837
WETLANDS, Vol. 24, No. 4, December 2004, pp. 837?847
q 2004, The Society of Wetland Scientists
INFLUENCE OF LAND USE ON THE INTEGRITY OF MARSH BIRD
COMMUNITIES OF CHESAPEAKE BAY, USA
William V. DeLuca1, Colin E. Studds1, Larry L. Rockwood2, and Peter P. Marra1
1Smithsonian Environmental Research Center
Edgewater, Maryland, USA 21037-0028
E-mail: delucaw@si.edu
2Department of Environmental Science and Policy
George Mason University
Fairfax, Virginia, USA 22030
Abstract: The landscape within the Chesapeake Bay watershed has been and continues to be impacted by
human modifications. Understanding if such anthropogenic disturbances influence organisms that are depen-
dent upon estuarine wetlands remains unclear. We developed an index of marsh bird community integrity
(IMBCI) to evaluate marsh bird communities and wetland condition. During the 2002 and 2003 summers,
we detected 30 bird species at 219 point count locations distributed among 96 wetlands. IMBCI scores for
each wetland were used to determine whether wetland habitat characteristics and urban/suburban develop-
ment, agriculture, and forest at three different spatial scales (watershed, 1000-m buffer, and 500-m buffer)
influenced marsh bird community integrity. We found no relationship between IMBCI scores and wetland
habitat characteristics, implying that marsh bird community integrity is not related to any single plant com-
munity. Nonparametric changepoint analysis indicated that marsh bird community integrity was significantly
reduced when the amount of urban/suburban development within 500 m and 1000 m of the marsh exceeded
14% and 25%, respectively. There was no effect of urban/suburban development on IMBCI scores at the
watershed scale. The results of our study demonstrate that marsh bird community integrity shows a threshold
response to urban/suburban development at local scales. IMBCI scores, combined with the identification of
a land-use threshold, can be easy to interpret and may help communicate complex ecological data to natural
resource managers and conservation planners.
Key Words: threshold response, specialist, generalist, index of marsh bird community integrity, estuarine
wetlands, Chesapeake Bay
INTRODUCTION
Virtually half of the world?s population resides
within 200 km of a coastline (Cohen et al. 1997). Be-
tween 1960 and 1990, the population of U.S. coastal
areas increased by 41%, with the Chesapeake Bay wa-
tershed having the fastest growing population in North
America (Culliton et al. 1990). Accompanying these
population increases are numerous anthropogenic
landscape modifications, including expansion of urban/
suburban development. The cumulative result of these
pressures is that the world?s coastal ecosystems, in-
cluding the Chesapeake Bay estuary, have become de-
graded (Nordstrom and Roman 1996, Edgar et al.
2000).
The increase in anthropogenic stressors in coastal
areas makes it imperative to devise sound methods for
assessing estuarine integrity and provide direction for
future conservation efforts. Methods have already been
developed for measuring the integrity of many com-
ponents of estuarine systems, such as water quality,
sediment quality, and the benthic community (Weis-
berg et al. 1997, Schimmel et al. 1999, Dauer et al.
2000). Methods for assessing estuarine wetland integ-
rity, however, remain elusive largely due to dynamic
environmental conditions, such as fluctuations in tidal
cycles, salinity, and dissolved oxygen levels. Addi-
tionally, marsh ecosystems are particularly challenging
to assess because they are vulnerable to a suite of land-
scape variables originating from surrounding areas
and/or entire watersheds (Pennings et al. 2002).
One approach that could provide insight on estua-
rine wetland condition may be an index capable of
measuring wetland community integrity. Thus far,
most indices have focused on benthic and fish com-
munities in freshwater (Angermeier and Karr 1986,
Weisberg et al. 1997), marine (Dauer et al. 2000), and
wetland ecosystems (Burton et al. 1999). The success
of this technique in aquatic ecosystems has led to its
expansion for use in other ecosystems. For example,
avian indices have now been developed for forests
838 WETLANDS, Volume 24, No. 4, 2004
(O?Connell et al. 1998, Canterbury et al. 2000), ripar-
ian areas (Bryce et al. 2002), and rangelands (Bradford
et al. 1998).
An effective community index should reveal a sig-
nal of integrity and indicate the presence of a distur-
bance (Karr 1991). A community with high integrity
is composed of species with attributes that are depen-
dent upon components of an undisturbed system
(O?Connell 2000). Such attributes are specialized and
are typically sensitive to anthropogenic disturbance,
while life history traits that are more generalized en-
able a species to use a variety of habitats. It is impor-
tant to consider multiple life history attributes because
it is possible for one species to have both specialist
and generalist attributes (Price 1984). For example,
red-winged blackbirds establish breeding territories
only in wetland habitats, a fairly specialized attribute.
In contrast, the foraging behavior of red-winged black-
birds can be characterized as a generalized attribute
because they feed in marshes, garbage dumps, prairies,
agricultural fields, and lawns (Yasukawa and Searcy
1995). By classifying the attributes of each species for
the entire community on a specialist-to-generalist gra-
dient, an index can be constructed that represents the
integrity of that community.
Birds are considered ideal for use in a community
index because they are easy to survey and avian life
histories are relatively well defined. Previous research
has shown that birds are linked to the overall ecolog-
ical integrity of their respective ecosystem (Bradford
et al. 1998, O?Connell 1998, Canterbury et al. 2000,
Bryce et al. 2002). This is true primarily because they
are sensitive to habitat fragmentation (Freemark and
Collins 1992), landscape composition (Rodewald and
Yahner 2001), and change in habitat structure (Riffell
et al. 2001). Birds may also be particularly good in-
dicators because species at high trophic levels can be
sensitive to disturbances at lower levels (Pettersson et
al. 1995). Therefore, it is unlikely that a marsh with
low ecological integrity can support a high-integrity
marsh bird community. Furthermore, a marsh that has
a degraded upper trophic level (i.e., birds) cannot be
considered to have high overall integrity.
We constructed a community index based on marsh
birds designed to estimate the integrity of the marsh
bird community and provide insight into the integrity
of the entire marsh ecosystem. Previous IBIs have used
sites that occur along a known reference to degradaded
continuum to test potential metrics. Environmental
stressors to estuarine wetlands remain relatively un-
described (Pennings et al. 2002). Therefore, identify-
ing marshes on an a priori basis as either pristine or
degraded can be problematic. We used sound ecolog-
ical principles to develop an index of marsh bird com-
munity integrity (IMBCI) and subsequently tested the
sensitivity of marsh bird community integrity to in-
dependently quantified land-use disturbances. Specifi-
cally, we scored marsh bird species attributes along a
generalist-to-specialist gradient to develop species
scores and, ultimately, site scores. Site scores were
then used to determine whether wetland habitat char-
acteristics and urban/suburban development, agricul-
ture, and forest at three different geographic scales
(watershed, 1000-m buffer, and 500-m buffer) affected
marsh bird community integrity.
METHODS
Study Area
Our study was conducted in tidal wetlands of Ches-
apeake Bay, USA (39E 239 N; 36E 489 N to 76E 459
W; 75E 449 W) (Figure 1). The Chesapeake Bay is
one of the largest and most productive estuaries in the
world. It is characterized by 7,400 kilometers of tidal
shoreline (Lippson and Lippson 1997), shallow waters,
approximately 101,000 hectares of estuarine wetlands
(Tiner and Burke 1995), and diverse floral and faunal
communities. Land use within the Chesapeake Bay
watershed is diverse. Industrial and high-density urban
development is concentrated on the western shore of
the bay near Baltimore, Maryland and Portsmouth,
Virginia. Other areas of the western shore of the bay
are primarily composed of oak-tulip poplar forests, ur-
ban/suburban development, and low-density agricul-
ture. Commercial agriculture dominates the eastern
shore of the bay and consists of row crops, chicken
farms, and pasture.
Wetlands in the northern part of the bay are slightly
more influenced by freshwater and have habitats char-
acterized by narrow-leaved cattail (Typha angustifolia
L.), common reed (Phragmites australis (Gav.) Trin.),
arrow arum (Peltandra virginica L.), and rose mallow
(Hibiscus moscheutos L.). Wetlands in the southern
part of the bay are more saline and are dominated by
smooth cordgrass (Spartina alterniflora Loisel.), salt
meadow grass (Spartina patens (Ait.) Muhl.), big
cordgrass (Spartina cynosuroides (L.) Roth), and
marsh elder (Iva fructescens L.).
Site Selection and Classification
To include a full range of land-use types represented
at multiple scales, we began by identifying five wa-
tershed categories across a land-use disturbance gra-
dient. These watershed categories were used solely
during site selection. We selected 30 watersheds of
third through fifth order tributaries of the Chesapeake
Bay distributed roughly evenly among five categories
(see Table 1 for watershed land use composition): 1)
DeLuca et al., MARSH BIRD COMMUNITY INTEGRITY 839
Figure 1. Map of Chesapeake Bay, USA showing the distribution and land-use classification of the 30 watersheds used in
our study. All watersheds are associated with a sub-estuary, each of which contains three study wetlands.
Table 1. Percent of urban development, agriculture, and forest within each watershed category used during site selection (mean 6
standard deviation (range)).
Watershed Category Urban Development Agriculture Forest
Developed
Low-intensity development
Agriculture
Low-intensity agriculture
Forest
65.7 6 5.1 (60.2?73.9)
23.8 6 7.7 (13.0?32.1)
4.1 6 4.5 (0.8?11.4)
4.1 6 3.3 (0.7?8.6)
4.2 6 2.5 (1.0?7.4)
8.0 6 4.7 (3.3?14.6)
15.8 6 7.7 (7.7?27.4)
62.4 6 9.4 (72.1?50.6)
35.7 6 4.0 (31.2?40.3)
20.6 6 4.9 (16.0?29.0)
20.0 6 1.9 (18.2?22.1)
53.8 6 13.0 (32.2?71.9)
31.1 6 5.1 (32.2?71.9)
53.6 6 6.1 (32.2?71.9)
72.8 6 6.1 (32.2?71.9)
urban/suburban development (hereafter, development);
2) low intensity development; 3) agriculture; 4) low-
intensity agriculture; and 5) forested. Landscape com-
position was measured using National Land Cover Da-
tabase (30-m raster coverage) and geographic infor-
mation systems (GIS) software ArcGIS and ArcView
(Environmental Systems Research Institute, Redlands,
California).
Three estuarine wetlands were chosen as study sites
within each watersheds? subestuary. A subestuary is a
smaller estuarine embayment within a larger estuary
(e.g., the Severn River within the Chesapeake Bay).
Six wetlands were selected outside of the original 30
watersheds, resulting in 96 study sites. These addition-
al six wetlands were selected due to a lack of large
wetlands with extensive local development and were
therefore included only in local land-use analysis.
Buffers were developed using GIS at 1000 m and 500
840 WETLANDS, Volume 24, No. 4, 2004
m and were exclusive for each study wetland. Land-
use composition was calculated at each scale.
Bird Community Sampling
We used 50-m radius, five-minute, double observer
point counts (Nichols et al. 2000) to sample the wet-
land bird community. Differences in species detect-
ability were not calculated because the IMBCI is based
on occurrence and estimates of abundance were not
needed. One point count location was used to sample
wetlands ,2 ha, two for wetlands between 2 and 7 ha,
and three for wetlands .7 ha. Point count locations
were selected to encompass representative portions of
the marsh and were .150 m from the nearest point
count and .50 m from the upland edge. Two counts
were conducted at each site from 15 May to 18 July
between 0600 and 1100 hours, with a minimum of 14
days between counts. Fifty-seven wetlands were sam-
pled in 2002, 66 wetlands were sampled in 2003, and
27 wetlands were sampled in both years.
At each wetland, a centrally located point count was
selected to conduct an unlimited radius secretive marsh
bird survey. At the conclusion of the point count, calls
of black rail (Laterallus jamaicensis), least bittern
(Ixobrychus exilis), common moorhen (Gallinula chlo-
ropus), Virginia rail (Rallus limicola), clapper rail
(Rallus longirostris), sora (Porzana carolina), Amer-
ican bittern (Botaurus lentiginosus), and pied-billed
grebe (Podilymbus podiceps) (obtained from Elliot et
al. 1997) were broadcast. Species? calls were arranged
from least intrusive to most intrusive (as listed above).
Each playback period had one minute of a species?
call, with a 30-second listening period between each
species? call, and ended with a one-minute listening
period as recommended by Gibbs and Melvin (1993)
and Ribic et al. (1999).
Wetland Habitat Measurements
To quantify plant species composition and vegeta-
tion structure at each site, four vegetation sampling
circles were established within each 50-m radius (0.79
ha) point count location. Vegetation circles were 12 m
in diameter, and the first circle was centered on the
point count area. The remaining three circles were 35
m from the point count center at 08, 1208, and 2408,
respectively. Each vegetation circle was divided into
four equal quadrants. Within each quadrant, plant spe-
cies? abundance was quantified using a modified
Braun-Blanquet scale (Leps and Hadincova 1992).
Percent cover values for each plant species were av-
eraged among the four quadrants of each circle, then
averaged again among the four circles, resulting in a
percent cover value for each plant species for each
point count location.
Vegetation structure was measured five meters from
the center of each vegetation circle on each of the lines
separating the four quadrants. Vertical structure was
measured using methods modified from Haukos et al.
(1998). A profile board (three meter pole with 0.1 m
marked intervals) was held vertically while an observ-
er, standing four meters away, determined the lowest
0.1 m increment not completely covered by vegetation.
Horizontal structure was measured using similar meth-
ods; however, the profile board was held horizontally
at 0.5 m, 1.0 m, 1.5 m, and 2.0 m above the ground
and an observer counted the number of 0.1 m sections
not covered by vegetation at each height increment.
Vegetation structure measurements for the entire point
count circle were summarized using the same proce-
dure as for percent cover values.
Index of Marsh Bird Community Integrity
The IMBCI developed in our study combined two
approaches for developing indices of bird community
integrity: 1) the guild-based community index (Croon-
quist and Brooks 1991, Weller 1995, Canterbury et al.
2000, O?Connell et al. 2000, Bryce et al. 2002), and
2) the indicator species approach (Landres et al. 1988,
Pendergast and Eversham 1997, Tardif and Des-
Granges 1998, Chase et al. 2000). The IMBCI incor-
porates components of the guild-based community in-
dex by measuring species? attributes on a specialist-
to-generalist gradient. Specialists have attributes that
are capable of exploiting specific resources of their
native ecosystem. If those resources are disturbed, spe-
cialists are less capable of switching to other resources
and are therefore a good indicator of habitat quality
(Odum 1971, Miller et al. 1998, O?Connell et al.
2000). The IMBCI also examines the community at
the species level and exploits the efficiency of includ-
ing indicator species. The indicator species approach
is capable of accurately linking a single species to a
specific portion of an ecosystem (Kushlan 1979). Thus
connections might be made that aid in the identifica-
tion of a specific stressor to marsh condition. Combin-
ing components of both the guild and indicator species
approaches, allows the IMBCI to be an accurate and
sensitive index for a marsh bird community that has a
low number of native species.
For the purpose of our study, we defined the term
species? attribute as a component of a species? behav-
ioral strategy or an inherent adaptation that a species
uses for survival. Species? attributes were used as the
basic component of IMBCI development. To represent
the major factors that determine a species? vulnerabil-
ity to disturbance, we used species? attributes that de-
DeLuca et al., MARSH BIRD COMMUNITY INTEGRITY 841
Table 2. Bird species attributes (Poole and Gill 1999) and scoring criteria used to develop Index of Marsh Bird Community Integrity
(IMBCI) scores.
Score
Generalist ? Specialist
Species
attributes 1 2 2.5 3 4
Foraging habitat habitat generalist marsh facultative marsh obligate
Nesting substrate non-marsh nesters marsh vegetation marsh ground nester
Migratory status resident short distance/temperate Neotropical
Breeding range North America North America?
east of Rocky
Mts.
coastal North
America
North America?east
coast only
scribed foraging, nesting, migration, and breeding
range (Table 2). The presence of specialized bird for-
aging and nesting strategies in an ecosystem are rep-
resentative of an undisturbed system (Canterbury et al.
2000, Bryce et al. 2002). The specialization of a spe-
cies? migratory behavior was selected as a species? at-
tribute in IMBCI development because it has been
shown to respond as an indicator of disturbance
(Croonquist and Brooks 1991, O?Connell 2000, Bryce
et al. 2002). Breeding range was selected because it is
likely that a bird with a limited range is adapted to
conditions that are specific to that geographic region;
thus, by measuring this attribute, we capture a species?
capacity to respond to regional disturbances. For ex-
ample, if coastal song sparrow (Melospiza melodia)
populations are impacted by anthropogenic disturbanc-
es specific to coastal areas, the species may persist due
to populations located inland. In contrast, the entire
seaside sparrow (Ammodramus maritimus) population
could be threatened because its range is confined to
coastal areas.
To develop IMBCI scores, we first calculated the
individual score for each species detected at a site
(SIMBCI):
S 5 L (1)
OIMBCI S
where LS is the score of each species? attribute on a
scale of 1 to 4 (Table 1). Once all species? scores were
calculated (Appendix I), scores for specific wetlands
(WIMBCI) and sub-estuaries (EIMBCI) were calculated:
W 5 S /S 1 MO 2 4 (2)
1O 2IMBCI IMBCI N N[ ]
E 5 W /M (3)
OIMBCI IMBCI N
where SN is the total number of species detected at a
wetland and MON is the number of marsh obligate spe-
cies detected at the wetland. The equation for WIMBCI
represents the mean species score for all species pre-
sent, plus the total number of marsh obligate species
at a wetland. These equations incorporate species at-
tributes and species richness, while preserving a signal
in the equation for what may be the most sensitive
group to disturbance, marsh obligates. We subtract
four while determining the WIMBCI score to ensure a
scoring scale that begins with zero and remains con-
stant. MN is the total number of marshes sampled in a
sub-estuary. Therefore, EIMBCI represents the mean
score for all marshes sampled within a sub-estuary.
An IMBCI score of zero indicated that only species
with generalist attributes were present (i.e., European
Starling). An unweighted score of 12 indicated that all
birds detected have specialist attributes that best indi-
cate a healthy marsh ecosystem.
Statistical Analyses
A species was considered present at a wetland and
given an IMBCI species score (SIMBCI) if it was de-
tected at least once at any of the wetland?s point count
stations during either of the two surveys. Five sites
contained no birds and, because IMBCI scores could
not be calculated for sites without birds, these sites
were not included in further analyses. Each wetland
was considered an independent sample (N 5 96) when
analyzing relationships between bird indices, wetland
habitat characteristics, and landscape composition at
the 500-m and 1000-m scales. Each EIMBCI score was
considered an independent sample for watershed-level
analysis (N 5 30).
We examined potential confounding variables to
isolate the relationship between land use, habitat char-
acteristics, and IMBCI scores. Regression analysis was
performed on a sample of 52 sites to determine if sa-
linity had an impact on the IMBCI. Study wetlands
ranged in size from 0.01 to 130 hectares, and a re-
gression analysis was used to determine if marsh patch
size influenced IMBCI scores. There are were few
large wetlands (.80 ha) associated with a defined wa-
tershed and within a subestuary in the Chesapeake
Bay; consequently, our study included only three. To
842 WETLANDS, Volume 24, No. 4, 2004
Table 3. Interpretation of axes from a principal component analysis of five vegetation structural measurements and 13 vegetation ground
cover variables.
Axis
Cumulative percent
variance explained
Eigen
values Biological interpretation of axes
PC 1 33.1 6.0 abundance of Phragmites australis, high vertical structure and dense, homo-
geneous horizontal structure
PC 2 45.5 2.2 abundance of Hibiscus moscheutos, Peltandra virginica, and Typha angusti-
folia, a marsh plant community indicative of lower salinity levels
PC 3 57.1 2.1 abundance of Spartina patens, Distichlis spicata, and Iva fructescens, a
marsh plant community with generally higher salinity levels
PC 4 67.7 1.9 high cover values for bare ground and Amaranthus cannabinus, marshes ex-
hibiting characteristics of erosion
be certain that these sites would not produce a spurious
correlation between IMBCI scores and wetland size,
we used Cook?s distance to measure whether regres-
sion coefficients would change if individual wetlands
were removed from the analysis. None of the individ-
ual wetlands had an extreme influence on the relation-
ship between IMBCI score and wetland size (Cook?s
distance ,1 in all cases).
To determine if results from testing the IMBCI for
area sensitivity were an artifact of passive sampling
(Horn and Fletcher 2000, Johnson 2001), we randomly
selected one point count from each marsh. Next, we
included secretive marsh birds detected at the marsh,
and then calculated the IMBCI score using only the
randomly selected point. After confirming the effect of
wetland size on IMBCI scores (r 5 0.38, P , 0.001),
we used the standardized residuals of the analysis,
which included all the point counts in a marsh as the
dependent variable in subsequent analyses of IMBCI
scores.
Upon exploring the relationship between IMBCI
scores and land use at the watershed, 1000-m, and
500-m scales, it was evident that the data did not meet
assumptions of linearity. Consequently, to test for a
non-linear relationship, a non-parametric changepoint
analysis (nCPA) (King and Richardson 2003, Qian et
al. 2003) was used to test for the presence of an eco-
logical threshold in the IMBCI due to percent devel-
opment, agriculture, and forest at the three geographic
scales. The nCPA detects changes in the mean and
variance of an ecological response variable (in this
case the IMBCI) due to changes of a disturbance (in
this case land use at three geographic scales). It ex-
amines every point along a continuum of predictor val-
ues (e.g., percent development) and determines the
probability that a value can split the data into two
groups that have the highest between-group variance
relative to within-group variance. The resulting cu-
mulative probability curve illustrates the probability of
a changepoint occurring at various levels of distur-
bance. When probabilities of a Type I error (P to be
shown in Figure 3) were , 0.1, we could be fairly
certain that cumulative probability curves accurately
assessed the likelihood of an ecological threshold oc-
curring.
To analyze wetland habitat variables, a principal
component analysis (PCA) was used to reduce the five
vegetation structural measurements and the 13 ground-
cover variables that accounted for 80% of total ground
cover to four principal components (Table 3) (Jong-
man et al. 1995). The standardized factor scores from
each principal component for each site were then com-
pared using multiple regression to determine if wetland
habitat characteristics influenced IMBCI scores.
RESULTS
Wetlands averaged 3.51 6 0.23(SE) bird species per
site, and thirty species were detected across all sites
and used for IMBCI development and testing (Appen-
dix I). The observed IMBCI scores for each wetland
ranged from 1 to 10.39, with a mean of 4.08 6 0.18.
Salinity did not have a significant effect on IMBCI
scores (r25 0.02, P 5 0.29). Wetland size had a sig-
nificant influence on IMBCI scores (Figure 2). Con-
sequently, the residuals from the analysis in Figure 2
were used to control for size in all subsequent analyses
between IMBCI scores and land use.
IMBCI and Land Use
nCPA determined that development had a signifi-
cant effect on IMBCI scores at the 500 m scale (P 5
0.04). The cumulative probability curve revealed a
95% chance of a changepoint occurring at #14% de-
velopment (Figure 3). This result indicates that when
.14% of the area within 500 m of the marsh is de-
veloped, IMBCI scores change significantly. Further-
more, there was a relatively low probability (25%) of
a threshold response with #2% development. How-
ever, there was a 60% probability of a changepoint
occurring when as little as 6% of the land within 500
DeLuca et al., MARSH BIRD COMMUNITY INTEGRITY 843
Figure 2. Results of a regression analysis showing the re-
lationship between the index of marsh bird community in-
tegrity (IMBCI) and wetland size for 91 wetlands in the
Chesapeake Bay, USA. IMBCI scores are calculated by
scoring species attributes on a generalist to specialist gra-
dient.
Figure 3. Results of non-parametric changepoint analyses
for percent development within 500 m and 1000 m of wet-
land study sites and index of marsh bird community integrity
(IMBCI) scores controlled for wetland size using the stan-
dardized residuals from Figure 2. The cumulative probability
curve represents the cumulative probability that a change-
point occurred at various levels of development. IMBCI
scores are calculated by scoring species? attributes on a gen-
eralist to specialist gradient. Dashed lines indicate the per-
cent of development within a wetland buffer required to pro-
duce a 95% cumulative probability of an ecological thresh-
old occurring.
m of a wetland was developed (Figure 3). Change-
points were not detected when agriculture and forest
were tested using nCPA against IMBCI scores at the
500 m scale (P 5 0.22 and P 5 0.21, respectively).
Development also had a significant effect on IMBCI
scores at the 1000-m buffer scale (P 5 0.09). There
was a 95% probability of a changepoint occurring at
#25% development (Figure 3). At this scale, there was
a 38% probability of a changepoint occurring with
#2% development. Also, there was a 60% chance of
a changepoint occurring when #8.5% of the 1000-m
buffer was developed (Figure 3). Changepoints were
not detected for agriculture and forest at the 1000-m
scale (P 5 0.25, P 5 0.18, respectively). We did not
detect changepoints in IMBCI scores for percent de-
velopment, agriculture, or forest at the watershed scale
(P 5 0.22, P 5 0.22, and P 5 0.26, respectively).
IMBCI and Wetland Habitat Characteristics
The PCA resulted in four axes that together ex-
plained 67.7% of the variation. Wetland plant com-
munities (Table 2) did not have a significant effect on
marsh bird community integrity (r2 5 0.03, df 5 90,
P 5 0.74).
DISCUSSION
IMBCI and Land Use
An index of community integrity should be capable
of measuring the response of a biological community
to an environmental stressor (Bradford et al. 1998,
O?Connell 2000). Our IMBCI identified thresholds
with 95% probability for the amount of development
that can occur within 500-m (14%) and 1000-m (25%)
buffers of an estuarine wetland before ecological in-
tegrity is significantly compromised (Figure 3). More-
over, we found a 60% probability that a changepoint
occurred in IMBCI scores with relatively low levels
of disturbance (6% development at the 500-m scale
and 8.5% development at the 1000-m scale). The
thresholds identified in our study indicate levels of lo-
cal disturbance, beyond which, marsh bird community
integrity is unattainable. This does not imply that
marshes with local percent development levels below
thresholds necessarily have high integrity. It does im-
ply, however, that marshes with local development be-
low threshold levels have the capacity for high integ-
rity, whereas marshes with local development above
threshold levels do not.
Other studies have also identified ecological thresh-
844 WETLANDS, Volume 24, No. 4, 2004
olds or non-linear responses to anthropogenic distur-
bances. Interestingly, most reported thresholds were
found between 10 and 20% disturbance, similar to the
results of our study. For example, Paul and Meyer
(2001) reported thresholds of stream degradation when
impervious surface in stream catchments exceeded 10
to 20%. Urban disturbance has also been linked to deg-
radation of fish communities with as little as 10 (Lim-
burg and Schmidt 1990) and 20% (Wang et al. 1997)
impervious surface area within stream catchments.
These studies, in concert with our findings, provide
strong evidence that the response of biological com-
munities dependant upon various aquatic systems re-
spond to anthropogenic disturbances in similar, non-
linear manners (i.e., threshold response).
Understanding the spatial scale at which land use
influences the integrity of an ecosystem may provide
some insight into the potential mechanisms driving the
response. In our study, the threshold response to de-
velopment decayed when we moved from local buffers
to the watershed scale. This finding implies that there
is a scale-dependent effect of development on estua-
rine wetlands. Findlay and Houlahan (1997) and Whit-
ed et al. (2000) both found that road density at local
scales had a pronounced negative effect on wetland
bird assemblages and concluded that reduced connec-
tivity between wetland patches caused by development
may have restricted wetland bird distribution. Distur-
bances in close proximity to wetlands also provides
habitat for an abundance of generalist birds (Blair
1996). Generalists are then capable of invading the
marsh and increasing interspecific competition with
marsh birds for available resources.
Fragmented, isolated, and altered forest patches are
known to be less suitable for terrestrial bird commu-
nities (Faaborg 1995). The primary mechanisms be-
hind forest fragmentation and isolation effects may
also apply to marsh bird communities. Small and
Hunter (1988) found that roads, power lines, and edg-
es, all characteristics of developed areas, provide path-
ways for potential predators to enter undisturbed hab-
itat and depredate bird nests. Pathways or corridors
such as these may act in similar ways near marshes to
increase nest predation and lower reproductive suc-
cess. Another explanation for the reduction in marsh
bird community integrity may be the transfer of pol-
lutants from adjacent land. Chemical pollutants and
nutrients transferred from developed areas through
point sources may reduce the food resources of marsh
birds (Poulin et al. 2002). For example, aquatic mac-
roinvertebrates might be impacted from such point-
source pollution. Interestingly, most secretive marsh
birds, marsh foraging specialists, feed primarily on
aquatic macroinvertebrates, and only three of the 45
lowest scoring wetland sites of our study had secretive
marsh birds present. Future research aimed at identi-
fying stressors may need to look at how the macro-
invertebrate community is affected by changes in soil
and water chemistry of a marsh.
IMBCI and Wetland Habitat Characteristics
Previous studies have shown that vegetation com-
position and structure can influence bird distribution
and abundance in wetlands (Burger 1982, Craig and
Beal 1992, Watts 1993, Reinert and Mello 1995, Rif-
fell et al. 2001). Our study did not address this ques-
tion directly, but instead examined how various attri-
butes of wetland plant communities influenced marsh
bird community integrity. We found that marsh bird
community integrity was not dependent on any one
marsh plant community type. This is not to say that
the marsh plant community does not influence bird
distribution. Rather, it implies that marsh bird com-
munities may reach high integrity in a number of
marsh plant communities. Such is the case when a spe-
cialist bird that inhabits one type of plant community
is replaced by other specialists in different plant com-
munities. For example, seaside sparrows and clapper
rails that inhabit marshes dominated by Iva fructescens
and Spartina alterniflora are replaced by marsh wrens
and virginia rails in a marsh dominated by Typha an-
gustifolia and Phragmites australis. Therefore, marsh
plant communities may not be good predictors of
marsh bird community integrity. Managers should use
caution however, because it is uncertain how dramatic
temporal changes in plant communities due to extreme
hydrological changes (Wilcox et al. 2002) may influ-
ence the IMBCI. Our findings illustrate that although
the IMBCI can accurately measure how specialized a
bird community is, it is less useful for recognizing how
marsh bird community composition varies among
marshes of similar integrity.
CONCLUSIONS
Our research demonstrates that anthropogenic dis-
turbance surrounding a wetland can have negative ef-
fects on marsh bird community integrity. Furthermore,
we identified a specific land-use disturbance threshold
at approximately 14%, similar to thresholds identified
in other studies. Such information, as provided by this
and other research (Findlay and Houlahan 1997, Find-
lay and Bourdages 2000, Whited et al. 2000), should
be incorporated into existing wetland assessments,
which tend to include primarily within-site character-
istics. Finally, our study has shown that marsh bird
community integrity is not dependent on any one type
of marsh plant community. Wetland managers should
consider that marsh bird communities with high integ-
DeLuca et al., MARSH BIRD COMMUNITY INTEGRITY 845
rity may be found in a variety of wetland community
types.
In summary, the IMBCI can be an accurate indicator
of marsh bird community integrity that may assist in
the assessment of the integrity of the entire marsh eco-
system. IMBCI scores, combined with the identifica-
tion of a land-use threshold, are easily interpreted, and
provide rapid assessment approaches for communicat-
ing complex ecological data to natural resource man-
agers and conservation planners. By helping to bridge
the gap between scientists and regional conservation
decision makers, the IMBCI could become a valuable
contribution to the ongoing efforts of restoring and
maintaining the ecological integrity of the Chesapeake
Bay. Finally, this approach could provide a template
for measuring ecological integrity in other marsh sys-
tems in North America and potentially throughout the
world.
ACKNOWLEDGMENTS
We appreciate the hard work of Anne Balogh, Suz-
anne Conrad, Sacha Mkheidze, Dan Mummert, Ryan
Peters, and Beth Wright. We also thank Ryan King for
providing statistical advice and improvements on the
manuscript and Tim O?Connell for assistance during
IMBCI development. This research was funded by a
grant from the U.S. Environmental Protection Agen-
cy?s Science to Achieve Results (STAR) Estuarine and
Great Lakes (EaGLe) program to the Atlantic Slope
Consortium. Although the research described in this
article has been funded wholly or in part by the U.S.
Environmental Protection Agency through cooperative
agreement #R-82868401 to the Smithsonian Environ-
mental Research Center, it has not been subjected to
the Agency?s required peer and policy review and
therefore does not necessarily reflect the views of the
Agency and no official endorsement should be in-
ferred.
LITERATURE CITED
Angermeier, P. L. and J. R. Karr. 1986. Applying an index of biotic
integrity based on stream-fish communities: considerations in
sampling and interpretation. North American Journal of Fisheries
Management 6:418?429.
Blair, R. B. 1996. Land use and species diversity along an urban
gradient. Ecological Applications 6:506?519.
Bradford, D. F., S. E. Franson, G. R. Miller, A. C. Neale, G. E.
Canterbury, and D. T. Heggem. 1998. Bird species assemblages
as indicators of biotic integrity in Great Basin rangeland. Envi-
ronmental Monitoring and Assessment 49:1?22.
Bryce, S. A., R. M. Hughes, and P. R. Kaufmann. 2002. Develop-
ment of a bird integrity index: using bird assemblages as indica-
tors of riparian condition. Environmental Management 30:294?
310.
Burger, J., J. Shisler, and F. H. Lesser. 1982. Avian utilization of
six salt marshes in New Jersey. Biological Conservation 23:187?
212.
Burton, T. M., D. G. Uzarski, J. P. Gathman, J. A. Genet, B. E.
Keas, and C. A. Stricker. 1999. Development of a preliminary
invertebrate index of biotic integrity for Lake Huron costal wet-
lands. Wetlands 19:869?882.
Canterbury, G. E., T. E. Martin, D. R. Petit, L. J. Petit, and D. F.
Bradford. 2000. Bird communities and habitat as ecological in-
dicators of forest condition in regional monitoring. Conservation
Biology 14:544?558.
Chase, M. K., W. B. Kristan III, A. J. Lynam, M. V. Price, and J.
T. Rotenberry. 2000. Single species as indicators o species rich-
ness and composition in California coastal sage scrub birds and
small mammals. Conservation Biology 14:474?487.
Cohen, J. E., C. Small, A. Mellinger, J. Gallup, and J. Sachs. 1997.
Estimates of coastal populations. Science 278:1211?1212.
Craig, R. J. and K. G. Beal. 1992. The influence of habitat variables
on marsh bird communities of the Connecticut River estuary. Wil-
son Bulletin 104:295?311.
Croonquist, M. J. and R. P. Brooks. 1991. Use of avian and mam-
malian guilds as indicators of cumulative impacts in riparian wet-
land areas. Environmental Monitoring 15:701?714.
Culliton, T. J., M. A. Warren, T. R. Goodspeed, D. G. Remer, C.
M. Blackwell, and J. J. McDonough, III. 1990. 50 Years of Pop-
ulation Change along the Nation?s Coasts, 1960?2010. U.S. De-
partment of Commerce, National Oceanic and Atmospheric Ad-
ministration. Rockville, MD, USA.
Dauer, D. M., J. A. Ranasinghe, and S. B. Weisberg. 2000. Rela-
tionsships between benthic community condition, water quality,
sediment quality, nutrient loads, and land use patterns in Chesa-
peake Bay. Estuaries 23:80?96.
Edgar, G. J., N. S. Barrett, D. J. Graddon, and P. R. Last. 2000. The
conservation significance of estuaries: A classification of Tasman-
ian estuaries using ecological, physical and demographic attributes
as a case study. Biological Conservation 92:282?397.
Elliot, L., D. Stokes, and L. Stokes. 1997. Stokes Field Guide to
Bird Songs. Time Warner AudioBooks, New York, NY, USA.
Faaborg, J., M. Brittingham, T. Donovan, and J. Blake. 1995. Hab-
itat fragmentation in the temperate zone. p. 357?380. In T. E.
Martin and D. Finch (eds.) Ecology and Management of Neotrop-
ical Migratory Birds: a Synthesis and Review of Critical Issues.
Oxford University Press, New York, NY, USA.
Findlay, C. S. and J. Houlahan. 1997. Anthropogenic correlates of
species richness in southeastern Ontario wetlands. Conservation
Biology 11:1000?1009.
Findlay, C. S. and J. Bourdages. 2000. Response time of wetland
biodiversity to road constuction on adjacent lands. Conservation
Biology 14:86?94.
Freemark, K. and B. Collins. 1992. Landscape ecology of birds
breeding in temperate forest fragments. p. 443?454. In J. M. Ha-
gan III and D. W. Johnston, (eds.) Ecology and Conservation of
Neotropical Migrant Landbirds. Smithsonian Institution Press,
Washington DC, USA.
Gibbs, J. P. and S. M. Melvin. 1993. Call-response surveys for mon-
itoring breeding waterbirds. Journal of Wildlife Management 57:
27?34.
Haukos, D. A., H. Z. Sun, D. B. Webster, and L. M. Smith. 1998.
Sample size, power, and Analytical considerations for vertical
structure data from profile boards in wetland vegetation. Wetlands
18:203?215.
Horn, D. J. and R. J. Fletcher, Jr. 2000. Detecting area sensitivity:
A comment on previous studies. American Midland Naturalist
144:28?35.
Johnson, D. H. 2001. Habitat fragmentation effects on birds in grass-
lands and wetlands: A critique of our knowledge. Great Plains
Research 11:211?231.
Jongman, R. G. H., C. J. F. Ter Braak, and O. F. R. Van Tongeren,
(eds.). 1995. Data Analysis in Community and Landscape Ecol-
ogy. Cambridge University Press, New York, NY, USA.
Karr, J. R. 1991. Biological integrity: a long-neglected aspect of
water resource management. Ecological Applications 1:66?84.
King, R. S. and C. J. Richardson. 2003. Integrating bioassessment
and ecological risk assessment: an approach to developing nu-
merical water-quality criteria. Environmental Management 31:
795?809.
846 WETLANDS, Volume 24, No. 4, 2004
Kushlan, J. A. 1979. Design and management of continental wildlife
reserves: Lessons from the Everglades. Biological Conservation
15:281?290.
Landres, P. B., J. Verner, J. W. Thomas. 1988. Ecological uses of
vertabrate indicator species: A critique. Conservation Biology 2:
316?328.
Leps, J. and V. Hadincova. 1992. How reliable are our vegetation
analysis? Journal of Vegetation Science 3:119?124.
Limburg, K. E. and R. E. Schmidt. 1990. Patterns of fish spawning
in Hudson River tributaries: response to an urban gradient? Ecol-
ogy 71:1238?1245.
Lippson, A. J. and R. L. Lippson. 1997. Life in the Chesapeake Bay.
The Johns Hopkins University Press. Baltimore, MD, USA.
Miller, S. G., R. L. Knight, and C. K. Miller. 1998. Influence of
recreational trails on breeding bird communities. Ecological Ap-
plications 8:162?169.
Nichols, J. D., J. E. Hines, J. R. Sauer, F. W. Fallon, J. E. Fallon,
and P. J. Heglund. 2000. A double-observer approach for esti-
mating detection probability and abundance from point counts.
The Auk 117:393?408.
Nordstrom, K. F. and C. T. Roman (eds.). 1996. Estuarine Shores.
John Wiley and Sons Ltd, Chichester, England.
O?Connell, T. J., L. E. Jackson, and R. P. Brooks. 1998. A bird
community index of biotic integrity for the Mid-Atlantic High-
lands. Environmental Monitoring Assessment 51:145?156.
O?Connell, T. J., L. E. Jackson, and R. P. Brooks. 2000. Bird guilds
as indicators of ecological condition in the central Appalachians.
Ecological Applications 10:1706?1721.
Odum, E. P. 1971. Fundamentals of Ecology. 3rd edition. W. B.
Saunders, Philadelphia, PA, USA.
Paul, M. J. and J. L. Meyer. 2001. Streams in the urban landscape.
Annual Review of Ecology and Systematics 32:333?365.
Pendergast, J. R. and B. C. Eversham. 1997. Species richness co-
variance in higher taxa: empirical tests of the biodiversity indi-
cator concept. Ecography 20:210?216.
Pennings, S. C., V. D. Wall, D. J. Moore, M. Pattanayek, T. L. Buck,
and J. J. Alberts. 2001. Assessing salt marsh health: a test of the
utility of five potential indicators. Wetlands 22:405?414.
Pettersson, R. P., J. P. Ball, K. Renhorn, P. Esseen, and K. Sjoberg.
1995. Invertabrate communities in boreal forest canopies as influ-
enced by forestry and lichens with implications for passerine
birds. Biological Conservation 74:57?63.
Poole, A., and F. Gill. 1999. The Birds of North America. The Birds
of North America, Inc., Philadelphia, PA, USA.
Poulin, B., G. Lefebvre, and A. Mauchamp. 2002. Habitat require-
ments of passerines and reedbed management in souther France.
Biological Conservation 107:315?325.
Price, P. W. 1984. Communities of specialists: vacant niches in eco-
logical and evolutionary time. p. 510?523. In D. R. Strong Jr., D.
Simberloff, L. G. Abele, and A. B. Thistle, (eds.) Ecological Com-
munities, Conceptual Issues and the Evidence. Princeton Univer-
sity Press, Princeton, NJ, USA.
Qian, S. S., R. S. King, and C. J. Richardson. 2003. Two statistical
methods for the detection of environmental thresholds. Ecological
Modelling 166:87?97.
Reinert, S. E. and M. J. Mello. 1995. Avian community structure
and habitat use in a southern New England estuary. Wetlands 15:
9?19.
Ribic, C. A., S. J. Lewis, S. Melvin, J. Bart, and B. Peterjohn. 1999.
Proceedings from the marsh bird monitoring workshop. U.S. Fish
and Wildlife Service, U.S. Geological Survey, National Wildlife
Visitor Center, Laurel, MD, USA.
Riffell, S. K., B. E. Keas, and T. M. Burton. 2001. Area and habitat
relationships in Great Lakes coastal wet meadows. Wetlands 21:
492?507.
Rodewald, A. D. and R. H. Yahner. 2001. Influence of landscape
composition on avian community structure and associated mech-
anisms. Ecology 82:3493?3504.
Schimmel, S. C., S. J. Benyi, and C. J. Strobel. 1999. An assessment
of the ecological condition of Long Island Sound, 1990?1993.
Environmental Monitoring and Assessment 56:27?49.
Small, M. F. and M. L. Hunter. 1988. Forest fragmentation and avian
nest predation in forested landscapes. Oecologia 76:62?64.
Tardif, B. and J-L. DesGranges. 1998. Correspondance between bird
and plant hotspots of the St. Lawrence River and the influence of
scale on their location. Biological Conservation 84:53?63.
Tiner, R. W. and D. G. Burke. 1995. Wetlands of Maryland. U.S.
Fish and Wildlife Service, Ecological Services, Region 5, Hadley,
MA and Maryland Department of Natural Resources, Annapolis,
MD. Cooperative publication 1?193.
Wang, L., J. Lyons, P. Kanehl, and R. Gatti. 1997. Influences of
watershed land use on habitat quality and biotic integrity in Wis-
consin streams. Fisheries 22:6?12.
Watts, B. D. 1993. Effects of marsh size on incidence rates and
avian community organization within the lower Chesapeake Bay
(Final report). Center for Conservation Biology, College of Wil-
liam and Mary:1?54.
Weisberg, S. B., J. A. Ranasingh, D. M. Dauer, L. C. Schaffner, R.
J. Diaz, and J. B. Frithsen. 1997. An estuarine benthic index of
biotic integrity (B-IBI) for Chesapeake Bay. Estuaries 20:149?
158.
Weller, M. W. 1995. Use of two waterbird guilds as evaluation tools
for the Kissimmee River restoration. Restoration Ecology 3:211?
224.
Whited, D., S. Galatowitsch, J. R. Tester, K. Schik, R. Lehtinen,
and J. Husveth. 2000. The importance of local and regional factors
in predicting effective conservation planning strategies for wet-
land bird communities in agricultural and urban landscapes. Land-
scape and Urban Planning 49:49?65.
Wilcox, D. A., J. E. Meeker, P. L. Hudson, B. J. Armitage, M. G.
Black, and D. G. Uzarski. 2002. Hydrologic variability and the
application of index of biotic integrity metrics to wetlands: a Great
Lakes evaluation. Wetlands 22:588?615.
Yasukawa, K. and W. A. Searcy. 1995. Red-winged Blackbird (Age-
laius phoeniceus). p. 1?28. In A. Poole and F. Gill (eds.) The
Birds of North America, No.184. The Birds of North America
Inc., Philadelphia, PA, USA.
Manuscript received 23 December 2003; revisions received 28 June
2004; accepted 22 August 2004.
DeLuca et al., MARSH BIRD COMMUNITY INTEGRITY 847
Appendix I. List of bird species detected during point count surveys in Chesapeake Bay marshes, including scores for each species
attribute used to calculate the total species score (SIMBCI) in IMBCI development.
Common name Scientific name
Forag-
ing
habitat
Nesting
sub-
strate
Migra-
tory
status
Breed-
ing
range SIMBCI
Least bittern*
Green heron*
Clapper rail*
Virginia rail*
Willet*
Ixbrychus exilis (Gmelin)
Butorides virescens (L.)
Rallus longirostris (Gmelin)
Rallus limicola (Viellot)
Catoptrophorus semipalmatus (Gmelin)
4
2.5
4
4
4
2.5
1
4
4
4
4
4
2.5
2.5
4
1
1
3
1
4
11.5
8.5
13.5
11.5
16
Chimney swift
Ruby-throated hummingbird
Eastern kingbird
American crow
Chaetura pelagica (L.)
Archilochus colubris (L.)
Tyrannus tyrannus (L.)
Corvus brachyrhynchos (Brehm)
1
1
2.5
1
1
1
1
1
4
4
4
1
2
2
1
1
8
8
8.5
4
Purple martin
Bank swallow
Tree swallow
Barn swallow
Marsh wren*
Pronge subis (L.)
Riparia riparia (L.)
Tachycineta bicolor (Vieillot)
Hirundo rustica (L.)
Cistothorus palustris (Wilson)
1
1
1
1
4
1
1
1
1
2.5
4
4
4
4
2.5
1
1
1
1
1
7
7
7
7
10
American robin
Gray catbird
European starling
Yellow warbler
Common yellowthroat
Turdus migratorius (L.)
Dumetella carolinensis (L.)
Sturnus vulgaris (L.)
Dendroica petechia (L.)
Geothlypis trichas (L.)
1
1
1
1
2.5
1
1
1
1
2.5
1
4
1
4
4
1
1
1
1
1
4
7
4
7
10
Indigo bunting
Chipping sparrow
Seaside sparrow*
Song sparrow
Passerina cyanea (L.)
Spizella passerina (Bechstein)
Ammodramus maritimus (Wilson)
Melospiza melodia (Wilson)
1
1
4
1
1
1
2.5
1
4
1
2.5
2.5
2
1
4
1
8
4
13
5.5
Swamp sparrow*
Northern cardinal
Brown-headed cowbird
Melospiza Georgiana (Latham)
Cardinalis cardinalis (L.)
Molothrus ater (Boddaert)
4
1
1
2.5
1
1
2.5
1
1
4
2
1
13
5
4
Red-winged blackbird
Common grackle
Boat-tailed grackle*
American goldfinch
Agelaius phoeniceus (L.)
Quiscalus quiscula (L.)
Quiscalus major (Vieillot)
Carduelis tristis (L.)
1
1
1
1
2.5
1
2.5
1
2.5
1
1
1
1
2
4
1
8.5
5
8.5
4
* Wetland obligates.