Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
74
2018 SOUTHEASTERN NATURALIST 17(1):74–94
Effects of Introduced Small Wood in a Degraded Stream on
Fish Community and Functional Diversity
Ken A. Sterling1,* and Melvin L. Warren Jr.1
Abstract - Though the effects of introduced wood on fishes is widely studied for salmonids
in upland coldwater streams, there are few studies on this topic conducted in the Coastal
Plain of the southeastern US. This research gap is problematic because the introduction of
wood is a critical component of efforts aimed at conserving the threatened fish diversity
of the Coastal Plain, but managers lack data on the effects of installed wood on fish communities.
Over a nearly 4-year study period, we contrasted the effects of introduced, small,
wood bundles on the fish community in a channelized and deeply incised sand-bed Coastal
Plain stream with an unmanipulated reference treatment. The central question was whether
or not stream reaches with introduced wood had greater taxonomic and functional diversity
than unmanipulated reference reaches within the same stream. The introduction of modest
amounts of small wood had measurable and biologically significant positive impacts on fish
community composition and perhaps functional diversity relative to stream reaches lacking
wood. However, species-specific responses varied among treatments, suggesting the design
of wood installations has an impact on whether or not managemen t goals are achieved.
Introduction
The effects of introduced wood on stream fishes is seemingly well studied, but
most publications have focused on salmonids in cold water, upland habitats. Few
published studies that explicitly examined the effects of introduced wood on fishes
are available from the Coastal Plain of the southeastern US (Schneider and Winemiller
2008; Shields et al. 2003, 2006; Warren et al. 2009). This research gap is
problematic for several reasons. Many Coastal Plain streams lack fallen wood due
to habitat alteration and subsequent degradation, making the introduction of wood
a critical component of preserving fish diversity (Warren 2012). Studies from sites
elsewhere in North America fundamentally differ from the Coastal Plain region in
factors like geology, hydrology, land-use history, and available habitat (Meffe and
Sheldon 1988; Montgomery et al. 2003; Shields et al. 1998, 2000); thus, results
from these studies cannot be assumed to apply to Coastal Plain streams. In addition,
the effects of wood on fish abundance and diversity are inconsistent (Roni et al.
2014, Stewart et al. 2009). Finally, efforts to mitigate the negative effects of human
development on fishes are hampered by a lack of data to inform effective strategies
to restore, enhance, and maintain fish diversity and stream habi tat.
The lack of information on fish–wood interactions in the region seems odd because
a high proportion of southeastern fishes are imperiled (Jelks et al. 2008), and
fishes in lowland Coastal Plain streams are likely more dependent on wood than
1USDA Forest Service, Southern Research Station, Stream Ecology Laboratory, 1000 Front
Street, Oxford, MS 38655. *Corresponding author - kennethsterling@fs.fed.us.
Manuscript Editor: Matthew Heard
Southeastern Naturalist
75
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
fishes in upland streams with rocky cover. Southeastern Coastal Plain streams generally
lack coarse rock substrate, and fishes are reliant on wood to provide habitat
complexity and stability as well as spawning substrate, and as a source of diverse
invertebrate prey, cover, and current and drought refugia (Crook and Robertson
1999, Monzyk et al. 1997, Montgomery et al. 2003, Pilotto et al. 2014, Shields et
al. 1994, Smock and Gilinsky 1992, Warren 2012, Warren et al. 2002).
Results from the few studies examining the effects of introduced wood on
fishes in warmwater streams in North America, including those outside the
Coastal Plain, show a gradient in responses among species from negative to positive
as well as changes in stream morphology due to altered hydrology (Angermeier
and Karr 1984, Gatz 2008, Hrodey and Sutton 2008, Warren et al. 2009).
Experimental reaches with introduced wood generally have more diverse sediments,
lower flow, and greater depths than reaches without wood (Angermeier
and Karr 1984, Webb and Erskine 2005). Overall results from manipulative studies
are generally consistent with observational studies from the Coastal Plain
showing fish and wood relationships (Meffe and Sheldon 1988, Scott and Angermeier
1998, Warren et al. 2002).
Studies of stream invertebrate communities provide direct and indirect evidence
for the effects of wood on fishes in lowland streams. The presence or introduction
of wood in streams influences invertebrate density, richness, and biomass (Benke
and Wallace 2015, Benke et al. 1985, Pilotto et al. 2014) and likely has invertebratemediated
effects on fishes (Benke et al. 1985, Gary and Hargrave 2017). However,
as for fish, the introduction of wood has variable effects on benthic communities
(Leps et al. 2016, Palmer et al. 2010), which renders invertebrate-mediated effects
on fishes uncertain.
The study of introduced wood as fish habitat has almost exclusively focused on
large wood (LW; here defined as ≥10 cm diameter, ≥100 cm long). Only 2 publications
have reported the effects of introduced small wood (SW, less than 10 cm diameter, 100
cm long) on fishes in lowland warmwater streams (Schneider and Winemiller 2008,
Warren et al. 2009). The study of SW deserves more attention because: (1) LW is often
removed from streams by humans (Benke et al. 1985, Hrodey and Sutton 2008,
Shields et al. 2000), (2) deforestation of riparian zones and poor land management
lower rates of recruitment of LW into streams (Hrodey and Sutton 2008, Keeton
et al. 2007, Warren 2012, Williams 1989), and (3) LW is often rapidly transported
out of streams by flashy flows resulting from stream channelization and incisement
(Shields et al. 1994, Warren 2012), a process that apparently occurs in confined
streams worldwide (Kramer and Wohl 2017, Wyżga et al. 2017). Under these conditions
(i.e., heavily modified, channelized, and incised, sand-bottomed Coastal
Plain streams), LW is often relatively rare and SW, when present, provides most of
the available structure and habitat for aquatic organisms (Hrodey and Sutton 2008;
Shields et al. 1994, 2006; Warren et al. 2002; Wohl 2004). Degraded, channelized,
and incised streams are common in the southeastern Coastal Plain (Schoof 1980,
Wohl 2004), especially for streams running through agricultural lands (Pierce et al.
2012). These streams generally also have depauperate fish communities relative to
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
76
less-disturbed streams (Etnier 1972, Lau et al. 2006, Roth et al. 1996, Sullivan et al.
2004). The study presented here used installed SW bundles in a degraded Coastal
Plain stream to investigate whether or not the fish community would show greater
taxonomic and functional diversity in reaches with installed wood than in unmanipulated
reaches that largely lacked wood.
Field-Site Description
The study site was located at West Cypress Creek, a 3rd-order Coastal Plain
stream typical of the Little Tallahatchie River system in north-central Mississippi
(Fig. 1). The area near West Cypress Creek consists of low rolling hills (maximum
relief about 130 m), and land use is a mix of Pinus (pine) plantations, pine–hardwood
forest, row crops, and scattered housing. Streams within the area have sand
as their primary substrate and substrate particles >16 mm diameter are exceedingly
rare (Warren et al. 2002). Natural stream-bed controls are uncommon (Shields et
al. 1997, Warren et al. 2002) and are not apparent in West Cypress Creek. In-stream
wood is also uncommon (Warren et al. 2002). Extensive channelization throughout
the Little Tallahatchie River watershed has caused West Cypress Creek to be
deeply incised (5–6 m) and channelized with unstable banks and a highly unstable
stream bed; the stream is wide and uniformly shallow and experiences flashy flows.
We selected West Cypress Creek as our study site because we hypothesized that a
Figure 1. Map of the general location
of the study in north-central Mississippi
and the location of the study
site in West Cypress Creek showing
channelized stream reaches from the
headwaters of West Cypress Creek
downstream to the old channel of
the Little Tallahatchie River and
the channelized Tallahatchie Canal
cut-off.
Southeastern Naturalist
77
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
highly degraded stream would be more likely to show a response to our treatments
than less-degraded streams that contained more wood.
Deforestation, channelization, and construction of headwater impoundments
are all common on streams in the Coastal Plain of the southeastern US. Managers
employ these tools to mitigate flooding, improve drainage, and to efficiently use
all available land for agricultural purposes. However, all 3 practices cause streams
to become incised and disconnected from floodplains (Wohl 2004). Channelized
and incised streams result in wide, shallow, homogenous stream habitat with unstable
banks and often highly mobile stream beds with limited stable structures,
including woody debris (Shields et al. 1994, Sullivan et al. 2004, Wohl 2004). West
Cypress Creek is apparently representative of many other highly degraded smallto
medium-sized streams throughout the southeastern Coastal Plain, especially in
agricultural areas.
The study reach was typical of the rest of West Cypress Creek and other
similar-sized streams in the area. Variation in the types and amounts of available
habitat was minimal. The study reach consisted almost entirely of a sand-bed run
with scattered undercut banks and a few small, shallow, ephemeral pools along
the bank. SW and organic debris were the primary available cover types. LW was
rare, and SW and detritus were ephemeral. Two headwater impoundments were
upstream of our study reach (~3.1 km and 3.3 km). A bridge located downstream
(~150 m) had a short segment of rip-rap crossing the stream bed that had a slightly
steeper gradient than the rest of the stream and partially isolated the study reach.
Watershed area upstream of the study reach was ~21 km2.
The fish fauna within degraded channelized streams in the region is generally
dominated by small cyprinids tolerant of harsh and variable environments (Adams
et al. 2004; Shields et al. 1994, 1998, 2003). Compared with other less-disturbed
streams in the area, fish assemblages tend to be less diverse (Shields et al. 1994,
Warren et al. 2002).
Methods
Study design and measurement of response variables
The study reach was about 455 m in length. We divided the reach among 3
treatments with 2 replicates in each: reference, patchy, and dense. We installed SW
brush bundles in the patchy and dense treatments. The patchy treatment consisted of
3 patches of bundles that occupied the wetted width of the stream. Each was ~12 m
long (Fig. S1 in Supplemental File 1, available online at http://www.eaglehill.us/
SENAonline/suppl-files/s17-1-S2388-Sterling-s1, and, for BioOne subscribers, at
http://dx.doi.org/10.1656/S2388.s1). There was ~12 m of unmanipulated stream
between patches (~60 m length total/treatment). The dense treatment was identical
to the patchy treatment, but we filled the two 12-m gaps between patches with more
bundles leaving no gaps. The wetted-width changed over time; thus, the number
of bundles in the stream varied and we replaced bundles as needed, but at least 2
weeks prior to sampling.
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
78
The third treatment consisted of 2 unmanipulated reference reaches. Reference
reaches were also about 60 m long and had small amounts of ephemeral cover,
small side-pools, and scattered undercut banks that changed rapidly through time.
We made no attempt to remove habitat from the reference reaches. We categorized
3 segments of the stream within the study reach as open and collected no data in
those reaches. From downstream to upstream, the order of the stream segments
within the study reach was: patchy 1 (60 m), open 1 (60 m), dense 1 (60 m), reference
1 (60 m), open 2 (10 m), reference 2 (60 m), patchy 2 (60 m), open 3 (25 m),
and dense 2 (60 m).
We constructed brush bundles from freshly cut ~1.5–2-m–tall deciduous shrubs.
Stem diameters were ~1.5–3.5 cm diameter. We used zip ties to bind together 3–4
shrubs and fastened them to steel rebar driven into the stream bed.
We sampled the study reach 9 times from July 2009 to May 2013 at ~4–7-month
intervals. We employed single-pass backpack electroshocking with 3 people dipnetting
in an upstream direction to collect fishes from the entire reach in 1 day. We
allocated sufficient effort to thoroughly sample the entire area of each treatment.
We identified fishes in the field and released them near the center of the capture
reach. We preserved in 5% buffered formalin and brought back to the laboratory for
identification all fish for which field identifications were imposs ible.
Taxonomic and functional diversity indices
To quantify fish community diversity, we calculated 3 indices for each treatment
for each of the 9 samples: rarefied species richness (Colwell et al. 2012);
Hurlbert’s probability of an interspecific encounter (PIE; Hurlbert 1971), which
is a measure of evenness (i.e., probability that 2 individuals drawn randomly from
the sample represent different species); and the Berger–Parker dominance index,
which is the proportion of the most common species for a given sample (Berger and
Parker 1970). To estimate rarefied species richness, we used the program EstimateS
(Colwell 2013) and the individual-based option (Colwell et al. 2012) to produce
estimates for each sample through time. Hurlbert’s PIE has the advantage of not
confounding richness and evenness in 1 number (e.g., Shannon index), which renders
estimates easy to interpret (Hurlbert 1971). Likewise, using the proportion of
the most abundant species in a sample is a straightforward and easily interpretable
estimate of dominance. We employed the formula function in a spreadsheet to calculate
evenness and dominance.
To quantify functional fish-community diversity, we calculated 2 indices for
each treatment for each of the 9 samples (Fig. S2 inSupplemental File 1, available
online at http://www.eaglehill.us/SENAonline/suppl-files/s17-1-S2388-Sterlings1,
and, for BioOne subscribers, at http://dx.doi.org/10.1656/S2388.s1): functional
richness and functional evenness (Cornwell et al. 2006; Mason et al. 2005; Villéger
et al. 2008, 2010). Each of the indices accommodates multiple functional traits and
is measured in a multidimensional hypervolume. Functional richness is the volume
of functional-trait space occupied by a given assemblage of species and was
estimated using convex hulls. We calculated functional evenness using a minimum
Southeastern Naturalist
79
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
spanning tree that connected all species and measured the uniformity of species’
distribution and abundance along the tree (Fig. S2 in Supplemental File 1, available
online at http://www.eaglehill.us/SENAonline/suppl-files/s17-1-S2388-Sterlings1,
and, for BioOne subscribers, at http://dx.doi.or g/10.1656/S2388.s1).
We employed a principal coordinates analysis based on Gower’s distance to create
the functional space from which functional diversity estimates were calculated
(Maire et al. 2015; Villéger et al. 2008, 2010). All calculations were carried out in R
ver. 3.3.1 (R Core Team 2016) using R scripts and functions available online at http://
villeger.sebastien.free.fr/homepage.html (Villéger 2016). We estimated 13 functional
traits for each species to consider various aspects of ecological function including
life history, physiology, habitat, and trophic variables (Table S1 in Supplemental File
1, available online at http://www.eaglehill.us/SENAonline/suppl-files/s17-1-S2388-
Sterling-s1, and, for BioOne subscribers, at http://dx.doi.org/10.1656/S2388.s1). We
obtained functional traits for each species from the online FishTraits database (Frimpong
and Angermeier 2009, 2013) and from Ross (2001).
Statistical analyses
To test for differences among treatments for each of the 5 diversity indices,
we used a repeated measures MANOVA as implemented in JMP ver. 13.0 (SAS
Institute, Cary, NC). For studies using repeated measures over time with multivariate
data, 2 of the most common methods of analysis are mixed-model approaches
and longitudinal MANOVA. We chose to use MANOVA because the approach and
output are familiar to a wide audience, results are easily interpretable, and, for
simple balanced designs such as ours with no missing data, MANOVA performs
as well as the mixed-model approach. As a parametric method, data is assumed
to be distributed normally, though the method is robust to moderate violations of
this assumption (O’Brien and Kaiser 1985, Sall et al. 2005). The method is limited
to balanced data with no missing values and with qualitative differences among
repeated samples (e.g., different seasons), and does not accommodate categorical
variables. We log10-transformed the rarefied richness data (McDonald 2014) and
Logit-transformed all other diversity indices (Warton and Hui 2011). Our visual
inspections of data histograms for each index confirmed that they were distributed
normally. Alpha was adjusted (α = 0.027) for pairwise comparisons among treatments
using a false-discovery–rate method (Narum 2006).
In some studies, functional richness is correlated with species richness (Villéger
et al. 2008); thus, we performed correlation analysis between the 2 indices for each
treatment. We square-root–transformed and relativized the data to the stand ard deviate
(McCune and Grace 2002).
Characterizing differences in fish communities
We employed several methods to examine differences in community composition
among treatments. We summarized rank abundance (catch/unit of effort
[CPUE], fish/s) of species by treatment. Two nonmetric multidimensional-scaling
(NMDS) ordinations were created using species and family abundance data in PCORD
ver. 6.21 (McCune and Mefford 2011) to array species and families in sample
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
80
(i.e., treatment) space (transpose analysis), which indicates the ecological preferences
of each species and family (McCune and Grace 2002). We quantified relative
proportional abundance of families with >1 representative species among treatments
(±95% confidence intervals) to produce a visual representation (bar graphs)
of fish-family responses. We estimated confidence intervals using 10,000 iterations
in the Resampling Stats add-in for Excel (Statistics.com LLC 2009). We combined
the data into a single “wood” treatment for ordinations to render them more easily
interpretable and present results for the rank-abundance data for each woody treatment
and the 2 woody treatments combined.
Results
Taxonomic and functional diversity indices: MANOVA and correlation
We detected differences among treatments for rarefied richness (MANOVA: F =
34.16, df = 2, P < 0.024). Pairwise comparisons showed that rarefied richness was
higher in dense than in reference reaches (F = 34.16, df = 1, P < 0.01), but patchy
and reference reaches were not significantly different (F = 12.98, df = 1, P < 0.037).
No differences occurred between patchy and dense reaches (F = 5.03, df = 1, P =
0.11) (Fig. 2). Likewise, there were no interactions between time and treatment (F
= 0.7, df = 16, P = 0.76).
For evenness, differences existed among treatments (MANOVA: F = 12.52,
df = 2, P < 0.035). Pairwise comparisons showed the dense treatment had higher
evenness than the reference treatment (F = 23.02, df = 1, P < 0.017), but the patchy
and the reference treatments were not significantly different (F = 13.19, df = 1, P less than
0.036) No differences occurred between the patchy and the dense treatments (F =
1.36, df = 1, P = 0.33; Fig. 2). There were no interactions between time and treatment
(F = 1.15, df = 16, P = 0.37).
We detected differences in species dominance among treatments (MANOVA:
F = 15.67, df = 2, P < 0.026). Pairwise comparisons showed dominance was lower
in the dense than in the reference treatments (F = 24.42, df = 1, P < 0.013) and
in the patchy than in the reference treatments (F = 17.16, df = 1, P < 0.026). No
differences were detected between the patchy and the dense treatments (F = 1.41,
df = 1, P = 0.32) (Fig. 2). There were no interactions between time and treatment
(F = 0.97, df = 16, P = 0.51).
For functional richness, we detected no significant differences in the overall
test (MANOVA: F = 5.02, df = 2, P = 0.11. Pairwise comparisons showed no
significant difference in functional richness between the dense and the reference
treatments (F = 9.37, df = 1, P < 0.053), between the patchy and the reference
treatments (F = 4.15, df = 1, P = 0.13) or between the patchy and the dense treatments
(F = 0.1.41, df = 1, P = 0.36) (Fig. 3). There were no interactions between
time and treatment (F = 0.83, df = 16, P = 0.64).
Mean functional evenness was highly similar among treatments and not significantly
different (F = 0.89, df = 2, P = 0.49; Fig. 3). Correlations between functional
richness and species richness showed mixed responses, but all were non-significant
(P > 0.31): reference, r = 0.13; patchy, r = -0.25; and dense, r = -0.07.
Southeastern Naturalist
81
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
Ranked abundance
Ranked abundance (CPUE) for each treatment showed that most species had 1
of 3 responses (Table 1): a positive response to the woody treatments and a negative
one to the reference; a positive response to the reference and a negative one to
the woody treatments; or a mixed response showing higher abundance in the reference
and dense treatments and lower abundance in the patchy treatment. The most
abundant species for all treatments was Notropis rafinesquei (Yazoo Shiner), which
showed a negative response to the woody treatments. Two commonly sampled sunfishes
Lepomis megalotis (Longear Sunfish) and Lepomis macrochirus (Bluegill)
were more abundant in the reference and the dense treatments than in the patchy
Figure 2. Mean values for each taxonomic
diversity index is shown (±
95% CIs) for each treatment.
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
82
treatment; however, Lepomis cyanellus (Green Sunfish) showed little response
to any of the treatments. Cyprinids varied in their responses with some showing a
pattern similar to Bluegills, e.g., Pimephales notatus (Bluntnose Minnow), others
showed a positive response to the woody treatments, e.g., Notemigonus crysoleucas
(Golden Shiner) and Cyprinella venusta (Blacktail Shiner), others showed a negative
response to the woody treatments, e.g., Yazoo Shiner, and still others showed
little response at all, e.g., Cyprinella camura (Bluntface Shiner).
Percina sciera (Dusky Darter) and Noturus miurus (Brindled Madtom) showed
among the strongest positive responses to the woody treatments. Two other darter
species, Etheostoma lynceum (Brighteye Darter) and Etheostoma artesiae (Redspot
Darter), also showed positive responses to the woody treatments. However, the
frequently sampled Noturus phaeus (Brown Madtom) only showed a weak positive
response to the woody treatments similar to another catfish, Ameiurus natalis (Yellow
Bullhead).
NMDS ordinations
Ordinations of species in treatment space described gradients in wood density.
The final NMDS ordination of fish species in treatment space recommended a 3-dimensional
solution and had moderate stress (10.2) and low instability (less than 0.0001)
(Fig. 4). Separate runs with random starting points returned very similar results;
however, only the third axis was meaningful. On axis 3, samples described a
gradient from woody treatments to the reference treatment. Yazoo Shiners were
Figure 3. Mean values for each functional
diversity index is shown (±
95% CIs) for each treatment.
Southeastern Naturalist
83
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
most strongly associated with the reference treatment. Etheostoma gracile (Slough
Darter), Luxilus chrysocephalus (Striped Shiner), and Semotilus atromaculatus
(Creek Chub) were also associated with reference reaches. Species showing the
strongest associations with wood were Redspot and Brighteye Darters, Ameiurus
melas (Black Bullhead), Micropterus salmoides (Largemouth Bass), and several
minnows: Hybognathus nuchalis (Mississippi Silvery Minnow), Lythrurus fumeus
(Ribbon Shiner), and Golden Shiner (Table 2).
Table 1. Ranked abundance (CPUE, fish/s) for each fish species and each treatment; Woody = patchy
and dense combined. Key: ANAT: Ameiurus natalis, AMEL: Ameiurus melas, NMIU: Noturus miurus,
NPHA: Noturus phaeus, PSCI: Percina sciera, EART: Etheostoma artesiae, EGRA: E. gracile,
ELYN: E. lynceum, LCYA: Lepomis cyanellus, LGUL: L. gulosus, LMAC: L. macrochirus, LMAR:
L. marginatus, LMEG: L. megalotis, MSAL: Micropterus salmoides, CCAM: Cyprinella camura,
CVEN: C. venusta, LCHR: Luxilus chrysocephalus, NCRY: Notemigonus crysoleucas, NRAF: Notropis
rafinesquei, NATH: Notropis atherinoides, LUMB: Lythrurus umbratilis, LFUM: Lythrurus
fumeus, PNOT: Pimephales notatus, SATR: Semotilus atromaculatus, HNUC: Hybognathus nuchalis,
ASAY: Aphredoderus sayanus, FOLI: Fundulus olivaceus, ECLA: Erimyzon claviformis, GAFF:
Gambusia affinis
Reference Patchy Dense Woody
Species CPUE Species CPUE Species CPUE Species CPUE
NRAF 0.1252 NRAF 0.0701 NRAF 0.0959 NRAF 0.0824
FOLI 0.0402 FOLI 0.0467 FOLI 0.0434 FOLI 0.0451
LMEG 0.0257 LMAC 0.0206 LMAC 0.0327 LMAC 0.0264
LMAC 0.0241 LMEG 0.0152 LMEG 0.0259 LMEG 0.0203
PNOT 0.0216 CCAM 0.0141 PNOT 0.0200 CCAM 0.0150
CCAM 0.0137 NPHA 0.0125 CCAM 0.016 PNOT 0.0146
SATR 0.0107 SATR 0.0122 NPHA 0.0115 NPHA 0.0120
NPHA 0.0096 PNOT 0.0096 LCYA 0.0100 SATR 0.0105
LCYA 0.0074 LCYA 0.0077 SATR 0.0086 LCYA 0.0088
GAFF 0.0059 NMIU 0.0037 NMIU 0.0053 NMIU 0.0045
ANAT 0.0028 GAFF 0.0032 GAFF 0.0051 GAFF 0.0041
LUMB 0.0022 PSCI 0.0032 ANAT 0.0041 ANAT 0.0034
CVEN 0.0019 ANAT 0.0028 PSCI 0.0036 PSCI 0.0034
NMIU 0.0019 CVEN 0.0028 CVEN 0.0032 CVEN 0.0030
ASAY 0.0016 ASAY 0.0023 LUMB 0.0029 LUMB 0.0021
ECLA 0.0010 ELYN 0.0016 NCRY 0.0023 ASAY 0.0021
LGUL 0.0010 ECLA 0.0015 ASAY 0.0018 NCRY 0.0018
MSAL 0.0008 NCRY 0.0015 ECLA 0.0018 ECLA 0.0017
ELYN 0.0006 LUMB 0.0014 MSAL 0.0016 ELYN 0.0014
LCHR 0.0006 EART 0.0009 LGUL 0.0012 MSAL 0.0010
NCRY 0.0006 LGUL 0.0007 ELYN 0.0011 LGUL 0.0009
PSCI 0.0006 MSAL 0.0005 NATH 0.0007 EART 0.0006
AMEL 0.0006 LFUM 0.0002 LCHR 0.0005 NATH 0.0003
EGRA 0.0002 AMEL 0.0002 EART 0.0004 LCHR 0.0002
EART 0.0000 LMAR 0.0001 LMAR 0.0002 AMEL 0.0002
LMAR 0.0000 EGRA 0.0000 AMEL 0.0002 LMAR 0.0002
HNUC 0.0000 LCHR 0.0000 HNUC 0.0001 LFUM 0.0001
LFUM 0;0000 HNUC 0.0000 EGRA 0.0000 HNUC 0.0001
NATH 0;0000 NATH 0.0000 LFUM 0.0000 EGRA 0.0000
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
84
Similarly, the ordination of families in treatment space revealed a wooddensity
gradient. The final NMDS ordination of fish families in treatment space
Figure 4. Results from NMDS ordination of fish species in sample space showing samples
from woody treatments and reference reaches; key for species abbreviations are in Table 1;
the portion of the graph containing sample points is enlar ged for clarity.
Southeastern Naturalist
85
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
recommended a 2-dimensional solution and had low stress (0.0003) and instability
(less than 0.00001) (Fig. 5). Separate runs with random starting points returned very similar
results; however, only the first axis was biologically interpretable. Samples showed
a gradient from woody treatments to the reference treatment that is similar to that
from the ordination of species data. Cyprinidae were associated with the reference
treatment. Centrachidae and Fundulidae showed little association with either
woody or reference treatments. Ictaluridae, Poeciliidae, Percidae, Aphredoderidae,
and Catostomidae were all associated with the woody treatments (Table 3).
Proportional data
Relative proportional abundances within families were generally consistent
with ordinations and ranked-abundance results. Darters (Percidae) and catfishes
(Ictaluridae) showed the strongest positive response to the woody treatments,
but differences were not apparent in response between those treatments (Fig. 6).
Proportional abundance of Cyprinidae and Centrarchidae showed a gradient in
response with apparent differences between all treatments. For all families, abundance
was lowest in the reference treatment (Fig. 6).
Table 2. NMDS ordination scores for each fish species on 3 axes.
Species Axis 1 Axis 2 Axis 3
Ameiurus natalis (Lesueur) (Yellow Bullhead) -0.1529 -0.0260 -0.2636
Ameiurus melas (Rafinesque) (Black Bullhead) -0.2803 0.8999 -0.7431
Noturus miurus Jordan (Brindled Madtom) -0.1697 -0.1529 -0.2282
Noturus phaeus Taylor (Brown Madtom) -0.0307 -0.3709 0.0979
Percina sciera (Swain) (Dusky Darter) 0.3398 0.2817 -0.2401
Etheostoma artesiae (Hay) (Redspot Darter) -0.5716 -0.8626 -0.9106
Etheostoma gracile (Girard) (Slough Darter) 1.4525 0.8174 0.7136
Etheostoma lynceum Hay (Brighteye Darter) 0.4855 0.1032 -0.7750
Lepomis cyanellus Rafinesque (Green Sunfish) -0.0840 -0.2030 0.2831
Lepomis gulosus (Cuvier) (Warmouth) -0.0489 0.6754 -0.1922
Lepomis macrochirus Rafinesque (Bluegill) 0.0735 -0.4776 0.5269
Lepomis marginatus (Holbrook) (Dollar Sunfish) -1.7251 -0.1657 0.1227
Lepomis megalotis (Rafinesque) (Longear Sunfish) 0.1446 -0.3790 0.6007
Micropterus salmoides (Lacepède) (Largemouth Bass) -0.8582 -0.2077 -0.5084
Cyprinella camura (Jordan and Meek) (Bluntface Shiner) 0.0745 -0.3013 0.3569
Cyprinella venusta Girard (Blacktail Shiner) -0.0672 0.1806 0.0482
Luxilus chrysocephalus Rafinesque (Striped Shiner) -0.7273 0.9575 0.6617
Notemigonus crysoleucas (Mitchell) (Golden Shiner) -0.4237 0.3195 -0.5558
Notropis rafinesquei Suttkus (Yazoo Shiner) 0.3561 -0.6992 0.8260
Notropis atherinoides Rafinesque (Emerald Shiner) 0.0552 1.3306 0.1178
Lythrurus umbratilis (Girard) (Redfin Shiner) -0.0509 0.5515 0.3992
Lythrurus fumeus (Evermann) (Ribbon Shiner) 1.3728 0.3397 -0.6950
Pimephales notatus (Rafinesque) (Bluntnose Minnow) 0.1787 -0.3662 0.4994
Semotilus atromaculatus (Mitchill) (Creek Chub) 0.0044 -0.2673 0.6275
Hybognathus nuchalis Agassiz (Mississippi Silvery Minnow) 1.148 -1.2687 -0.7757
Aphredoderus sayanus (Gilliams) (Pirate Perch) -0.5887 -0.0557 -0.3304
Fundulus olivaceus (Storer) (Blackspotted Topminnow) 0.1802 -0.5908 0.5824
Erimyzon claviformis (Girard) (Western Creek Chubsucker) 0.3771 0.2357 -0.5210
Gambusia affinis (Baird and Girard) (Western Mosquitofish) -0.4636 -0.2982 0.2752
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
86
Discussion
Examination of P-values from the MANOVA tests show that the results contrasting
the patchy with the reference treatment were either marginally significant
or insignificant after adjustment for multiple pairwise tests. In contrast, the dense
treatment was consistently significantly different from the reference treatment,
but the dense treatment was no different than the patchy treatment. To interpret
biological differences between treatments at P-values of less than 0.027 (i.e., dense versus
Figure 5. Results from NMDS ordination of fish families in sample space showing samples
from woody treatments and reference reaches; the portion of the graph containing sample
points is enlarged for clarity.
Table 3. NMDS ordination scores for each fish family on 2 axes.
Family Axis 1 Axis 2
Percidae -0.689 0.247
Ictaluridae 0.139 -0.012
Centrarchidae 0.904 0.029
Cyprinidae 1.378 0.238
Fundulidae 0.629 0.199
Catostomidae -1.294 0.717
Aphredoderidae -0.998 -0.564
Poeciliidae -0.069 -0.854
Southeastern Naturalist
87
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
reference) but not when P-values were less than 0.04 (i.e., patchy versus reference) is unreasonable
and lacking in biological rationale (Gelman and Stern 2006). Additional
samples would undoubtedly have increased our ability to detect differences, even
with adjustments for multiple comparisons, by reducing variance and increasing
power. We interpret our results as showing biologically significant differences
between woody reaches and reference reaches for each of the 3 taxonomic diversity
indices and perhaps a difference between the dense and reference treatments for the
functional richness index.
Results showing higher species richness and evenness and lower dominance
coupled with suggested higher functional richness in the woody-treatment reaches
indicate that within the habitat-starved streams of channelized or incised watersheds
(Shields et al. 1994, Warren et al. 2002), even a modest input of wood may
have a positive impact on fish community and functional diversity. This conclusion
is supported by the proportional-abundance data (Fig. 6). Results were consistent
with general expectations (Moulliot et al. 2013, Roni et al. 2014, Warren 2012, but
see Stewart et al. 2009), as well as with the few experimental studies from warmwater
streams in the US (Angermeier and Karr 1984, Gatz 2008, Hrodey and Sutton
2008, Warren et al. 2009, but see Schneider and Winemiller 2008).
Two factors may have had important impacts on the results of this study. The
first is that the downstream road crossing may have been a finer filter to fish
passage than we anticipated. Fish species diversity was higher downstream of
Figure 6. Proportional abundances (± 95% CIs) of fish families with more than 1 species are
shown for each treatment.
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
88
the crossing (K.A. Sterling and M.L. Warren Jr., unpubl. data), and though we
anticipated immigration from this larger species pool, it did not occur or at least
was highly limited, and may have decreased the effect of the treatments on the
fish community. A second factor was that our woody-bundle installations were
ephemeral. Though we maintained the bundles through time, we lost a substantial
proportion of the bundles after every flood event, which then had to be replaced.
The lack of stable habitat and cover through time likely decreased the effect of
the treatments on the fish community. This contention is supported by evidence
within the Cypress Creek drainage showing that fish assemblages are highly unstable
and reactive to flashy flows, likely because fishes lack stable habitat and
cover (Adams et al. 2004). Had our wood-bundle installations been permanent, it
is likely we would have observed a more pronounced response.
Mean values for the 3 taxonomic diversity indices and the proportional abundance
data suggest there may be a gradient in fish community response to the
treatments, with the dense treatment showing the largest response. However, this
result was likely due to more than just the wood introduced into the stream. The
installation of wood bundles certainly provided complex cover, but also created
areas of higher and slower current velocities as well as areas with deeper and
shallower water around the bundles relative to the reference treatment, which had
more or less homogenous flow and water depths. These observations are consistent
with changes observed in another study (Angermeier and Karr 1984). Differences
in flow velocity were also apparent between the dense and patchy treatments. The
patchy treatments mostly created areas of swifter currents, and the dense treatments
created areas of both slow and swift currents. This flow pattern was reflected in the
responses of species as shown in the ranked abundance data. For example, Longear
Sunfish and Bluegills had lower abundances in the patchy treatment compared
with the dense and reference treatments, likely due to the presence of swift water.
Unfortunately, we did not measure water depths and velocities across treatments
through time to investigate the effects of these variables on fish populations. Even
so, it is likely that the overall positive response of the fish community to the input
of wood was due at least in part to the creation of more-variable current velocities
and water depths (Dolloff and Warren 2003).
As summarized in the Introduction, fishes in lowland, soft-bottom streams rely
on wood to provide habitat complexity and stability, cover, spawning substrate,
current and drought refugia, and diverse invertebrate prey. We believe most, if
not all, of these factors were likely influential in explaining our results, and other
fish–wood relationship studies from northern Mississippi support this view. For
example, Brown Madtoms and Aphredoderus sayanus (Pirate Perch) used woody
cover with species-specific structural characteristics, showing that fishes select for
certain characteristics and perhaps partition woody resources, allowing for greater
species diversity (Monzyk et al. 1997). Sunfishes ( Lepomis spp.) showed close associations
with LW, perhaps because of increased pool volumes, refuge from stronger
currents, greater invertebrate food resources, and cover from predation within
degraded streams (Shields et al. 1998, Warren et al. 2002). A study that installed 3
Southeastern Naturalist
89
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
types of SW or detritus bundles showed that use by a diverse fish assemblage was
higher in streams that were more degraded and lacked woody cover, and also that
use increased after a spate in degraded streams, but not in a less degraded stream,
which suggests cover was limiting in the streams that were more degraded and
fishes used the installed bundles as a high-flow refuge (W arren et al. 2009).
These regional results and our own are consistent with the literature across the
Coastal Plain and other warmwater, lowland streams (Angermeier and Karr 1984,
Gatz 2008, Hrodey and Sutton 2008, but see Schneider and Winemiller 2008);
however, published studies, including ours, are mostly limited to describing patterns,
and actual processes are understudied. Study results are consistent and our
study stream is typical of many lowland, soft-bottom streams; thus, we believe
our results are applicable to other highly degraded lowland streams, especially
channelized streams running through agricultural lands. Confirmation of our
findings is warranted, and we plan on replicating this study in other streams. Our
study was limited in scope, but further research could include measurement of
physical stream-habitat variables, flow, changes in the volume of wood through
time, biomass and length of fishes, different spatial scales, and other components
of the stream community (e.g., presence and/or abundance of invertebrate insects,
crayfishes, amphibians, or fungi).
Mean functional richness, but not mean functional evenness, showed an apparent
gradient in response to the addition of wood—reference reaches had the lowest
mean values, values for the patchy treatment were intermediate, and the dense
treatment had the highest values (Fig. 3). However, the lack of a clear positive or
negative correlation between functional and rarefied species richness across samples
and treatments precludes any definitive conclusion regarding why functional
richness was apparently higher in the dense treatment. This result deserves to be
explored further. The lack of differences among treatments for functional evenness
is likely due to functional redundancy and relatively small changes in abundance
among species and treatments.
General agreement emerged among our descriptive results (ranked abundance,
ordinations, and proportional abundances) that darters and madtom catfishes
showed the strongest response to the introduction of SW. This finding is consistent
with results from 2 studies that included streams in the Little Tallahatchie River
system and that examined, among other factors, the effects of LW (Warren et al.
2002) and SW (Warren et al. 2009) on fish communities.
Species-specific responses differed within families, and there were apparent
positive, negative, mixed, and neutral responses to the woody treatments. This result
shows that introducing SW to streams may not benefit all species and that the
design of SW installations may greatly affect whether or not management goals are
achieved (Angermeier and Karr 1984, Gatz 2008, Langford et al. 2012, Warren et
al. 2009).
Overall, though results point to significant differences for 3 of 5 taxonomic
and functional-diversity indices, numerical differences among treatments were
moderate. However, an apparently small or moderate numerical shift in abundance
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
90
or number of species present may have a great impact on stream communities
(Dangles and Malmqvist 2004, Jackson et al. 2001, Taylor and Warren 2001) and
especially on ecosystem functioning (Leitão et al. 2016, Martin et al. 2016, Moulliot
et al. 2013). Investigation of the effects of fish community changes in response
to the addition of wood on stream ecosystem function in lowland streams is an area
ripe for investigation (e.g., Gary and Har grave 2017).
Our results indicate that the introduction of modest amounts of SW in degraded
Coastal Plain streams can increase fish community and perhaps functional diversity
within stream reaches with introduced wood relative to reaches without installed
wood. Notably, this increase was accomplished using only a few hand tools and
limited personnel. A next step in researching the effects of SW on Coastal Plain
stream fishes is to estimate whether or not the addition of SW increases species
diversity, abundance, and ecosystem function over a larger spatiotemporal scale
using stable installations of wood.
Acknowledgments
We thank the many people who contributed to this project by assisting in the field and the
laboratory, sharing ideas and information, providing logistical support, and offering numerous
other professional courtesies: S. Adams, Z. Barnett, M. Bland, A. Commens-Carson, W.
Haag, G. Henderson, A. Jacobs, C. Jenkins, E. McGuire, G. McWhirter, A. Reitl, V. Reithel,
J. Ryndock, and D. Warren. This study was supported by the Stream Ecology Laboratory,
Center for Bottomland Hardwoods Research, USDA Forest Service, Oxford, MS.
Literature Cited
Adams, S.B., M.L. Warren Jr., and W.R. Haag. 2004. Spatial and temporal patterns in
fish assemblages of Upper Coastal Plain streams, Mississippi, USA. Hydrobiologia
528:45–61.
Angermeier, P.L., and J.R. Karr. 1984. Relationships between woody debris and fish
habitat in a small warmwater stream. Transactions of the American Fisheries Society
113:716–726.
Benke, A.C., and J.B. Wallace. 2015. High secondary- production in a Coastal Plain river is
dominated by snag invertebrates and fueled mainly by amorphous detritus. Freshwater
Biology 60:236–255.
Benke, A.C., R.L. Henry III, D.M. Gillespie, and R.J. Hunter. 1985. Importance of snag
habitat for animal production in southeastern streams. Fisherie s 10:8–13.
Berger, W.H., and F.L. Parker. 1970. Diversity of planktonic Foraminifera in deep-sea sediments.
Science 168:1345–1347.
Colwell, R.K. 2013. EstimateS: Sstatistical estimation of species richness and shared species
from samples. Version 9.1 Available online at http://viceroy.eeb.uconn.edu/estimates/
index.html. Accessed 19 December 2016.
Colwell, R.K., A. Chao, N.J. Gotelli, S.-Y. Lin, C.X. Mao, R.L. Chazdon, and J.T. Longino.
2012. Models and estimators linking individual-based and sample-based rarefaction,
extrapolation, and comparison of assemblages. Journal of Plant Ecology 5:3–21.
Cornwell, W.K., D.W. Schwilk, and D.D. Ackerly. 2006. A trait-based test for habitat filtering:
Convex hull volume. Ecology 87:1465–1471.
Southeastern Naturalist
91
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
Crook, D., and A. Robertson. 1999. Relationships between riverine fish and woody debris:
implications for lowland rivers. Marine and Freshwater Research 50:941–953.
Dangles, O., and B. Malmqvist. 2004. Species richness–decomposition relationships depend
on species dominance. Ecology Letters 7:395–402.
Dolloff, C.A., and M.L. Warren Jr. 2003. Fish relationships with large wood in small
streams. American Fisheries Society Symposium 37:179–193.
Etnier, D.A. 1972. The effect of annual rechanneling on a stream fish population. Transactions
of the American Fisheries Society 101:372–375.
Frimpong, E.A., and P.L. Angermeier. 2009. FishTraits: A database of ecological and lifehistory
traits of freshwater fishes of the United States. Fisher ies 34:487–495.
Frimpong, E.A., and P.L. Angermeier. 2013. FishTraits Database. Available online at http://
www.fishtraits.info/. Accessed 19 December 2016.
Gary, K.P., and C.W. Hargrave. 2017. In-stream structure alters density-dependent fish effects
in stream ecosystems. Ecology of Freshwater Fish 26:403–4 14.
Gatz, A.J., Jr.. 2008. The use of floating overhead cover by warmwater stream fishes. Hydrobiologia
600:307–310.
Gelman, A., and H. Stern. 2006. The difference between “significant” and “not significant”
is not itself statistically significant. The American Statistician 60:328–331.
Hrodey, P.J., and T.M. Sutton. 2008. Fish community responses to half-log additions in
warmwater streams. North American Journal of Fisheries Management 28:70–80.
Hurlbert, S.H. 1971. The nonconcept of species diversity: A critique and alternative parameters.
Ecology 52:577–585.
Jackson, D.A., P.R. Peres-Neto, J.D. Olden. 2001. What controls who is where in freshwater
fish communities: The roles of biotic, abiotic, and spatial factors. Canadian Journal of
Fisheries and Aquatic Sciences 58:157–170.
Jelks, H.L., S.J. Walsh, N.M. Burkhead, S. Contreras-Balderas, E. Diaz-Pardo, D.A. Hendrickson,
J. Lyons, N.E. Mandrak, F. McCormick, J.S. Nelson, S.P. Platania, B.A. Porter,
C.B. Renaud, J.J. Schmitter-Soto, E.B. Taylor, and M.L. Warren Jr. 2008. Conservation
status of imperiled North American freshwater and diadromous fishes. Fisheries
33:372–407.
Keeton, W.S., C.E. Kraft, and D.R. Warren. 2007. Mature and old-growth riparian forests:
Structure, dynamics, and effects on Adirondack stream habitats. Ecological Applications
17:852–868.
Kramer, N., and E. Wohl. 2017. Rules of the road: A qualitative and quantitative synthesis
of large-wood transport through drainage networks. Geomorphology 279: 74–97.
Langford, T., J. Langford, and S.J. Hawkins. 2012. Conflicting effects of woody debris
on stream fish populations: Implications for management. Freshwater Biology
57:1096–1111.
Lau, J.K., T.E. Lauer, and M.L. Weinman. 2006. Impacts of channelization on stream
habitats and associated fish assemblages in east central Indiana. The American Midland
Naturalist 156:319–330.
Leitão, R.P., J. Zuanon, S. Villéger, S.E. Williams, C. Baraloto, C. Fortunel, F.P. Mendonça,
and D. Mouillot. 2016. Rare species contribute disproportionately to the functional
structure of species assemblages. Proceedings of the Royal Soci ety B 283:20160084.
Leps, M., A. Sundermann, J.D. Tonkin, A.W. Lorenz, and P. Haase. 2016. Time is no healer:
Increasing restoration age does not lead to improved benthic invertebrate communities
in restored river reaches. Science of the Total Environment 557:722–732.
Maire, E., G. Grenouillet, S. Brosse, and S. Villéger. 2015. How many dimensions are
needed to accurately assess functional diversity? A pragmatic approach for assessing the
quality of functional spaces. Global Ecology and Biogeography 2 4:728–740.
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
92
Martin, E.C., K.B. Gido, N. Bello, W.K. Dodds, and A. Veach. 2016. Increasing fish taxonomic
and functional richness affects ecosystem properties of small headwater prairie
streams. Freshwater Biology 61:887–898.
Mason, N.W.H., D. Mouillot, W.G. Lee, and J.B. Wilson. 2005. Functional richness,
functional evenness and functional divergence: The primary components of functional
diversity. Oikos 111:112–118.
McCune, B., and J.B. Grace. 2002. Analysis of Ecological Communities. MjM Software,
Gleneden Beach, OR. 300 pp.
McCune, B., and M.J. Mefford. 2011. PC-ORD, ver. 6.21, Multivariate Analysis of Ecological
Data. MjM Software, Gleneden Beach, OR. Available online at https://www.pcord.
com/. Accessed 19 December 2016.
McDonald, J.H. 2014. Handbook of Biological Statistics, 3rd Edition. Sparky House Publishing,
Baltimore, MD. 299 pp.
Meffe, G.K., and A.L. Sheldon. 1988. The influence of habitat structure on fish-assemblage
composition in southeastern blackwater streams. American Midland Naturalist
120:225–240.
Montgomery D.R., B.D. Collins, J.M. Buffington, and T.B. Abbe. 2003. Geomorphic effects
of wood in rivers. American Fisheries Society Symposium 37:21–47.
Monzyk, F.R., W.E. Kelso, and D.A. Rutherford. 1997. Characteristics of woody cover
used by Brown Madtoms and Pirate Perch in coastal plain streams. Transactions of the
American Fisheries society 126:665–675.
Moulliot, D., D.R. Bellwood, C. Baraloto, J. Chave, R. Galzin, M. Harmelin-Vivien, M.
Kulbicki, S. Lavergne, S. Lavorel, and N. Mouquet. 2013. Rare species support vulnerable
functions in high-diversity ecosystems. PLoS Biology 1 1:e1001569.
Narum, S.R. 2006. Beyond Bonferroni: Less conservative analyses for conservation genetics.
Conservation Genetics 7:783–787.
O’Brien, R.G., and M.K. Kaiser. 1985. MANOVA method for analyzing repeated measures
designs: An extensive primer. Psychological Bulletin 97:316–333.
Palmer, M.A., H.L. Menninger, and E. Bernhardt. 2010. River restoration, habitat heterogeneity,
and biodiversity: A failure of theory or practice? Freshwater Biology 55:205–222.
Pierce, S.C., R. Kröger, and R. Pezeshki. 2012. Managing artificially drained low-gradient
agricultural headwaters for enhanced ecosystem functions. Biolo gy 1:794–856.
Pilotto, F., A. Bertoncin, G.L. Harvey, G. Wharton, and M.T. Pusch. 2014. Diversification
of stream invertebrate communities by large wood. Freshwater Biology 59:2571–2583.
R Core Team. 2016. R: A language and environment for statistical computing. R Foundation
for Statistical Computing, Vienna, Austria. Available online at https://www.R-project.
org/. Accessed 19 December 2016.
Roni, P., T. Beechie, G. Pess, and K. Hanson. 2014. Wood placement in river restoration:
Fact, fiction, and future direction. Canadian Journal of Fisheries and Aquatic Sciences
72:466–478.
Ross, S.T. 2001. Inland Fishes of Mississippi. University Press of Mississippi, Jackson,
MS. 624 pp.
Roth, N.E., J.D. Allan, and D.L. Erickson. 1996. Landscape influences on stream biotic
integrity assessed at multiple spatial scales. Landscape Ecolog y 11:141–156.
Sall, J., L. Creighton, and A. Lehman. 2005. JMP Start Statistics: A Guide to Statistics Using
JMP and JMP IN Software, 3rd Edition. SAS Institute, Inc. Cary, NC. 560 pp.
Schneider, K.N., and K.O. Winemiller. 2008. Structural complexity of woody debris patches
influences fish and macroinvertebrate species richness in a temperate floodplain-river
system. Hydrobiologia 610:235–244.
Southeastern Naturalist
93
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
Schoof, R. 1980. Environmental impact of channel modification. Journal of the American
Water Resources Association 16:697–701.
Scott, M.C., and P.L. Angermeier. 1998. Resource use by 2 sympatric black basses in impounded
and riverine sections of the New River, Virginia. North American Journal of
Fisheries Management 18:221–235.
Shields, F.D., Jr., S.S. Knight, and C.M. Cooper. 1994. Effects of channel incision on baseflow
stream habitats and fishes. Environmental Management 18:43–5 7.
Shields, F.D., Jr., S.S. Knight, and C.M. Cooper. 1997. Rehabilitation of warmwater stream
ecosystems following channel incision. Ecological Engineering 8 :93–116.
Shields, F.D., Jr., S.S Knight, C.M. Cooper. 1998. Rehabilitation of aquatic habitats
in warmwater streams damaged by channel incision in Mississippi. Hydrobiologia
382:63–86.
Shields, F.D., Jr., S.S. Knight, and C.M. Cooper. 2000. Cyclic perturbation of lowland
river channels and ecological response. Regulated Rivers: Research and Management
16:307–325.
Shields, Jr., F.D., S.S. Knight, N. Morin, and J. Blank. 2003. Response of fishes and
aquatic habitats to sand-bed stream restoration using large woody debris. Hydrobiologia
494:251–257.
Shields, F.D., Jr., S.S. Knight, and J.M. Stofleth. 2006. Large wood addition for aquatic
habitat rehabilitation in an incised, sand-bed stream, Little Topshaw Creek, Mississippi.
River Research and Applications 22:803–817.
Smock, L.A., and E. Gilinsky. 1992. Coastal plain blackwater streams. Pp. 272–311, In C.T.
Hackney, S.M. Adams, and W.H. Martin (Eds.). Biodiversity of the Southeastern United
States Aquatic Communities. J. Wiley and Sons, Inc., New York, NY. 779 pp.
Statistics. Com LLC. 2009. Resampling Stats add-in for Excel version 4.0. Arlington, VA.
Available online at http://www.resample.com/. Accessed 1 August 2017.
Stewart, G.B., H.R. Bayliss, D.A. Showler, W.J. Sutherland, and A.S. Pullin. 2009. Effectiveness
of engineered in-stream–structure mitigation measures to increase salmonid
abundance: A systematic review. Ecological Applications 19:931–941.
Sullivan, B.E., L.S. Rigsby, A. Berndt, M. Jones-Wuellner, T.P. Simon, T. Lauer, and M. Pyron.
2004. Habitat influence on fish community assemblage in an agricultural landscape
in 4 east-central Indiana streams. Journal of Freshwater Ecolog y 19:141–148.
Taylor, C.M., and M.L. Warren Jr. 2001. Dynamics in species composition of stream fish
assemblages: Environmental variability and nested subsets. Ecol ogy 82:2320–2330.
Villéger, S. 2016. Online webpage for downloading R scripts to calculate functional diversity
indices. Available online at http://villeger.sebastien.free.fr/homepage.html. Accessed
2 August 2017.
Villéger, S., N.W.H. Mason, and D. Mouillot. 2008. New multidimensional functional diversity
indices for a multifaceted framework in functional ecology. Ecology 89:2290–2301.
Villéger, S., J.R. Miranda, D.F. Hernández, and D. Mouillot. 2010. Contrasting changes in
taxonomic vs. functional diversity of tropical fish communities after habitat degradation.
Ecological Applications 20:1512–1522.
Warren, M.L., Jr., 2012. Forest landscape restoration: Linkages with stream fishes of the
southern United States. Pages 221–264, In J. Stanturf, P. Madsen, and D. Lamb (Eds.).
A Goal-Oriented Approach to Forest Landscape Restoration. Springer, Dordrecht, Netherlands.
476 pp.
Warren, M.L., Jr., W.R. Haag, and S.B. Adams. 2002. Forest linkages to diversity and abundance
in lowland stream fish communities. Pp. 169–182, In M.W. Holland, M.L. Warren
Jr., and J.A. Stanturf (Eds.). Proceedings of a Conference on Sustainability of Wetlands
and Water Resources: How Well Can Riverine Wetlands Continue to Support Society
into the 21st century? US Forest Service General Technical Report SRS-50. Washington,
DC. 191 pp.
Southeastern Naturalist
K.A. Sterling and M,L. Warren Jr.
2018 Vol. 17, No. 1
94
Warren, M.L., Jr., A.L. Sheldon, and W.R. Haag. 2009. Constructed woody microhabitat
bundles for sampling fishes and cray fishes in Coastal Plain streams. North American
Journal of Fisheries Management 29:330–342.
Warton, D.I., and F.K.C. Hui. 2011. The arcsine is asinine: The analysis of proportions in
ecology. Ecology 92:3–10.
Webb, A.A., and W.D. Erskine. 2005. Natural variability in the distribution, loading, and
induced scour of large wood in sand-bed forest streams. River Research and Applications
21:169–185.
Williams, M. 1989. Americans and Their Forests. Cambridge University Press, New York,
NY. 624 pp.
Wohl, E.E. 2004. Disconnected Rivers: Linking Rivers to Landscapes. Yale University
Press, New Haven, CT. 301 pp.
Wyżga, B., P. Mikuś, J. Zawiejska, V. Ruiz-Villanueva, R.J. Kaczka, and W. Czech. 2017.
Log transport and deposition in incised, channelized, and multithread reaches of a
wide mountain river: Tracking experiment during a 20-year flood. Geomorphology
279:98–111.