessof where the
cores originated.These net N mineralizationratescontradictboth the litterdecomposition results from the transplantexperimentdiscussed above and expectations
derivedfrom measurementsof soil C pools.
In contrast,when net N-mineralizationrates were measuredfor mixed 0, A,
andB horizonmaterial(15 cm depth),the totalinorganicN pool accumulatedwas
higher in ruralthan in urbansites, althoughless NOy was accumulatedin rural
samples (Goldmanet al. 1995). The mechanismsbehindthese results have yet to
be elucidated,though it has been hypothesizedthat methane consumptionrates
and the biomass of methanotrophsare importantregulatorsin overall soil C and
N dynamics(Goldmanet al. 1995).
Intense soil modificationsresulting from urbanizationmay potentially alter
soil C andN dynamics.To assess the potentialeffect on soil organicC, datafrom
"made"soils (1 m depth)fromfive differentcities, andsurface(0-15 cm) soils from
several land use types in Baltimorewere analyzed (R.V. Pouyat, P.M. Groffman,
I. Yesilonis & L. Hemandez,submittedfor publication).Soil pedons from the five
cities showed the highest soil organicC densities in loamy fill (28.5 kg m-2) with
the lowest in clean fill andold dredgematerials(1.4 and6.9 kg m-2, respectively).
Soil organic C for residentialareas (15.5 ? 1.2 kg m-2) was consistent across
cities. A comparisonof land-use types showed that low density residentialand
institutionalland had 44% and 38% higher organicC densities than commercial
land, respectively. Therefore, made soils, with their physical disturbancesand
inputsof variousmaterialsby humans,can greatlyalterthe amountof C storedin
urbansystems.
The complex patternsof C andN dynamicsthathave emergedfrom the studies
reviewed above indicateinteractionsbetween key soil and organismprocesses in
urbanenvironments.Simple predictionsbased on trends in pollution, stress, or
exotic species alone are inadequateto understandthe complex feedbacksbetween
134
ETAL.
PICKETT
these three governing factors of urban soil dynamics. Studies of soil C and N
dynamicsin unmanagedurbanforests, highly disturbedsoils, and surfacesoils of
variousurbanland-usetypes all show thaturbanizationcan directlyand indirectly
affect soil C pools andN-transformationrates.Ourreview also suggeststhatsoil C
storagein urbanecosystems is highly variable.How generalizablethese resultsare
acrosscities locatedin similaranddissimilarlife zones needs to be investigated.In
addition,more dataareneeded on highly disturbedsoils, such as landfill,managed
lawns, and covered soils to make regional and global estimates of soil C storage
andN-transformationratesin urbanecosystems. Specific uncertaintiesincludethe
quality of the C inputs governed by the input of litter from exotic plant species
and by stress effects on native species litter, the fate of soil C in covered soils,
measurementsof soil C densities at depthsgreaterthan I m, particularlyin made
soils, and the effects of specific managementinputson N-transformationrates.
The assessmentof vegetationin urbanlandscapes has a long history.For example, in Europe,studiesby De Rudder& Linke
(1940) documentedthe flora and fauna of cities duringthe early decades of the
twentiethcentury.After WorldWarII, Salisbury(1943) examinedvegetationdynamics of bombsites in cities. At the same time, ecologists in the United States
focused on describingflorain areasof cities minimallyalteredby humans,such as
parksand cemeteries.One of the firstcomprehensivestudies of urbanvegetation
and environmentswas conductedby Schmid (1975) in Chicago.
The structureandcompositionof vegetationhas been one of the foci of ecological studiesin cities, andthese studieshavedocumentedlargeeffects of urbanization
on forest structure.Urban standstend to have lower stem densities, unless those
standsare old-growthremnantsin largeparksor formerestates (Lawrence1995).
In the Chicago region, street trees and residential trees tend to be larger than
those in forestpreserves,naturalareas,andwild lands (McPhersonet al. 1997), althoughindividualtreesin urbansites areoften stressed,especially on ornearstreets
(Ballachet al. 1998). Streettreesin Chicagoaccountfor 24%of the totalLeaf Area
Index (LAI) and 43.7% of the LAI in residentialareas (Nowak 1994). In mesic
forest regions of the United States, tree cover of cities is approximately31%,
comparedto the nearly continuousforest cover the areas would have supported
before settlement.Brooks & Rowntree(1984) quantifiedthe forest cover in counties classified as nonmetropolitan,peripheralto a centralcity, and encompassing
a centralcity. They found thatthe proportionof total land areain forest decreased
fromnonmetropolitanto centralcity counties,withthe steepestreductionsbetween
anddesert
nonmetropolitanandperipheralcounties.In contrast,forprairie-savanna
regions, tree cover in cities is greaterthan in pre-urbanconditions (Nowak et al.
1996).
There are also local effects on urbanvegetation. For example, forest patches
adjacentto residential areas have increased edge openness (Moran 1984), and
their marginshave retreatedbecause of recreationaluse, especially by children,
anddamageto regeneration(Bagnall 1979). Suchregenerationfailureis frequentin
VEGETATIONAND FLORAIN CITIES
TERRESTRIAL
URBANECOLOGY
135
urbanand suburbanstands,owing to reducednaturaldisturbanceor gap formation,
substitutionfor naturaldisturbancesof unfavorableanthropogenicdisturbances
such as frequentgroundfires, trampling,and competitionwith exotics (Guilden
et al. 1990).
The composition of urbanand suburbanforests differs from that of wild and
ruralstandsin severalways. Species richness has increasedin urbanforests as a
whole, butthis is becauseof the increasedpresenceof exotics (Zippereret al. 1997).
Urbanareas show a preponderanceof trees of wetland or floodplainprovenance,
owing to the loweroxygen tensionssharedby wetlandsandimperviousurbansoils
(Spirn 1984). Even when the tree compositionremainssimilarbetweenurbanand
ruralforests, the herbaceousflora of urbanforests is likely to differ between the
two types of forest (Wittig 1998). Such compositionaltrendsreflect the context
and configurationof the forest standsin urbanareas.For example, vascularplant
diversityincreases with the area of the stand (lida & Nakashizuka1995, Hobbs
1988). Furthermore,in some urbanstands, the adjacentland use affects species
composition.Forexample,the interiorof forestsin residentialareasoften has more
exotics thanforests abuttingeitherroads or agriculturalzones (Moran1984).
The role of exotic species has receivedparticularattentionin urbanstudies.The
percentageof the florarepresentedby nativespecies decreasedfromurbanfringes
to centercity (Kowarik1990). Along the New YorkCity urban-to-ruralgradient,
the numberof exotics in the seedling and sapling size classes of woody species
was greaterin urbanand suburbanoak-dominatedstands(Rudnicky& McDonnell
1989). Rapoport(1993) foundthe numberof noncultivatedspecies decreasedfrom
fringe towardurbancenters in several Latin Americancities for various reasons
exemplified in different cities. In Mexico City, there was a linear decrease in
the number of species from 30-80 ha-1 encounteredin suburbsto 3-10 ha-1
encounteredin the city center.The social contextalso influencedspecies richness.
At a given housing density,more affluentneighborhoodshad more exotic species
thanless affluentones in Bariloche,Argentina.In Villa Alicura,Argentina,exotic
species increasedwithincreasinglocal site alterationby humans.Nearhomes there
were 74%exotics, while therewere 48% along riverbanks.Althoughexotics were
presentalong all roads, the numberdecreasedon roads less frequentlyused. No
exotic species were found outside the town. Pathways in ruralrecreationareas
(Rapoport1993) and in urbanparks (Drayton & Primack 1996) have enhanced
the presenceof exotics. In an urbanparkin Boston, of the plant species presentin
1894, 155 were absentby 1993, amountingto a decreasefrom 84%to 74% native
flora.Sixty-fourspecies were new. In additionto trails,Drayton& Primack(1996)
blamedfire and tramplingfor the change in exotics.
Structuraland compositionalchangesarenot the only dynamicsin urbanbiota.
Plants and animals in cities have evolved in response to the local conditions in
cities (Sukopp 1998, Bradshaw& McNeilly 1981). The famous populationgenetic differentiationin copper and zinc tolerancein creeping bent grass (Agrostis
stolonifera)and roadsidelead toleranceof ribwortplantain(Plantago lanceolata)
in urbansites is an example(Wu& Antonovics 1975a, b). Industrialmelanismand
136
ETAL.
PICKETT
other forms of melanism are also urbanevolutionaryresponses (Bishop & Cook
1980).
A majorfeatureof urbanvegetationis the spatialheterogeneityin urbanlandscapes createdby the arrayof buildingdensitiesandtypes, differentlanduses, and
different social contexts. To characterizethis heterogeneity,plant habitatshave
been subjectedto various classification systems. Forest Steams (Steams 1971),
one of the firstAmericanecologists to call for researchin urbanlandscapes,identified three majorvegetationtypes-ruderal, managed,and residual. In addition
to assessing spatialheterogeneity,classificationsystemsrecognizethe importance
of characterizingnaturalhabitatsin urbanlandscapes.Rogers & Rowntree(1988)
developed a system to classify vegetationusing life forms. This process was used
to assess naturalresourcesin New YorkCity (Sisinni & Emmerich1995). Based
on site histories,Zippereret al. (1997) classified tree-coveredhabitatsas planted,
reforested,or remnant.These approachesallow for spatiallyexplicit comparisons
between vegetationand othervariablesof interestfor a single point in time.
The vegetationandfloristicstudies of naturein cities sharekey characteristics.
They are largely descriptive,but illustrate spatial heterogeneity as a source of
diversity, and suggest a functional role for landscape structure(Rebele 1994).
Althoughit is legitimateto view the city as an open and dynamicecosystem, few
of the plant ecological studies document successional processes (Matlack 1997)
or expose the functionalrelationshipsof the vegetation(Mucina 1990).
Douglas (1983), Gilbert(1989), VanDruffet al. (1994),
andNilon & Pais (1997) have summarizedthe literatureon the animallife of cities.
Theyreveala long historyof researchon the faunaof EuropeanandNorthAmerican
cities. Althoughmuchof the researchhas been descriptive,severalstudiesfocused
on processes thatare also majorfoci of researchin ecology as a whole. We review
animal studies startingwith coarse scale or gradientcomparisons,followed by
patch-orientedstudies, including some that incorporatesocioeconomic aspects,
then show how wildlife biologists have contributedto urbanecology, andend with
severalexamples of process studies involving animals.
Birds, mammals, and terrestrialinvertebratesare the best studied taxonomic
groups,withaquaticfauna,reptiles,andamphibiansless studied(Luniak& Pisarski
1994). In addition, the fauna of green spaces are relatively well studied, with
built-upand derelict areasand waterbodies less well known. Andrzejewskiet al.
(1978) and Klausnitzer& Richter(1983) describedhow urban-to-ruralgradients
definedby humanpopulationdensity and buildingdensity impactthe occurrence
and abundanceof various animal species. Among mammals, there is a shift to
medium sized, generalist predatorssuch as racoons and skunks in urban areas
(Nilon & Pais 1997). The predatorfauna contrastsbetween cities and nonurban
environments.Forexample,antswere the most importantpredatorsin some urban
areas althoughvertebrateswere more significantin nonurbanhabitats(Wetterer
1997). Changesin faunamay influencevegetation.For example,invertebratepest
densities increasedin urbantrees comparedto wild forest stands, reflecting the
ANIMALSAND WILDLIFE
TERRESTRIAL
URBANECOLOGY
137
additionalstressanddamageto whichurbantreesaresubjected(Nowak& McBride
1992).
Animal ecologists have recognized spatialheterogeneityin a variety of ways.
Heterogeneityis typically expressedas patches or contrastingbiotopes. The approachto spatialheterogeneityof urbanfauna in Germanyhas been shapedby a
nationalprogramof biotope mapping(Sukopp 1990, Werner1999). The German
biotopeapproachis basedon phytosociologic-floristicandfaunalcharacteristicsof
sites in cities (Sukopp& Weiler 1988). Biotope mappingcan be used to document
dynamicsof urbanfaunasuch as the changes in the birdspecies composition,and
abundanceof differenttypes of biotopes (Witt 1996). Bradyet al.'s (1979) general
typology of urbanecosystems and Matthewset al.'s (1988) habitatclassification
scheme for metropolitanareas in New York State are examples of schemes that
recognize how patternsof land-use history and resulting changes in land cover
create ecologically distincthabitatpatches.
Use of patch-orientedapproachesis particularlywell developed in studies of
mammals. For example, land use and cover in 50-ha areas surroundingpatches
in Syracuse, New York,were the best predictorsof small mammal species composition (VanDruff& Rowse 1986). Similarly, patch configurationin Warsaw,
Poland,affectedmammalpopulations.Urbanizationblocked the dispersalof field
mice and altered population structure and survivorship in isolated patches
(Andrzejewskiet al. 1978). Largermouse populationswere associated with increasedpercentagesof builtandpavedareas,barriersto emigration,and decreases
in patch size (Adamczewska-Andrzejewskiet al. 1988). In addition,mammalian
predatorsof seeds may be increased in urban patches comparedto rural areas
(Nilon & Pais 1997). The change in patch configurationresulting from suburban sprawlhas led to the widespreadincreasein deer densities on urbanfringes.
Suburbanizationhas caused increasedjuxtaposition of forest habitats in which
deer shelterwith field or horticulturalpatches in which they feed (Alversonet al.
1988). The positive feedback of changingpatch configurationon deer population
densityis magnifiedby reducedhuntingandpredatorpressureassociatedwith suburbanization.Greaterdeer densities have the potentialto change both the animal
communityand vegetation(Bowers 1997).
Some biologistshavelookedto social causesof animalandplantdistributionand
abundancein cities. Studiesof floraandfaunaof metropolitanLiverpool,England,
showed that changes in land use and technology have influencedhabitatchange
since the industrialrevolution(Greenwood1999), and studiesof faunain suburban
Warsawincludedland-use informationdating from the 1700s (Mackin-Rogalska
et al. 1988). Key to the Warsawstudywas the recognitionthatsuburbanareashave
differentsocioeconomicprocessesthando eitherurbanorruralareas,andthatthese
processes influence land use, land-cover characteristics,and habitats for biota.
Examples of the relationshipsof animal populationsto land-covertypes include
those for birds. For a city as a whole, exotic generalists such as pigeons, starlings, and sparrowscan constitute80%of the birdcommunityin the summer,and
95% in winter (Wetterer1997). At finer scales, contrastingland-coverpatches of
138
PICKETTET AL.
residential areas, commercial sites, and parks supporteddifferent avian assemblages (Nilon & Pais 1997). Even differences in plant cover among residential
neighborhoodsaffectedbirdcommunitycomposition.Likewise, the land uses adjacent to urbangreen spaces strongly influence bird communities (Nilon & Pais
1997).
In Europeand North America, much of the researchon the faunaof cities has
been conductedby appliedecologists to supportconservationandnaturalresource
management.This workrecognizedthatcities areareasworthyof study,andmore
importantly,that people and their activities create a unique context for natural
resourcemanagementin andaroundcities (Waggoner& Ovington 1962, Noyes &
Progulske1974). In particular,the patchdynamicsapproachwas foreshadowedby
wildlife ecologists in cities. The activity of urbanwildlife biologists is illustrated
by the conferences of the NationalInstitutefor UrbanWildlife (Adams & Leedy
1987, 1991, Adams & VanDruff1998). These conferencesfocused on (a) ecology
of urbanwildlife, (b) planningand design, (c) managementissues and successes,
and (d) publicparticipationand education.The participatoryapproachto studying
animals in cities involves urbanresidentsin the process (Adams & Leedy 1987)
and serves as a model for other ecological research in cities. Furthermore,the
conferences focused on planning and managementas activities that can change
habitatsat both citywide and local scales.
An additionalstresson animalpopulationsin cities is the director indirecteffect
of domesticpets. Buildingdensityis associatedwith an increasein the numbersof
pets in an area(Nilon & Pais 1997). The amountof food energy availableto freeranging domestic cats depends on the affluence of the immediateneighborhood
(Haspel & Calhoon 1991). Cats have a significantimpact on bird populationsin
suburbs(Churcher& Lawton 1987).
An exampleof an ecological processfamiliarin wild andproductionlandscapes
thatalso appearsin cities is animalsuccession. After the establishmentof the new
townof Columbia,Maryland,birdssuchas bobwhiteandmourningdove, which are
associatedwith agriculture,gave way to starlingsand house sparrows,which had
been absentbefore urbanization(Hough 1995). In an urbanparknear Dortmund,
Germany,a 35-year study found an increase in generalist species richness and
density at the expense of specialists.The species turnoverrate of birdsin the park
between 1954 and 1997-42.1 %-was higherthanin a forestdistantfromthe city
over the same period (Bergen et al. 1998).
This overview of animals in urbanareashas confirmedthe diversityof exotic
and native species in cities. These species, along with the planted and volunteer
vegetationin cities, are in some cases an importantamenity,and in othersa significanthealthor economic load. They can serve to connectpeople with naturalprocesses througheducationalactivities.Althoughthereis a greatdeal of descriptive
knowledgeaboutthe biota of cities, thereis the need to comparefood web models
for green andbuiltpartsof the city, to link these datawith ecosystem function,and
to quantifythe relationshipsbetweeninfrastructural
andhumanbehavioralfeatures
of the metropolis(Flores et al. 1997). In addition,long-termstudies are required.
TERRESTRIALURBAN ECOLOGY
139
The case studies we reporthere are a foundationfor generalizationsconcerning
the structureand functionof biota within cities.
Ecology of the City
The knowledgeof naturein cities is a firmfoundationfor understandingecological
processes in metropolitanareas.Yet it is not sufficient(Flores et al. 1997). If scientists,planners,anddecision makersareto understandhow the social, economic,
andecological aspectsof cities interact,the feedbacksanddynamicsof the ecological linkages must be assessed. We thereforeturnto a review of systems-oriented
approachesto urbanecology. These representa shift to the perspectiveof ecology
of cities, as contrastedwith the literaturewe have reviewed so far, which focused
on ecology in cities.
The diversespatialmosaics of metropolitanareaspresenta varietyof ecological
situationsin which to examine ecological structureand dynamics. For example,
severalof the conditionsin cities are analogousto majorpredictionsof global climate change. Increasedtemperatures,alteredrainfallpatterns,and dryingof soils
anticipatetrendsprojectedfor some wild lands.Examinationof existing urbanassemblagesor experimentationwith novel assemblagesof nativeandexotic species
may be useful for assessing the effects of climate change on biodiversity.The
strandedriparianzones of urbansites, resultingfrom the downcuttingof streams
associatedwith impervioussurfaces,can be used to examine alteredenvironmental driversof system function.Plant communityregenerationand the response of
ecosystems to soil nitrogencycling underalteredmoistureandtemperatureconditions may be investigatedin such areas.Examiningsuccession in vacantlots may
informpracticalvegetationmanagementand suggest strategiesfor changingland
use as the density of humansand buildingsdecline in some city centers. Finally,
patternsof adjacencyof managed and wild patches in and aroundcities can be
used to examine landscapefunction.
The ecology of the entire city as a system is representedby researchrelating
species richness to the characteristicsof cities. For instance,the numberof plant
species in urbanareascorrelateswith the humanpopulationsize. Species number
increaseswith log numberof humaninhabitants,and thatrelationshipis stronger
than the correlationwith city area (Klotz 1990). Small towns have from 530 to
560 species, while cities having 100,000 to 200,000 inhabitantshave upwardsof
1000 species (Sukopp 1998). The age of the city also affects the species richness;
large,oldercities havemoreplantspecies thanlarge,youngercities (Sukopp1998,
Kowarik 1990). These plant assemblages are characteristicthroughoutEurope,
with 15%of species sharedamong cities (Sukopp 1998).
BIOGEOPHYSICALBUDGETS One of the earliest modern ecological approaches
to urbansystems was the assessment of biogeochemical budgets of whole cities
(Odum& Odum1980). It is clearthaturbanareasareheterotrophicecosystemsthat
dependon the productivityfromelsewhere(Collinset al. 2000). Cities in industrial
140
PICKETT ET AL.
countriesmay use between 100,000 and 300,000 Kcal m-2yr-1,whereas natural
ecosystemstypicallyexpendbetween 1,000 and 10,000Kcal m-2yr-1(Odum1997).
The energy budgetsof cities are drivenby fossil fuel subsidies, which contribute
to the urbanheat island effect.
How the green componentof cities affects biogeochemicalprocessing for the
city as a whole has been examined. For example, in Chicago, trees have been
estimatedto sequester5575 metric tons of air pollution, and 315,800 metric tons
of C per year at an average rate of 17 metric tons ha-1 y-1 (McPhersonet al.
1997). In the Mediterraneanclimate of Oakland,California,trees sequester 11
metric tons of C ha-l y-1 (Nowak 1993). In contrastto urban forests, natural
forests on averagesequester55 metric tons ha-1 y-1 (Zippereret al. 1997). The
capacityof treesto filterparticulatesfromurbanairis basedon leaf size andsurface
roughness(Agrawal1998).
Urbanareas also concentratematerialsfrom elsewhere. For example, the elevation of the surfacein old cities is generally higher than the surroundingareas
as a result of importingconstructionmaterials(Sukopp 1998). A carbondioxide
dome accumulatesovercities in associationwith combustionof fossil fuels (Brazel
et al. 2000). Anthropogenicallyproducedforms of N are concentratedin (Lovett
et al. 2000) anddownwindof cities (Chamiedeset al. 1994). Nitrogenaccumulated
fromhumanmetabolismand fertilizersconcentratesdownstreamof cities. Hence,
populationdensity of the watershedsis statisticallycorrelatedwith N loading in
the majorriversof the world (Caraco& Cole 1999).
A useful way to quantify the dependence of urban systems on ecosystems
beyond their bordersis the concept of the ecological footprint.The ecological
footprintof an urbanareaindexes the amountof land requiredto producethe materialand energeticresourcesrequiredby, and to process the wastes generatedby,
a metropolis(Rees 1996). The city of Vancouver,Canada,requires180 times more
landto generateandprocessmaterialsthanthe city actuallyoccupies. The concept
is highly metaphorical,becausethe actualnetworksfromwhich anyparticularcity
drawsresources,andthe areasaffectedby its waste, may extend aroundthe globe.
An analysis of the growth of Chicago (Cronon 1991) showed that a network of
resourceacquisitionextendedthroughoutthe westernregions of the United States
in the late nineteenthcentury.The metropolis in the postindustrial,information
age in nations enjoying a high fossil fuel subsidy has differentconnections with
the hinterlandthan did the industriallyand agriculturallyanchoredChicago of a
centuryago (Bradley 1995). Telecommuting,materiallyand energeticallysubsidized recreation,andthe alterationof landvalues for urbanuses in the countryside
representa footprintbased on urbancapital.
ECOSYSTEM
ANDPROCESSTherearethreeways in which contemporary
PATTERN
studiesof ecology of cities, ormoreproperlyentiremetropolitanareas,areprepared
to move beyond the classical ecological approachesand to supportintegration
with social andphysical sciences. 1. The contemporaryassumptionsof ecosystem
TERRESTRIALURBAN ECOLOGY
141
function are more inclusive than the classical assumptions.2. The net effects
approachto ecosystem budgets has evolved to consider multiple processes and
spatial heterogeneity.Finally, 3. the narrow theories that were used to bridge
disciplines in the past have been broadenedor replaced.
The firsttool for integrationis the contrastbetween contemporaryand classical assumptionsabout ecosystem function. Classically, ecologists based studies
of urbanareas on the assumptionsthat ecosystems were materiallyclosed and
homeostaticsystems. Such assumptionswere a partof the ecosystem theoryused
by many early geographers(Zimmerer1994). These assumptionshave been replaced(Zimmerer1994, Pickettet al. 1992). Consequently,thereis a new theoryof
ecosystems thatwas not availableto those who pioneeredthe budgetaryapproach
to urbansystems (e.g., Boyden et al. 1981). Whatremainsis the basic concept of
the ecosystem as a dynamic, connected, and open system (Likens 1992), which
can servethe variousdisciplines(Rebele 1994) thatneed to be integratedto form a
more comprehensivetheoryto supportjoint ecological, social, and physical study
of urbansystems.
Contemporaryecology propoundsa systems view that builds on the rigorous
budgetaryapproachto ecosystems (e.g., Jones & Lawton 1995). Of course ecologists necessarilycontinueto exploit the laws of conservationof matterandenergy
to generatebudgetsfor ecosystem processes. However,many contemporarystudies of ecosystem budgets do not treat systems as though they were black boxes.
Rather,the structuraldetails and richness of processes that take place within the
boundariesof the system area majorconcernof contemporaryecosystem analysis.
Contemporaryecosystem ecology exposes the roles of specific species and interactionswithincommunities,flows betweenpatches,andthe basis of contemporary
processes in historicalcontingencies.These insightshave not been fully exploited
in urbanecological studies.
The thirdfeatureof contemporaryecology is the breadthof key theories that
can be used in integration.Classical ecological theory provided social scientists
with only a narrowstructurefor integrationwith ecology. The social scientists of
the ChicagoSchool, which was activein the earlydecadesof the twentiethcentury,
used ecologically motivatedtheoriesof succession and competition,for example.
At the time, only the relatively general, deterministic,and equilibriumversions
of those theorieswere available.Contemporarytheoriesof interactionaccountfor
bothpositiveandcompetitiveeffects, andpredictionsarebasedon the actualmechanisms for interactionratherthannet effects. In contemporarysuccession theory,
mechanismsotherthanfacilitationareincluded,andthe sequenceof communities
may not be linearor fixed. In addition,ecologists have come to recognizethatthose
theorieshave hierarchicalstructure,which allows them to addressdifferentlevels
of generalityand mechanistic detail depending on the scale of the study or the
scope of the researchquestions (Pickettet al. 1994). Hence, social scientists and
ecologists now can select the most appropriatelevels of generalityin a theoretical
area for integration.No longer must integrationrely on general theories of net
effects of such processes as competitionand succession.
142
PICKETTET AL.
URBANECOLOGYAS A PLANNINGAPPROACH
Althoughthe firstvolume of the journalEcology containeda scientificpaperdevoted to the effect of weatheron the spreadof pneumoniain New YorkandBoston
(Huntington1920), the interactionsof humanswith the urbanenvironmenthave
been primarilythe province of plannersand landscape architects.For example,
CentralParkin New YorkCity and otherurbanparksdesigned by FrederickLaw
Olmstedseem intuitivelyto link environmentalpropertiesto humanwell being in
cities. In particular,Olmsted's design for the Boston Fens and Riverway shows
ecological presciencein its sophisticatedcombinationof wastewatermanagement
and recreationalamenity (Spirn 1998). Ian McHarg's(1969) Design with Nature
alertedplannersand architectsto the value of incorporatingknowledge of ecological and naturalfeatures among the usual engineering, economic, and social
criteriawhen developing a regional plan. In McHarg'sapproach,environmental
risks and amenitiesof differenttypes are mappedon separatelayers. The composite map suggests where certaintypes of developmentshould or should not occur.
This approachpresagedthe technologyof GeographicInformationSystems (GIS),
which has become an importanttool to incorporatemultiple criteriain planning
(Schlutnik1992, Grove 1997) such as those proposedby McHarg(1969). A more
explicit ecological approachis that of Spirn (1984), who examined how natural processes are embedded in cities, and how the interactionbetween the built
environmentand naturalprocesses affected economy, health, and human community. For instance, she showed how the forgotten environmentaltemplate of
drainagenetworkscontinuedto affect infrastructureand the social structureof a
Philadelphianeighborhood.
The planning perspective of human ecology is especially strong in Europe
(Sukopp 1998). Planningin Germanyhas been heavily influencedby a national
programof biotope mapping that includes cities (Sukopp 1990, Werner 1999).
This programincludes descriptionsof the flora and fauna of biotopes as a key to
identifyingtypes of habitatsthatare significantfor 1. protectingnaturalresources,
2. qualityof life, and 3. a sense of place and identityin the city (Werner1999). In
additionto identifyingspecificbiotopes,researchersin Mainzhavemappedthe distributionof floraand fauna,naturalphenomena,and recreationalactivitieswithin
the biotopes (Frey 1998). Similar researchby the Polish Academy of Sciences
has focused on urbanand suburbanareas. The researchincluded studies of soils
and abiotic ecosystem components,and researchby social scientists in a mosaic
of habitatswith differentdegrees of development(Zimny 1990). Building upon
the foundationof vegetationclassificationin cities, Brady et al. (1979) proposed
a continuumof habitatsfrom the naturalto the highly artificial.Dorney (1977),
using a similarapproach,proposedan urban-ruralcontinuumfrom a planningperspectiveandidentifiedsix representativelandzones-central businessdistrict,old
subdivisions,new subdivisions,urbanconstructionzones, urbanfringe, andrural.
Each zone was characterizedby threecomponentsor subsystems:culturalhistory,
PICKETTET AL C- 1
Figure 1 Patchiness in the Rognel Heights neighborhoodof Baltimore,Maryland.The
spatial heterogeneityof urbansystems presents a rich substratefor integratingecological, socioeconomic, and physical patternsusing GeographicInfonnation Systems. The
patch mosaic discriminatespatches based on the structuresof vegetation and the built
environment.The base image is a false color infraredorthorectifiedphoto from October
1999. (M.L. Cadenasso,S.T.A Pickett & W.C. Zipperer,unpublisheddata.)
TERRESTRIAL
URBANECOLOGY
143
abiotic characteristics,and biotic features.We review the status and implication
of such integratedclassificationslater.
Urban ecology manifest as city planning is contrastedwith spatial planning
(de Boer & Dijst 1998) in whichprimarymotivationsarethe degreeof segregation
or aggregationof differenteconomicand social functions,efficiency of transportation and deliveryof utilities,andefficientfilling of undevelopedspace. Additional
components of urbanplanning said to have ecological foundationsinclude life
cycle analysis of products,utilityplanningbased on use ratherthanmedium,efficiency of resourceuse, exploitationof green infrastructure,and requirementsfor
monitoringof the results (Breusteet al. 1998).
Although the planningdescribedabove is ecologically motivated,and it relies
on mapping to describe environmentalamenities, it is rarelybased on data concerning ecological function. It thereforerelies on general ecological principles
and assumptions,and on the success of prior case histories (Flores et al. 1997).
The insights of urbanecology as planningare summarizedin manualsand codified in zoning andplanningpractice.However,like otherenvironmentalpractices,
these insights may not be applicablein novel ecological circumstances.Given the
changing forms of cities in both Europeand the United States, novel ecological
circumstancesmay be in the offing.
AN INTEGRATEDFRAMEWORK
FORURBAN
ECOLOGICAL
STUDIES
We see threeopportunitiesfor improvingthe theoryto understandurbansystems.
First,ratherthanmodelinghumansystems andbiogeophysicalsystems separately,
understandingwill be improvedby using integratedframeworksthatdeal with social andbiogeophysicalprocesses on an equalfooting (Groffman& Likens 1994).
Second, knowing thatthe spatialstructureof biogeochemical systems can be significantfor theirfunction(Pickettet al. 2000), we hypothesizethatthe spatialheterogeneityso obvious in urbansystems also has ecological significance(Figure 1,
see color insert).Third,insights from hierarchytheorycan organizeboth the spatial models of urbansystems and the structureof the integratedtheorydeveloped
to comprehendthem. We explore and combine these themes below.
Social Ecologyand Social Differentiation
The studyof social structures,and how those structurescome to exist, are the key
social phenomenato supportintegratedstudy of the ecology of urban systems.
It is increasinglydifficultto determinewhere biological ecology ends and social
ecology begins (Golley 1993). Indeed,the distinctionbetween the two has diminished throughthe convergenceof related concepts, theories, and methods in the
biological, behavioral,andsocial sciences. Social ecology is a life science focusing
144
ETAL.
PICKETT
on the ecology of various social species such as ants, wolves, or orangutans.We
may also study Homo sapiens as an individual social species or comparatively
with the ecology of othersocial species. The subjectmatterof social ecology, like
thatof biological ecology, is stochastic,historic,andhierarchical(Grove& Burch
1997). In otherwords, living systems arenot deterministic;they exhibit historical
contingencies that cannot be predictedfrom physical laws alone (Botkin 1990,
Pickettet al. 1994).
The underlyingbasis for this life science approachto the study of humanecological systems dependsupon threepoints (Grove & Burch 1997):
1. Homosapiens, like all otherspecies, arenot exemptfromphysical,chemical,
or biological processes.Biophysicaland social characteristicsof humansare
shapedby evolution and, at the same time, shape the environmentin which
Homo sapiens live;
2. Homo sapiens, like some other species, exhibit social behaviorand culture;
and
3. Social and cultural traits are involved fundamentallyin the adaptationof
social species to environmentalconditions.
Humanecology must reconcile social and biological facts to understandthe
behaviorof Homo sapiens over time (Machlis et al. 1997). Such a biosocial approachto humanecological systems (Burch 1988, Field & Burch 1988, Machlis
et al. 1997) standsin contrastto a more traditionalgeographicor social approach
(see Hawley 1950, Catton 1994). This is not to say that social sciences such as
psychology, geography,anthropology,sociology, economics, andpolitical science
are not importantto social ecology. They are, because the most fundamentaltrait
thatdistinguisheshumansandtheirevolutionaryhistoryfrom otherspecies-both
social and nonsocial-is that human social developmenthas enabled the species
to escape local ecosystem limitationsso thatlocal ecosystems no longer regulate
humanpopulationsize, structure,or genetic diversity(Diamond 1997). Nowhere
is this more apparentthanin urbanecosystems.
One of the majortools for integrationbetween social and biogeophysical sciences is in the phenomenonof social differentiation.All social species arecharacterizedby patternsandprocessesof social differentiation(vandenBerghe 1975). In
the case of humans,social differentiationor social morphologyhas been a central
focus of sociology since its inception(Grusky1994). In particular,social scientists
have used concepts of social identity (i.e., age, gender,class, caste, and clan) and
social hierarchiesto study how and why human societies become differentiated
(Burch& DeLuca 1984, Machlis et al. 1997).
Social differentiationis importantfor humanecological systems because it affects the allocationof criticalresources,includingnatural,socioeconomic,andculturalresources.In essence, social differentiationdetermines"whogets what,when,
how and why" (Lenski 1966, Parker& Burch 1992). Being rarelyequitable,this
allocation of criticalresourcesresults in rankhierarchies.Unequal access to and
controlover criticalresourcesis a consistentfact within and between households,
TERRESTRIALURBAN ECOLOGY
145
communities,regions, nations, and societies (Machlis et al. 1997). Five types of
socioculturalhierarchiesarecriticalto patternsandprocessesof humanecological
systems:wealth,power,status,knowledge, and territory(Burch& DeLuca 1984).
Wealthis access to and control over materialresourcesin the form of naturalresources, capital, or credit. Power is the ability to alter others' behavior through
explicit or implicitcoercion(Wrong1988). The powerfulhave access to resources
thataredeniedthe powerless.One exampleis politicianswho make land-usedecisions or provideservicesfor specific constituentsat the expense of others.Statusis
access to honorandprestigeandthe relativeposition of an individual(or group)in
aninformalhierarchyof social worth(Lenski 1966). Statusis distributedunequally,
even within small communities,but high-statusindividualsmay not necessarily
have access to eitherwealth or power.For instance,a ministeror an imam may be
respectedand influentialin a communityeven thoughhe or she is neitherwealthy
nor has the ability to coerce other people's behavior.Knowledge is access to or
control over specialized types of information,such as technical, scientific, and
religious. Not everyonewithin a social system has equal access to all types of information.Knowledgeoften providesadvantagesin termsof access to andcontrol
over the critical resourcesand services of social institutions.Finally, territoryis
access to andcontrolover criticalresourcesthroughformaland informalproperty
rights (Burchet al. 1972, Bromley 1991).
Social differentiationof human ecological systems has a spatial dimension
characterizedby patternsof territorialityand heterogeneity(Morrill 1974, Burch
1988). As Burch(1988) noted, "Intimateand distantsocial relations,high andlow
social classes, favoredand despised ethnic, occupational,and caste groupingsall
have assignedandclearlyregulatedmeasuresas to when andwherethose relations
should and should not occur."When ecosystem and landscape approachesare
combined, the researchchanges from a question of "who gets what, when, how
and why?" to a question of "who gets what, when, how, why and where?"and,
subsequently,what are the reciprocalrelationshipsbetween spatial patternsand
socioculturalandbiophysicalpatternsandprocesses of a given area(Grove 1997)?
Variousprocesses of social differentiationoccur at different scales and have
correspondingspatialpatternsand biophysicaleffects (Grove& Hohmann1992).
Based on existing social and ecological theory,examples include global and regional urban-ruralhierarchies(Morrill1974), the distributionof land uses within
urbanareas(Guest 1977), the stratificationof communitieswithin residentialland
uses (Logan & Molotch 1987), and the social differentiationof ownershipsand
householdswithin communities(Burch& Grove 1993, Grove 1995).
SpatialHeterogeneity
A human landscape approachmay be understoodas the study of the reciprocal
relationshipsbetween patternsof spatialheterogeneityand socioculturaland biophysical processes. Further,when human ecosystem and landscape approaches
are combined, human ecosystem types are defined as homogeneous areas for a
146
ETAL.
PICKETT
specified set of socioculturalandbiophysicalvariableswithin a landscape.Analyses then focus on two primaryissues: 1. the developmentand dynamicsof spatial
heterogeneity,and 2. the influences of spatial patternson cycles and fluxes of
critical ecosystem resources (e.g., energy, materials,nutrients,genetic and nongenetic information,population,labor, capital, organizations,beliefs, or myths).
For instance,the developmentanddynamicsof heterogeneityin a watershedspanning urbanto ruralconditions may influence and be influencedby sociocultural
and biophysical processes. Patches within the watershedmay function as either
sourcesor sinksas well as to regulateflows andcycles of criticalresourcesbetween
otherpatches.The delineationand classificationof these relativelyhomogeneous
patches is based on a limited numberof representativesocioculturaland biophysical indicators(Burch & DeLuca 1984, Parker& Burch 1992), and the patches
are studied as black boxes with fluxes and cycles of critical resources between
areas (Zonneveld 1989). The spatial linkages between the social and ecological
differentiationof the watershedand the relationshipof the linkages to different
types of allocationmechanismsat differentscales areimportantfor understanding
the flows and cycles of criticalresourceswithin the watershed.
Hydrologistshave recognized mosaics of spatialheterogeneityin the variable
source area (VSA) approach.They examine how the abiotic attributesof different patches within a watershed-suchas temperatureand physical characteristics including topography,soil properties,water table depth, and antecedentsoil
moisture-contributevariable amounts of water and nutrientsto streamflow,depending upon their spatial location in the watershed (Black 1991). This VSA
approachcan be integratedwith a delineation of patches based upon the biotic
attributesof the watershed,such as vegetation structureand species composition
(Bormann& Likens 1979), and the social attributesof the watershed,such as indirecteffects from land-usechange and forest/vegetationmanagement,and direct
effects frominputsof fertilizers,pesticides,andtoxins, to examinehow the abiotic,
biotic, and social attributesof differentpatcheswithina watershedcontributevariable amountsof water and nutrientsto streamflow,dependingupon their spatial
location in the watershed(Grove 1996). This integratedVSA approachcombines
nested hierarchiesof land use and land cover, sociopolitical structures,and watershed heterogeneity(Figure2). GIS is a useful tool for analyzingnestedhierarchies
of spatialheterogeneity.
VSA approachescan be linked to additional social processes. For example,
catchmentscan be examined via hydrological, land-use, and economic models
(Costanzaet al. 1990). The threecomponentscan be combinedinto an ecosystem
model composedof grid cells within a catchment.The integratedmodel is builtup
from a basic model thathas hydrologicandecologic components,butno economic
components. Therefore, in the basic model, human behavior causing land-use
changemustbe consideredas a factorexternalto the focal catchment.To construct
truly integratedecological-economic models, majornutrient,water,productivity,
and successionalcomponentsof the basic model must be combinedwith land-use
and economic valuations(Bockstael et al. 1994). The integratedmodel calculates
TERRESTRIALURBAN ECOLOGY
147
Land Cover - Ecological
Hierarchy
Land Cover Type
Patch Configuration
Soil
Permeability Class
Socio-Political
Hierarchy
Political Units
Land Use
Neighborhood
Household
Hydrological
Hierarchy
Watershed
Sub-watershed
Sub-catchment
Hillslope
Floodplain
Channel
Figure 2 Threeinteractingnestedhierarchiesof spatialheterogeneityrepresenting
keydisciplinaryperspectivesusedin modelingwatershedfunctionin urbanareas.Land
coverrepresentsnestedecologicalstructuresthatcontrolinterceptionandrunoff.The
decisionmakingand
sociopoliticalhierarchycontainsnestedunits of environmental
resourceuse.Thehydrologicalhierarchyindicatesthenestingof spatiallydifferentiated
unitsconnectedby runoffandrunondynamics.
land-use designation througha habitat-switchingmodule that determineswhen,
throughnaturalsuccession or weather-drivenecological catastrophe(e.g., floods),
the habitatshifts from one type to another.Hypotheticalhuman-causedland-use
changes can be imposed exogenously using the integratedmodel. Recognizing
thatthe ecological effects of humanactivityaredrivenby the choices people make
concerningstocksof naturalcapital,theeconomic modelinguses an understanding
of how land-usedecisions aremadeby individualsandhow they arebased on both
the ecological andeconomic featuresof the landscape.Again, GIS serves as a tool
for manipulatingthe spatial data on which the model depends, and for assessing
the model outputover space.
The Human EcosystemFramework
and Urban EcologicalSystems
An integratedframeworkfor analyzing urbansystems as social, biological, and
physical complexes now emerges. Social scientists have focused on interactions
between humans and their environmentssince the self-conscious origins of their
148
PICKETTET AL.
disciplines. However,the explicit incorporationof the ecosystem concept within
the social sciences dates to Duncan's (1961, 1964) articles "FromSocial System
to Ecosystem"and "Social Organizationand the Ecosystem."Recently,the gocial
sciences have focused increasinglyon the ecosystem concept because it has been
proposed and used as an organizing approachfor naturalresource policy and
management(Rebele 1994).
The ecosystem concept and its applicationto humansis particularlyimportant
because of its utility as a frameworkfor integratingthe physical, biological, and
social sciences. The ecosystem concept owes its origin to Tansley (1935), who
noted that ecosystems can be of any size, as long as the concern is with the
interactionof organismsand their environmentin a specified area. Further,the
boundariesof an ecosystem are drawnto answera particularquestion.Thus, there
is no set scale or way to bound an ecosystem. Rather,the choice of scale and
boundaryfor definingany ecosystem dependsupon the questionasked and is the
choice of the investigator.In addition,each investigatormay place more or less
emphasis on the chemical transformationsand pools of materials drawn on or
createdby organisms;or on the flow, assimilation,and dissipationof biologically
metabolizableenergy;or on the role of individualspecies or groupsof species on
flows and stocks of energy and matter.The fact thatthereis so much choice in the
scales and boundariesof ecosystems, and how to study and relate the processes
within them, indicates the profounddegree to which the ecosystem representsa
researchapproachratherthan a fixed scale or type of analysis.
Althoughthe ecosystem concept is flexible enough to accountfor humansand
their institutions(Tansley 1935, Rebele 1994), the applicationof an ecosystem
approachto the study of human ecosystems requiresadditionalanalyticalcomponents. The analyticalframework(Figure 3) we use here (see Burch & DeLuca
1984, Machlis et al. 1997, Pickett et al. 1997) is not itself a theory.As Machlis
et al. (1997: p. 23) noted,
"This humanecosystem model is neitheran oversimplificationnor caricature
of the complexity underlyingall types of human ecosystems in the world. Parts
of the model are orthodoxto specific disciplines and not new. Otherportions of
the model are less commonplace-myths as a culturalresource,justice as a critical institution.Yet we believe that this model is a reasonably coherent whole
and a useful organizing concept for the study of human ecosystems as a life
science.''
Several elements are critical to the successful applicationof this framework.
First, it is importantto recognize that the primarydriversof human ecosystem
dynamicsare both biophysical and social. Second, thereis no single determining
driverof anthropogenicecosystems. Third,the relativesignificanceof driversmay
vary over time. Fourth,componentsof this frameworkneed to be examinedsimultaneously in relationshipto each other (Machlis et al. 1997). Finally, researchers
need to examine how dynamicbiological and social allocationmechanismssuch
as ecological constraints,economic exchange, authority,tradition,andknowledge
affect the distributionof critical resourcesincluding energy,materials,nutrients,
TERRESTRIALURBAN ECOLOGY
149
Human Ecosystem Framework
Human Social System
Social Institutions
Social Cycles
Social Order
Institutional
Functions
Origins
Kinds
Resource System
Cultural Resources
Socio-Economic
Reource
Resources
Organizations
Beliefs
Myths
Information
Population
Capital
Labor
Ecosystem Structure and Processes
Energy Water Nutrients Materials Biodiversity Soil Patch mosaic Vegetation
Figure 3 A human ecosystem framework for integrating biogeophysical and social
structures and processes. This conceptual structure shows the most general components of any ecosystem that includes or is affected by humans. The nesting of the
boxes indicates the inclusion of specific structures or processes within more general
phenomena. The social system contains social institutions which function for provision of government, administration of justice, delivery of health services, provision of
sustenance, etc. Social order is determined by factors of individual and group identity,
formal and informal norms of behavior, and hierarchies that determine allocation of
resources. The resource system is founded on bioecological structures and processes
that include the standard subjects of ecology texts. The resource system also includes
cultural and socioeconomic resources, which interact with bioecological resources in
determining the dynamics of the social system. Based loosely on the work of Machlis
et al. (1997). Further ecological details added by Pickett et al. (1997).
population, genetic and nongenetic information, labor, capital, organizations,
liefs, and myths within any human ecosystem (Parker & Burch 1992).
be-
CONCLUSIONS
Although there is a wealth of information on the terrestrial components of urban
ecological sytstems, much of it is organized from the perspective of ecology in
cities. This perspective stands in contrast to a more comprehensive perspective
identified as the ecology of cities (Grimm et al. 2000). Studies of ecology in cities
150
ETAL.
PICKETT
have exposed the environmentalstresses, subsidies, and constraintsthat affect
urbanbiotaandhave documentedthatthe biotic componentsof metropolitanareas
have considerablepredictability.In addition,the capacityof certainorganismsto
adaptto urbanenvironmentsresults in characteristicassemblages.
An alternativeapproachto urbanecology exists in landscapearchitectureand
planning.This professionalpractice is motivatedby a desire to incorporateecological principles,to make environmentalamenitiesavailableto metropolitanresidents, and to decreasethe negative impactsof urbanresourcedemandand waste
on environmentselsewhere. Althoughfloristicand faunisticdescriptionsfrom urban sites are frequentlyused in design and planning,there are few data available
on ecological functions in cities that can inform such practice.Furthermore,the
rapidlychanging spatial forms of urbangrowthand change, and the complex of
environmentalfactorsthatinteractin and aroundcities, make simple environmental extrapolationsrisky. Although most of the urbanecological researchthat has
been motivatedby planning is of the sort that can be labeled ecology in cities,
the field often takes a more comprehensiveapproachthat expresses the ideal of
ecology of cities.
In basic ecological researchthe ecology of the city was firstaddressedby budgetary studies. Classically this approachhas been informedby a biogeochemical
perspectivebased on closed, homeostatic systems. Materialand energy budgets
of urbansystems have been estimatedunderthis rubric.Of course the budgetary
approachworks whethersystems are closed or not, or externallyregulatedor not.
However,assumptionsaboutspatialuniformityand thatsocial agents are external
to the ecological processes are questionable.
The distinct ecological approachesclassically applied to urbanresearch,and
the parallelplanningapproach,point to the need for integrationof differentdisciplinary perspectives.We have presented a frameworkthat uses middle level
theoriesfrom ecology and social sciences to identify key factorsthat should govern the structureand function of biotic, abiotic, and socioeconomic processes in
and aroundcities. This frameworkis identifiedas a humanecosystem model, in
which both social and ecological processes are integral.Furthermore,the spatial
heterogeneityin the biogeophysicaland social componentsof urbansystems can
be portrayedas patch dynamics. Because patch dynamics can be addressedover
many nested scales, structure-functionrelationshipsin the humanecosystem can
be examined from household to region. The integrativetools and existing data
prepareurbanecological studies to continue to benefit from the naturein cities
approach,as well as to expoit the contemporaryconcernsof ecology with spatially
heterogeneous,adaptive,self-organizingnetworksof entiremetropolitansystems.
ACKNOWLEDGMENTS
We are gratefulto the National Science Foundation,DEB 97-48135 for support
of the BaltimoreEcosystem Study (BES) Long-TermEcological Research program.We also thankthe EPA STARGrantProgramfor supportthroughthe Water
TERRESTRIAL
URBANECOLOGY
151
and Watershedsprogram,GAD R825792. This paperhas benefitedfrom insights
gained throughinteractionwith generouscollaborators,students,and community
partnersin New YorkCity and in Baltimoreover the last decade.
Visit the Annual Reviewshome page at www.AnnualReviews.org
LITERATURE
CITED
Adamczewska-Andrzejewska K,
MackinRogalskaR, Nabaglo L. 1988. The effect of
urbanizationon densityandpopulationstructure of Apodemus agrarius (Pallas, 1771).
Pol. Ecol. Stud. 14:171-95
Adams LW, Leedy DL, eds. 1987. Integrating
Man and Nature in the MetropolitanEnvironment.Columbia, MD: Natl. Inst. Urban
Wildl.
Adams LW, Leedy DL, eds. 1991. Wildlife
Conservationin MetropolitanEnvironments.
Columbia,MD: Natl. Inst. UrbanWildl.
Adams LW, VanDruffLW. 1998. Introduction.
UrbanEcosyst. 2:3
Agrawal M. 1998. Relative susceptibility of
plants in a dry tropical urban environment.
See Breusteet al. 1998, pp. 603-7
Airola TM, Buchholz K. 1984. Species structureandsoil characteristicsof five urbansites
along the New JerseyPalisades.UrbanEcol.
8:149-64
AlversonWS, WallerDM, Solheim SL. 1988.
Forests too deer: edge effects in northern
Wisconsin. Conserv.Biol. 2:348-58
AndrzejewskiR, Babinska-WerkaJ, Gliwicz
J, Goszczynski J. 1978. Synurbizationprocesses in populationof Apodemusagrarius.
I. Characteristicsof populations in an urbanization gradient.Acta Theriol. 23:34158
Bagnall RG. 1979. A study of human impact
on an urbanforest remnant:Redwood Bush,
Tawa,near Wellington,New Zealand.N. Z.
J.Bot. 17:117-26
Ballach H-J, Goevert J, KohlmannS, Wittig R. 1998. Comparativestudies on the size
of annualrings, leaf growthand structureof
See
treetopsof urbantreesin Frankfurt/Main.
Breusteet al. 1998, pp. 699-701
Bergen F, Schwerk A, Abs M. 1998. Longterm observationof the fauna of two manmade naturehabitats in cities of the RuhrValleyarea.See Breusteet al. 1998, pp. 61822
Bishop JA, Cook LM. 1980. Industrialmelanism and the urbanenvironment.Adv.Ecol.
Res. 11:373-404
Black PE. 1991. WatershedHydrology.Englewood Cliffs, NJ: PrenticeHall
Bockstael N, CostanzaR, StrandI, Boynton
W, Bell K, WaingerL. 1995. Ecological economic modeling and valuation of ecosystems. Ecol. Econ. 14:143-59
BormannFH, Likens GE. 1979. Pattern and
Process in a ForestedEcosystem.New York:
Springer-Verlag
Botkin DB. 1990. Discordant Harmonies: a
New Ecology for the Twenty-firstCentury.
New York:OxfordUniv. Press
Botkin DB, Beveridge CE. 1997. Cities as environments.UrbanEcosyst. 1:3-19
Bowers MA. 1997. Influenceof deer and other
factors on an old-field plant community.In
The Science of Overabundance:Deer Ecology and Population Management, ed. WJ
McShea, BH Underwood, JH Rappole, pp.
310-26. Washington, DC: Smithson. Inst.
Press
Boyden S, Millar S, Newcombe K, O'Neill
B. 1981. The Ecology of a City and its People: the Case of Hong Kong. Canberra:Aust.
Natl. Univ. Press
Bradley GA, ed. 1995. Urban Forest Landscapes: Integrating Multidisciplinaty Perspectives. Seattle:Univ. Wash.Press
BradshawAD, McNeilly T. 1981. Evolution
and pollution. London:EdwardArnold
Brady RF, Tobias T, Eagles PFJ, OhrnerR,
152
PICKETTET AL.
Micak J et al. 1979. A typology for the urban ecosystem and its relationshipto larger
biogeographicallanduse units. Urban Ecol.
4:11-28
Brazel A, Selover N, Vose R, Heisler G.
2000. The tale of two cities-Baltimore and
Phoenix urban LTER sites. Climate Res.
15:123-35
BreusteJ, FeldmanH, Uhlmann0, eds. 1998.
UrbanEcology. Berlin: Springer-Verlag
Bromley DW. 1991. Environmentand Economy. Property Rights and Public Policy.
Cornwall,UK: TJ
Brooks RT, Rowntree RA. 1984. Forest area
characteristics for metropolitan and nonmetropolitancounties of three northeastern
states of the United States. Urban Ecol.
8:341-46
Bullock P, GregoryPJ. 1991. Soils: a neglected
resourcein urbanareas.In Soils in the Urban
Environment,ed. P Bullock, PJ Gregory,pp.
1-5. Oxford,UK: Blackwell Sci.
Burch WR Jr. 1988. Humanecology and environmentalmanagement.In EcosystemManagementfor Parks and Wilderness,ed. JK
Agee, J Darryll, pp. 145-59. Seattle: Univ.
Wash.Press
Burch WR Jr, Cheek NH Jr, Taylor L, eds.
1972. Social Behavior, Natural Resources,
and the Environment.New York:Harper&
Row
BurchWR Jr, DeLuca D. 1984. Measuringthe
Social Impactof Natural ResourcePolicies.
Albuquerque:Univ. New Mexico Press
Burch WR Jr, Grove JM. 1993. People, trees,
and participationon the urbanfrontier.Unasylva 44: 19-27
CaracoNF, Cole JJ. 1999. Human impact on
nitrateexport:an analysisusing majorworld
rivers.Ambio28:167-70
CarreiroMM, Howe K, ParkhurstDF, Pouyat
RV. 1999. Variationsin quality and decomposability of red oak litter along an urbanruralland use gradient.Biol. Fert. Soils 30:
258-68
CattonWR Jr.1994. Foundationof humanecology. Sociol. Perspect. 37:75-95
Chameides WL, KasibhatlaPS, Yienger J,
Levy H. 1994. Growth of continental-scale
metro-agro-plexes,regionalozone pollution,
and worldfood production.Science 264:7477
ChurcherPB, LawtonJH. 1987. Predationby
domestic cats in an English village. J. Zool.
212:439-55
Collins JP, Kinzig A, GrimmNB, FaganWF,
Hope D et al. 2000. A new urban ecology.
Am. Sci. 88:416-25
Costanza R, Sklar FH, White ML. 1990.
Modeling coastal landscape dynamics.BioScience 40:91-107
CrononW. 1991.Nature'sMetropolis:Chicago
and the Great West.New York:Norton
de Boer J, Dijst M. 1998. Urban development and environmentalpolicy objectivesan outline of a multi-disciplinaryresearch
program. See Breuste et al. 1998, pp. 3842
Deelstra T. 1998. Towardsecological sustainable cities: strategies,models and tools. See
Breuste et al. 1998, pp. 17-22
De Rudder B, Linke F. 1940. Biologie der
Gro&stadt.Dresden:Leipzig
Diamond JM. 1997. Guns, Germs, and Steel:
the Fate of Human Societies. New York:
Norton
Domey RS. 1977. Biophysical and culturalhistoric land classification and mapping for
Canadianurbanand urbanizinglandscapes.
In Ecologicallbiophysical Land Classification in UrbanAreas, ed. Environ.Can., pp.
57-71. Ottawa:Environ.Can.
Douglas I. 1983. The UrbanEnvironment.Baltimore,MD: EdwardArnold
Dow CL, DeWalleDR. 2000. Trendsin evapotranspirationandBowen rationon urbanizing
watersheds in eastern United States. Water
Resour.Res. 7:1835-43.
Drayton B, PrimackRB. 1996. Plant species
lost in an isolated conservation area in
metropolitan Boston from 1894 to 1993.
Conserv.Biol. 10:30-39
Duncan OD. 1961. From social system to
ecosystem. Sociol. Inq. 31:140-49
Duncan OD. 1964. Social organization and
the ecosystem. In Handbook of Modern
TERRESTRIALURBAN ECOLOGY
Sociology, ed. REL Faris, pp. 37-82.
Chicago: Rand McNally
Effland WR, Pouyat RV. 1997. The genesis,
classification,and mappingof soils in urban
areas. UrbanEcosyst. 1:217-28
Field DR, BurchWR Jr.1988.RuralSociology
and the Environment.Middleton, WI: Soc.
Ecol.
Flores A, Pickett STA, ZippererWC, Pouyat
RV, Pirani R. 1997. Adopting a modem
ecological view of the metropolitanlandscape: the case of a greenspace system for
the New York City region. Landsc. Urban
Plan. 39:295-308
Frey J. 1998. Comprehensivebiotope mapping
in the city of Mainz-a tool for integratednatureconservationandsustainableurbanplanning. See Breusteet al. 1998, pp. 641-47
GarreauJ. 1991. Edge City: Life on the New
Frontier.New York:Doubleday
GilbertOL. 1989. TheEcology of UrbanHabitats. New York/London:Chapman& Hall
GoldmanMB, GroffmanPM, PouyatRV, McDonnell MJ, Pickett STA. 1995. CH4 uptake and N availabilityin forest soils along
anurbanto ruralgradient.SoilBiol. Biochem.
27:281-86
Golley FB. 1993. A History of the Ecosystem
Concept in Ecology: More than the Sum of
the Parts. New Haven:Yale Univ. Press
GreenwoodEF, ed. 1999. Ecology and Landscape Development:a History of the Mersey
Basin. Liverpool:LiverpoolUniv. Press
GrimmNB, GroveJM, PickettSTA, Redman
CL. 2000. Integrated approaches to longterm studies of urban ecological systems.
BioScience 50:571-84
GroffmanPM, Likens GE, eds. 1994. Integrated Regional Models: Interactions BetweenHumansand TheirEnvironment.New
York:Chapman& Hall
GroffmanPM, Pouyat RV, McDonnell MJ,
Pickett STA, ZippererWC. 1995. Carbon
pools and trace gas fluxes in urban forest soils. In Soil Management and Greenhouse Effect, ed. R Lal, J Kimble, E Levine,
BA Stewart,pp. 147-58. Boca Raton: CRC
Lewis
153
Grove JM. 1995. Excuse me, could I speak to
the propertyownerplease?Comm.Prop.Resour.Dig. 35:7-8
GroveJM. 1996. The relationshipbetweenpatterns and piocesses of social stratification
and vegetationof an urban-ruralwatershed.
PhD dissertation.YaleUniv.,New Haven.pp.
109
GroveJM. 1997. New tools for exploringtheory
and methods in humanecosystem and landscape analyses: computermodeling, remote
sensingandgeographicinformationsystems.
InIntegratingSocial Sciences andEcosystem
Management,ed. HK Cordell,JCBergstrom.
Champaign:Sagamore
Grove JM, Burch WR Jr. 1997. A social ecology approach and application of urban
ecosystem and landscape analyses: a case
study of Baltimore, Maryland.Urban Ecosyst. 1:259-75
GroveJM, HohmannM. 1992. GIS and social
forestry.J. For.90:10-15
Grusky DB, ed. 1994. Social Stratification:
Class, Race, and Genderin Sociological Perspective. Boulder:Westview
GuestAM. 1977. Residentialsegregationin urban areas.In ContemporaryTopicsin Urban
Sociology, ed. KP Schwirian,pp. 269-336.
Morristown:Gen. Learn.
Guilden JM, Smith JR, Thompson L. 1990.
Standstructureof anold-growthuplandhardwood forest in OvertonPark,Memphis,Tennessee. In Ecosystem Management. Rare
Species and Significant Habitats, ed. RS
Mitchell, CJ Sheviak, DJ Leopold, pp. 6166. Albany:New YorkState Mus.
Haspel C, CalhoonRE. 1991. Ecology andbehaviorof free-rangingcats in Brooklyn,New
York.See Adams & Leedy 1991, pp. 27-30
Hawley D. 1950. HumanEcology: A Theoryof
CommunityStructure.New York:Ronald
HobbsER. 1988. Species richnessin urbanforest patches and implicationsfor urbanlandscape diversity.Landsc. Ecol. 1:141-52
Hough M. 1995. Cities and NaturalProcesses.
London:Routledge
HuntingtonE. 1920. The controlof pneumonia
an influenzaby weather.Ecology 1:1-23
154
PICKETTET AL.
Iida S, NakashizukaT. 1995. Forest fragmentation and its effect on species diversity in
suburbancoppice forests in Japan.For.Ecol.
Manage. 73:197-210
Jones CG, Lawton JH, eds. 1995. Linking
Species and Ecosystems. New York: Chapman & Hall
Katz B, Bradley J. 1999. Divided we sprawl.
Atl. Mon. 284:26-42
KlausnitzerB, Richter K. 1983. Presence of
an urbangradientdemonstratedfor carabid
associations.Oecologia 59:79-82
Klotz S. 1990. Species/area and species/
inhabitantsrelationsin Europeancities. See
Sukopp 1990a, pp. 100-12
KowarikI. 1990. Some responses of flora and
vegetationto urbanizationin CentralEurope.
See Sukopp 1990a, pp. 45-74
LawrenceHW. 1995. Changingforms andpersistent values: historicalperspectiveson the
urban forest. See Bradley 1995, pp. 1740
Lenski GE. 1966. Power and Privilege: a
Theory of Social Stratification.New York:
McGraw-Hill
LikensGE. 1992. Excellencein Ecology, Vol.3.
TheEcosystemAppioach:Its Use andAbuse.
Oldendorf/Luhe:Ecol. Inst.
LoganJR, MolotchHL. 1987. UrbanFortunes:
the Political Economy of Place. Berkeley:
Univ. Calif. Press
LovettGM, TraynorMM, PouyatRV, Carreiro
MM, Zhu W, Baxter J. 2000. Nitrogen deposition along an urban-ruralgradientin the
New York City metropolitanarea. Environ.
Sci. Technol.34:4294-300
LuniakM, PisarskiB. 1994. State of research
into the faunaof Warsaw(up to 1990). Mem.
Zool. 49:155-65
MachlisGE, ForceJE, BurchWRJr.1997. The
human ecosystem partI: the humanecosystem as an organizing concept in ecosystem management. Soc. Nat. Res. 10:34767
Mackin-RogalskaR, Pinowski J, Solon J,
WojickZ. 1988. Changesin vegetation,avifauna, and small mammals in a suburban
habitat.Pol. Ecol. Stud. 14:239-330
Makse HA, Havlin S, Stanley HE. 1995. Modelling urbangrowth patterns.Nature 377:
608-12
MatlackGR. 1997. Landuse and forest habitat
distributionin the hinterlandof a large city.
J. Biogeogr.24:297-307
MatthewsMJ, O'Connor S, Cole RS. 1988.
Database for the New York State urban
wildlife habitat inventory. Landsc. Urban
Plan. 15:23-37
McDonnell MJ, Pickett STA, eds. 1993. Humans as Components of Ecosystems: the
Ecology of Subtle Human Effects and PopulatedAreas. New York:Springer-Verlag
McDonnell MJ, Pickett STA, Groffman P,
Bohlen P, PouyatRV et al. 1997. Ecosystem
processes along an urban-to-ruralgradient.
UrbanEcosyst. 1:21-36
McDonnell MJ, Pickett STA, Pouyat RV.
1993. The applicationof the ecological gradient paradigmto the study of urbaneffects.
See McDonnell & Pickett 1993, pp. 17589
McHargI. 1969. Design with Nature. Garden
City, NJ: Doubleday/Nat.Hist.
McPhersonEG, Nowak D, Heisler G, GrimmondS, Souch C et al. 1997. Quantifyingurban forest sturcture,function,and value: the
Chicago urbanforest climate project.Urban
Ecosyst. 1:49-61
Medley KE, McDonnell MJ, Pickett STA.
1995. Humaninfluenceson forest-landscape
structurealong an urban-to-ruralgradient.
Prof. Geogr.47:159-68
Moran MA. 1984. Influence of adjacentland
use on understoryvegetation of New York
forests. UrbanEcol. 8:329-40
MorrillRL. 1974. The Spatial Organizationof
Society. Duxbury,MA: Duxbury
Mucina L. 1990. Urban vegetation research
in European COMECON countries and
Yougoslavia: a review. See Sukopp 1990a,
pp. 23-43
Nilon CH, Pais RC. 1997. Terrestrialvertebrates in urbanecosystems: developing hypotheses for the Gwynns Falls Watershedin
Baltimore,Maryland.UrbanEcosyst. 1:24757
TERRESTRIALURBAN ECOLOGY
Nowak DJ. 1993. Atmosphericcarbonreduction by urban trees. J. Environ. Manage.
37:207-17
Nowak DJ. 1994. Urban forest structure:the
stateof Chicago's urbanforest. In Chicago's
Urban Forest Ecosystem: Results of the
Chicago Urban Forest Climate Project, ed.
EG McPherson,D Nowak,RA Rowntree,pp.
140-64. Radnor,PA: USDA ForestServ.
Nowak DJ, McBride JR. 1992. Differences in
Montereypine pest populationsin urbanand
naturalforests. For;Ecol. Manage. 50:13344
Nowak DJ, RowntreeR, McPhersonEG, Sisinni SM, KerkmannE, Stevens JC. 1996.
Measuring and analyzing urbantree cover.
Landsc. UrbanPlan. 36:49-57
Noyes JH, Progulske DR. 1974. Wildlife in
an UrbanizingEnvironment.Amherst:Mass.
Coop. Ext. Serv., Univ. Mass.
Odum EP. 1997. Ecology: a Bridge BetweenScience and Society. Sunderland,MA:
Sinauer
OdumHT, OdumEC. 1980. EnergyBasis for
Man and Nature.New York:McGraw-Hill
Oke TR. 1973. City size and urbanheat island.
Atmos.Environ.7:769-79
Oke TR. 1995. The heat island of the urban
boundarylayer:charcteristics,causes andeffects. In WindClimate in Cities, ed. JE Cermak,pp.81-107. Netherlands:KluwerAcad.
ParkerJK, BurchWR Jr.1992. Towarda social
ecology for agroforestryin Asia. In Social
Science Applicationsin Asian Agroforestiy,
ed. WR BurchJr,JK Parker,pp. 60-84. New
Delhi: IBH
PaulMJ, MeyerJL.2001. Riverineecosystems
in an urbanlandscape.Annu.Rev.Ecol. Syst.
32: In press
PickettSTA, BurchWR Jr, Dalton SD, Foresman TW. 1997. Integratedurbanecosystem
research.UrbanEcosyst. 1:183-84
PickettSTA, CadenassoML, Jones CG. 2000.
Generationof heterogeneity by organisms:
creation, maintenance, and transformation.
In Ecological Consequencesof HabitatHeterogeneity,ed. M Hutchings,L John,A Stewart,pp. 33-52. New York:Blackwell
155
Pickett STA, Kolasa J, Jones CG. 1994. Ecological Understanding:TheNatureof Theory
and the Theoryof Nature. San Diego: Academic
PickettSTA, ParkerVT, FiedlerPL. 1992. The
new paradigmin ecology: implications for
conservationbiology abovethe species level.
In ConservationBiology: The Theoty and
Practice of Nature Conservation,Preservation, and Management,ed. PL Fiedler, SK
Jain,pp. 65-88. New York:Chapman& Hall
PouyatRV, McDonnellMJ. 1991. Heavy metal
accumulationin forest soils along an urban
to ruralgradientin southernNY, USA. Water
Soil Air Pollut. 57/58:797-807
PouyatRV, McDonnellMJ, PickettSTA. 1997.
Litter decomposition and nitrogen mineralization along an urban-ruralland use gradient. UrbanEcosyst. 1:117-31
Pouyat RV, McDonnell MJ, Pickett STA,
Groffman PM, CarreiroMM et al. 1995.
Carbonand nitrogendynamicsin oak stands
along an urban-rural gradient. In Carbon Formsand Functions in ForestSoils, ed.
JM Kelly, WW McFee, pp. 569-87. Madison, WI: Soil Sci. SocAm.
PouyatRV, ParmeleeRW, CarreiroMM. 1994.
Environmental effects of forest soil-invertebrateand fungal densities in oak stands
landuse gradient.Pedoalong an urban-rural
biologica 38:385-99
RapoportEH. 1993. The process of plant colonizationin small settlementsandlarge cities.
See McDonnell & Pickett 1993, pp. 190207
Rebele F. 1994. Urbanecology and special featuresof urbanecosystems. GlobalEcol. Biogeogr.Lett. 4:173-87
Rees WE. 1996. Revisiting carrying capacity: area-basedindicators of sustainability.
Popul. Environ.17:195-215
Rogers GF, RowntreeRA. 1988. Intensivesurveys of structureandchangein urbannatural
areas.Landsc. UrbanPlan. 15:59-78
Rowntree RA. 1995. Towardecosystem management:shifts in the core andthe contextof
urbanforest ecology. See Bradley 1995, pp.
43-59
156
PICKETTET AL.
Rudnicky JL, McDonnell MJ. 1989. Fortyeight years of canopychangein a hardwoodhemlock forest in New YorkCity.Bull. TorreyBot. Club 116:52-64
SalisburyEJ. 1943. The flora of bombed areas
Proc. R. Inst. G. B. 32:435-55
Schlutink G. 1992. Integrated remote sensing, spatial information systems, and applied models in resource assessment, economic development, and policy analysis.
Photogramm.Eng. Remote Sens. 58:122937.
Schmid JA. 1975. Urban Vegetation:a Review and Chicago Case Study. Chicago:
Univ.ChicagoDep. Geogr.
Sissini SM, EmmerichA. 1995. Methodologies, results and applicationsof naturalresource assessment in New York City. Nat.
Areas J. 15:175-88
Spim AW. 1984. The Granite Garden: Urban
NatureandHumanDesign. New York:Basic
Books
SprinAW. 1998. The Language of Landscape.
New Haven:Yale Univ. Press
Steams FW. 1971. Urban botany-an essay
on survival. Univ. Wis.Field Stn. Bull. 4:16
Steinberg DA, Pouyat RV, Parmelee RW,
GroffmanPM. 1997. Earthwormabundance
and nitrogen mineralizationrates along an
urban-ruralland use gradient. Soil Biol.
Biochem. 29:427-30
Sukopp H. 1998. Urban ecology-scientific
andpracticalaspects.See Breusteet al. 1998,
pp. 3-16
Sukopp H, ed. 1990a. Urban Ecology: Plants
and Plant Communitiesin Urban Environments.The Hague: SPB Acad.
Sukopp H. 1990b. Urban ecology and its application in Europe. See Sukopp 1990a, pp.
1-22
Sukopp H, Weiler S. 1988. Biotope mapping
and nature conservationstrategiesin urban
areas of the Federal Republic of Germany.
Landsc. UrbanPlan. 15:39-58
TansleyAG. 1935. The use andabuseof vegetational concepts and terms.Ecology 16:284307
United Nations. 1993. WorldPopulation Prospects. New York:United Nations
van den Berghe PL. 1975. Man in Society: a
Biosocial View.New York:Elsevier
VanDruffLW, Bolen EG, San JulianGJ. 1994.
Managementof urbanwildlife. In Research
and Management Techniquesfor Wildlife
and Habitats,ed. TA Bookhout,pp. 507-30.
Bethesda,MD: Wildl. Soc.
VanDruffLW, Rowse RN. 1986. Habitatassociation of mammalsin Syracuse,New York.
UrbanEcol. 9:413-34
VelguthPH, White DB. 1998. Documentation
of genetic differences in a volunteer grass,
Poa annua(annualmeadowgrass),underdifferentconditionsof golf course turf,and implicationsfor urbanlandscapeplantselection
andmanagement.See Breusteet al. 1998, pp.
613-17
WaggonerPE, Ovington JD, eds. 1962. Proceedings of theLockwoodConferenceon the
SuburbanForest and Ecology. New Haven:
Conn. Agric. Exp. Stn.
WernerP. 1999. Why biotope mappingin populatedareas?Deinsea 5:9-26
WettererJK. 1997. UrbanEcology.Encyclopedia of EnvironmentalSciences. New York:
Chapman& Hall
White CS, McDonnell MJ. 1988. Nitrogen
cycling processes and soil characteristicsin
an urbanversusruralforest.Biogeochemistry
5:243-62
Witt K. 1996. Species/plot turnoversfrom reapeated atlas mapping of breeding birds in
southernBerlin 1980 and 1990. Acta Orn.
31:81-84
Wittig R. 1998. Urban development and the
integrationof nature:reality or fiction? See
Breuste et al. 1998, pp. 593-99
WrongDH. 1988. Power: its Forms,Bases, and
Uses. Chicago:Univ. Chicago Press
Wu L, Antonovics J. 1975a. Zinc and copper uptake by Agrostis stolonifera tolerant
to zinc and copper. New Phytol. 75:23137
Wu L, AntonovicsJ. 1975b.Experimentalecological genetics in Plantago. II. Lead tolerance in Plantago lanceolata and Cynodon
TERRESTRIALURBAN ECOLOGY
dactylon from a roadside.Ecology 57:2058
ZimmererKS. 1994. Humangeographyandthe
"new ecology:" the prospectandpromise of
integration.Ann.Assoc. Am. Geogr;84:10825
Zimny H. 1990. Ecology of urbanizedsystemsproblems and researchin Poland. In Urban
Ecological Studies in Central and Eastern
157
Europe, ed. M Luniak, pp. 8-18. Warsaw:
Pol. Acad. Sci. Inst. Zool.
Zipperer WC, Foresman TW, Sisinni SM,
Pouyat RV. 1997. Urbantree cover: an ecological perspective. Urban Ecosyst. 1:22947
Zonneveld IS. 1989. The land unit-a fundamental concept in landscapeecology and its
applications.Landsc.Ecol. 3:67-89