143
Influence of spatial positioning within stream
networks on fish assemblage structure in the
Kansas River basin, USA
Darren J. Thornbrugh and Keith B. Gido
Abstract: We found that riverine confluences had localized effects (within 20 km) on stream fish assemblages of the Kansas River basin. The majority of variation in fish assemblages occurred from east to west and along a stream size gradient.
After controlling for the influences of longitude and stream size, distance of sample sites from streams ‡ 5th order accounted for a small proportion of taxonomic variability. However, species richness was significantly higher and assemblage structure was different in tributary stream segments directly connected to larger-ordered streams, suggesting that the
effects of spatial position within this stream network were greatest in close proximity to tributary–mainstem confluences.
Fish collections from three intensively sampled tributaries also indicated an abrupt change in species assemblages between
mainstem river sites and tributary sample sites above confluences, followed by a gradual taxonomic change with increasing
distance up to 20 km from the confluence. Changes in fish assemblages were associated with the reduced abundance of
adult stream species near the confluence with the mainstem, rather than the occurrences of riverine species in the tributary.
Résumé : Les confluences des cours d’eau ont des effets bien localisés (sur une distance de moins de 20 km) sur les peuplements de poissons d’eau courante dans le bassin versant de la rivière Kansas. La plus grande partie de la variation dans
les peuplements de poissons se produit d’est en ouest et suivant un gradient de taille des cours d’eau. Une fois qu’on a
tenu compte des influences de la longitude et de la taille des cours d’eau, la distance entre le point d’échantillonnage et
les cours d’eau d’ordre ‡ 5 explique un petit pourcentage de la variabilité taxonomique. Cependant, la richesse spécifique
est significativement plus élevée et la structure du peuplement différente dans les sections de tributaires qui sont reliées directement à des cours d’eau d’ordre supérieur, ce qui indique que les effets de la position spatiale dans le réseau hydrographique sont maximaux près des confluences des tributaires et du cours principal. Des récoltes de poissons dans trois
tributaires fortement échantillonnés montrent aussi un changement abrupt dans les peuplements de poissons entre les sites
du cours principal de la rivière et les sites d’échantillonnage des tributaires en amont de la confluence, suivi d’un changement taxonomique graduel en fonction de la distance pouvant se manifester jusqu’à 20 km de la confluence. Les changements dans les peuplements de poissons s’expliquent par une abondance réduite d’espèces adultes de ruisseau près de la
confluence avec le cours principal plutôt que par la présence d’espèces de rivière dans le tributaire.
[Traduit par la Rédaction]
Introduction
Stream fish assemblages are influenced by the availability
of aquatic habitats across a variety of spatial scales and the
ability of individuals to move among those habitats
(Schlosser 1982; Angermeier and Winston 1998; Grossman
et al. 1998). Accordingly, measures of local fish assemblage
structure are often predicted by the species composition of
the surrounding drainage basin, as well as the position of
that locality within the drainage (Osborne and Wiley 1992;
Matthews and Robison 1998; Schaefer and Kerfoot 2004).
Such spatial dependencies of fish populations and communities among stream habitats (e.g., streams of different size or
basins with different geology) have been conceptualized by
several authors. Recently, Fausch et al. (2002) developed
the riverscapes perspective, which posits that dendritic
stream networks cannot be viewed in the context of disjunct
parts but should be considered in their entire interconnected
network architecture. Campbell-Grant et al. (2007) further
emphasized the importance of the structure of dendritic networks in shaping biotic assemblages and highlighted the importance of movement among branches and the
heterogeneity of habitats associated with a branch–node
structure. These studies are important because they conceptualize how regional factors operating at spatial scales of
tens to hundreds of kilometres can influence local assemblage structure, which is typically measured at spatial scales
of <0.5 km. Despite this increasing recognition that landscape and stream network architecture influence fish assemblages, empirical data on the relative importance of
Received 24 December 2008. Accepted 23 September 2009. Published on the NRC Research Press Web site at cjfas.nrc.ca on
11 December 2009.
J20956
D.J. Thornbrugh1 and K.B. Gido. Kansas State University, Division of Biology, Manhattan, KS 66506, USA.
1Corresponding
author (e-mail: dthor@msu.edu).
Can. J. Fish. Aquat. Sci. 67: 143–156 (2010)
doi:10.1139/F09-169
Published by NRC Research Press
144
spatial position of a habitat within a dendritic stream network and its connectivity to other habitats is limited (Benda
et al. 2004a; Lowe et al. 2006; Campbell-Grant et al. 2007).
Below, we review studies that address these issues and describe an experiment to test their importance in a prairie
stream network.
Physical structuring of stream habitats and associated
changes in biota are best documented along gradients of
stream size (Horwitz 1978; Vannote et al. 1980; Matthews
1998). However, because dendritic stream networks include
abrupt transitions in stream size at tributary confluences,
these habitats are likely important in mediating the exchange
of biota among branches of these networks (Benda et al.
2004b; Lowe et al. 2006; Campbell-Grant et al. 2007).
Whereas some species may favor habitats that vary with
stream size, it is likely that many stream organisms move
among tributary and mainstem habitats at different life
stages for resource use and refugia (Schlosser and Angermeier 1995; Fausch et al. 2002). Landscape complementation
or supplementation may occur around these tributary confluences because desired resources (e.g., food, spawning
habitat, refugia) in different habitat patches are in close
proximity, thus allowing for larger populations to be supported, increased species richness, and potentially longer
persistence of species (Dunning et al. 1992). According to
the above-mentioned studies, we predict that dispersal patterns of stream fishes should vary along gradients of stream
size and movement through confluence zones should be influenced by stream network architecture (Figs. 1a and 1b).
We define confluence zones as sections of tributary streams
within the floodplain or an adjoining river that has a notably
different habitat than sections further upstream of the confluence. These tributary habitats near mainstem rivers typically have lower gradient and wider channels that result in
a reduction in the transport of sediment and wood (Benda et
al. 2004a).
Spatial positioning within a stream network can affect immigration and extinction risk (Fagan 2002; Campbell-Grant
et al. 2007). For example, fishes in prairie rivers have
greater colonization probabilities from surrounding habitats
in downstream sites and higher extinction risks in upstream
or more isolated sites (Gotelli and Taylor 1999; Franssen et
al. 2006). Small adventitious streams (site 3 in Fig. 1b) may
be isolated from other tributaries if there is little movement
of fishes through large rivers, resulting in lower species
richness in these streams. Colonization opportunities are
also influenced by proximity to unique habitats. For example, Hitt and Angermeier (2008) concluded that proximity
to mainstem river confluences influenced tributary fish assemblage structure in mid- to large stream sites (i.e., basin
areas > 1000 ha) and extended up to 20 stream km (skm)
from the confluence. However, assemblages at sites on the
smallest streams (i.e., basin areas < 1000 ha) were not related to stream network topology. Within medium streams
(i.e., basin area 1000 ha – 5000 ha), proximity to a confluence increased total species richness, river species richness,
catostomid species richness, and cyprinid species richness
for up to 20 skm, but this effect was greatest near mainstem
rivers. Smith and Kraft (2005) also found that confluence
link (watershed size of stream segments immediately downstream of sites) was a significant predictor of fish assem-
Can. J. Fish. Aquat. Sci. Vol. 67, 2010
Fig. 1. Conceptual diagram representing (a) hypothesized movement patterns and (b) levels of connectivity within stream networks. In (a), arrows represent species or size classes of fish and
their potential movements through confluence zones and among
streams of different size. In (b), sites 1, 2, and 3 are examples of
stream habitats that vary in stream size (1 and 3 versus 2) and connectivity (1 versus 3) to a mainstem river.
(a) Movement patterns
Mainstem
Confluence
Medium
tributary
Confluence
Small
tributary
Limited exchange
Confluence exchange
Network dispersal
(b) Levels of connectivity with mainstem
blage structure in the Beaverkill–Willowemoc watershed in
New York. Similarly, Falke and Gido (2006a) showed that
confluence sites between streams and reservoirs (i.e., stream
sites within the floodplain of the reservoir) had higher total
nonnative and reservoir species richness than sites 4–16 km
upstream of those confluences, suggesting a localized effect
of reservoirs. Close proximity to a large river also can influence variability of stream fish assemblages, as sites closer to
interface zones were found to be more temporally variable
in an adventitious Illinois stream than sites more distant to
those interfaces (Schaefer and Kerfoot 2004). Gorman
(1986) presented arguments that fish assemblages on tributary streams within close proximity to larger mainstem
streams may be strongly influenced by the discontinuity in
the stream continuum and the adventitious nature of the
streams spatial position in the stream network. This adventitious nature of some streams is an ideal example of mixedhabitat patches within close proximity to mainstem rivers
(Dunning et al. 1992).
Our study focused on patterns of fish species richness,
abundance and size structure in prairie streams in relation
to proximity to mainstem rivers (i.e., stream order ‡ 5th),
and the connectance to streams of different sizes. We defined proximity as distance (along the stream channel) between a sample site on a tributary and the junction of that
Published by NRC Research Press
Thornbrugh and Gido
145
Fig. 2. (Upper panel) Fish collection sites by the Kansas Department of Wildlife and Parks (KDWP) during their annual stream monitoring
and assessment survey of the Kansas River basin during summer in 1995–2006. (Lower panel) Site locations within three intensively monitored tributary streams (Clarks Creek, McDowell Creek, and Wildcat Creek).
tributary with a mainstem river. Associations between proximity and fish assemblage structure were tested at both large
(up to 200 km) and small (up to 20 km) spatial scales. At
larger spatial scales (across stream segments), fish assemblages within tributaries are predicted to diverge from mainstem river assemblage with increasing distance from the
mainstem as a result of decreasing probabilities of biotic exchange (i.e., network dispersal; Fig. 1a). However, biotic exchange may attenuate rapidly as stream size decreases and
habitat (e.g., shallow riffle depth) limits movement. Within
stream segments connected to mainstem rivers, we predicted
that there would be an abrupt change in fish assemblage
structure between tributary sites and sites on mainstem rivers because of rapid changes in habitat in the confluence
zones. Finally, we predicted that headwater tributaries (site
1 in Fig. 1b) would have fewer riverine species and overall
lower species richness than mainstem tributaries (site 3 in
Fig. 1b) because of reduced biotic exchanges with mainstem
rivers (Osborne and Wiley 1992).
Considering the physical and ecological processes associated with stream networks, we hypothesize that stream
fishes exhibit the following movement patterns: (i) limited
movement — fishes are restricted to large rivers or small
streams because of habitat or resource limitation; (ii) confluence exchange — fishes move between large or small
streams and confluences but do not move through confluence boundaries; and (iii) network dispersal — fishes move
through confluence zones but the probability of moving
Published by NRC Research Press
146
Can. J. Fish. Aquat. Sci. Vol. 67, 2010
Table 1. Descriptions of habitat variables used to quantify stream size or connectivity within the Kansas
basin, Kansas.
Habitat variable
Stream order
Downstream order
Distance
Basin area
Variable description
Strahler order of stream segment (order)
Strahler order of downstream segment (dorder)
Distance to downstream mainstem (stream km)
Upstream catchment area of stream segment (km2)
Fig. 3. Number of species collected at sites as a function of basin
area for eastern (Big Blue and Kansas basins) and western (Republican and Smoky Hill basins) portions of the Kansas River basin, as
well as several tributaries to the Missouri River (Lower Missouri–
Blackwater and Missouri–Nishnabotna). Least-squared regression
lines are drawn for significant relationships.
away from a habitat (e.g., from large river into tributary) declines with distance. If confluence exchange is prevalent,
fish assemblages should vary among streams that are connected to different-sized streams (e.g., sites 1 and 3 in
Fig. 1b). If network dispersal is prevalent, we would expect
fish assemblage structure to vary at headwater tributary
stream sites that occur at different stream distances from
mainstem rivers (e.g., sites 1 and 2 in Fig. 1b). Because juvenile fishes can have different habitat preferences than
adults (Schlosser 1995), we also expect patterns of movement to vary among species, as well as among size classes.
Materials and methods
Study area
We evaluated the influence of proximity and connectance
to mainstem rivers on stream fish assemblages at two spatial
scales: basin-wide surveys and targeted sampling of three
tributaries of the Kansas River (Kansas, USA). The Kansas
River basin covers 156 286 km2 and includes the Smoky
Hill and Republican rivers as main tributaries in western
Kansas and the Big Blue and Delaware rivers in the east. In
addition, we included a few tributaries of the Missouri River
in the Missouri–Nishnabotna and Lower Missouri subregions
(e.g., level 2 HUCs) in our basin-wide study because they
represented similar habitats to those in the nearby Kansas
River basin (Fig. 2; Seaber et al. 1994). Hereafter, these
Missouri tributaries and the Kansas River Basin proper are
Reference
Strahler 1957
Grenouillet et al. 2004
Horwitz 1978
Horton 1945
collectively referred to as the Kansas basin. This region contains seven EPA level III ecoregions (Omernik 1987), with
land cover primarily comprised of grasses (49%) and agriculture (46%), with a small proportion of the basin forested
(<3%) and the remainder urban, water, wetlands, or barren
lands (US Geological Survey (USGS) 2006a). Streams at
the collection sites ranged from 1st to 8th order (Strahler
1957), and basin area of study reaches ranged from 0.3 to
116 978 km2. Twenty major reservoirs with surface areas between 399 ha and 5911 ha occurred within the study area
along with numerous smaller impoundments.
Three local tributaries of the Kansas River were targeted
to evaluate fine-scale variation in fish assemblage structure
along a gradient of increasing distance from their confluence
with the mainstem (Clarks Creek, McDowell Creek, and
Wildcat Creek; Fig. 2). Five 150 m reaches within the first
20 skm of the Kansas River on each of the three study
streams were sampled along with two reaches on the mainstem Kansas River (Fig. 2). Land cover in the three study
basins was dominated by grasslands (68%–81%) and agriculture (14%–24%), with the remainder comprised of shrub,
forest, urban, water, and wetlands (USGS 2006a). McDowell Creek and Wildcat Creek were similar in basin area,
mean width, mean depth, and substrates. Clarks Creek had
a basin area approximately twice that of McDowell Creek
and Wildcat Creek, and as such, it also had a greater mean
width and depth than the other two streams.
Experimental design
Fish sampling
Collections at 413 localities made by the Kansas Department of Wildlife and Parks (KDWP) during their annual
stream monitoring and assessment program between 1995
and 2006 were used in our analyses (Fig. 2). These records
did not include sites within 20 skm upstream of reservoirs
because those sites may have been influenced by reservoir
fish assemblages (Falke and Gido 2006b). The KDWP sampling protocol followed that of the US Environmental Protection Agency’s (EPA) Environmental Monitoring and
Assessment Program (EMAP; Lazorchak et al.1998). Each
sample reach was 40 times the average wetted width of the
stream (reach length range: 150–300 m). Sites were sampled
using a combination of straight and bag seines (4.7 mm
mesh) and pulsed-DC backpack electrofishing. One upstream pass was made with the electrofishing gear, and one
downstream pass was made seining all suitable habitats.
Fishes were identified to species, and each site was georeferenced with a geographical positioning system (GPS).
Vouchers were deposited at the University of Kansas or
Fort Hays State University Natural History Museums.
Fishes also were collected at the 17 sites in three tributaPublished by NRC Research Press
Thornbrugh and Gido
147
Fig. 4. Principal coordinates analysis (PCoA) summarized variation in fish assemblage structure across sites in the Kansas basin (a and c)
and associated species loadings (b and d). Axis 1 was graphed against the second and third axes and sites were coded by stream order.
Species codes are as presented in Table 2.
ries and the mainstem Kansas River using similar methods
over three seasons in 2006 and 2007: 6–12 July (summer),
21 August – 9 October (fall), and 11 May – 21 June
(spring). Species with notable length classes were separated
into small and large size classes. For most taxa (e.g., cyprinids and darters), the size classes represented juveniles (typically, <30 mm total length, TL) and adults (>30 mm TL),
whereas in other taxa (e.g., catostomids and ictalurids), the
large size class (>60 mm TL) represented adults and subadults (hereafter referred to as adults).
Habitat
Stream habitat was quantified at a variety of spatial
scales. Stream segments, classified as units of stream between tributary confluences that ranged in length from 0.1
skm to 33.4 skm, were quantified from a stream network derived from a modified version of the national hydrography
data set (NHD; USGS 2006b) using ArcGIS 9.2 software
(ESRI 2006). Strahler stream order and basin area for stream
segments were used to quantify stream size and spatial position within a drainage network. Downstream order (dorder)
and distance from the mainstem were used to quantify
downstream stream size and downstream spatial position
within a drainage network and connectivity to other habitats
(Table 1). For the Kansas basin, the mainstem river was de-
fined as ‡5th-order stream and distances were calculated as
the distance (skm) between the collection site and the nearest mainstem stream segment. At the basin scale, Universal
Transverse Mercator easting coordinates represented an east
to west gradient of precipitation, land use, and geography
that correspond to changes in fish assemblage structure
(Metcalf 1966).
Habitat quantification at the three local tributary sites and
the mainstem Kansas River was similar to the EPA-EMAP
physical habitat protocol (Lazorchak et al. 1998). At each
collection site, width measurements were collected along 10
transects that represented the area sampled, except for Kansas River sites, which only had five transects due to homogeneous habitat. Depth, width, and substrate were quantified
along 10 transects in the sampling site. Depth and dominant
substrate were recorded every metre. Proximity to the Kansas River was measured from the NHD as the distance (skm)
from each sampling site to the Kansas River.
Data analyses
Relations between fish assemblages, stream size, and
season
As a first step in our analysis, we quantified patterns of
assemblage structure across a gradient of stream size (i.e.,
Published by NRC Research Press
Summer
Species
Yellow bullhead
Freshwater drum
Central stoneroller
River carpsucker
White sucker
Common carp
Red shiner
Gizzard shad
Johnny darter
Orangethroat darter
Western mosquitofish
Channel catfish
Green sunfish
Orangespotted sunfish
Bluegill
Longear sunfish
Longnose gar
Common shiner
Redfin shiner
Smallmouth bass
Spotted bass
Largemouth bass
White bass
Shorthead redhorse
Published by NRC Research Press
Slender madtom
Stonecat
Carmine shiner
Sand shiner
Logperch
Suckermouth minnow
Southern redbelly
dace
Bluntnose minnow
Species code
AMENAT
APLGRU
CAMANO
Total
abundance
1
68
2 112
Relative
abundance
<0.0
0.1
2.2
Abundance
1
1
885
Sites
occupied
1
1
14
Spring
Abundance
—
19
1 193
Sites
occupied
—
4
14
Abundance
—
48
34
Sites
occupied
—
3
10
CARCAR
CATCOM
681
23
0.7
<0.0
608
20
11
6
26
3
9
2
47
—
10
—
CYPCAR
CYPLUT
DORCEP
102
46 973
693
0.1
48.8
0.7
32
13 845
63
5
16
4
7
20 950
577
4
17
10
63
12 178
53
4
17
3
599
1 562
5 237
113
227
3 330
232
1 262
67
232
3 889
15
64
9
8
183
0.6
1.6
5.4
0.1
0.2
3.5
0.2
1.3
0.1
0.2
4.0
<0.0
0.1
<0.0
<0.0
0.2
339
541
1 387
51
55
164
16
406
61
19
1 346
13
12
3
—
61
15
11
16
11
15
10
4
15
14
4
12
3
5
3
—
7
236
745
3 641
52
51
2 446
45
607
1
77
1 991
2
10
5
—
119
14
12
17
11
11
17
8
15
1
8
14
1
5
3
—
13
24
276
209
10
121
720
171
249
5
136
552
—
42
1
8
3
8
10
14
5
10
17
15
12
2
7
15
—
9
1
2
3
NOTATH
NOTCRY
3
1
<0.0
<0.0
—
—
—
—
2
—
1
—
1
1
1
1
NOTEXI
NOTFLA
NOTPER
NOTSTR
PERCAP
PHEMIR
PHOERY
111
662
4 873
8 742
167
1 155
94
0.1
0.7
5.1
9.1
0.2
1.2
0.1
64
22
287
1 460
126
587
1
9
5
10
13
13
14
1
24
12
4 209
2 795
35
394
1
8
4
15
15
10
13
1
23
628
4 246
618
6
174
92
6
5
17
10
5
11
3
PIMNOT
9 049
9.4
801
15
6 415
17
1 833
16
ETHNIG
ETHSPE
GAMAFF
ICTPUN
LEPCYA
LEPHUM
LEPMAC
LEPMEG
LEPOSS
LUXCOR
LYTUMB
MICDOL
MICPUN
MICSAL
MORCHR
MOXMAC
Can. J. Fish. Aquat. Sci. Vol. 67, 2010
Emerald shiner
Golden shiner
Scientific name
Ameiurus natalis
Aplodinotus grunniens
Campostoma anomalum
Carpiodes carpio
Catostomus commersonii
Cyprinus carpio
Cyprinella lutrensis
Dorosoma cepedianum
Etheostoma nigrum
E. spectabile
Gambusia affinis
Ictalurus punctatus
Lepomis cyanellus
L. humilis
L. macrochirus
L. megalotis
Lepisosteus osseus
Luxilus cornutus
Lythrurus umbratilis
Micropterus dolomieu
M. punctulatus
M. salmoides
Morone chrysops
Moxostoma macrolepidotum
Notropis atherinoides
Notemigonus crysoleucas
Noturus exilis
N. flavus
Notropis percobromis
N. stramineus
Percina caprodes
Phenacobius mirabilis
Phoxinus erythrogaster
Pimephales notatus
Fall
148
Table 2. Number of fish collected and their relative abundance, percentage of total individuals, and occurrences across 17 sample sites and three seasons from three tributaries of the
Kansas River (Clarks, McDowell, and Wildcat creeks).
23 098
17
16
23 886
Abundance
—
2 624
18
5
19
Abundance
1
537
62
3
6
49 356
Abundance
128
366
8
2
22
Spring
Fall
Sites
occupied
—
17
6
4
5
Sites
occupied
1
16
5
3
4
96 340
Total
Species
Fathead minnow
Bullhead minnow
White crappie
Flathead catfish
Creek chub
Table 2 (concluded).
Scientific name
P. promelas
P. vigilax
Pomoxis annularis
Pylodictis olivaris
Semotilus atromaculatus
Species code
PIMPRO
PIMVIG
POMANN
PYLOLI
SEMATR
Total
abundance
129
3 527
88
10
47
Relative
abundance
0.1
3.7
0.1
<0.0
<0.0
Summer
17
149
Sites
occupied
12
17
3
1
5
Thornbrugh and Gido
basin area) within the Kansas River basin using regression
analysis. This characterization was a necessary step because
it allowed us to identify species that showed an affinity for
mainstems or tributary habitats. For analyses of fish assemblage structure, rare fish species (i.e., occurring at <5% of
sample sites) were excluded because they can mask variation in the more abundant species in the assemblage. In addition, we reduced the influence of extreme high abundances
with log(x + 1) transformations. To characterize relationships among sites, we calculated a matrix of assemblage dissimilarities among sites using a Bray–Curtis index and
visualized gradients of fish assemblage composition with a
principal coordinates analysis (PCoA).
In the three intensively sampled tributaries, we characterized spatial and temporal variation in assemblage structure
to evaluate the influence of mainstem rivers within the first
20 skm of their confluence. We first tested for differences in
fish species richness among three sampling periods
(summer, fall, and spring) and streams (Clarks, McDowell,
and Wildcat creeks) with analysis of variance (ANOVA).
As with the basin-wide analyses, a PCoA was performed to
summarize variation in the fish assemblage structure within
and among tributaries and among seasons based on a Bray–
Curtis dissimilarity matrix. Analysis of similarity
(ANOSIM) was used to test for differences in assemblage
structure among sample periods and streams. These analyses
were conducted using the R statistical package (R Development Core Team 2007).
Relations between fish assemblage structure and distance
from mainstem rivers
We used multiple regression models and constrained ordination to test if proximity to mainstem rivers explained a
significant amount of variation in assemblage structure,
while controlling for longitudinal and stream size gradients
within the Kansas River basin. The influence of proximity
to mainstem rivers on species richness was made by comparing adjusted R2 values from models that included stream
order and easting (i.e., universal transverse mercator easting
coordinates), but differed in the inclusion of proximity to the
mainstem river. To evaluate the influence of proximity to a
mainstem river on fish assemblage structure, we used a partial constrained analysis of principal coordinates (CAP; Anderson and Willis 2003). CAP is similar to redundancy
analysis but allowed us to evaluate the match between environmental variables and variation in fish assemblage structure based on a Bray–Curtis dissimilarity matrix. As in the
multiple regression, we controlled for stream order and easting and tested for significance with a permutation test. CAP
analyses were performed in the R statistical package version
2.7.2 using the function ‘‘capscale’’ in the vegan library (R
Development Core Team 2007). In addition, we ran models
with mean abundance of riverine and stream fish as dependent variables to evaluate the use of tributaries by fishes that
typically occupy the mainstem river or the avoidance of
confluence zones by species that occupy smaller streams.
We classified species based on their affinity for stream size
(e.g., riverine or stream fishes) using our basin-wide database. Species were classified as riverine if their mean abundance was greater at sites on stream segments ‡5th order
than at smaller-ordered sites, and classified as stream fish if
Published by NRC Research Press
150
Can. J. Fish. Aquat. Sci. Vol. 67, 2010
Fig. 5. Principal coordinates analysis (PCoA) summarized variation in the fish assemblage structure across sites in three tributaries of the
Kansas River (a and c) and associated species loadings (b and d). Axes 1 and 2 plots were coded for streams and distance from mainstem,
and axes 1 and 3 plots were coded for streams and sampling periods. Species codes are as presented in Table 2.
Table 3. Results from multiple regression analyses testing the
association between distance from a mainstem river (i.e., ‡5th
order) on local fish species richness and abundances of river and
stream fishes in the Kansas basin.
Dependent variable
Local species richness
log river fish abundance
log stream fish abundance
Adjusted R2
with
distance
0.19
0.26
0.21
Adjusted R2
without distance
0.19
0.26
0.20
Note: All models contained independent variables stream order and
easting but differed in the inclusion of distance to a mainstem (>5th order) river.
their mean abundance was greatest on streams <5th order.
Prior to analyses, normality of habitat variables was tested
and non-normal variables were log(x + 1)-transformed to reduce the effects of outliers and heterogeneity of variances
Regression analysis also was used to quantify the association between fish assemblage structure (richness and abundance of riverine and stream fishes) and habitat measures
with distance from the Kansas River at the three intensively
sampled tributaries. In addition, we contrasted dissimilarity
of sample sites on each of the three tributaries (Clarks,
McDowell, and Wildcat creeks) with that of the Kansas
River for all sampling periods using a Bray–Curtis dissimilarity index. Means and standard deviations of Bray–Curtis
indices were calculated by combining data from all sampling
periods for each site and examined as a function of distance
from the mainstem Kansas River.
We separately tested the association in abundance of
adults and juveniles of 18 species with distance from the
mainstem using regression analyses. We were unable to consider other species because we did not capture both life
stages in high enough abundance to conduct the regression
analysis. Coefficients of determinations (r2) were used to
characterize the deviation in longitudinal structuring between adult and juvenile fishes.
Connectivity
To test the prediction that streams sharing their confluences with larger-order streams would have greater mean species richness and different assemblage structure than streams
that shared their confluence with smaller-order tributaries,
we evaluated differences in mean species richness, abundances of riverine and stream fishes, and assemblage structure
among sites on stream segments with the same stream order,
but differing downstream orders, using a one-way ANOVA
for richness and ANOSIM for assemblage structure. If the
ANOVA results were significant, a post hoc least significant
difference (LSD) test was run on the comparison of mean
Published by NRC Research Press
Thornbrugh and Gido
Fig. 6. Mean Bray–Curtis dissimilarity of fish assemblages between
sites within three tributary streams (*) and sites on the mainstem
Kansas River. Error bars represent standard deviation of mean
Bray–Curtis dissimilarity scores across sampling periods and means
and standard deviations calculated. Open circles (*) represent dissimilarity of fish assemblage between the Kansas River sites. (a)
Sites on Clarks Creek, (b) sites on McDowell Creek, and (c) sites
on Wildcat Creek. skm, stream kilometres.
fish species richness to test for differences between species
richness.
Results
Effects of stream size and season on fish assemblage
structure
Sixty-eight species of fish representing 39 genera in 14
families were collected from the 413 sites in the Kansas ba-
151
sin. Cyprinids were the dominant family, comprising 84%
of the total individuals captured, followed by centrarchids
(8%) and percids (4%). The number of sites visited ranged
from four in 8th-order streams to 154 in 3rd-order streams.
There was a significant increase in the number of fish species with drainage area for both eastern and western portions of the Kansas basin (P = <0.01, r2 = 0.14, r2 = 0.13,
respectively; Fig. 3). A similar trend was observed for the
Missouri tributary sites, but this was not significant (P =
0.14, r2 = 0.03).
Local fish assemblage structure also varied with stream
size, as characterized with PCoA (Fig. 4; 1st axis eigenvalue, 21.4; 2nd, 13.9; and 3rd, 10.9). The first axis represented a gradient of stream size with sites in larger streams
having a positive association with this axis and small
streams having a negative association (Fig. 4). Examples of
fishes associated with larger streams were Aplodinotus grunniens, Pylodictis olivaris, and Dorosoma cepedianum, and
fishes associated with small streams were Phoxinus erythrogaster, Etheostoma spectabile, and Luxilus cornutus. Axis
2 likely represented an east to west gradient because of the
large positive loadings of Fundulus kansae, a species restricted to western Kansas, in contrast to negative loadings
of species such as Percina caproides, which are more abundant in eastern Kansas. Axis 3 was likely a gradient from
low to high gradient streams, as inferenced by the high loadings of fishes typical of small streams (e.g., P. erythrogaster) in contrast to larger stream species (e.g.,
Notemigonus crysoleucas).
Thirty-nine species of fish representing 29 genera in nine
families were collected during the intensive sampling of the
three local tributaries (Table 2). Repeated-measures ANOVA indicated that species richness varied by season (P =
0.03), with richness generally lower in spring than in
summer and fall. There was no significant difference in richness among tributaries or interaction among seasons and
tributaries. ANOSIM results indicated that assemblages in
McDowell and Clarks creeks were distinct from those in
Wildcat Creek (Global R = 0.192, P = 0.08) and that there
were distinct assemblages among the three sample seasons
(Global R = 0.461, P = 0.001). A PCoA (Fig. 5; 1st axis eigenvalue, 1.20; 2nd, 0.82; and 3rd, 0.65) illustrated these
differences and indicated a relationship between proximity
to the mainstem and assemblage structure. The 1st axis described a gradient of proximity to the Kansas River having a
positive association with this axis, whereas streams that are
farther from the Kansas River had a negative association
with this axis. The species most closely associated with the
Kansas River was A. grunniens. Fishes having a positive association with axis 1 (i.e., further away from the mainstem)
were typical stream species such as Campostoma anomalum,
Notropis percobromis, Noturus exilis, E. spectabile, and E.
nigrum. Fish species that had a negative association with
axis 2 occurred in higher abundance in confluence zones
(e.g., Micropterus punctulatus, Pimephales promelas, and
Lepomis macrochirus), with the one exception being juvenile P. erythrogaster, which occurred after a spring flood
event downstream of a tributary confluence. The 3rd axis
was not associated with proximity to mainstem, rather it described a gradient across sampling periods, with the summer
and fall sampling periods having negative associations with
Published by NRC Research Press
152
Can. J. Fish. Aquat. Sci. Vol. 67, 2010
Fig. 7. Regression analysis quantifying the relationship between distance (in stream kilometres (skm)) from mainstem Kansas River and
log(x + 1)-transformed fish abundance of riverine and stream fishes by sampling periods. Species were classified as riverine or stream fish
based on comparisons of mean abundance in stream segments ‡5th order (river) or <5th order (stream).
Fig. 8. Comparisons of coefficients of determination (r2) for juveniles and adults from regression analyses testing the association between 18 dominate fish species abundance in local tributaries
(Clarks, McDowell, and Wildcat creeks) and distance from the
mainstem Kansas River. Species codes are as presented in Table 2.
Points below the broken line indicate that abundance of adults was
more strongly associated with distance from the mainstem than juveniles.
Fig. 9. Regression analysis quantifying the relationship between
distance (in stream kilometres (skm)) from mainstem Kansas River
and the change in percent coarse substrate for the three local tributaries (i.e., Clarks Creek, *; McDowell Creek, ~; and Wildcat
Creek, &).
100
Percent coarse substrate
80
60
40
20
0
0
5
10
15
20
Distance from Kansas River (skm)
this axis and the spring sampling period having a positive
association. Fishes collected in higher abundance in the
summer were Lepisosteus osseus, Carpiodies carpio, and
D. cepedianum. Fishes that occurred in higher abundance
during spring were P. erythrogaster, P. promelas, and
L. cornutus.
Effects of distance from mainstem rivers on fish
assemblage structure
Multiple regression models that predicted fish species
richness in the Kansas basin indicated that stream order and
easting were significantly associated with richness, whereas
proximity to a mainstem river explained little additional
variation (<2% change in adjusted R2; Table 3). A partial
CAP analysis indicated a significant effect of proximity to
mainstem rivers after controlling for stream order and easting (F = 4.36, P = 0.005). However, proximity only explained 1.1% of the unconstrained variation in fish
assemblage structure (full constrained ordination including
stream order and easting explained 14.7% of unconstrained
variation).
Consistent with this large-scale pattern, fish assemblages
changed abruptly between sites on the mainstem Kansas
River and those on the three intensively sampled tributary
streams. Bray–Curtis dissimilarity of fish assemblages in
tributary sites compared with Kansas River sites increased
abruptly with distance from the Kansas River. However,
Published by NRC Research Press
Thornbrugh and Gido
153
Fig. 10. Mean fish species richness, mean principal coordinates analysis (PCoA) axis 1 and axis 3 scores, and mean abundance of river and
stream fishes between streams of similar stream order with different downstream order. Different letters represent significant (a = 0.05)
difference among groups based on analysis of variance (ANOVA; richness and abundance) and analysis of similarity (ANOSIM; assemblage
composition) post hoc comparisons. PCoA axis 1 and axis 3 scores are shown to represent changes in fish assemblage composition, but
letters represent significant differences based on ANOSIM of the Bray–Curtis dissimilarity matrix. The numbers of replicates for each
stream type are given in the first column of graphs.
after this abrupt shift in assemblage structure, there was only
a slight divergence from the mainstem fish assemblage with
increasing distance from the Kansas River (Fig. 6). Fishes
driving these changes as distance increased from the Kansas
River were stream fishes that were collected only at the sites
farthest away. Increases in the abundance of stream fish
with distance from the mainstem Kansas River were significant in the summer and fall sampling periods (P < 0.01 and
P < 0.01, respectively), but this same pattern was not seen
in the spring sampling period (Fig. 7). The relationships between distance from the mainstem Kansas River and abundance of 18 fish species, with both adults and juveniles
collected during sampling periods, suggested that adults
were more structured along this gradient than juveniles
(Fig. 8). Coefficients of determinations (r2) describing this
relationship ranged from approximately 0.00 to 0.14 with a
mean of 0.03 for juvenile fishes and approximately 0.00 to
0.42 with a mean of 0.12 for adults.
The most notable association between physical habitat
and distance from the Kansas River was in the proportion
of coarse substrate (r2 = 0.285, P = 0.040), which was generally <40% in the first 2 skm of local tributaries and increased to around 70% at sites >2 skm from the mainstem
(Fig. 9). There was no significant difference in mean depth
(range: Clarks Creek, 20.6–16.3 cm; McDowell Creek,
21.1–93.2 cm; Wildcat Creek, 26.2–110.7 cm) or width
(range: Clarks Creek, 5.1–52.8 m; McDowell Creek, 4.4–
23.0 m; Wildcat Creek, 10.0–21.3 m) with distance from
the mainstem (P values > 0.440).
Effects of connectivity on fish assemblage structure
Not all combinations of joining streams of different orders
were represented in the basin-wide data set, thus we only
considered network combinations for which we had adequate data to statistically test patterns (i.e., N > 4). Of
those, there was a general trend that stream segments sharing their confluences with larger mainstem tributaries had
higher mean species richness than segments that shared their
confluence with smaller-order tributaries (Figs. 10a, 10e,
10i). Fish assemblages also differed among streams with different levels of connectivity. Assemblage structure in 1storder streams was dependent on connection with 2nd-, 3rd-,
and 4th-order streams (ANOSIM Global R = 0.315). This
can be visualized by increasing PCoA axis 1 scores for 1storder sites as the stream order of connecting streams increased (Fig. 10b). Assemblages in sites on 2nd-order
streams connected to 3rd-order streams were different than
sites on 2nd-order streams connected to 2nd-, 4th-, or 7thorder streams (Global R = 0.081), and this is reflected with
PCoA axis 3 scores (Fig. 10f). Fish assemblages in 3rd-order
streams were marginally associated with connection to other
streams (Global R = 0.064), and there was a trend (P >
0.059) for differences in fish assemblage structure of 3rdorder streams connected to 4th-order streams than those connected to 3rd- or 8th-order streams (Fig. 10j). First-order
streams that share their confluences with 3rd- or 4th-order
tributaries had significantly (P < 0.01) higher riverine fish
abundance than streams that share their confluences with
2nd-order tributaries (Fig. 10c). Third-order streams that
Published by NRC Research Press
154
share their confluences with 3rd- and 4th-order tributaries
had significantly (P < 0.01) different abundances of stream
fishes (Fig. 10l). There was no difference in mean stream
fish abundance in 1st- or 2nd-order sites for sites with differing downstream orders.
Discussion
By characterizing patterns of assemblage structure at the
basin and individual tributary scales, we developed a thorough evaluation of factors influencing local fish assemblage
structure in the Kansas River basin. The basin-wide analysis
was coarse but gave us a greater number of samples and a
wide variety of network combinations (e.g., probability of
different-order streams having their confluences with one
another) to quantify the influence of connectivity and proximity on fish assemblage structure. When considering both
stream segments that shared a confluence with mainstem
rivers and those that do not, we only detected a marginal influence of proximity to the mainstem on fish assemblage
structure, thus rejecting our hypothesis of network dispersal.
However, the significant influence of downstream order on
fish assemblage structure in stream segments directly connected to mainstem rivers supports the hypothesis of confluence exchange and a more localized effect of downstream
habitats. Moreover, when evaluated at a finer spatial scale
(i.e., during intensive sampling of sites in three mainstem
tributaries that were <20 skm from the confluence), we
found fish assemblages at sites near the Kansas River to differ from the mainstem assemblage, but those differences associated with proximity to the source were primarily related
to the reduced abundance of small stream fishes near tributary confluences. Combining these analyses suggests a relatively localized exchange of biota (i.e., confluence
exchange) between mainstem rivers and their tributaries.
This is not to say that periodic events or migrations that are
important to the persistence of the species do not occur, but
long-term monitoring at a finer temporal scale may be necessary to evaluate the importance of these rare events.
Our results are consistent with previous research that
noted an association between fish assemblage structure and
the presence of mainstem rivers (Hitt and Angermeier 2006;
Hitt and Angermeier 2008) and reservoirs (Falke and Gido
2006b) downstream from sample sites. In these studies,
however, downstream habitats had the most notable influence on fish assemblages in medium-sized tributaries. Hitt
and Angermeier (2006) suggested that increased fish species
richness and increased riverine species richness in 3rd-order
and larger tributaries of larger rivers were mediated through
microhabitat complexity and channel shape. We found larger
differences in assemblage structure of 1st-order streams
draining into larger streams and hypothesize that the isolation of these small streams may result in a reduction of obligate small-stream fishes, while allowing the immigration of
small-bodied larger river fishes, such as red shiners. The latter is probably the case in the Kansas River basin, as abundance of larger river species also increased as 1st-order
streams drained into larger streams.
The presence of riverine fishes in smaller streams above
tributary junctions could be associated with reproduction
and spawning migrations. Hitt and Angermeier (2008) noted
Can. J. Fish. Aquat. Sci. Vol. 67, 2010
that tributaries were used for nursery and spawning habitat
by riverine species such as catostomids and cyprinids, which
were higher in abundance and more species-rich in mainstem tributaries than in headwater tributaries. In contrast to
this result, abrupt changes in assemblage structure in the intensively sampled tributaries were primarily driven by the
lack of small stream fish (e.g., Luxilus cornutus, E. spectabile, and C. anomalum) near mainstem rivers, rather than
the presence of large river fishes. The primary exception to
this was P. vigilax, which generally decreased in abundance
with distance from the mainstem river. The absence of small
stream species near confluences with mainstem rivers may
be due to increased turbidity, changes in habitat (e.g., substrate), or biotic interactions. Indeed, changes in fish assemblage structure coincided with abrupt changes in habitat
between mainstem rivers and their tributaries, with the lower
portions made up of more homogeneous floodplain habitat
types with finer substrates. These deeper homogenous floodplain ecotones also might allow predatory fishes to persist in
the assemblage for a short distance in these transitional
zones of tributaries and their mainstem. As our data do not
show a marked increase in these predatory species, they may
not have been sampled effectively.
Within our three intensively monitored tributary streams,
we also found a slight decrease in variability and a decrease
in dissimilarity to the mainstem with increasing distance
from the Kansas River and variability across sampling periods in fish assemblage. Gorman (1986) speculated that the
impact of mainstem fish assemblages on adventitious tributaries is probably significant but noted that these influences
were probably temporally variable and synchronized with
reproductive seasonal pulses of migrant fishes. Schaefer and
Kerfoot (2004) provided empirical evidence of greater species diversity and variability in fish assemblages over time
at interface sites of Piasa Creek, Illinois, with declining variability and diversity with increasing distance from the
mainstem river over their approximate 60 skm study area.
Our study indicated a high degree of temporal variation in
fish assemblage structure among sampling periods that was
confounded by variable abundance of juveniles and adults.
Moreover, the abundance of adult fishes was more structured along a gradient of distance from the mainstem than
juvenile fishes. More randomly distributed juveniles along a
longitudinal gradient could indicate that they are more prone
to dispersal, which may be necessary to locate suitable rearing habitat (e.g., shallow, predator-free areas; Harvey 1987).
Adults, however, may respond to habitat at larger spatial
scales (i.e., longitudinal zonation) that is driven by physical
habitat (e.g., depth) or biotic interactions.
Ecological characteristics of fish species are likely to mediate their ability to disperse through stream networks. Many
of the species that we captured in mainstem rivers that
quickly disappeared in collections upstream of tributary confluences were relatively large-bodied as adults (e.g.,
A. grunniens, P. olivaris) and may avoid small tributary
streams because shallow, clear water may expose them to
terrestrial predators (e.g., Power et al. 1989). Other common
mainstem species such as N. stramineus and Cyprinella lutrensis are small-bodied but may avoid streams with low turbidity (Cross and Collins 1995). Although these two species
can sporadically occur in mainstem tributaries, sometimes in
Published by NRC Research Press
Thornbrugh and Gido
high abundance, it is not clear why these species are more
dominant in turbid mainstem rivers. Nevertheless, they
seem to be at a disadvantage compared with other stream
minnow species with similar ecological characteristics (e.g.,
Lythrurus umbratilis and P. notatus) that are dominant in
tributaries. The avoidance of mainstem rivers and potentially
confluence zones by tributary specialists is likely related to
habitat, particularly turbidity and substrate. For example,
many darters and minnow species require coarse substrates
to successfully reproduce.
Conservation implications
Quantifying the linkage between habitats within a stream
network can help managers evaluate the relative importance
of these disjunct habitats when making decisions on the
scale at which to conserve stream organisms. Although our
results suggest that many prairie stream fish populations are
localized and possibly independent from mainstem river
habitats during normal conditions, a number of studies have
reported the importance of connectance to refugia habitats
after a disturbance in small- to medium-order streams
(Bayley and Osborne 1993; Sheldon and Meffe 1994; Lonzarich et al. 1998). Recolonization of fish assemblages in
lotic systems can be quite rapid (£1year; Detenbeck et al.
1992; Bayley and Osborne 1993), but recolonization rate is
dependent on distance from source populations (Sheldon
and Meffe 1994; Lonzarich et al. 1998) and the presence of
barriers (Detenbeck et al. 1992). Most of these studies cited
above were in small- to medium-order streams in which fish
assemblages are isolated from source populations and fishes
were likely more adapted to rapid recolonization due to the
frequent disturbances from floods and droughts (Schlosser
1987). Larger tributary fish assemblages are more apt to be
recolonized from source populations of fishes that are not
adapted to highly stochastic habitats. These mainstem tributary fish assemblages may be more resilient to extinction
risk by being less isolated and closer to source pools and
rapidly recolonized via emigration as predicted by island
biogeography theory (MacArthur and Wilson 1967).
The importance of connectivity and location within a
drainage network is likely to vary among species with different ecological characteristics. In the Great Plains, USA,
many species that characterize mainstem river assemblages
are in need of conservation, but the importance of tributaries
to their conservation is not well understood. Given the recent threats to mainstem rivers, tributaries may be important
refugia habitat from pulsed disturbance events in mainstem
rivers. Likewise, the role of mainstem rivers as corridors for
tributary species is also equivocal, but recent studies suggest
that these habitats allow connectance of metapopulations
and maintenance of allelic diversity (Skalski et al. 2008).
Although our results suggest that biotic exchanges such as
these may not be sufficient to result in strong shifts in assemblage structure, the temporal scale of our study was not
adequate to evaluate the long-term importance of these network linkages. Long-term management and conservation of
fishes will require further evaluation of the influence of connectivity in stream networks and how this influences metapopulation dynamics and overall resiliency of stream biota.
155
Acknowledgements
Nathan Franssen, Justin Bengtson, David Hoeinghaus, Katie Bertrand, Tyler Pilger, Deb Walks, and Eric Kightley
helped with fieldwork. Chris Mamoliti, Kristen Hase, Mark
Van Scoyoc, Ryan Waters, and Mark Shaw with the Kansas
Department of Wildlife and Parks were instrumental in making fish collections and providing data. Walter Dodds, Craig
Paukert, James Whitney, and two anonymous reviewers provided valuable input to an earlier version of this manuscript.
This work was supported by grants from the US Geological
Survey National Gap Analysis Program and the Kansas Department of Wildlife and Parks. The National Science Foundation Research Experiences for Undergraduates Program
provided support for E. Kightley. This is contribution No.
10-088-J from the Kansas State University Agricultural Experiment Station.
References
Anderson, M.J., and Willis, T.J. 2003. Canonical analysis of principal coordinates: a useful method of constrained ordination for
ecology. Ecology, 84(2): 511–525. doi:10.1890/00129658(2003)084[0511:CAOPCA]2.0.CO;2.
Angermeier, P.L., and Winston, M.R. 1998. Local vs. regional influences on local diversity in stream fish communities of Virginia. Ecology, 79: 911–927.
Bayley, P.B., and Osborne, L.L. 1993. Natural rehabilitation of
stream fish populations in an Illinois catchment. Freshw. Biol.
29(2): 295–300. doi:10.1111/j.1365-2427.1993.tb00765.x.
Benda, L., Proff, N.L., Miller, D., Dunne, T., Reeves, G., Pess, G.,
and Pollock, M. 2004a. The network dynamics hypothesis: how
channel networks structure riverine habitats. Bioscience, 54(5):
413–427. doi:10.1641/0006-3568(2004)054[0413:TNDHHC]2.0.
CO;2.
Benda, L., Andras, K., Miller, D., and Bigelow, P. 2004b. Confluence effects in rivers: interactions of basin scale, network geometry, and disturbance regimes. Water Resour. Res. 40(5):
W05402. doi:10.1029/2003WR002583.
Campbell Grant, E.H., Lowe, W.H., and Fagan, W.F. 2007. Living
in the branches: population dynamics and ecological processes
in dendritic networks. Ecol. Lett. 10(2): 165–175. doi:10.1111/j.
1461-0248.2006.01007.x. PMID:17257104.
Cross, F.B., and Collins, J.T. 1995. Fishes in Kansas. 2nd ed. (revised). University of Kansas Press, Lawrence, Kansas.
Detenbeck, N.E., DeVore, P.W., Niemi, G.J., and Lima, A. 1992.
Recovery of temperate-stream fish communities from disturbance: a review of case studies and synthesis of theory. Environ.
Manage. 16(1): 33–53. doi:10.1007/BF02393907.
Dunning, J.B., Danielson, J.B., and Pulliam, H.R. 1992. Ecological
processes that effect populations in complex landscapes. Oikos,
65(1): 169–175. doi:10.2307/3544901.
ESRI. 2006. ArcGIS version 9.2. [Computer program.] Environmental Systems Research Institute, Redlands, California.
Fagan, W.F. 2002. Connectivity, fragmentation, and extinction risk
in dendritic metapopulations. Ecology, 83: 3243–3249.
Falke, J.A., and Gido, K.B. 2006a. Effects of reservoir connectivity
on stream fish assemblages in the Great Plains. Can. J. Fish.
Aquat. Sci. 63(3): 480–493. doi:10.1139/f05-233.
Falke, J.A., and Gido, K.B. 2006b. Spatial effects of reservoirs on
fish assemblages in great plains streams in Kansas, USA. River
Res. Appl. 22(1): 55–68. doi:10.1002/rra.889.
Fausch, K.D., Torgersen, C.E., Baxter, C.V., and Li, H.W. 2002.
Landscapes to riverscapes: bridging the gap between research
Published by NRC Research Press
156
and conservation of stream fishes. Bioscience, 52(6): 483–498.
doi:10.1641/0006-3568(2002)052[0483:LTRBTG]2.0.CO;2.
Franssen, N.R., Gido, K.B., Strakosh, T.R., Bertrand, K.N., Franssen, C.M., Paukert, C.P., Pitts, K.L., Guy, C.S., Tripe, J.A., and
Shrank, S.J. 2006. Effects of floods on fish assemblages in an
intermittent prairie stream. Freshw. Biol. 51(11): 2072–2086.
doi:10.1111/j.1365-2427.2006.01640.x.
Gorman, O.T. 1986. Assemblage organization of stream fishes: the
effect of rivers on adventitious streams. Am. Nat. 128(4): 611–
616. doi:10.1086/284592.
Gotelli, N.J., and Taylor, C.M. 1999. Testing metapopulation models with stream-fish assemblages. Evol. Ecol. Res. 1: 835–845.
Grenouillet, G., Pont, D., and Hérissé, C. 2004. Within-basin fish
assemblage structure: the relative influence of habitat versus
stream spatial position on local species richness. Can. J. Fish.
Aquat. Sci. 61(1): 93–102. doi:10.1139/f03-145.
Grossman, G.D., Ratajczak, R.E., Crawford, M., and Freeman,
M.C. 1998. Assemblage organization in stream fishes: effects of
environmental variation and interspecific interactions. Ecol.
Monogr. 68: 395–420.
Harvey, B.C. 1987. Susceptibility of young-of-the-year fishes to
downstream displacement by flooding. Trans. Am. Fish. Soc.
116(6):
851–855.
doi:10.1577/1548-8659(1987)116<851:
SOYFTD>2.0.CO;2.
Hitt, N.P., and Angermeier, P.L. 2006. Effects of adjacent streams
on local fish assemblage structure in western Virginia: implications for biomonitoring. In Landscape Influences on Stream Habitats and Biological Assemblages: Proceedings of the
Symposium on Influences of Landscape on Stream Habitat and
Biological Communities, Madison, Wisconsin, USA, 25–26 August 2004. Edited by R.M. Hughes, L. Wang, and P.W. Seelbach. American Fisheries Society Symp. No. 48, Bethesda, Md.
pp. 75–86.
Hitt, N.P., and Angermeier, P.L. 2008. Evidence for fish dispersal
from spatial analysis of stream network topology. J. N. Am.
Benthol. Soc. 27(2): 304–320. doi:10.1899/07-096.1.
Horton, R.E. 1945. Erosional development of streams and their
drainage basins: hydrophysical approach to quantitative morphology. Geol. Soc. Am. Bull. 56(3): 275–370. doi:10.1130/
0016-7606(1945)56[275:EDOSAT]2.0.CO;2.
Horwitz, R.J. 1978. Temporal variability patterns and the distributional patterns of stream fishes. Ecol. Monogr. 48(3): 307–321.
doi:10.2307/2937233.
Lazorchak, J.M., Klemm, D.J., and Peck, D.V. 1998. Environmental Monitoring and Assessment Program — surface waters: field
operations and methods for measuring the ecological condition
of wadeable streams. No. EPA/620/R-94/004F, US Environmental Protection Agency, Washington, D.C.
Lonzarich, D.G., Warren, M.L., Jr., and Lonzarich, M.R.E. 1998.
Effects of habitat isolation on the recovery of fish assemblages
in experimentally defaunated stream pools in Arkansas. Can. J.
Fish. Aquat. Sci. 55(9): 2141–2149. doi:10.1139/cjfas-55-92141.
Lowe, W.H., Likens, G.E., and Power, M.E. 2006. Linking scales
in stream ecology. Bioscience, 56(7): 591–597. doi:10.1641/
0006-3568(2006)56[591:LSISE]2.0.CO;2.
MacArthur, R.O., and Wilson, E.O. 1967. The theory of island biogeography. Princeton University Press, Princeton, N.J.
Matthews, W.J. 1998. Patterns in freshwater fish ecology. Chapman
and Hall, New York.
Matthews, W.J., and Robison, H.W. 1998. Influence of drainage
connectivity, drainage area and regional species richness on fishes
Can. J. Fish. Aquat. Sci. Vol. 67, 2010
of the interior highlands in Arkansas. Am. Midl. Nat. 139(1): 1–
19. doi:10.1674/0003-0031(1998)139[0001:IODCDA]2.0.CO;2.
Metcalf, A.L. 1966. Fishes of the Kansas River system in relation
to zoogeography of the Great Plains. Univ. Kansas Publ. Mus.
Nat. Hist. 17: 23–189.
Omernik, J.M. 1987. Ecoregions of the conterminous United States.
Ann. Assoc. Am. Geogr. 77(1): 118–125. doi:10.1111/j.14678306.1987.tb00149.x.
Osborne, L.L., and Wiley, M.J. 1992. Influence of tributary spatial
position on the structure of warmwater fish communities. Can. J.
Fish. Aquat. Sci. 49(4): 671–681. doi:10.1139/f92-076.
Power, M.E., Dudley, T.L., and Cooper, S.D. 1989. Grazing catfish, fishing birds, and attached algae in a Panamanian stream.
Environ. Biol. Fishes, 26(4): 285–294. doi:10.1007/
BF00002465.
R Development Core Team. 2007. R: a language and environment
for statistical computing. [Computer program.] R Foundation for
Statistical Computing, Vienna, Austria.
Schaefer, J.F., and Kerfoot, J.R. 2004. Fish assemblage dynamics
in an adventitious stream: a landscape perspective. Am. Midl.
Nat.
151(1):
134–145.
doi:10.1674/0003-0031(2004)
151[0134:FADIAA]2.0.CO;2.
Schlosser, I.J. 1982. Fish assemblage structure and function along
two habitat gradients in a headwater stream. Ecol. Monogr.
52(4): 395–414. doi:10.2307/2937352.
Schlosser, I.J. 1987. The role of predation in age- and size-related
habitat use by stream fishes. Ecology, 68(3): 651–659. doi:10.
2307/1938470.
Schlosser, I.J. 1995. Critical landscape attributes that influence fish
population dynamics in headwater streams. Hydrobiologia, 303:
71–81.
Schlosser, I.J., and Angermeier, P.L. 1995. Spatial variation in demographic processes in lotic fishes: conceptual models, empirical evidence, and implications for conservation. In Proceedings
of the Symposium on Evolution and the Aquatic Ecosystem: Defining Unique Units in Population Conservation, Monterey, California, USA, 23–25 May 1994. Edited by J.L. Nielsen.
American Fisheries Society Symp. No. 17, Bethesda, Md. pp.
392–401.
Seaber, P. R., Kapinos, P., and Knapp, G.L. 1994. Hydrologic units
maps. US Geological Survey Water-Supply Paper No. 2294,
Federal Center, Box 25425, Denver, CO 80225, USA.
Sheldon, A.L., and Meffe, G.K. 1994. Short-term recolonization by
fishes of experimentally defaunated pools of a Coastal Plain
stream. Copeia, 1994(4): 828–837.
Skalski, G.T., Landis, J.B., Grose, M.J., and Hudman, S.P. 2008.
Genetic structure of creek chub, a headwater minnow, in an impounded river system. Trans. Am. Fish. Soc. 137(4): 962–975.
doi:10.1577/T07-060.1.
Smith, T.A., and Kraft, C.E. 2005. Stream fish assemblages in relation to landscape position and local habitat variables. Trans.
Am. Fish. Soc. 134(2): 430–440. doi:10.1577/T03-051.1.
Strahler, A.N. 1957. Quantitative analysis of watershed geomorphology. Trans. Am. Geophys. Union, 38: 913–920.
US Geological Survey. 2006a. USGS land use and land cover
(LULC). Available online at http://edc.usgs.gov/products/
landcover/lulc.html.
US Geological Survey. 2006b. National hydrography dataset.
Available online at http://nhd.usgs.gov/index.html.
Vannote, R.L., Minshall, G.W., Cummins, K.W., Sedell, J.R., and
Cushing, C.E. 1980. The river continuum concept. Can. J. Fish.
Aquat. Sci. 37(1): 130–137. doi:10.1139/f80-017.
Published by NRC Research Press