Landscape and Urban Planning 99 (2011) 115–122

Contents lists available at ScienceDirect

Landscape and Urban Planning

journal homepage: www.elsevier.com/locate/landurbplan

Flooding-induced landscape changes along dendritic stream networks

and implications for wildlife habitat
Humberto L. Perotto-Baldivieso a,∗ , X. Ben Wu a , Markus J. Peterson b , Fred E. Smeins a ,
Nova J. Silvy b , T. Wayne Schwertner c,1
a
Department of Ecosystem Science and Management, Texas A&M University, College Station, TX 77843-2138, United States
b
Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843-2258, United States
c
Texas Parks and Wildlife Department, Mason, TX 76856, United States

a r t i c l e i n f o a b s t r a c t

Article history: Severe low frequency natural disturbances along stream networks can substantially alter urban and rural
Received 19 February 2010 landscapes and impact habitat and population dynamics of wildlife species. In 1978, severe flooding along
Received in revised form 22 July 2010 the North Prong of the Medina River significantly altered the habitat for the Rio Grande wild turkey and
Accepted 19 September 2010
may have contributed to the decreased abundance of this species observed during recent decades in the
Available online 3 November 2010
southeastern Edwards Plateau, TX, USA. The objective of our study was to examine the changes in land-
scape structure caused by this flooding event and their potential impact on wild turkey habitat. Aerial
Keywords:
photography from 1972, 1984, and 1995 was used to quantify habitat changes in riparian zones and
Dendritic network
Flooding disturbance
adjacent bottomlands along the Medina River. We documented substantial reductions in habitat suit-
Habitat fragmentation ability and connectivity caused by the flooding, followed by a partial recovery over 17 years. Analysis
Landscape change using patch-level metrics in conjunction with class-level metrics, provided insights to the pattern and
Meleagris gallopavo intermedia possible mechanisms of the landscape changes. Habitat along higher-order streams was most affected,
Semi-arid landscapes reducing not only the suitable habitat locally, but also the habitat connectivity throughout the riparian
network. This loss of connectivity rendered numerous habitat patches along lower-order streams unavail-
able to Rio Grande wild turkeys as this species depends on riparian corridors for dispersal and movement
among habitat patches. Our results illustrate the critical importance of multiple-scale analysis based on
hierarchical dendritic structures of river networks when assessing habitat changes and their impact on
populations of terrestrial wildlife species dependent on riparian habitats in semi-arid landscapes.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction in rural and urban landscapes, provide evidence that connectivity

Spatial and temporal habitat changes are critical to the distri-
bution and abundance of wildlife (Fahrig, 1997, 2001). Changes
in the type, size, and spatial arrangement of habitat patches can
influence avian population dynamics (Ambuel and Temple, 1983;
Estades, 2001; Holmes and Sherry, 2001; Saveraid et al., 2001;
Stephens et al., 2003). Connectivity between habitat patches and
the presence and the quality of dispersal routes also influences
avian numbers (Van Dorp and Opdam, 1987). Recent studies, both
∗ Corresponding author. Present address: Natural Resources Department, Cran-
field University, Cranfield, Bedfordshire, MK43 0AL, United Kingdom.
Tel.: +44 01324 750111; fax: +44 01234 752970.
E-mail addresses: h.perotto@cranfield.ac.uk (H.L. Perotto-Baldivieso),
xbw@tamu.edu (X. Ben Wu), mpeterson@tamu.edu (M.J. Peterson),
f-smeins@tamu.edu (F.E. Smeins), n-silvy@tamu.edu (N.J. Silvy),
tschwertner@bio-west.com (T. Wayne Schwertner).
1
Present address: BIO-WEST, Inc., 1812 Central Commerce Court, Round Rock, TX
78664-8546, United States.
at multiple spatial scales is critical to avian diversity (Fernandez-
Jurisic, 2000; Viles and Rosier, 2001; Fitzsimons and Wescott, 2008;
Parker et al., 2008; Hepcan et al., 2009; Leibenath et al., 2010).
Changes in habitat structure are critical factors influencing wild
turkey (Meleagris gallopavo) population dynamics in the United
States (Lindzey and Wanless, 1973; Weinstein et al., 1995). Sev-
eral studies have addressed the relationship between landscape
spatial structure, especially woody vegetation, and the abundance
of eastern wild turkeys (M. g. silvestris). Thogmartin (1999) found
that woody patch size was positively related to nesting success
in Arkansas. Habitat factors such as topographic position, amount
of edge, and patch type (hardwood and mixed pine–hardwood
forest patches) also affect eastern wild turkey nesting success
(Thogmartin and Schaeffer, 2000; Thogmartin, 2001). Miller et al.
(1999) found habitat-use patterns were similar for males and
females across spatial scales in Mississippi. Chamberlain et al.
(2000) noted patch type (pine and hardwood sawtimber patches)
was key to eastern wild turkey habitat use, and the presence of
tall, mature tree stands was critical for roosting. Lack of suitable

0169-2046/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.landurbplan.2010.09.002
116 H.L. Perotto-Baldivieso et al. / Landscape and Urban Planning 99 (2011) 115–122
roosting areas could limit turkey distribution, particularly in semi-
arid regions that support few trees. These woody patches typically
were located near streams and rivers and were often used by wild
turkeys for nesting, dispersal, and roosting (Thomas et al., 1966;
Gore, 1973; Texas Parks and Wildlife Department, 1977; Palmer
and Hurst, 1995; Miller et al., 2000).
In the semi-arid regions of Texas (USA), winter roosting habi-
tat is particularly important for Rio Grande wild turkeys (M. g.
intermedia; hereafter ‘wild turkeys’) (Haucke, 1975). Roost-site
preservation is essential for maintaining wild turkey populations
and roost removal and/or disturbance has led to a reduction of
wild turkey numbers in Texas (Cook, 1973; Texas Parks and Wildlife
Department, 1995). Roosting sites are typically located near creeks,
rivers, and intermittent drainages and are composed primarily of
patches of large, tall trees with low-growing brush and herbaceous
vegetation, both under the roost trees and along the approach to
the roost (Thomas et al., 1966; Gore, 1973; Texas Parks and Wildlife
Department, 1977). These patterns were observed for streams and
drainage networks in Kansas (Capel, 1973; Hennen and Lutz, 2000),
Iowa (Wigal, 1973), and Oregon (Keegan and Crawford, 2000).
Wild turkey is a gregarious, nomadic bird (Glazener, 1967;
Beasom and Wilson, 1992). Its native range includes the states of
Kansas, Oklahoma, Texas, Northeastern New Mexico, and north-
ern Mexico. Prior to European settlement, perhaps 2 million
wild turkeys dwelled within their native ranges (Schorger, 1966;
Beasom and Wilson, 1992). By the end of the nineteenth century,
this subspecies had been extirpated from New Mexico, Oklahoma
and Kansas due to hunting and habitat changes, and approximately
100,000 turkeys remained in Texas (Walker, 1950). The first efforts
to restore wild turkey populations in Texas were taken by the state’s
legislature in 1881 (Beasom and Wilson, 1992). Over time, restock-
ing proved to be one of the most effective strategies (Beasom and
Wilson, 1992), with most birds originating directly or indirectly
from remnant flocks in semi-arid rangelands used for livestock
grazing from the Edwards Plateau and South Texas (Peterson et al.,
2002).
The Edwards Plateau of Texas traditionally was considered
excellent habitat for wild turkeys (Beasom and Wilson, 1992), and
directly or indirectly served as the source for most translocations of
this subspecies (Peterson et al., 2002). Beginning in the late 1970s,
however, wild turkey abundance declined in the southeastern por-
tion of the Plateau while it exhibited no long-term changes in
population trends elsewhere in this physiographic region (Randel
et al., 2005; Schaap et al., 2005; Collier et al., 2007; Randel et al.,
2007). The reasons for this decline remain unclear. Possible factors
include habitat changes (Hubbard et al., 1999), increased human
population (Beasom and Wilson, 1992), decreased availability of
foraging resources (Thogmartin, 2001), and predation, disease, and
natural disturbances (Peterson et al., 2002).
During the summer of 1978, remnants of tropical storm Amelia
produced a 500-year flood along the Sabinal, Guadalupe, and Med-
ina rivers, causing heavy loss of human life and property. During
this powerful, but rare storm, over 500 mm of rainfall in less than
24 h fed the headwaters of these rivers causing flashfloods. Massive
bald cypress (Taxodium distichum) up to 1.8 m in diameter at breast
height were pulled from the ground or snapped off along riparian
zones (Bomar, 1995). Because the flooded areas were within the
region where wild turkey abundance had declined since the late
1970s, it is possible that landscape changes caused by the flood may
have contributed to their decline. The flooding may have altered
the spatial configuration of woody vegetation suitable for roosting,
breeding, and dispersal along the streams and bottomlands associ-
ated with these rivers and potentially contributed to the decline in
wild turkey numbers.
The goal of this study was to quantify landscape changes that
resulted from the flooding of 1978 along the North Prong of the
Medina River and its tributaries, and to determine the poten-
tial impact on wild turkey habitat. Specific objectives were (1) to
quantify over time the amount and spatial configuration of woody
vegetation cover at 2 spatial scales: the bottomlands and riparian
zones of the North Prong of the Medina River and its tributaries, and
(2) to evaluate the changes occurred in woody vegetation before
and after the flooding of 1978 and their implications to wild turkey
habitat connectivity in dendritic stream networks. Our hypotheses
were that (1) the amount and spatial pattern of woody vegetation
were altered significantly due to the flood of 1978, and (2) suit-
able habitat was fragmented and connectivity reduced, resulting in
decreased overall habitat suitability of the area for wild turkeys.
2. Methods
2.1. Study area
The study area was located in semi-arid landscapes in south-
ern North America, along the watershed of the North Prong of
the Medina River near the boundary of Bandera and Kerr coun-
ties, TX, USA (Fig. 1). This area is located within the region where
there has been a decline in wild turkey abundance since the 1970s.
Detailed analysis and trends of wild turkey numbers have been
reported in previous studies (Randel et al., 2005, 2007; Schaap et al.,
2005; Collier et al., 2007). This area was comprised of privately held
ranches that supported livestock grazing and management to ben-
efit hunting for native and exotic wild species. The upper reaches
of the watershed are dominated by the Eckrant rock outcrop soil
association, which includes limestone derived soils that vary from
shallow stony (undulating) or rocky (steep) clays (NRCS, 2000). The
bottomlands are composed of 11 soil series, including deep loams,
clay loams, or loams. Woody vegetation primarily consists of pecan
(Carya illinoensis), Texas oak (Quercus buckleyi), shin oak (Q. sin-
uata), post oak (Q. stellata), live oak (Q. virginiana), Ashe juniper
(Juniper ashei), and bald cypress. Grasses include switchgrass (Pan-
icum virgatum), bluestem (Andropogon spp.), gramas (Bouteloua
spp.), Indiangrass (Sorghastrum nutans), curly mesquite (Hilaria
belangeri), and buffalograss (Buchloe dactyloides) (Van Auken, 1988;
Randel et al., 2005). Riparian forests are richer in woody species
composition and support trees that are taller and have larger basal
areas than other areas in the Edwards Plateau (Van Auken, 1988;
Texas Parks and Wildlife Department, 1991).
2.2. Data collection
We processed and classified available aerial photography from
1972 and 1984 (Aerial Photography Field Office, Farm Services
Agency, United States Department of Agriculture), and color
infrared digital ortho-quadrangles from 1995 (Texas Natural
Resources Information System), at 1-m resolution (scale 1:4800)
into 3 cover categories (woody, non-woody, water) using an
unsupervised classification in ERDAS 8.6. Overall classification
accuracies were 92 (1972), 93 (1984), and 93% (1995). We per-
formed analyses at 2 spatial scales pertinent to wild turkey ecology:
the bottomlands of the North Prong of the Medina River and its trib-
utaries where the impact of flooding on turkey habitat potentially
extended, and, in more detail, the riparian zones within the bot-
tomlands where roosting habitat typically was concentrated and
the affects of flooding the greatest.
2.3. Bottomland analysis
We defined bottomlands as areas below 585 m altitude and with
<12% slope based on a digital elevation model (national elevation
dataset) for the North Prong of the Medina River. We included a
 H.L. Perotto-Baldivieso et al. / Landscape and Urban Planning 99 (2011) 115–122
Fig. 1. Location of the North Prong of the Medina River and its tributaries, TX, USA. The dark grey represents the bottomland areas and the light grey corresponds to the
watershed.
60-m buffer surrounding all bottomland water bodies and an addi-
tional 200-m buffer, as bottomland edge habitat based on half the
average daily distance travelled by wild turkeys on our study areas
(Schaap, 2005). The bottomland section (including buffer areas) of
the watershed had an area of 2381 ha (21% of the watershed).
Using the bottomland classified images (1975, 1984, and 1995),
we performed a moving window analysis (diameter 400 m, step
50 m) for each period (Rho, 2003). For each moving window, we
used Fragstats (U.S. Forest Service, 1995) to calculate a set of
landscape metrics that would quantify landscape structure and
compare differences among the three time periods (Wu et al., 2000;
Perotto-Baldivieso et al., 2009). We used percentage of woody
vegetation to measure overall woody vegetation change in the
bottomlands, and mean patch size, patch density, edge density
(total edge length of a class per unit area), mean shape index,
area-weighted mean shape index (AWMSI), mean nearest neigh-
bor distance, and mean proximity index to compare the spatial
characteristics of woody vegetation patches (Gustafson et al., 1994;
U.S. Forest Service, 1995; McIntyre, 1995). We compared frequency
distributions for each landscape metric between years using the
117
Kolmogorov–Smirnov Z goodness of fit test with a 0.05 level of
significance. Patch-level shape index and patch density related to
patch size were compared among years using power (slope a, expo-
nent b, and regression coefficient r2 ), and inverse first order (slope
a and regression coefficient r2 ) regression curves respectively (Ott,
1992).
2.4. Riparian zone analysis
Using ArcView 3.2a (ESRI), we manually digitized streams based
on aerial photography, and classified them into stream orders.
Within each stream order, we randomly selected 400-m samples
and analyzed 18 first-order, 15 second-order, 11 third-order, and
3 fourth-order stream samples. For each sample, we created 50-
m buffers up to 200 m from the stream, and clipped the classified
images. For each clipped image, we calculated landscape metrics
that describe the spatial attributes of woody vegetation relevant to
turkey habitat; percent woody vegetation cover, mean patch size,
patch size standard deviation (PSSD), largest patch index, patch
density, edge density, mean nearest neighbor, and mean proxim-
118 H.L. Perotto-Baldivieso et al. / Landscape and Urban Planning 99 (2011) 115–122

50

Percent of total area

40

30

20

10

0
0 20 40 60 80 100 0.0 0.1 0.2 0.3 0.4 0.5 0.6
Percent of woody cover Mean patch size
25
Percent of total area

20

15

10

0
0 20 40 60 80 100 0 1000 2000 3000
Patch density (patches/ha) Edge density (m/ha)
40
Percent of total area

30

20

10

0
1.0 1.2 1.4 1.6 1.8 2.0 0 5 10 15 20 25
Mean shape index Area weighted mean shape index

50
Percent of total area

40

30

20

10

0
0 2 4 6 8 10 12 0 2000 4000 6000 8000 10000
Mean nearest neighbor distance (m) Mean proximity index

1972
1984
1995

Fig. 2. Frequency distributions of woody vegetation metrics for the bottomlands in 1972, 1984, and 1995 based on moving-window analysis for the North Prong of the
Medina River, TX, USA.

ity index (Gustafson et al., 1994; U.S. Forest Service, 1995; Lausch
and Herzog, 2002). Based on these metrics, we compared the differ-
ences in woody vegetation patterns between 1972, 1984, and 1995
for each stream order using Kruskal–Wallis test with a significance
level of 0.05.
3. Results
3.1. Bottomland analysis
The percent woody vegetation cover in the bottomlands did not
significantly differ by year. However, in 1972, 87% of the moving
windows had woody patches with a mean patch size ≤0.03 ha,
while in 1984, only 31% had a mean patch size ≤0.03 ha. Density
of woody patches also decreased in range and number between
1972 (24 patches/ha, range 136) and 1984 (6 patches/ha, range 35).
The frequency distribution of mean patch size and patch density in
1995 appears between those in 1972 and 1984 (Fig. 2).
Woody patch shape complexity in the bottomlands differed sig-
nificantly between years (Fig. 2). Edge density of woody patches in
the moving windows decreased between 1972 (mode = 1800 m/ha,
range = 3857) and 1984 (mode = 700 m/ha, range = 2044). Other
shape complexity metrics (mean shape index and AWMSI), how-
ever, exhibited trends opposite of those expected during this
 H.L. Perotto-Baldivieso et al. / Landscape and Urban Planning 99 (2011) 115–122
Fig. 3. Frequency distributions of patch shape index and patch density by patch size in the bottomland areas of the North Prong of the Medina River (TX, USA) for 1972, 1984,
and 1995.
period. For example, the frequency distributions for mean shape
index suggested an increase in woody vegetation complexity
between 1972 and 1984 (Fig. 2), but the significant decrease in small
woody patches could explain this difference. Similarly, AWMSI val-
ues decreased from 1972 (mode = 9, range = 26) to 1984 (mode = 4,
range = 17). It is likely that smoothing of edges in larger patches due
to flooding may have contributed to this difference. In 1995, the
frequency distributions for the complexity metrics (edge density,
mean shape index, and AWMSI) fell between 1972 and 1984 values.
The distance between neighboring patches increased between
1972 and 1984 (Fig. 2). Mode and range for mean nearest neigh-
bor increased between 1972 (2 m and 180 m, respectively) and
1984 (4 m and 233 m, respectively). Although the mean proximity
index mode was 500 for all periods, the proportion of bottom-
land with this value or lower changed from 19.3% (1972) to
23.6% (1984) and to 20.5% (1995). The frequency distributions for
the mean nearest neighbor and mean proximity index in 1995
were intermediate between those in 1972 and 1984. Shape index
and patch density regression parameters (slope a, the exponent
b, and regression coefficient r2 ) were higher in 1972 (a = 0.137,
b = 0.441, r2 = 0.88; a = 2295, r2 = 0.96, respectively) than in 1984
(a = 0.098, b = 0.399, r2 = 0.83; a = 536, r2 = 0.94, respectively), and
1995 (a = 0.141, b = 0.387, r2 = 0.84; a = 700, r2 = 0.95, respectively)
(Fig. 3).
119
3.2. Riparian zone analysis
Changes in percent woody vegetation cover between years var-
ied among stream orders (Fig. 4). The amount of woody vegetation
along fourth-order streams and within 50 m of third-order streams
decreased significantly after the flooding of 1978, while no signifi-
cant changes occurred in areas beyond 50 m of third-order streams
or along first- and second-order streams.
Spatial pattern of woody vegetation in riparian zones was sig-
nificantly different before and after the flooding of 1978. In 1972,
riparian zones had small woody patches in high densities for all
streams (Fig. 4). In 1984, mean woody patch size and patch size
variability (PSSD) increased, while the patch density decreased
as compared to 1972. Although no significant differences were
observed for largest patch index values in first- and second-order
streams, these significantly decreased in third- and fourth-order
streams. Changes in mean patch size and patch density in first and
second-order streams between 1972 and 1984 were due to a coa-
lescence of contiguous patches, whereas in third- and fourth-order
streams these changes accounted for decreased numbers of small
woody patches. In 1995, most spatial distribution metrics of woody
vegetation fell between the values found in 1972 and 1984 (Fig. 4).
Woody patch edge density and distance between neighbor-
ing woody patches changed along riparian zones between 1972
120 H.L. Perotto-Baldivieso et al. / Landscape and Urban Planning 99 (2011) 115–122

60

% woody
a a
a
a
40 a
b
b ab ab
b
20 b b b
b
c

0
0.3
MPS (Ha)

0.2
a ab a a a
0.1 a ab aa aa a a
b b b b ab b b b
b b b
0.0
a ab
0.6 a
ab
PSSD

a
0.4 b ab
b b
0.2 a a
bb bb
0.0
a
50
a
40
LPI

a
30
20 bb a
ab
10 bb bb b
0
60 a
a a
a a a a a a a a a
40 a
PD

a a
a a ab ab
b ab
20 b b b b b
b b b b bb bb b b b b
b b b b b b b b b b

0
2500 a
ED (m/ha)

a
a a a a a
a a a a a a
2000 a a a
1500 b b b b
b b b b b b b b
b b b b b b b b
b b b
1000 b b b b b
c b b
500 c

0
10
MNN (m)

8 a a
a a
6 a
a a a a a a a b a b
a a a
ab b ab ab ab b b a b
4 b
b b b b b b b b b c
c c c c c b c
2
0
15000
MPI

10000 a
a a
a
5000 a
bb
bb bb bb bb
0
50 100 150 200 50 100 150 200 50 100 150 200 50 100 150 200
Stream order 1 Stream order 2 Stream order 3 Stream order 4

1972 1984 1995

Fig. 4. Riparian zones’ woody-cover metrics [percent woody vegetation cover (% woody), mean patch size (MPS), patch size standard deviation (PSSD), largest patch index
(LPI), patch density (PD), edge density (ED), mean nearest neighbor (MNN), mean proximity index (MPI)] by distance from stream and stream order, in 1972, 1984, and 1995
for the North Prong of the Medina River, TX, USA. Error bars represent the standard error, and letters (a, b, c) indicate statistical differences with a significance level of 0.05.

and 1984 (Fig. 4). Edge densities before the flooding (1972; 4. Discussion
x̄ = 1750 m/ha) were greater than after the flooding (1984; x̄ =
727 m/ha) for all stream orders. The decrease in edge density Although anthropogenic changes often are key factors influ-
was greater in third- and fourth-order streams as compared to encing wildlife habitat and populations, low-frequency natural
first- and second-order streams. Distance between neighboring disturbances, such as severe flooding events, can significantly
patches increased for all stream orders between 1972 and 1984. impact wildlife habitat over large areas and lead to long-term
For example, mean nearest neighbor values increased by 2 m in changes in wildlife abundance. Flooding associated with trop-
first-, second-, and third-order streams, and by 4 m in fourth-order ical storm Amelia in 1978 significantly altered the landscape
streams. There were no significant changes in mean proximity structure along the North Prong of the Medina River in Texas,
indices between 1972 and 1984 for first- and second-order streams, USA. Fragmentation of woody vegetation along the riparian areas
and areas beyond 50 m of third-order streams, but mean proximity resulted in a substantial reduction of roosting, breeding, and feed-
indices significantly decreased in areas within 50 m of third- ing habitats that were critical to sustaining wild turkey populations
order streams and fourth-order streams (82.5 and 96.1% decreases, (Cook, 1973; Haucke, 1975). Travel ways along the stream net-
respectively). work, another important habitat element (Texas Parks and Wildlife
 H.L. Perotto-Baldivieso et al. / Landscape and Urban Planning 99 (2011) 115–122
Department, 1977), also were reduced, especially along higher-
order streams, which limited the connectivity between suitable
habitat areas. These changes were likely to be responsible, at least
in part, for the decline in wild turkey abundance in the region. Stud-
ies of eastern wild turkeys also have shown the effect of flooding
on reproduction, habitat use, and demography (Cobb et al., 1993;
Cobb and Doerr, 1997), although these researchers did not explore
the impact of flooding on spatial pattern and connectivity of the
habitat.
Even though reductions in habitat area and connectivity due
to the flooding were substantial, it appeared that system resilience
led to a partial recovery of woody vegetation in the bottomland and
riparian areas. Seventeen years after the flooding (1995), new small
woody patches emerged in areas significantly impacted (higher-
order streams) by the 1978 flood. Connectivity between suitable
habitat areas also increased between 1984 and 1995, but still
remained lower than in 1972. These changes suggest improvement
in wild turkey habitat since the flooding in 1978, but it is too soon
to determine whether they will lead to the recovery of wild turkey
numbers in this region.
Class- and landscape-level metrics (U.S. Forest Service, 1995;
McGarigal, 2002) are effective for quantifying landscape changes
and are most often used in landscape studies (Gustafson, 1998).
Analyses based on patch-level metrics, however, can be impor-
tant or even critical in understanding the mechanisms of landscape
change. We demonstrated that after the 1978 flood, woody vegeta-
tion mean shape index increased possibly due to the loss of many
small patches. We expected an increase in the area-weighted mean
shape index with the reduction of small patches, but instead it
decreased. Further analyses based on patch-level metrics revealed
the flooding had a significant impact at the patch scale by smooth-
ing the edges of large patches, which lowered the shape index, in
addition to the landscape-scale impact of reducing the density of
small patches (Fig. 3). The elimination of most of the small woody
patches and making the shapes of larger woody patches more com-
pact (loosing irregular extensions) along the higher order streams
effectively reduced the habitat connectivity for the wild turkeys.
These additional analyses, based on patch-level data, led to a bet-
ter understanding of how the landscape and turkey habitat were
affected by the flooding that could not be gained from analyses
using class- and landscape-level metrics alone.
Riparian areas, as network structures, serve as critical move-
ment and dispersal corridors for upland wildlife species within
semi-arid landscapes. Although wild turkey habitat extends well
beyond riparian areas, they depend upon riparian corridors for dis-
persal and movement among habitat patches. We found suitable
habitat areas for wild turkeys in the upper watershed (low order
streams) remained largely intact after a 500-year flood in 1978,
but flooding caused significant loss of suitable habitat in the lower
part of the watershed (high order streams). This differential spa-
tial distribution of disturbance within the stream network greatly
reduced habitat connectivity for wild turkeys by rendering intact
riparian habitat (along low order streams) inaccessible. Previous
studies have shown the value of hydrologic connectivity for ecolog-
ical integrity of the landscape and the role of aquatic populations on
terrestrial fauna (Pringle, 2003); spatial pattern and composition
richness at local and regional scales (Renofalt et al., 2005); vege-
tation structure (Sponseller and Fischer, 2006); refugia, breeding,
and dispersal between upstream and downstream forest fragments
for riparian/forest specialist (Rykken et al., 2007); links between
stream hierarchies and land use hierarchies (Huang et al., 2007);
and local gradients on banded damselflies (Calopteryx splendens)
morphology (Chaput-Bardy et al., 2007). Geometric and metapopu-
lation models have also been used to highlight how a small amount
of fragmentation can significantly reduce dispersal and survival of
aquatic species in dendritic landscapes such as river creek systems
121
(Fagan, 2002). Our study demonstrates the importance of consider-
ing hierarchical dendritic structures, such as river networks, for the
analysis of vegetation fragmentation and its impact on terrestrial
species that depend on riparian areas for survival and dispersion
in semi-arid landscapes. While local habitats in the upper areas of
the watershed remained adequate, lower areas did not provide suit-
able conditions for the species and reduced the overall suitability at
the landscape scale, which likely contributed to the decline in wild
turkey numbers. Severe disturbances, such as the 500-year flood
of 1978, can pose significant challenges to wildlife and land man-
agers. Although limited aerial photography data prevented direct
examination of the correlation between wild turkey abundance in
the area and habitat changes associated with the flooding events,
spatial analysis of habitat changes based on well-documented habi-
tat requirements in this study enabled exploration of the impact of
the severe flooding on habitat availability and connectivity. Our
study illustrates the importance of riparian networks both at local
and at broader scales (i.e. watersheds) to ensure adequate habi-
tat for wild turkey. It also highlights the broader implication that
considerations at multiple spatial scales and of connectivity over
heterogeneous landscapes, as well as longer term planning cog-
nizant of the impact of low frequency events, are essential for
successful wildlife conservation and management efforts.
5. Conclusion
Low frequency natural disturbances, such as severe flooding
events (500-year flood), can be major factors influencing long-term
changes in habitat and population dynamics of wildlife species at
landscape to regional scales in semi-arid landscapes. Landscape
changes caused by flooding associated with the tropical storm
Amelia in 1978 substantially reduced the suitable habitat along
higher-order streams and the accessibility of the suitable habi-
tat along the lower-order streams. This likely contributed to the
decline in wild turkey abundance in the southeastern Edwards
Plateau of Texas; the region that once supported a source popu-
lation for turkey translocations across the continent. The impact
of flooding was long lasting, and habitat only recovered par-
tially after 17 years. Although class- and landscape-level metrics
are often used and effective in studying the pattern of land-
scape changes associated with disturbances, the use of patch-level
metrics can be critical to better understanding the pattern and pos-
sible mechanisms of the landscape changes. Local habitat changes
along higher-order streams can reduce the habitat connectivity
at landscape scales and render otherwise abundant habitat along
lower-order streams unavailable for terrestrial species that depend
on riparian areas for survival and dispersion. Multiple-scale analy-
sis based on the hierarchical dendritic structures of river networks
is essential in assessing habitat changes in semi-arid landscapes
and their impact on terrestrial species populations.
Acknowledgments
Financial support for this project was provided by the Turkey
Stamp Fund of Texas Parks and Wildlife Department and Texas
A&M University College of Agriculture and Life Sciences through
a Tom Slick Fellowship to Perotto-Baldivieso. This manuscript was
improved by comments from E.B. Hollister, H. Hood, K.B. Melton,
J.N. Schaap, M.E. Simmons, and 3 anonymous reviewers.
References
Ambuel, B., Temple, S.A., 1983. Area-dependent changes in bird communities and
vegetation of southern Wisconsin forests. Ecology 64, 1057–1068.
Beasom, S.L., Wilson, D., 1992. Rio Grande turkey. In: Dickson, J.G. (Ed.), The Wild
Turkey: Biology and Management. Stackpole Books, Harrisburg, PA, pp. 306–330.
Bomar, G.W., 1995. Texas Weather. University of Texas Press, Austin, TX.
122 H.L. Perotto-Baldivieso et al. / Landscape and Urban Planning 99 (2011) 115–122
Capel, S.W., 1973. Introduction of Rio Grande turkeys into Kansas. In: Sanderson, G.C.,
Schultz, H.C. (Eds.), Wild Turkey Management: Current Problems and Programs.
University of Missouri Press, Columbia, MO, pp. 11–17.
Chamberlain, M.J., Leopold, B.D., Burger, L.W., 2000. Characteristics of roost sites of
adult wild turkey females. Journal of Wildlife Management 64, 1025–1032.
Chaput-Bardy, A., Pays, O., Lodé, T., Secondi, J., 2007. Morphological clines in den-
dritic landscapes. Freshwater Biology 52, 1677–1688.
Cobb, D.T., Doerr, P.D., 1997. Eastern wild turkey reproduction in an area subjected
to flooding. Journal of Wildlife Management 61, 313–317.
Cobb, D.T., Doerr, P.D., Seamaster, M.H., 1993. Habitat use and demography of a
wild turkey population subjected to human-induced flooding. In: Proceedings
of the Annual Conferences of the Southeastern Association of Fish and Wildlife
Agencies, vol. 47, pp. 148–162.
Collier, B.A., Jones, D.A., Schaap, J.N., Randel, C.J., Willsey, B.J., Aguirre, R., Schwertner,
T.W., Silvy, N.J., Peterson, M.J., 2007. Survival of Rio Grande wild turkeys on the
Edwards Plateau of Texas. Journal of Wildlife Management 71, 82–86.
Cook, R.L., 1973. A census technique for the Rio Grande turkey. In: Sanderson, G.C.,
Schultz, H.C. (Eds.), Wild Turkey Management: Current Problems and Programs.
University of Missouri Press, Columbia, MO, pp. 279–283.
Estades, C.F., 2001. The effect of breeding-habitat patch size on bird population
density. Landscape Ecology 16, 161–173.
Fagan, W.F., 2002. Connectivity, fragmentation, and extinction risk in dendritic
metapopulations. Ecology 88, 3243–3249.
Fahrig, L., 1997. Relative effects of habitat loss and fragmentation on population
extinction. Journal of Wildlife Management 61, 603–610.
Fahrig, L., 2001. How much habitat is enough? Biological Conservation 100, 65–74.
Fernandez-Jurisic, E., 2000. Avifaunal use of wooded streets in an urban landscape.
Conservation Biology 14, 513–521.
Fitzsimons, J.A., Wescott, G., 2008. The role of multi-tenure reserve networks in
improving reserve design and connectivity. Landscape and Urban Planning 85,
163–173.
Glazener, W.C., 1967. Management of Rio Grande turkey. In: Hewitt, O.H. (Ed.), Wild
Turkey and its Management. The Wildlife Society, Washington, DC, pp. 453–492.
Gore, H.G., 1973. Land-use practices and Rio Grande turkeys in Texas. In: Sander-
son, G.C., Schultz, H.C. (Eds.), Wild Turkey Management: Current Problems and
Programs. University of Missouri Press, Columbia, MO, pp. 253–263.
Gustafson, E.J., 1998. Quantifying landscape spatial pattern: what is the state of the
art? Ecosystems 1, 143–156.
Gustafson, E.J., Parker, G.R., Backs, S.E., 1994. Evaluating spatial pattern of wildlife
habitat: a case study of the wild turkey (Meleagris gallopavo). American Midland
Naturalist 131, 24–33.
Haucke, H.H., 1975. Winter roost characteristics of the Rio Grande wild turkey in
South Texas. In: Proceedings of the National Wild Turkey Symposium, vol. 3, pp.
164–169.
Hennen, R., Lutz, S., 2000. Rio Grande wild turkey female survival in southcentral
Kansas. In: Proceedings of the National Wild Turkey Symposium, vol. 8, pp.
117–122.
Hepcan, Ş., Hepcan, C.C., Bouwma, I.M., Jongman, R.H.G., Özkan, M.B., 2009. Ecologi-
cal networks as a new approach for nature conservation in Turkey: a case study
o Izmir Province. Landscape and Urban Planning 90, 143–154.
Holmes, R.T., Sherry, T.W., 2001. Thirty-year bird population in an unfragmented
temperate deciduous forest: importance of habitat change. Auk 118, 589–609.
Huang, S.L., Budd, W.W., Chan, S.L., Lin, Y.C., 2007. Stream order, hierarchy, and
energy convergence of land use. Ecological Modelling 205, 255–264.
Hubbard, M.W., Garner, D.L., Klaas, E.E., 1999. Factors influencing wild turkey hen
survival in southcentral Iowa. Journal of Wildlife Management 63, 731–738.
Keegan, T.W., Crawford, J.A., 2000. Seasonal habitat use and home ranges of Rio
Grande turkeys in Oregon. In: Proceedings of the National Wild Turkey Sympo-
sium, vol. 8, pp. 109–116.
Lausch, A., Herzog, F., 2002. Applicability of landscape metrics for the monitoring
of landscape change: issues of scale, resolution and interpretability. Ecological
Indicators 2, 3–15.
Leibenath, M., Blum, A., Stutzriemer, S., 2010. Transboundary cooperation in estab-
lishing ecological networks: the case of Germany’s external borders. Landscape
and Urban Planning 94, 84–93.
Lindzey, J.S., Wanless, D.D., 1973. Problems in wild turkey management and
research. In: Sanderson, G.C., Schultz, H.C. (Eds.), Wild Turkey Management:
Current Problems and Programs. University of Missouri Press, Columbia, MO,
pp. 229–232.
McGarigal, K., 2002. Landscape pattern metrics. In: El-Shaarawi, A.H., Piegorsch,
W.W. (Eds.), Encyclopedia of Environmentrics, vol. 2. John Wiley & Sons, Sussex,
England, pp. 1135–1142.
McIntyre, N.E., 1995. Effects of forest patch size on avian diversity. Landscape Ecol-
ogy 10, 85–99.
Miller, D.A., Hurst, G.A., Leopold, B.D., 1999. Habitat use of eastern wild turkey in
Central Mississippi. Journal of Wildlife Management 63, 210–222.
Miller, D.A., Leopold, B.D., Hurst, G.A., Gerard, P.D., 2000. Habitat selection models
for eastern wild turkey in central Mississippi. Journal of Wildlife Management
64, 765–776.
NRCS, 2000. Bandera and Kerr Counties Soil Reports. SSURGO V2.2000. Retrieved
from http://www.ftw.nrcs.usda.gov/ssur data.html.
Ott, R.L., 1992. An Introduction to Statistical Methods and Data Analysis. Duxbury
Press, Belmont, CA.
Parker, K., Head, L., Chisholm, L.A., Feneley, N., 2008. A conceptual model of ecolog-
ical connectivity in the Shellharbour local government area, New South Wales,
Australia. Landscape and Urban Planning 86, 47–59.
Palmer, W.E., Hurst, G.A., 1995. Drainage systems as minimum habitat manage-
ment units for wild turkey hens. In: Proceedings of the National Wild Turkey
Symposium, vol. 7, pp. 97–101.
Peterson, M.J., Aguirre, R., Ferro, P.J., Jones, D.A., Lawyer, T.A., Peterson, M.N., Silvy,
N.J., 2002. Infectious disease survey of Rio Grande wild turkey in the Edwards
Plateau of Texas. Journal of Wildlife Diseases 38, 826–833.
Perotto-Baldivieso, H.L., Meléndez-Ackerman, E., García, M.A., Leimgruber, P.,
Cooper, S.M., Martínez, A., Calle, P., Ramos Gónzales, O.M., Quiñones, M., Chris-
ten, C.A., Pons, G., 2009. Spatial distribution, connectivity, and the influence of
scale: habitat availability for the endangered Mona Island rock iguana. Biodiver-
sity and Conservation 18, 905–917.
Pringle, C., 2003. What is hydrologic connectivity and why is it ecologically impor-
tant? Hydrological Processes 17, 2685–2689.
Randel, C.J., Aguirre, R., Jones, D.A., Schaap, J.N., Willsey, B.J., Peterson, M.J., Silvy, N.J.,
2005. Nesting ecology of Rio Grande wild turkey in the Edwards Plateau of Texas.
In: Proceedings of the National Wild Turkey Symposium, vol. 9, pp. 237–243.
Randel, C.J., Aguirre, R., Peterson, M.J., Silvy, N.J., 2007. Invertebrate abundance at
Rio Grande wild turkey brood locations. Journal of Wildlife Management 71,
2417–2420.
Renofalt, B.M., Nilsson, C., Jansson, R., 2005. Spatial and temporal patterns of species
richness in a riparian landscape. Journal of Biogeography 32, 2025–2037.
Rho, P., 2003. GIS-based multiple-scale study on scale quail. Ph.D. Dissertation. Texas
A&M University, College Station, TX.
Rykken, J.J., Moldenke, A.R., Olson, D.H., 2007. Headwater riparian forest-floor inver-
tebrate communities associated with alternative forest management practices.
Ecological Applications 17, 1168–1183.
Saveraid, E.H., Debinski, D.M., Kindscher, K., Jakubauskas, M.E., 2001. A comparison
of satellite data and landscape variables in predicting bird species occurrences
in the Greater Yellowstone Ecosystem, USA. Landscape Ecology 16, 71–83.
Schaap, J.N., 2005. Ranges, movements and spatial distribution of radio-tagged Rio
Grande wild turkeys in the Edwards Plateau of Texas. M.Sc. Thesis. Texas A&M
University, College Station, TX.
Schaap, J.N., Silvy, N.J., Peterson, M.J., Aguirre, R., Perotto-Baldivieso, H.L., 2005.
Spatial distribution of female Rio Grande wild turkeys during the reproduc-
tive season. In: Proceedings of the National Wild Turkey Symposium, vol. 9, pp.
231–235.
Schorger, A.W., 1966. The Wild Turkey: Its History and Domestication. University of
Oklahoma Press, Norman, OK.
Sponseller, R.A., Fischer, S.G., 2006. Drainage size, stream intermittency, and ecosys-
tem function in a Sonoran desert landscape. Ecosystems 9, 344–356.
Stephens, S.E., Koons, D.N., Rotella, J.J., Willey, D.W., 2003. Effects of habitat frag-
mentation on avian nesting success: a review of the evidence at multiple scales.
Biological Conservation 115, 101–110.
Texas Parks and Wildlife Department, 1977. Food Habits of the Rio Grande Wild
Turkey in the Permian Basin of Texas. Technical Series 18. Austin, TX.
Texas Parks and Wildlife Department, 1991. Vegetation of the Kerr Wildlife Man-
agement Area and its Preference by White-tailed Deer. Federal Aid Report 30.
Austin, TX.
Texas Parks and Wildlife Department, 1995. Rio Grande Turkey Habitat Manage-
ment. PWD RP W7100-263. Austin, TX.
Thogmartin, W.E., 1999. Landscape attributes and nest-site selection in wild turkeys.
Auk 116, 912–923.
Thogmartin, W.E., 2001. Home-range size and habitat selection of female wild
turkeys (Meleagris gallopavo) in Arkansas. American Midland Naturalist 145,
247–260.
Thogmartin, W.E., Schaeffer, B.A., 2000. Landscape attributes associated with mor-
tality events of wild turkeys in Arkansas. Wildlife Society Bulletin 28, 865–874.
Thomas, J.W., Van Hoozer, C., Marburger, R.G., 1966. Wintering concentrations and
seasonal shifts in range in the Rio Grande turkey. Journal of Wildlife Manage-
ment 30, 34–49.
U.S. Forest Service, 1995. FRAGSTATS Spatial Pattern Analysis Program for Quanti-
fying Landscape Structure. Technical Report PNW-GTR-351. Portland, OR.
Van Auken, O.W., 1988. Woody vegetation of the southeastern escarpment and
Plateau. In: Amos, B.B., Ghelbach, F.R. (Eds.), Edwards Plateau Vegetation. Baylor
University Press, Waco, TX, pp. 43–56.
Van Dorp, D., Opdam, P., 1987. Effects of patch size, isolation and regional abundance
on forest bird communities. Landscape Ecology 1, 59–73.
Viles, R.L., Rosier, D.J., 2001. How to use roads in the creation of greenways: case
studies in three New Zealand landscapes. Landscape and Urban Planning 55,
15–27.
Walker, E.A., 1950. Tips on wild turkey. Texas Game and Fish 9 (2–3), 28.
Weinstein, M., Miller, D.A., Connor, L.M., Leopold, B.D., Hurst, G.A., 1995. What affects
turkeys? A conceptual model for future research. In: Proceedings of the National
Wild Turkey Symposium, vol. 7, pp. 135–142.
Wigal, D.D., 1973. Status of the introduced Rio Grande turkey in northeastern Iowa.
In: Sanderson, G.C., Schultz, H.C. (Eds.), Wild Turkey Management: Current Prob-
lems and Programs. University of Missouri Press, Columbia, MO, pp. 35–43.
Wu, X.B., Thurow, T.L., Whisenant, S.G., 2000. Fragmentation and changes in hydro-
logic function of tiger bush landscapes, south west Niger. Journal of Ecology 88,
790–800.