evaluating wetlands within an urban context · ecological engineering 15 (2000) 253–265...

Ecological Engineering 15 (2000) 253–265

Evaluating wetlands within an urban context

Joan G. Ehrenfeld *Department of Ecology, E6olution, and Natural Resources, Cook College, 14 College Farm Road, Rutgers Uni6ersity,

New Brunswick, NJ 08901, USA

Received 25 November 1998; accepted 10 March 2000

Abstract

Coastal regions are among the most rapidly urbanizing places on earth. The numerous effects of urbanization onhydrology, geomorphology, and ecology make wetlands in urban regions function differently from wetlands innon-urban lands. Furthermore, wetlands in urban regions may take on human-related values that they lack innon-urban areas, as they provide some contact with nature, and some opportunities for recreations that are otherwiserare in the urban landscape. Evaluations of the success of restorations in urban regions require criteria first todetermine the kinds, and intensities of urban influence on the site, and secondly to assess functional performance. Thedevelopment of success criteria, at both the levels of assessment, depends on the proper definition of a referencedomain (the set of wetlands to which success criteria will apply), and the documentation of a set of reference siteswithin the domain; both must be based within the urban context appropriate for the region of interest. An exampleis presented from a study of urban wetlands in New Jersey of a procedure for establishing the reference domain, thereference set of wetlands, and criteria for the assessment of urban influence. © 2000 Elsevier Science B.V. All rightsreserved.

Keywords: Cities; Urban wetlands; Urban ecology; Coastal wetlands; Restoration; Wetland assessment; Functional assessment

www.elsevier.com/locate/ecoleng

1. Introduction

The growth of urban, and suburban areas hasbeen a dominant demographic characteristic ofthe 20th century. During this time urban popula-tion has increased ten-fold, and the proportion ofthe human population living in urban areas hasrisen from 14 to over 50% (Platt, 1994). Much ofthis expansion of urban land, and citizenry hasoccurred along coasts, as port cities have ex-

panded, coalesced, and engulfed neighboring un-developed lands. Between 1960 and 1990, coastalcounties in the US increased in population by43%, a faster rate of growth than in the countryas a whole. Likewise, between 1970 and 1989,nearly half of all building activities took placealong the coasts (Anon., 1994). As of 1981, 28%of municipal areas were coastal, but they ac-counted for 55% of the US population (Walker,1990). Elsewhere in the world, the story is similar:of cities with populations over 1 million, 100% ofthose in South America are coastal, as are 75% ofthose in Asia and Africa (Berry, 1990).

* Tel.: +1-732-9321081; fax: +1-732-9328746.E-mail address: [email protected] (J.G. Ehrenfeld).

0925-8574/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.

PII: S 0925 -8574 (00 )00080 -X

J.G. Ehrenfeld / Ecological Engineering 15 (2000) 253–265254

Not surprisingly, the effects of this burgeoningcoastal development on natural resources havebeen profound (Walker, 1990; Nordstrom, 1994).Damage to and loss of wetlands have been exten-sive (Tiner, 1984; Dahl and Johnson, 1991). Arecent survey by the US Department of Agricul-ture found that urbanization was implicated inwetland loss in nearly all surveyed watersheds(96%) and may account for as much as 58% of thetotal wetland loss (Anon., 1997; Opheim, 1997).Yet wetlands remain an integral part of the devel-oped landscape, particularly along coasts, wheresalt marshes surround oil tanks and abut summerhomes, and riparian forests persist along streamsas they meander through dense development andindustrial campuses (see, for example, Greiling,1993).

Restoration of coastal wetlands often must takeplace within urban regions. In order to evaluatethe success of such restorations, it is necessary toestablish criteria that reflect both the ecologicalqualities of the wetland and the realities of theurban context. Below, I discuss several ways inwhich the context of urban environments affectsthe process of determining success criteria inrestoration. These comments are based on a con-sideration of the nature of urban environments,supplemented by the observations of urban/subur-ban wetlands in New Jersey.

2. Adapting success criteria to cities

While the human dimension in restoration hasbeen stressed repeatedly, (Jordan, 1994), its im-portance is best expressed in urban regions.Restoration projects necessarily affect many peo-ple in their daily lives by altering the places inwhich they walk the dog, commute to work, do abit of fishing. Such places often constitute theonly ‘natural’ habitats that urban residents experi-ence. The multiple aesthetic, emotional, and prac-tical values that people ascribe to trees have beenably described by Dwyer et al. (1994), many ofthe same values accrue to natural landscapes.Since the commitment and involvement of localcitizens are critical to the success of a project,these values must be taken into account in urban

restoration more explicitly and more fully than innon-urban projects.

This point is well illustrated by recent contro-versies surrounding ecological restoration inChicago (Shore, 1997; Gobster, 1997). A long-term collaborative effort by many agencies andorganizations sought to restore prairie and oaksavanna. The effort provoked harsh public oppo-sition when some newspapers and local residentsrealized that re-establishment of these ecosystemsinvolved removing large numbers of trees. Confl-icting values — the amenities provided by treesversus the re-creation of the pre-settlement ‘natu-ral’ landscape, local determination of the fate ofland versus the science-based wisdom of profes-sionals, the use of public funds for restorationversus acquisition of land for conservation hasthrown this extensive urban restoration projectinto confusion. Its goals and success criteria weredeveloped primarily, if not exclusively, in terms ofthe ecological characteristics of native prairie andoak savanna. But the local opposition suggeststhat the goals of a restoration project in a denselypopulated area must take into account factorsunique to that area, including the historical andcontemporary landscape to which the local citi-zens are accustomed. Urban restoration projectsin Florida and New York City have generatedsimilar controversies (Shore, 1997).

Urban environments place a number of con-straints on evaluating the success of restoration inurban regions (Table 1). These include the consid-erable differences in the physical conditions thatlimit the kinds of restoration that are possible andmay preclude some kinds of ecosystem from beingthe goal of a restoration program. For example,intense disturbance events such as major floodsand hot forest fires obviously cannot be toleratedin densely settled areas; ecosystems that dependon such disturbances may thus be impossible torestore. Air pollution and storm-water runoff in-troduce high concentrations of nitrogen; speciesand communities characteristic of very nutrient-poor habitats would be unrealistic targets forrestoration in N-enriched sites. Species near thelimits of their geographic ranges may be intoler-ant of the altered climate within cities. Thus, thecultural and physical environment in cities places

J.G. Ehrenfeld / Ecological Engineering 15 (2000) 253–265 255

some a priori constraints on the development ofgoals and success criteria for restoration.

Criteria for the evaluation of success usuallybased on qualitative or semi-quantitative expres-sions of wetland functions. Brinson et al. (1995),for example, have shown how simple equationscan be derived to evaluate a variety of biologicaland physical functions such as wildlife support,nutrient cycling, flood storage, or biodiversity,from readily observable and measurable proper-ties. Zedler (1996) similarly describes assessmentmethods that depend on measurements of indica-tors of identified functions in California tidalmarshes. Many other assessment methodologieshave been proposed and utilized (e.g. Hruby et

al., 1995; Weinstein et al., 1997); all are based onthe assumptions that separate wetland functionscan be identified and evaluated relative to mea-sured performance in unaltered reference sites,and that indicator variables reliably reflect therate or magnitude of these functions. These ap-proaches to determining restoration success makesense for non-urban areas, and they are alsouseful for regulatory situations in which the re-placement of specific functions is required by apermit.

In urban areas, however, a two-tiered approachto evaluation of restoration success may be moreappropriate. Appraisal of functional capacity andreplacement is clearly one necessary facet of eval-uation, but this level of appraisal should takeplace within a framework of social expectations,wetland capacities, needs for active management,and values unique to the particular urban context.This enveloping framework affects the definitionof the reference domain, the identification of ref-erence sites, and the establishment of indicatorvariables for function and their expected ranges ofvalues.

3. The establishment of ‘reference domains’specific to urban wetlands

The development of success criteria dependsfundamentally on the establishment of the refer-ence domain. This concept (developed by Kentulaet al., 1992; Brinson, 1993; Rheinhardt et al.,1997) refers to the identification of those at-tributes that define a particular class of wetland.A wetland class is usually defined by the plantcommunity composition in conjunction with thehydrogeomorphic setting; for example, a referencedomain might be defined as wetlands occurring indepressions, with mineral soils, a canopy of hard-wood trees, and a dense understory of woodyshrubs. Criteria for evaluating the condition orfunctional properties of a given wetland are thendeveloped, which are specific for this class and arebased on analyses of a reference set of sites thatare deemed to be most representative of the ‘best’ecological state currently possible for this class.Various objective and subjective criteria can be

Table 1Constraints on the specification of success criteria in naturalvs. urban environments

Natural Urban

Watershed-based approach is Municipality-basedideal approach is often necessary

Ecological characteristics and Ecological functions may befunctions are readily less important than humanidentified and are primary values, which may be

difficult to specifyNatural disturbance regimesNatural disturbance regimes

are critical may be impossible torestore

Restoration work is Volunteers are extensivelyimplemented by involvedprofessionals orconsultants, possiblysupplemented byvolunteers

Nutrient limitations are the Nutrients are often presentin abundant ornormover-abundant amounts,and cannot be reducedHabitat patches are oftenHabitat patches can vary

greatly in size and small and isolated;connections are difficult orconnectednessimpossible to re-establish

Climate and microclimate Climate and microclimatereflect regional geography are significantly altered

from the geographicallybased expectations

Hydrology is a function of Hydrology is usually highlyregional climate, geology, altered, in amounts,

sources, and flow rates ofphysiographywater


Table 2Likely effects of urbanization on wetland hydrology and ge-omorphology

HydrologyDecreased surface storage of stormwater results in

increased surface runoff (= increased surface water inputto wetland)

Increased stormwater discharge relative to baseflowdischarge results in increased erosive force within streamchannels, which results in increased sediment inputs torecipient coastal systems

Changes occur in water quality (increased turbidity,increased nutrients, metals, organic pollutants, decreasedO2, etc)

Culverts, outfalls, etc. replace low-order streams; thisresults in more variable baseflow and low-flowconditions

Decreased groundwater recharge results in decreasedgroundwater flow, which reduces baseflow and mayeliminate dry-season streamflow

Increased flood frequency and magnitude result in morescour of wetland surface, physical disturbance ofvegetation

Increase in range of flow rates (low flows are diminished;high flows are augmented) may deprive wetlands ofwater during dry weather

Greater regulation of flows decreases magnitude of springflush

GeomorphologyDecreased sinuousity of wetland/upland edge reduces

amount of ecotone habitatDecreased sinuosity of stream and river channels results in

increased velocity of stream water discharge to receivingwetlands

Alterations in shape of slopes (e.g. convexity) affectswater-gathering or water-disseminating properties

Increased cross-sectional area of stream channels (due toerosional effects of increased flood peak flow) increaseserosion along banks

The urban environment is physically and bio-logically different from the non-urban environ-ment (Gilbert, 1989; Adams and Dove, 1989;Platt, 1994) Tables 2 and 3 summarize some ofthe salient differences. The hydrology of regions ischanged by urbanization, as Leopold (1968) firstdocumented. All these hydrological changes haveindirect effect on wetland structure and function(Table 2). In addition, direct hydrological changesin wetlands commonly occur by filling, ditching,diking, draining, and damming. For example,90% of a sample of suburban cedar swamps inNew Jersey had been so modified (Ehrenfeld andSchneider, 1991). Physical changes to the shape ofthe land, wrought by massive land movementassociated with road and building construction,also cause geomorphological changes both in thewetlands and adjacent areas (Table 2).

Climate and air quality are also altered byurbanization. In addition to increased concentra-tions of oxidants (O3, SO2) and nutrients (nitrates,cations in dust), net radiation and average windspeed decrease, cloudiness and precipitation in-crease, and temperature rises by 1–3°C (‘heat

Table 3Likely effects of urbanization on wetland ecology.

VegetationLarge numbers of exotic species present; large and

numerous sources for continuous re-invasion of exoticsLarge amounts of land with recently disturbed soils

suitable for weedy, invasive speciesDepauperate species poolRestricted pool of pollinators and fruit dispersersChemical changes and physical impediments to growth

associated with the presence of trashSmall remnant patches of habitat not connected to other

natural vegetationHuman-enhanced dispersal of some speciesTrampling along wetland edges and periodically unflooded

areas

FaunaSpecies with small home ranges, high reproductive rates,

high dispersal rates favoredLarge predators virtually non-existent‘Edge’ species favored over forest-interior speciesAbsence of upland habitat adjacent to wetlandsAbsence of wetland/upland ecotonesHuman presence disruptive of normal behaviors

used to identify the reference sites; these includehistorical accounts of the ecosystem of interest byearly naturalists, existing ecological data, criteriaderived from ecological theories such as islandbiogeography, and the judgment of ecologistswith extensive experience in the region. Thus, thedefinition of a reference domain circumscribes thechoice of sites that can be used to set boundaryconditions for indicating ecosystem status, andalso the choice of variables used to indexfunction.


island’ effect), (Berry, 1990). These changes implythat urban sites cannot be directly compared withnon-urban sites, any more than sites in differentclimatic zones.

Biological and ecological changes accompanythe physical effects of urbanization (Table 3). Thepool of species available to form communities,dispersal ability, and mutualistic interactions (e.g.pollination, mycorrhizae) may all be limited oraltered. Rare or unusual types of microhabitats,especially forest or wetland interiors, may be verylimited or non-existent in extent. Both animalbehavior and plant reproductive ecology may bestrongly affected by the size, shape and hetero-geneity of habitat patches. Thus, the possiblestates for the flora and fauna in urban areas arequalitatively different from those of non-urbanareas.

Physical and biological changes in urban areascan be both large in scope and permanent, whichin turn may constrain the kinds of communitiesthat can be restored and the level of successattainable. A good example is the HackensackMeadowlands in northern New Jersey. It occupiesan area of about 7000 ha in close proximity toNew York City, and is currently brackish tosaline marsh. It is traversed by numerous largehighways and railroads, has many closed andactive landfills, contains a large sports stadiumcomplex, hotel/condominium developments andmega-stores, and is surrounded by industry. Asrecently as 1896, the Meadowlands supportedfreshwater forested wetlands, including Atlanticwhite-cedar (Chamaecyparis thyoides) swamps(Heusser, 1963; Sipple, 1972). Mixed with thesetrees was a rich flora of bog species. In 1819, thenoted botanist John Torrey said, ‘‘Few placeshave afforded us more plants, than in the vicinityof Hoboken and Weehawk, and the neighboringmarshes’’ (Torrey, 1819). Freshwater marshes,supporting large stands of wild rice (Zizaniaaquatica), were also present. Extensive ditchingand diking occurred early in the 20th century,which, combined with the rising sea level and theimpoundment of freshwater in several large reser-voirs upstream, converted the area to brackishand saline marsh. The freshwater marshes weretaken over by Phragmites australis and other salt

and brackish marsh species by 1919; today purePhragmites stands fill the basin. Although, therehave been discussions and some attempts to re-store freshwater cedar swamps in the region(Schmid, 1987), most of the current restorationwork aims to establish Phragmites-free saltmarsh; freshwater wetlands are no longer physi-cally realistic. In an analogous way, cottonwoods(Populus spp.) cannot be restored on floodplainsof western cities because permanent hydrologicalchanges associated with urbanization preventmovement of river channels, and therefore, thecreation of suitable habitat for tree seedlings(Auble et al. 1997).

These considerations lead to the conclusion thaturban wetlands need to be defined as referencedomains separate from non-urban wetlands. Fur-thermore, local reference sites need to be estab-lished within each urban region as the basis forestablishing criteria and standards for judgingrestoration success. For example, the structureand function of coastal marshes within large portcities may be very different from those of marshesoccurring on coasts intensively developed forsummer recreation.

4. Criteria for defining reference domains andselecting reference sites

Reference sites for evaluating restoration suc-cess can be chosen based on criteria that reflectthe factors discussed above (Table 4). I illustratethe process by describing how a reference domainand reference set of sites has been developed fornortheastern New Jersey, an extensive urban/sub-urban region.

Clearly, the first step in defining a referencedomain should be the delineation of urban land-use (Table 4). In northern New Jersey, the urban/suburban core is easily definable by the density ofmajor highways, municipalities with populationsizes \10 000, and a known long history ofdense settlement. Surrounding this core is a bandof rapidly developing suburbia which fringes In-terstate 287 (Fig. 1). This road also delimits majorphysiographic regions. The northeast/southwestleg of the road (Fig. 1) follows the boundary


Table 4Proposed procedure and criteria for identifying reference sites within an urban context

Definition of reference domain, based on a combination of municipal or demographic criteria and hydrogeomorphic1criteria

2 Survey of wetland resources (given available databases) to identify appropriate wetland classes (e.g. vegetation types) andto locate all the possible sites of each class within the domainCharacterization of each site with respect to size (area), proximity to urban land use, likelihood of hydrological3modifications (either on-site or off-site), evidence of physical disturbance (e.g. bike trails, large amounts of trash)Stratification of the study region to ensure representative sampling4Selection of the largest unfragmented wetland in each subsection of the region that is surrounded by urban land-use but5that has the least evidence of direct disturbance or hydrological alterations.

between the Piedmont physiographic province(east) and the Highlands Province (west), whilethe east/west leg of the road separates the moreurban northern portion of the state from the lessdensely settled areas south of the Raritan River. Itwas assumed that the hydrogeomorphic settingsof wetlands would be similar within each physio-graphic province, so the interstate highway wasused as a convenient delimiter of the referencedomain.

The delineation of an urban reference domainnecessarily involves arbitrary decisions aboutboundaries between urban and suburban, andbetween suburban and rural areas. Urban andsuburban land-use form a continuum, becausemuch urban land area is occupied by dense-to-moderate residential housing. In New Jersey, thiscontinuum extends from high densities of housingwithin the oldest developed areas, along the Hud-son River and the Arthur Kill, to low densities inthe wealthy suburbs in the central and westernparts of the study area (Fig. 1). Formerly agricul-tural land, within the matrix of roads and towns,has filled with housing, commercial, and office-park development during the past three decades.Thus, there are no clear physical boundaries be-tween suburban lands and rural lands, or betweenthe urban cores of the larger cities and suburbia.However, the region west of Interstate 287 hasonly been subjected to intense development pres-sure in the past few years, so the highway forms aconvenient if somewhat arbitrary boundary be-tween older suburbs and rural/newly suburbanland.

The second step in the process (Table 4) shouldbe a survey of the wetland resources within the

defined domain, for which existing maps and in-formation can be used. From the New JerseyDepartment of Environmental Protection Fresh-water Wetland Maps, a database was created byidentifying 20 largest areas of deciduous forestedwetlands (wetland class ‘PFO1’) within each ofthe 80 maps which covered the region. Sites B300 m across were excluded, except in the areaaround the city of Newark, which contained B20sites per map, all very small.

Preliminary inventory of the range of character-istics of the mapped sites can provide a basis forselecting the reference set. Each wetland that wasidentified on the maps was characterized in termsof the number and types of Cowardin wetlandclasses occurring within the contiguous wetlandarea, its location relative to roads, railroad beds,town centers, or other landmarks and urban fea-tures, the occurrence of open water within oradjacent to the site, and the presence of publicopen space (municipal, county or state-ownedlands) within or adjacent to the wetland area. TheNJDEP GIS system was then used to determinethe area and perimeter of each wetland polygon,and the total area of each wetland patch. Two-hundred-sixty-two forested wetland areas werethus identified and characterized.

These data were then used to determine thefrequency distribution of wetland sizes (Fig. 2),the distribution of wetland classes (Fig. 3), andthe number of wetland classes within each site(Fig. 4). In addition, about 120 of the sites werebriefly inspected in the field. These analyses sug-gested that although many classes of wetland werepresent in the region, most sites were dominatedby saturated wetlands (PFO1B), and many con-


Fig. 1. Map of northeastern New Jersey, showing the density of municipalities, roads, and city centers. The urban wetland study areais delimited by Interstate 287.


Fig. 2. Frequency distributions of areas (ha) in (a) the 262wetlands identified in the survey of forested wetlands in thestudy area and (b) the 20 reference sites.

the site, (1) the larger the area of ‘interior’ habitat;(2) the more protected the interior areas would befrom direct human impact (e.g. road runoff andstorm sewer outfalls, roadkill of animals, dumpingof trash, etc.); (3) the more likely that it has notbeen substantially altered in the past, since it is stilla wetland habitat; (4) the greater the amount ofmicrohabitat heterogeneity, and therefore, thelarger the species pool that may be present and (5)the greater the likelihood that animal species withlarge home ranges could be present. This reasoningis similar to that used for other area-based criteriafor restoration planning (Zedler 1996).

The heterogeneity within an urban region in thehistory and intensity of development necessitatesome method of stratification to ensure representa-tive sampling of the range of conditions within thedefined domain. The USGS quadranglesoffer a convenient device for blocking large areas.For northeast New Jersey, 20 quandrangles coverthe region, and one reference site is establishedwithin each quadrangle. This ensures that all por-tions of the area are represented by a site that isthe largest wetland in that block. The inclusion inthe reference set of small sites within the heavilyurbanized areas as well as large sites from more

Fig. 3. Number of sites containing each class of forestedwetland. Classification follows Cowardin (1979); see text fordetails.

tained either seasonally flooded (PFO1C), season-ally saturated (PFO1E), or temporarily flooded(PFO1A) forested wetlands (Fig. 3). These classescould be used as a primary filter to select referencesites by assuming that sites with similar generalhydrology (as mapped) would have broadly similarhydrogeomorphic settings. This assumption will betested with well and piezometer measurements inthe course of characterizing the reference set.

This preliminary work suggested that the culturaland physical settings of the wetlands were ex-tremely heterogeneous, and could not be readilyclassified using a priori criteria. Therefore, in theabsence of detailed information about the biota,hydrology or function of these wetlands, wetlandsize was chosen as the most useful criterion foridentifying the set of reference sites. This decisionwas based in part on the assumptions that the larger


Fig. 4. Frequency distributions of the number of wetlandclasses found in each site, for (a) forested wetland classes and(b) non-forested wetland classes.

levels of assessment: first, an assessment of theextent of urban influence which sets bounds onthe potential for the performance of the conven-tionally defined functions and identifies social val-uation, and second, within that context, amodified assessment for functional capacity.Table 5 presents possible components of an as-sessment protocol for urban influence. It reliesprimarily on the available geographic data todescribe the urban environment and utilizes read-ily observable indicators within the wetland siteitself (e.g. presence of trash, ditches, etc.). Thevariables listed in Table 5 include continuous,quantitative data (e.g. wetland area, distance tonearest other habitat), categorical data (e.g. pres-ence/absence of contiguous wetland or uplandhabitat; presence/absence of ditches, dams,drains), and qualitative data (residents’ stated val-ues; current use). However, an overall index ofurban influence could be constructed from thedata from the reference set of sites. The potentialurban influence for a new site can then be indexedrelative to the range of impact observed for thereference wetlands within the same urban region.

The assessment of ecological functions can fol-low the models developed for non-urban wetlands(e.g. Brinson et al., 1995). The same range offunctions is possible within urban wetlands, withthe same constraints set by overall hydrogeomor-phic setting (Brinson, 1993; Smith et al., 1995).However, the range of values of indicator vari-ables may differ in several ways from their coun-terparts in non-urban settings (Table 6). First, therange of variability, both within and betweenwetlands, is likely to be much higher in urbanareas than non-urban areas. This pattern wasobserved in a range of biotic and environmentaldescriptors of suburban and undisturbed wetlandsin the New Jersey Pinelands (Ehrenfeld andSchneider 1993; Zampella 1994).

Second, indicator values may receive weightingsin an urban context that are different from thosein a non-urban area. For example, isolated wet-lands often support fewer species than sites withina regional complex of wetlands (Weller, 1990).But in an urban area, an isolated patch mayprovide essential habitat in a large region with nonatural habitat. Small wetlands have been shown

suburban areas provides the range of values thatcould be considered characteristic of urban wet-lands in the study region. As Kentula et al. (1992)and many others emphasize, temporal and spatialvariability can be high even among sites initiallythought to be uniform, so it is crucially importantto capture the potential range of variation withinthe reference set of sites.

5. Assessment protocols in an urban context

Assessment protocols are intended to identifypotential or replacement functional capacity. Suchprotocols for urban wetlands, should include two


Tab

le5

Pro

pose

dco

mpo

nent

san

din

dica

tor

vari

able

sfo

ras

sess

ing

the

degr

eeof

urba

nim

pact

onw

etla

nds,

prio

rto

func

tion

alas

sess

men

t

Indi

cato

rva

riab

les

tobe

asse

ssed

Sign

ifica

nce

Cat

egor

y

Are

a,pe

rim

eter

,fr

acta

ldi

men

sion

rank

ing

ofsi

tew

ith

resp

ect

toSi

zeE

colo

gica

lex

pect

atio

ns,

give

npa

rcel

wit

hgi

ven

dim

ensi

ons

refe

renc

ese

tC

onne

cted

ness

/fra

Lik

elih

ood

ofan

imal

and

plan

tdi

sper

sal,

pres

ence

ofa

spec

ies

pool

,P

rese

nce

ofco

ntig

uous

othe

rha

bita

t(w

etla

ndan

d/or

upla

nd)

habi

tat

hete

roge

neit

y;re

gion

alsi

gnifi

canc

eas

avai

labl

eha

bita

tis

land

gmen

tati

ondi

stan

ceto

near

est

othe

rha

bita

t(b

oth

wet

land

and

upla

nd)

Hyd

roge

omor

phic

On-

site

:di

tche

s,da

ms,

dive

rsio

ns,

drai

ns,

fill

off-

site

:pe

rcen

tage

ofL

ikel

ihoo

dth

atth

ehy

drol

ogy

ofth

esi

teha

sbe

enal

tere

d;w

hat

impe

rvio

ussu

rfac

e;gr

ound

wat

erw

ithd

raw

als;

upst

ream

alte

rati

ons

perm

anen

ceof

hydr

olog

ical

chan

ges

impo

undm

ents

and

flow

cont

rol

stru

ctur

es;

chan

ges

inba

sin

drai

nage

dens

ity,

chan

nel

shap

e,st

orm

hydr

ogra

phs

Per

cent

age

ofla

ndin

resi

dent

ial,

com

mer

cial

,in

dust

rial

use

road

Res

iden

tial

:lik

elih

ood

oftr

ampl

ing,

tras

h;co

mm

erci

al:

likel

ihoo

dof

Urb

anse

ttin

gtr

ash;

indu

stri

al:

likel

ihoo

dof

pollu

tant

s;ro

adde

nsit

y:lik

elih

ood

ofde

nsit

ypo

pula

tion

dens

ity

rate

ofpo

pula

tion

grow

thpr

esen

ceof

tras

h,tr

ails

wit

hin

the

wet

land

,pe

rcen

tage

impe

rvio

ussu

rfac

ero

adru

noff

;po

pula

tion

dens

ity

and

grow

thra

te:

inde

xof

urba

nin

tens

ity

and

likel

ihoo

dof

chan

ging

impa

ctin

the

futu

reR

ange

ofcu

rren

tan

dfu

ture

hum

anus

es(p

osit

ive

and

nega

tive

);H

uman

valu

esC

urre

ntan

dpo

tent

ial

futu

rehu

man

uses

ofsi

teva

lues

ascr

ibed

tosi

teby

neig

hbor

ing

resi

dent

sva

lues

ascr

ibed

tosi

teby

prof

essi

onal

pote

ntia

lfo

rco

nflic

tsbe

twee

nlo

cal

resi

dent

san

dpr

ofes

sion

als

reso

urce

man

ager

s,sc

ient

ists


to have high habitat value for some wetland fauna(Gibbs 1993; Semlitsch and Bodie, 1998). Con-versely, species richness of both plant and animalgroups may be higher than in comparable non-ur-ban wetlands, due to the incursion of exotic andweedy species, the presence of microsites thatallow more upland and facultative-upland speciesto become established, and the removal of nutri-ent limitations due to pollutants in both air andwater (Ehrenfeld and Schneider 1993). Weightingof species richness variables must be done inconjunction with an analysis of the species list inorder to determine to what extent the value of thevariable indicates ecological health versusdegradation.

Third, urban wetlands often have unusual phys-ical features which are the result of human inter-vention. Disturbed soil profiles, mounds of soilfrom old ditch-digging or channel-widening activi-ties, heterogeneous materials, etc., are not uncom-mon. Sandy surficial deposits from storm waterrunoff may overlie organic sediments. Provisionneeds to be made in sampling protocols to docu-ment the presence and extent of these features.

Fourth, indicator variables for direct humaninfluence are essential. Measures of the presenceof trash provide useful indicators of human prox-

imity and pathways of human influence. For ex-ample, stream corridors, even small first-ordercreeks, often serve as conduits of both trash andsediments, which are deposited in a narrow beltalong the channel. The presence of a stream corri-dor, can thus, indicate the potential for extendedhuman influence into the interior of a site. Inaddition, trails may become established onslightly higher ground, giving access to the wet-land. Finally, indicators of positive human usesare also needed; the existence of local ‘Save ourSwamp’ organizations are an example.

Functional assessment using a set of referencewetlands is currently accomplished by setting thevalue of the mathematical expression describingeach function equal to 1.0 for the reference site(s),and evaluating the expression for a sample siterelative to that value (Brinson et al., 1995; Rhein-hardt et al., 1997). In urban regions, it may benecessary to modify this approach by subdividingthe reference set into classes, based on the charac-teristics of a particular urban region, and develop-ing separate mathematical expressions for eachfunction within each class. Such flexibility wouldallow differential weighting of variables withineach expression, to take into account the varia-tions within the urban region in question. Forexample, the New Jersey metropolitan region cur-rently being studied contains areas with largewetland complexes resulting from two extensiveglacial lake basins, and other areas outside thelake basins with only small wetland areas. Theregion also contains areas that have been denselysettled since the 1600s, and areas that have be-come densely settled only in the last 20–30 years.These geological and demographic subsets areoverlapping but not completely congruent. Theset of reference sites being developed for the entireregion, described above, will capture this range ofvariation for the region, but may need to besubdivided to establish assessment criteria for thedifferent subregions.

In summary, urban areas may retain large num-bers of wetlands, both as remnants of the naturalenvironment, and as the inadvertent result ofhuman activities (for example, fringing wetlandson dredge spoil deposits). These wetlands con-tinue to provide important ecological services and

Table 6Characteristics of indicator variables and of variable values inurban settings

High variability both within and between sitesIndicator values differently weighted in urban and

non-urban settingsFrequent occurrence of outlier values (both qualitative and

quantitative)Presence of unusual, anthropogenic landforms that have

no counterparts in non-urban settingsPresence of soil profiles that are unlike those in non-urban

wetlandIncreased presence of upland species,

disturbance-associated speciesIncreased dominance of nitrophilous, fertile-site species,

and absence or paucity of poor-site speciesTotal species richness possibly higher than expectedRiverine-influenced areas sometimes different from adjacent

areasIndicators of function include indices of human usage

(presence of trash; observations of people, etc.)


may be of particular importance to both peopleand wildlife because of their remaining presencewithin concrete landscapes. Measures of restora-tion success and functional performance muststart with an appreciation and assessment of theparticular conditions imposed by the urban envi-ronment. These conditions can be identified, mea-sured, and incorporated into assessment protocolsfor individual wetland functions.

Acknowledgements

I acknowledge the support of EPA CooperativeAgreement No. 824358010, and the enthusiasticsupport and intellectually stimulating assistanceof Dr Mary Kentula, project officer. The datareported here were collected with the assistance ofMichael Ciarlo, Tara Bowers and Peter Kourtev.

References

Adams, L.W., Dove L.E., 1989. Wildlife reserves and corri-dors in the urban environment. Natl. Inst. Urban Wildlife,Shepherdstown, WV.

Anon., 1994. World Resources 1994–1995. A Report by theWorld Resources Institute. Oxford University Press, NewYork.

Anon., 1997. Agricultural Resources and Environmental Indi-cators, 1996–1997. US Department of Agriculture, Wash-ington, DC.

Auble, G.T., Scott, M.L., Friedman, J.M., Back, J., Lee, V.J.,1997. Constraints on establishment of plains cottonwoodin an urban riparian preserve. Wetlands 17, 138–148.

Berry, B.J.L. II, 1990. Urbanization. In: Turner, B.L., Clark,W.C., Kates, R.W., Richards, J.F., Matthews, J.T., Meyer,W.B. (Eds.), The Earth as Transformed by Human Action.Cambridge University Press, New York, pp. 103–120.

Brinson, M., 1993. A Hydrogeomorphic Classification forWetlands. Wetlands Research Program TR-WRP-DE-4.US Army Waterways Exp. Station, Vicksburg, MS.

Brinson, M.M., Hauer, F.R., Lee, L.C., Nutter, W.L., Rhein-hardt, R.D., Smith, R.D., Whigham, D., 1995. A guide-book for application of hydrogeomorphic assessments toriverine wetlands. Wetlands Research Program TR-WRP-DE-11. US Army Waterways Exp. Station, Vicksburg,MS.

Dahl, T.E., Johnson, C.E., 1991. Status and Trends of Wet-lands in the Conterminous United States, Mid 1970’s toMid-1980’s. US Dept. Int., Fish and Wildlife Serv., Wash-ington, DC.

Dwyer, J.F., Schroeder, H.W., Gobster, P.H. Jr, 1994. Thedeep significance of urban trees and forests. In: Baldwin,A.D. Jr, DeLuce, J., Pletsch, C. III (Eds.), Beyond Preser-vation: Restoring and Inventing Landscapes. University ofMinnesota Press, Minneapolis, MN, pp. 137–150.

Ehrenfeld, J.G., Schneider, J.P., 1991. Chamaecyparis thyoideswetlands and suburbanization: effects of nonpoint sourcewater pollution on hydrology and plant community struc-ture. J. Appl. Ecol. 28, 467–490.

Ehrenfeld, J., Schneider, J.P., 1993. Responses of forestedwetland vegetation to perturbations of water chemistry andhydrology. Wetlands 13, 122–129.

Gibbs, J.P., 1993. Importance of small wetlands for the persis-tence of populations of wetland-associated animals. Wet-lands 13, 25–31.

Gilbert, O.L., 1989. The Ecology of Urban Habitats. Chap-man & Hall, London.

Gobster, P.H., 1997. The Chicago wilderness and its critics.III. The other side. Rest. Manage. Notes 15, 32–37.

Greiling, D.R., 1993. Greenways to the Arthur Kill: a green-way plan for the Arthur Kill tributaries. NJ ConservationFoundation, Mendham, NJ.

Heusser, 1963. Pollen diagrams from three former cedar bogsin the Hackensack tidal marsh, northeastern New Jersey.Bull. Torrey Bot. Club 90, 16–28.

Hruby, T., Cesanek, W.E., Miller, K.E., 1995. Estimatingrelative wetland values for regional planning. Wetlands 15,93–108.

Jordan, W.R. III, 1994. ‘Sunflower Forest’: ecological restora-tion as the basis for a new environmental paradigm. In:Baldwin, A.D. Jr, DeLuce, J., Pletsch, C. (Eds.), BeyondPreservation: Restoring and Inventing Landscapes. Univer-sity of Minnesota Press, Minneapolis, MN, pp. 17–35.

Kentula, M.E., Brooks, R.P., Gwin, E., Holland, C.C., Sher-man, A.D., Sifneos, J.C., 1992. In: Hairston, A.J. (Ed.),An Approach To Improving Decision Making in WetlandRestoration and Creation. US Environmental ProtectionAgency, Env. Res. Lab, Corvallis, OR.

Leopold, L., 1968. Hydrology for urban land planning — aguidebook on the hydrologic effects of urban land use. USGeol. Survey Circular 554.

Nordstrom, K., 1994. Beaches and dunes of human-alteredcoasts. Prog. Phys. Geog. 18, 497–516.

Opheim, T., 1997. Wetland losses continue but have slowed.Natl. Wetlands Newslett. 19, 7–9.

Platt, R., 1994. The ecological city: introduction and overview.In: Platt, R.H., Rowntree, R.A., Muick, P.C. (Eds.), TheEcological City. University of Massachusetts Press,Amherst, MA, pp. 1–20.

Rheinhardt, R.D., Brinson, M.M., Farley, P.M., 1997. Apply-ing reference wetland data to functional assessment, miti-gation and restoration. Wetlands 17, 195–215.

Schmid, J.A., 1987. Atlantic white cedar in the HackensackMeadowlands, New Jersey: its historic extirpation andfuture reestablishment. In: Laderman, A. (Ed.), AtlanticWhite Cedar Wetlands. Westview Press, Boulder, CO, pp.317–321.


Semlitsch, R.D., Bodie, J.R., 1998. Are small, isolated wet-lands expendable? Conserv. Biol. 12, 1129–1133.

Shore, D., 1997. The Chicago Wilderness and its critics. II.Controversy erupts over restoration in Chicago area. Rest.Manage. Notes 15, 25–31.

Sipple, W., 1972. The past and present flora and vegetation ofthe Hackensack Meadows. Bartonia 4, 4–25.

Smith, R.D., Ammann, A., Bartoldus, C., Brinson, M.M.,1995. Approach for assessing wetland functions using hy-drogeomorphic classification, reference wetlands, and func-tional indices. US Army Corps of Engineers Tech. Rep.,WRP-DE-9.

Tiner Jr., R.W., 1984. Wetlands of the United States: currentstatus and recent trends. US Fish and Wildlife Serv. Natl.Wet. Inv., Newton Corner, MA.

Torrey, J.C., 1819. A Catalogue of Plants Growing Sponta-neously Within Thirty Miles of the City of New York.Websters and Skinners, New York.

Walker, H.J., 1990. The coastal zone. In: Turner, B.L., Clark,

W.C., Kates, R.W., Richards, J.F., Matthews, J.T., Meyer,W.B. (Eds.), The Earth as Transformed by HumanAction. Cambridge University Press, New York, pp. 271–294.

Weinstein, M.P., Balletto, J.H., Teal, J.M., Ludwig, D.F.,1997. Success criteria and adaptive management for alarge-scale wetland restoration project. Wetland Ecol.Manage. 4, 111–127.

Weller, M.W., 1990. Waterfowl management techniques forwetland enhancement, restoration and creation useful inmitigation procedures. In: Kusler, J., Kentula, M. (Eds.),Wetland Creation and Restoration. Island Press, Washing-ton, DC, pp. 517–528.

Zampella, R., 1994. Characterization of surface water qualityalong a watershed disturbance gradient. Wat. Res. Bull. 30,605–611.

Zedler, J., 1996. Tidal wetland restoration: a scientific perspec-tive and Southern California focus. California Sea GrantCollege System, University of California, La Jolla, CA.Report No. T-038.

.

evaluating wetlands within an urban context · ecological engineering 15 (2000) 253–265...

Documents