Fish Assemblages in the Upper Esopus Creek, NY:
Current Status, Variability, and Controlling Factors
Barry P. Baldigo, Scott D. George, and Walter T. Keller
Northeastern Naturalist, Volume 22, Issue 2 (2015): 345–371
Full-text pdf (Accessible only to subscribers. To subscribe click here.)
Access Journal Content
Open access browsing of table of contents and abstract pages. Full text pdfs available for download for subscribers.
Current Issue: Vol. 30 (3)
Check out NENA's latest Monograph:
Monograph 22
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
345
2015 NORTHEASTERN NATURALIST 22(2):345–371
Fish Assemblages in the Upper Esopus Creek, NY:
Current Status, Variability, and Controlling Factors
Barry P. Baldigo1,*, Scott D. George1, and Walter T. Keller2
Abstract - The Upper Esopus Creek receives water diversions from a neighboring basin
through the Shandaken Tunnel (the portal) from the Schoharie Reservoir. Although the portal
is closed during floods, mean flows and turbidity of portal waters are generally greater
than in Esopus Creek above their confluence. These conditions could potentially affect local
fish assemblages, yet such effects have not been assessed in this highly regulated stream.
We studied water quality, hydrology, temperature, and fish assemblages at 18 sites in the
Upper Esopus Creek during 2009–2011 to characterize the effects of the portal input on
resident-fish assemblages and to document the status of the fishery resource. In general,
fish-community richness increased by 2–3 species at mainstem sites near the portal, and
median density and biomass of fish communities at sites downstream of the portal were
significantly lower than they were at sites upstream of the portal. Median densities of Salmo
trutta (Brown Trout) and all trout species were significantly lower than at mainstem sites
downstream from the portal—25.1 fish/0.1 ha and 148.9 fish/0.1 ha, respectively—than at
mainstem sites upstream from the portal—68.8 fish/0.1 ha and 357.7 fish/0.1 ha, respectively—
yet median biomass for Brown Trout and all trout did not differ between sites from
both reaches. The median density of young-of-year Brown Trout at downstream sites (9.3
fish/0.1 ha) was significantly lower than at upstream sites (33.9 fish/0.1 ha). Waters from the
portal appeared to adversely affect the density and biomass of young-of-year Brown Trout,
but lower temperatures and increased flows also improved habitat quality for mature trout
at downstream sites during summer. These findings, and those from companion studies,
indicate that moderately turbid waters from the portal had few if any adverse impacts on
trout populations and overall fish communities in the Upper Esopus Creek during this study.
Introduction
The Upper Esopus Creek is a historic trout fishery and recreational stream in
the Catskill Mountains of southeastern New York state. The Shandaken Tunnel (the
portal) delivers water that is usually cool during the summer from the Schoharie
Reservoir to the Upper Esopus Creek (hereafter, the Esopus), and maintains a minimum
flow in the Esopus as stipulated by state regulations (CCE 2007). On average,
these releases are more turbid than the waters of the Esopus and have been the focus
of controversy because of the perception that the turbidity of the portal’s inflow
waters negatively affects water quality and trout populations in the upper basin
(CCE 2007). Any stressors that adversely affect Salmo trutta L. (Brown Trout) and
Oncorhynchus mykiss (Walbaum) (Rainbow Trout) populations might negatively
1US Geological Survey, New York Water Science Center, 425 Jordan Road, Troy, NY 12180.
2New York State Department of Environmental Conservation (Retired), 65561 State Highway
10, Stamford, NY 12167. *Corresponding author - bbaldigo@usgs.gov.
Manuscript Editor: Rudolf G. Arndt
Northeastern Naturalist
346
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
affect the local economy because these fisheries are important resources. Both the
concentration of suspended sediment and the duration of turbid episodes can adversely
affect the health of lotic species and thus aquatic communities (Newcombe
and Macdonald 1991, Shaw and Richardson 2001). Turbidity has been linked with
decreased visual acuity and growth rates in trout under laboratory conditions (Shaw
and Richardson 2001; Sigler et al. 1984; Sweka and Hartman 2001a, b), reduced
density and biomass of macroinvertebrates in the wild (Wagener and LaPerriere
1985), reduced biomass and primary productivity in periphyton (Bilotta and Brazier
2008, Quinn et al. 1992), and siltation-related habitat degradation in all 3 communities
(Henley et al. 2000). Field and mesocosm studies, moreover, have found a range
of neutral to negative impacts of turbidity on feeding, growth, density, and biomass
of resident trout or their populations (Redding 1987, Sweka and Hartman 2001a,
White and Harvey 2007). Because there are few data on fish communities and turbidity
in the Esopus, the potential effects of turbidity on local fish communities and
trout populations in this basin are largely speculative and subject to debate.
Discharge of relatively cool water from the portal could also have beneficial effects
on cold-water species. Water temperatures in parts of the Esopus frequently
exceed upper thermal limits for growth (~19 °C) and survival (~25 °C) of Brown
Trout (Hasnain et al. 2013, Wehrly et al. 2007) for long low-flow periods that occur
during most summers (Ross 2012). The portal provides supplemental flows that can
greatly exceed natural flows in the mainstem Esopus (at their ju ncture) by as much
as 1 order of magnitude. The temperature of portal water often is below the upper
lethal limits for Brown Trout during warm months, and may allow the species to
survive, if not grow, in nearby reaches immediately downstream from the portal that
otherwise might be uninhabitable.
Several factors other than the portal input may affect water quality and impact
stream ecosystems in the Upper Esopus Basin including (1) effluent discharges from
sewage-treatment plants within the watershed; (2) runoff from a ski area; (3) withdrawals
from streams for various purposes, including snow-making, water supply,
and bottling; and (4) areas with high concentrations of septic-systems adjacent to
tributaries. These factors, combined with water releases from the portal, subject
local fish populations and communities to a range of thermal, hydrologic, and
waste-water stresses in the upper basin. Unbiased contemporary data are needed to
quantify the status of water quality, fish assemblages, and potential impairments in
the Esopus.
In 2009, the US Geological Survey (USGS), in collaboration with the New
York State Department of Environmental Conservation (NYSDEC), New York City
Department of Environmental Protection (NYCDEP), and Cornell Cooperative
Extension of Ulster County (CCEUC), began a comprehensive assessment of the
biological condition (fish, invertebrates, and diatoms) and water quality (turbidity
and nutrients) in the Esopus. Primary objectives of this study were to assemble
contemporary data to define the effects of discharges from the portal on resident
fish assemblages and to foster better decision-making regarding the management of
the fishery, water quality, and water quantity (allocated water from the Schoharie
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
347
Reservoir to the Ashokan Reservoir) in the Esopus. Specific goals of this effort
were to characterize the natural variability and its relationship with water quality
for the biological communities, and to assess the potential effects of point sources
of turbidity on fish assemblages. We gave special attention to trout populations
because of their economic value and location at the top of the aquatic food chain,
as well as the perception that the local fishery and other natural resources within
the basin have declined due to the turbid waters delivered to the upper basin via the
portal. In this paper, we describe the status of resident fish assemblages and trout
populations at 18 sites in the Esopus that we sampled annually, with a few exceptions
at selected sites, from June through August during 2009–2011.
Study Area
The Upper Esopus Creek Basin is located in the south-central Catskill Mountains
of southeastern New York and follows a 67.3-km semi-circular course from its
headwaters at Winnisook Lake, to its terminus at the Ashokan Reservoir (Fig. 1). Its
497-km2 watershed is contained entirely within the Catskill Park and drains some of
the most rugged and mountainous terrain in the Park. Forested land comprises over
95% of the watershed, and commercial and residential development occupies the
rest (CCE 2007). The Shandaken Portal and the Esopus confluence occur in the village
of Shandaken just west of Phoenicia (Fig. 1). As of about 2005, water discharged
from the portal had the highest median turbidity (8.8 Nephelometric Turbidity Units
[NTU]) of any tributary previously assessed in the upper basin; however, its contribution
to the total annual sediment load of the Esopus was not fully defined (CCE
Figure 1. Location of 20 sites on the Upper Esopus Creek surveyed 2009–2011.
Northeastern Naturalist
348
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
2007). Channels for several Esopus tributaries cut into clay-rich till and also contribute
a significant portion of the total annual sediment load during rain and snow-melt
events (CCE 2007). The portal can discharge more than 2 million m3 of water per
day into the Esopus when at maximum output. Since 1977, water releases have been
regulated by Part 670 of Title 6 of New York Compilation of Codes, Rules and Regulations
(NYCRR; CCE 2007). In 2006, a NY State Pollutant Discharge Elimination
System (SPDES) permit placed additional limits on flow, turbidity, and temperature
in order to protect the health of aquatic biota in the Esopus (CCE 2007).
Methods
We collected fisheries and water-quality data at 18 sites in 2009 and 2010 and at
16 sites in 2011 (Fig. 1). Of the 18 sites, 10 were on tributaries with drainage areas
of 10.3–83.9 km2, 1 (esop0) is a small headwater mainstem site of 30.3 km2, and 7
have drainages larger than 100 km2 and are located on the mainstem of the Upper
Esopus (Table 1). Elevations of sites range from 189 m to 455 m (Table 1). At the
time, there were no natural or man-made barriers to fish passage in the study area.
We distributed mainstem sites evenly above and below the portal in an attempt to
distinguish differences in fish communities caused by the portal from those that
occurred naturally due to normal shifts in stream conditions between successively
larger stream sites.
In general, we chose study sites (reaches) so that they encompassed 2 or more
complete geomorphic channel-unit sequences (Fitzpatrick et al. 1998, Meador et al.
2003, Simonson et al. 1994). Most study reaches were 20–35 mean-channel widths
long (maximum of 100 m). At each reach, we collected fish by electrofishing up to
a 100-m-long section using a backpack shocker (Smith-Root Model 12B) and 3 or
4 netters. After blocking each reach with seines, quantitative surveys consisted of
3 electrofishing passes through each reach; sample-area dimensions were measured
on site. At sites with narrow channels (less than 15 m), we placed the blocking
seines completely across the channel at the upstream and downstream ends of the
study section. At sites with wider channels, we completed 3 replicate surveys in relatively
small near-shore subreaches. At each subreach, we affixed 1 blocking seine
to the bank, stretched it perpendicular to the bank, and attached it to a rock or rebar
6–8 m from shore. We placed a second 25-m seine oriented downstream and parallel
to shore, attached it to a second rock or rebar, and placed a third seine between
the second rebar and shore. Hence, each subreach was 6–8 m wide x 25 m long and
was blocked by seines on 3 sides and by shoreline on the fourth. We processed fish
collected from each sub-reach and pass separately. We identified each fish to species
and recorded the lengths and weights of all individuals longer than 150 mm. For
very abundant species (some minnows), which were usually shorter than 150 mm,
we measured subsamples and obtained lengths and weights from 40–50 individuals;
thereafter, we recorded total weights and counts by species in batches of 10–50. We
returned all fish to the stream after processing.
We characterized several key habitat features within each study reach and subreach
(see Table A in Supplemental File 1, available online at http://www.eaglehill.
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
349
us/NENAonline/suppl-files/n22-2-N1280-Baldigo-s1, and for BioOne subscribers,
at http://dx.doi.org/10.1656/N1280.s1). We recorded total reach length and 10
measurements of reach width, and employed a modified point and transect method
(Fitzpatrick et al. 1998) to measure depth and velocity and to estimate dominant
substrate-size categories at 3 points—center, 25%, and 75% of crossection—along
each of 10 evenly spaced transects. We used the total length and mean width to
calculate sample areas, and included all measurements in mean depth and velocity
calculations. We determined the dominant and subdominant particle sizes from the
frequency of each category. Data for mean and median water-year (1 October–30
September) temperature, discharge, suspended-sediment concentration, and turbidity,
along with annual suspended-sediment loads at all study sites, were from
McHale and Siemion (2014) and USGS (2014).
We determined standard metrics for entire fish communities and selected species
populations for each survey (each site and year that they were sampled). We used
the number of fish captured during each pass to estimate density and biomass (and
Table 1. Site names, codes, and USGS site numbers for sites surveyed in the Upper Esopus Creek,
2009–2011. Elev. = elevation.
Drainage
Site USGS site Latitude Longitude area Elev.
Stream and site name code number (°N) (°W) (km2) (m)
Tributary sites
Fox Hollow fox 01362199 42.1161111 74.380556 10.3 309
Peck Hollow peck 01362215 42.1255556 74.376389 12.3 351
Broadstreet Hollow broad 01362232 42.1125556 74.358694 23.7 296
Bushnellsville Creek bush 01362197 42.1247222 74.401139 29.5 336
Esopus Creek at OlivereaA esop0 0136219203 42.0525000 74.456222 30.3 455
Birch Creek birch 013621955 42.1089787 74.451818 32.4 377
Little Beaverkill lbeav 01362497 42.0195364 74.266258 42.7 205
Woodland Valley Creek wood 0136230002 42.0797222 74.334583 53.4 268
Beaverkill beav 01362487 42.0467580 74.276814 64.7 213
Stony Clove Creek at Chichester stoc0 01362370 42.1020278 74.310889 80.0 292
Stony Clove Creek at Phoenicia stoc1 01362398 42.0830556 74.315833 83.9 245
Upstream mainstem sites
Esopus Creek at Big Indian esop2 0136219565 42.1041667 74.435833 111.9 355
Esopus Creek at Shandaken esop3 0136219710 42.1194444 74.397500 152.0 317
Esopus Creek at Allaben esop3a 01362200 42.1170341 74.380149 165.0 305
Downstream mainstem sites
Esopus Creek downstream esop3b 0136223005 42.1133333 74.361889 181.0 287
of portal
Esopus Creek upstream of esop4 01362250 42.0925000 74.335972 215.7 268
Phoenicia
Esopus Creek at Phoenicia esop4a 01362405 42.0819444 74.312028 357.4 238
Esopus Creek downstream esop4b 01362420 42.0636111 74.306389 365.2 225
of Phoenicia
Esopus Creek at Mt Tremper esop5 01362430 42.0468889 74.280000 373.0 207
Esopus Creek at Boiceville esop6 01362500 42.0142588 74.270425 497.3 189
ASite esop0 was also treated as an upstream mainstem site for selected analyses.
Northeastern Naturalist
350
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
95% confidence intervals [CI]) for the entire fish community and for each species
population with the Moran-Zippin method of proportional reduction (Van Deventer
and Platts 1985, Zippin 1958). These values were divided by the total area sampled
at each site to estimate the number or the biomass (g) of fish in the local community
or species population per unit area. Two components of ecosystem diversity,
breadth or size (total species richness) and heterogeneity (Simpson’s diversity
index), were also generated for each survey (Simpson 1949, Whittaker 1975). Richness
(S) is the number of different fish species collected in the sampled area at each
site. Simpson’s diversity index is commonly used to characterize community biodiversity,
and it employs the number of species present and the relative abundance of
each species to calculate a metric that ranges from 0 to 1. When calculated as 1 - D,
zero indicates no diversity (e.g., 0–1 species), and values close to 1 indicate a large
number of species and proportionally similar numbers of each.
We described the status of resident-fish populations and communities, and spatial
and temporal variations in key metrics graphically. We evaluated the potential
effects of the portal on fish assemblages with 4 related methods. Longitudinal and
annual trends in metrics (total community density, biomass, richness, and diversity,
as well as density and biomass for trout populations) were evaluated through
graphical analyses. The significance of changes or differences in metrics between or
among individual sites and groups of sites within (a) tributaries, (b) upstream of the
portal, and (c) downstream of the portal were evaluated using the upper and lower
95% CIs (which equal 1.96 times the standard errors), and parametric analysis of
variance (ANOVA) if variance and normality assumptions were met, or non-parametric
Kruskal-Wallis tests if data were not normally distributed. We considered
differences in means and medians to be significant when the CIs did not overlap
(Cumming et al. 2007) or when P-values for appropriate statistical tests were ≤0.05.
We completed additional analyses of spatial patterns in fish-community composition
and classifications (groupings of sites with similar assemblages) through nonmetric
multidimensional scaling (MDS) ordination of taxon-density data that were
square-root transformed (Kruskal 1964, Shepard 1962). MDS ordination generates
an arrangement of samples in “species space” according to the non-parametric ranks
of their Bray-Curtis similarities. We employed a similarity profile test (SIMPROF)
to identify statistically significant (a priori unstructured) clusters within the larger
dendrogram (Clarke and Warwick 2001). The significant groups were labeled as
such on a non-metric MDS ordination to visually express the relationships among
dissimilar groups. Subsequently, we used a similarity-percentage (SIMPER) analysis
to determine the contribution of individual species to the overall dissimilarity
between groups of samples (Clarke and Warwick 2001). The SIMPER breaks down
the original Bray-Curtis similarities between samples into percent contributions of
each species to those sample similarities, and identifies the species that are most
responsible for defining site and group dissimilarities. Finally, we used an analysis
of similarities (ANOSIM) to test a priori hypotheses that species assemblages in
various reaches (e.g., sites from tributary, upstream, or downstream reaches) did not
differ significantly (Clarke and Warwick 2001).
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
351
Results
Site names and the codes we used for our analyses and to report our results are
provided in Table 1. Study sites on the mainstem of the Esopus upstream from the
portal are collectively referred to as upstream sites and those downstream from
the portal are referred to as downstream sites in the remainder of this manuscript.
Fish communities
Species richness ranged from as low as 4 species at 3 of the small drainage-area
(DA) tributaries (fox, peck, and bush) during 1 or 2 years, to a maximum of 15 species
at the largest DA site (esop6) during 2009 and 2010 (Tables 1, 2; Fig. 2A). Total community
richness was strongly related to the drainage area of study sites (Fig. 3). Mean
richness at all tributary sites (DA < 100 km2) over all years was 7.0 species, whereas
mean richness at mainstem sites (DA > 100 km2) averaged 9.2 species upstream from
the portal and 11.1 species downstream from the portal (Table 3). Median richness
differed significantly among tributary, upstream, and downstream sites (Table 3).
With one exception (birch during 2010), richness at the 6 tributary sites with DAs less than 40
km2 ranged from 4 to 7 species. Richness reached an asymptote at about 10 species
as DA increased above 40 km2. Influx of nonresident species from Schoharie Reservoir
through the portal may have increased richness at 2 sites closest to the portal.
Richness at esop6 was larger than anticipated and inconsistent with the asymptotic
pattern—a finding probably related to the influx of several species from the nearby
Ashokan Reservoir. Mean richness across all sites decreased slightly during each year
of the study, yet median richness was relatively consistent at 8.0, 7.5, and 8.0 species
in 2009, 2010, and 2011, respectively, and did not differ between years.
Table 2. Fish-community metrics from all sites surveyed in the Upper Esopus Creek during 2009–2011.
Density Biomass Index of diversity
Site code Richness (S) (fish/0.1ha) (g/0.1ha) (1 - D) Evenness (J')
2009 surveys
fox 4 1707 13,517 0.35 0.48
peck 5 1396 9180 0.56 0.67
broad 7 881 4837 0.71 0.71
bush 4 2485 14,451 0.52 0.66
birch 7 1525 12,799 0.59 0.58
lbeav 9 657 4830 0.67 0.63
wood 7 2964 15,336 0.79 0.87
beav 11 1985 9891 0.72 0.65
stoc0 10 1790 10,749 0.74 0.66
stoc1 8 1440 7192 0.77 0.78
esop0 6 3474 9372 0.40 0.47
esop2 9 1517 40,843 0.79 0.80
esop3 8 2655 14,984 0.76 0.79
esop3a 10 2428 8307 0.80 0.79
esop3b 13 1260 3308 0.79 0.72
esop4 8 879 8743 0.72 0.72
esop4a 10 821 4696 0.81 0.79
esop6 15 1168 4993 0.69 0.62
Northeastern Naturalist
352
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
Community diversity, estimated using 1 - D, generally reflected the same patterns
observed for species richness (Fig. 2B). Diversity scores ranged from 0.15 at
bush (DA = 29.5 km2) in 2011 to 0.83 at esop6 (DA = 497.3 km2) in 2011 (Table 2).
Mean diversity scores for all 3 years were lower at all tributary sites (0.56) than
they were at all downstream (0.62) and upstream (0.74) mainstem sites (Table 3).
Like richness, mean diversity increased between tributary and mainstem sites, but
unlike richness, the median values at upstream sites (0.76) did not differ significantly
from median values at downstream sites (0.79). Furthermore, the average
diversity at 3 sites (esop2, esop3, and esop3a) immediately upstream from the portal
Table 2, continued.
Density Biomass Index of diversity
Site code Richness (S) (fish/0.1ha) (g/0.1ha) (1 - D) Evenness (J')
2010 surveys
fox 6 1567 147,06 0.66 0.68
peck 4 792 5638 0.57 0.77
broad 7 922 4248 0.77 0.81
bush 7 2375 9028 0.45 0.47
birch 9 1417 16,445 0.62 0.57
lbeav 10 1118 4943 0.77 0.73
wood 7 1215 7544 0.75 0.82
beav 9 2902 4533 0.69 0.63
stoc0 6 834 12,526 0.66 0.71
stoc1 7 817 3197 0.72 0.77
esop0 6 2219 11,032 0.22 0.30
esop2 9 1496 14,339 0.71 0.71
esop3 8 2436 11,205 0.70 0.69
esop3a 10 1301 5703 0.67 0.64
esop3b 11 533 1106 0.80 0.73
esop4 7 731 14,125 0.70 0.77
esop4a 9 2432 6854 0.79 0.77
esop6 15 850 2447 0.82 0.75
2011 surveys
fox 4 516 3758 0.40 0.52
peck 5 703 4732 0.32 0.41
broad 6 622 4304 0.53 0.59
bush 7 3483 10,247 0.15 0.19
birch 6 1115 9554 0.53 0.60
lbeav na na na na na
wood 7 497 2216 0.66 0.64
beav 10 2440 4643 0.50 0.45
stoc0 8 1029 4376 0.18 0.23
stoc1 8 438 2498 0.58 0.62
esop0 6 785 2801 0.48 0.48
esop2 8 1103 13,007 0.75 0.77
esop3 10 2570 9690 0.76 0.72
esop3a 11 1020 4222 0.76 0.73
esop3b 13 1072 2439 0.80 0.68
esop4 8 693 16,536 0.72 0.76
esop4a na na na na na
esop6 13 680 3175 0.83 0.75
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
353
was comparable to the 4 downstream sites (Fig. 2B). Average and median speciesdiversity
values at all Esopus sites were 0.59 and 0.72 in 2009, 0.67 and 0.70 in
2010, and 0.56 and 0.56 in 2011, respectively; medians did not differ significantly
among years (Table 3).
Figure 2. Mean richness (A), diversity (B), density (C), and biomass (D) and corresponding
95% confidence intervals (CIs) for resident-fish communities at all sites surveyed in
the Upper Esopus Creek, 2009–2011. The locations of the sites (i.e., in a tributary and in the
mainstem, either upstream or downstream from the portal) are denoted below the site identification
codes; site esop0 functions as a tributary and a mainstem site for selected analyses.
Northeastern Naturalist
354
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
Figure 3. Relationship between community richness and drainage area, and 95% confidence
intervals (CIs) for all sites surveyed in the Upper Esopus Cree k, 2009–2011.
Table 3. Mean and median fish-community and trout-population metrics grouped by site type and
survey year and the significance (P-values) for the differences in medians as determined by Kruskall-
Wallis non-parametric tests. Significantly different group-medians are denoted by non-similar letters
below each median. YOY = young-of-year, Trib = tributary sites, US = sites upstream from portal,
DS = sites downstream from portal.
Site type Survey year
Metric P-value Trib US DS P-value 2009 2010 2011
Community richness
mean 7.0 9.2 11.1 8.4 8.2 8.1
median less than 0.001 7.0 9.0 11.0 0.9243 8.0 7.5 8.0
similar a b c a a a
Community diversity (D)
mean 0.56 0.74 0.62 0.59 0.67 0.56
median less than 0.001 0.59 0.76 0.79 0.3487 0.72 0.70 0.56
similar a b b a a a
Community density (fish/0.1 ha)
mean 1503 1836 1011 1724 1442 1173
median 0.0313 1306 1517 850 0.0253 1521 1258 903
similar ab a b a ab b
Community biomass (g/0.1 ha)
mean 7973 13,589 6220 11,002 8312 6137
median 0.0508 7368 11,205 4696 0.0310 9276 7199 4340
similar a b a a ab b
All trout density (fish/0.1 ha)
mean 165.0 319.0 158.0 280.6 205.2 72.1
median 0.0697 138.6 357.7 148.9 0.0001 264.0 183.3 45.9
similar a b a a a b
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
355
Table 3, continued.
Site type Survey year
Metric P-value Trib US DS P-value 2009 2010 2011
All trout biomass (g/0.1 ha)
mean 2621 3953 2326 3629 2926 1689
median 0.1791 2031 2804 3006 0.0037 3012 2031 1972
similar a a a a ab b
Brown Trout density (fish/0.1 ha)
mean 91.1 101.7 24.9 137.7 65.9 27.6
median 0.0069 64.0 68.8 25.1 0.0001 138.0 54.6 21.3
similar a a b a b b
Brown Trout biomass (g/0.1 ha)
mean 2124 2941 1863 3175 2314 1007
median 0.2395 1399 2747 857 0.0014 2742 1399 581
similar a a a a ab b
Mature Brown Trout density (fish/0.1 ha)
mean 41.5 21.7 13.2 42.2 41.8 9.9
median 0.0112 36.4 22.2 11.8 0.0001 39.0 36.9 10.7
similar a ab b a a b
YOY Brown Trout density (fish/0.1 ha)
mean 49.6 80.0 11.7 95.5 24.1 17.7
median 0.0157 20.3 33.9 9.3 0.0003 81.6 9.5 8.2
similar ab a b a b b
Percent YOY Brown Trout (%)
mean 38.0 75.2 36.9 62.6 28.9 40.8
median 0.007 38.0 83.0 30.0 0.0053 65.9 31.2 34.7
similar a b a a b b
Rainbow Trout density (fish/0.1 ha)
mean 71.3 213.3 132.5 140.4 135.8 43.1
median 0.0254 42.8 210.8 127.3 0.0075 114.9 127.8 27.5
similar a b ab a a b
Rainbow Trout biomass (g/0.1 ha)
mean 475.7 975.0 461.9 432.5 595.2 661.1
median 0.5264 414.8 815.3 257.0 0.5769 372.3 507.8 376.7
similar a a a a a a
Mature Rainbow Trout density (fish/0.1 ha)
mean 13.8 18.2 5.1 6.4 18.5 13.4
median 0.0309 12.2 16.6 2.0 0.0525 2.7 21.7 10.4
similar ab a b a b ab
YOY Rainbow Trout density (fish/0.1 ha)
mean 57.5 195.2 127.4 134.0 117.3 29.8
median 0.0137 26.5 193.4 125.7 0.0030 105.3 111.8 13.9
similar a b b a a b
Percent YOY Rainbow Trout (%)
mean 62.1 76.7 92.6 77.4 80.2 53.6
median less than 0.001 71.4 92.1 15.1 0.0083 94.4 86.4 50.0
similar a ab b a a b
Northeastern Naturalist
356
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
The spatial trend in density of fish communities was not as strongly related to
DA as were richness and diversity (Fig. 2C). Total density ranged from 438 fish/0.1
ha at stoc1 in 2011 to 3483 fish/0.1 ha at bush in 2011 (Table 2). Mean density
at all tributary sites (1503 fish/0.1 ha) differed little from that at upstream (1836
fish/0.1 ha) and downstream (1011 fish/0.1 ha) sites; however, the median density at
downstream sites (850 fish/0.1 ha) was significantly lower than the median density
at upstream sites (1517 fish/0.1 ha) (Table 3). Total community density declined at
many sites (individually) and overall throughout the 3-year study. Mean and median
density for all surveys at all sites ranged from 1724 and 1521 fish/0.1 ha in 2009,
1142 fish/0.1 ha and 1258 fish/0.1 ha in 2010, and 1173 fish/0.1 ha and 903 fish/0.1
ha in 2011, respectively; median density at all sites was significantly lower during
2011 than during 2009. Community density decreased from 2009 to 2011 in about
80% of the tributaries and half of the mainstem sites (see Fig. B in Supplemental
File 1, available online at https://www.eaglehill.us/NENAonline/suppl-files/n22-
2-N1280-Baldigo-s1, and for BioOne subscribers, at http://dx.doi.org/10.1656/
N1280.s1). Based on 95% CIs, the 2011 estimates were significantly lower than in
2009 at most sites, and densities for 2010 were generally intermediate and not different
from the 2009 estimates.
Community biomass (Fig. 2D), like density, was not strongly related to DA.
Biomass ranged from 1106 g/0.1 ha at esop3b in 2010 to 40,843 g/0.1 ha at esop2
in 2009 (Table 2). Biomass averages and medians (g/0.1 ha) were, respectively:
7973 and 7368 in tributaries, 13,589 and 11,205 at upstream sites, and 6220 and
4696 at downstream sites. Estimates for median biomass at downstream sites were
significantly lower than that for upstream sites (Table 3). Total biomass averaged
about 9019 g/0.1 ha at sites esop3 and esop3a, immediately above the portal, and
7710 g/0.1 ha at sites esop3b and esop4, immediately below the portal. The lowest
mean biomass for any site over most years was found at site esop3b, just downstream
of the portal (Table 2; see also Fig. B in Supplemental File 1, available
online at https://www.eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigos1,
and for BioOne subscribers, at http://dx.doi.org/10.1656/N1280.s1). The low
mean biomass (across all years) and the narrow CIs (±1255 g/0.1 ha) at this site
(Fig. 2D) suggest that community biomass was significantly lower than at the
next 2 upstream and next 2 downstream sites. Community biomass declined at
many sites throughout the study period: it averaged 11,002 g/0.1 ha, 8312 g/0.1
ha, and 6137 g/0.1 ha for all surveys in 2009, 2010, and 2011, respectively.
However, median biomass at upstream and downstream sites did not differ significantly
(Table 3; see also Fig. B in Supplemental File 1, available online at http://
www.eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigo-s1, and for
BioOne subscribers, at http://dx.doi.org/10.1656/N1280.s1). Biomass estimates
were lowest at 11 sites during 2011. Total biomass decreased from 2009 to 2011 in
all of the tributaries, and in about 80% of the mainstem sites; these declines were
significant between 2009 and 2011, and data for 2010 were generally intermediate
or not different from 2009 or 2011.
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
357
Species populations
Changes in the distribution of several fish species accounted for most of the
differences in community metrics across sites. Species distributions (richness)
appeared to be influenced largely by stream size, which can be classified by DA.
Small DA sites (tributaries with DAs of less than 40 km2) tended to be dominated
by Cottus cognatus Richardson (Slimy Sculpin), trout, and Rhinichthys cataractae
Valenciennes (Longnose Dace) (see Fig. C in Supplemental File 1, available online
at https://www.eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigo-s1,
and for BioOne subscribers, at http://dx.doi.org/10.1656/N1280.s1), which are
characteristic species in cold-water communities. Few other species contributed
strongly to community density (see Table D in Supplemental File 1, available online
at https://www.eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigo-s1,
and for BioOne subscribers, at http://dx.doi.org/10.1656/N1280.s1) or biomass at
small DA sites (see Table E in Supplemental File 1, available online at http://www.
eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigo-s1, and for BioOne
subscribers, at http://dx.doi.org/10.1656/N1280.s1). Trout became more sporadically
distributed and Slimy Sculpin was partially replaced by minnow and sucker
species at tributary sites with moderately sized (40 km2–100 km2) DAs (see Fig. C
in Supplemental File 1, available online at https://www.eaglehill.us/NENAonline/
suppl-files/n22-2-N1280-Baldigo-s1, and for BioOne subscribers, at http://dx.doi.
org/10.1656/N1280.s1). With minor exceptions, species density and biomass at most
mainstem sites with DAs >100 km2 generally decreased from upstream to downstream
sites. Biomass of trout and Slimy Sculpin populations were very high at the 2
upper-most mainstem sites, esop2 and esop3, and declined at the sites immediately
above and below the portal (at esop3a and esop3b). Seven species—Luxilus cornutus
(Mitchill) (Common Shiner), Etheostoma olmstedi Storer (Tessellated Darter),
Noturus insignis (Richardson) (Margined Madtom), Notemigonus crysoleucas
Mitchill (Golden Shiner), Lepomis cyanellus Rafinesque (Green Sunfish), Micropterus
salmoides Lacépède (Largemouth Bass), and Alosa pseudoharengus (Wilson)
(Alewife)—were first encountered at these 2 sites, i.e., they were absent from all
mainstem sites upstream of esop3a (see Table D in Supplemental File 1, available
online at https://www.eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigos1,
and for BioOne subscribers, at http://dx.doi.org/10.1656/N1280.s1). Species
populations at esop4 also did not reflect the expected community composition given
its drainage area. Trout and Slimy Sculpin density and biomass increased markedly
and density and biomass of minnow species declined at esop4 relative to the 2 prior
upstream sites (see Fig. C in Supplemental File 1, available online at http://www.
eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigo-s1, and for BioOne
subscribers, at http://dx.doi.org/10.1656/N1280.s1). Increases in water velocity and
decreases in temperature (related to inputs from the portal) may be partially responsible
for observed community changes. Species diversity and richness increased and
additional centrarchids, as well as other minnow species, were present at the 2 lower
mainstem sites, esop4a and esop6. Four species—Micropterus dolomieu Lacépède
Northeastern Naturalist
358
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
(Smallmouth Bass), Ambloplites rupestris (Rafinesque) (Rock Bass), Semotilus
corporalis Mitchill (Fallfish), and Pimephales notatus (Raphinesque) (Bluntnose
Minnow)—were only collected at esop6.
Trout populations
Understanding spatial and temporal patterns in trout populations was a primary
focus of this study. Spatial trends in total density and biomass of trout populations
at all tributary and mainstem sites were not generally predictable based on
DA (Fig. 4). In fact, DA only explained 2–3% of the variability in the density or
biomass of all trout (pooled) populations at all sites in the basin. Trout density averaged
200 fish/0.1 ha at the 6 small tributary sites (DA < 40 km2), 120 fish/0.1 ha at
the 5 large tributary (DA = 40 km2–100 km2), and 230 fish/0.1 ha at the 7 mainstem
(DA > 100 km2) sites (excluding esop0) (Fig. 4A). Trout biomass averaged 3015,
2113, and 3058 g/0.1 ha at the small tributary, large tributary, and mainstem sites,
respectively (Fig. 4B). Median trout density was significantly lower at downstream
sites (149 fish/0.1 ha) than at upstream sites (358 fish/0.1 ha) (Table 3). Median
trout biomass, however, did not differ significantly between mainstem sites located
upstream and downstream from the portal.
We collected Brown Trout and Rainbow Trout at most sites. Estimates of total
biomass for both species collected from all sites each year during 2009–2011 are
summarized in Figure 5. The average density of Rainbow Trout populations at
all sites (169 fish/0.1 ha) was about 3-fold higher than that of Brown Trout (59
fish/0.1 ha), however average biomass of Brown Trout populations at all sites
Figure 4. Estimates of total density (A) and biomass (B) and 95% confidence intervals (CIs)
from all trout at sites surveyed in the Upper Esopus Creek during 2009, 2010, and 2011.
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
359
(2347 g/0.1 ha) was about 3-fold greater than that of Rainbow Trout populations
(693 g/0.1 ha) (Table 3).
The portal may affect Brown Trout and Rainbow Trout populations indirectly
by altering survival of their early life stages (e.g., eggs, swim-up fry, and youngof-
year), thus altering the proportions of juvenile and mature fish. We used nonparametric
Kruskal-Wallis tests to assess differences in median density and biomass
for both species and for differences in median densities and percentage of youngof-
year (YOY) Brown Trout and Rainbow Trout at 3 upstream, 4 downstream, and
11 tributary sites (Table 3). The median density of Brown Trout populations was
significantly lower at downstream sites (25 fish/0.1 ha) than at upstream sites (69
fish/0.1 ha), but median biomass did not differ significantly among the 3 site types.
Median densities of mature Brown Trout did not differ significantly between downstream
and upstream sites. Median density of YOY Brown Trout was significantly
lower at downstream sites (9 fish/0.1 ha) than upstream sites (34 fish/0.1 ha). The
YOY constituted 38%, 75%, and 37% of Brown Trout populations at tributary,
upstream, and downstream sites, respectively, and the median percentages were
significantly lower at downstream sites (30%) than at upstream sites (83%). Median
density and biomass estimates for Rainbow Trout populations did not differ
significantly between downstream and upstream sites (Table 3). Median density of
mature Rainbow Trout was significantly lower at downstream sites (2 fish/0.1 ha)
than at upstream sites (17 fish/0.1 ha). Median densities of YOY Rainbow Trout did
not differ significantly between downstream and upstream sites. The YOY constituted
62%, 77%, and 93% of Rainbow Trout populations at tributary, upstream, and
downstream sites, respectively, and median percentages did not differ significantly
between downstream and upstream sites.
Figure 5. Estimates of Brown Trout biomass (A) and Rainbow Trout biomass (B) and 95%
confidence intervals (CIs) at sites surveyed in the Upper Esopus Creek during 2009–2011.
Northeastern Naturalist
360
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
Temporal and spatial variability in fish assemblages
The temporal trends in fish communities and trout populations were relatively
similar. Community density and biomass decreased significantly at many sites
(e.g., stoc1, fox, and esop3a) (see Fig. B in Supplemental File 1, available online at
https://www.eaglehill.us/NENAonline/suppl-files/n22-2-N1280-Baldigo-s1, and for
BioOne subscribers, at http://dx.doi.org/10.1656/N1280.s1), and overall at all sites
(Table 3) between 2009 and 2011. Temporal trends in density and, to some degree,
biomass of trout populations were also significant and relatively consistent at many
sites (Table 3, Fig. 4). The density of trout populations at most sites was generally
highest in 2009, lower in 2010, and lowest in 2011; many of these differences
were significant between years (Fig. 4A). The changes in trout biomass over the
3 periods were comparable to density changes at several sites (e.g., esop3, wood,
and stoc0), yet the observed decreases were less regular, often not significant, and
sometimes interrupted by increases during 2010 (Fig. 4B). The few exceptions
to these temporal trends were generally related to collection of a few large trout,
which strongly influenced biomass estimates at individual sites such as esop2
during 2011 and esop4 during 2010. Median density of all trout was significantly
lower in 2011 (46 fish/0.1 ha) than in 2009 (264 fish/0.1 ha) and 2010 (183 fish/0.1
ha), and median biomass was significantly lower in 2011 (1972 g/0.1 ha) and 2010
(2031g/0.1 ha) than in 2009 (3012 g/0.1 ha) (Table 3). In general, all mean and
median population metrics for Brown Trout decreased significantly from 2009 to
2011; most decreases were also significant between 2009 and 2010 (Fig. 5A). Three
median Rainbow Trout population metrics decreased significantly between 2009
and 2011, but only the decrease in density of mature Rainbow Trout was significant
between 2009 and 2010 (Fig. 5B).
Multidimensional scaling ordinations and cluster analysis using species-density
data identified 5 unique site groupings distributed roughly along a stream-size or
drainage-area gradient (Fig. 6). The composition of fish assemblages at all sites
(and survey years) within each group was 58–72% similar to each other and differed
significantly (P < 0.05) among groups. The groups consisted of (1) only small
headwater tributaries, which were 67% similar; (2) large tributaries, Broadstreet
Hollow Brook (broad), and mainstem sites mostly upstream of esop4a, which were
53% similar; (3) large tributaries and one upstream site (esop3a), which were 58%
similar; (4) esop6 during 2010 and 2011, which were 64% similar; and (5) esop6
and esop3a during 2009, which were 72% similar. The last 2 groups were also 55%
similar and are treated as 1 group in the discussion. The overall groupings confirm
major differences among fish assemblages observed at small tributaries, the largest
downstream site, and all other sites, yet they identify no consistent differences
between fish assemblages at sites upstream and downstream of the portal. Density
bubble-plots for Slimy Sculpin and Exoglossum maxillingua Lesueur (Cutlips Minnow)
illustrated pronounced trends along the horizontal axis (Fig. 7). Slimy Sculpin
populations were generally most dense at small headwater tributaries (1800–3100
fish/0.1 ha at bush) and essentially absent at large downstream mainstem sites during
all years (Fig. 7A), whereas Cutlips Minnow populations were largest (600–800
fish/0.1 ha) at upstream (esop3a) and downstream (esop6) mainstem sites during
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
361
2009 (Fig. 7B). A SIMPER analysis confirmed that Slimy Sculpin was responsible
for as much as 32.4% of the dissimilarity among small tributary, large tributary, and
mainstem groups. Cutlips Minnow was almost as influential as Slimy Sculpin and
contributed as much as 19.6% to the dissimilarities among groups. These findings
indicate that Slimy Sculpin and Cutlips Minnow may be suitable indicator species
for classifying community types in this basin.
Discussion
In general, the results from this investigation indicate that fish assemblages
in the Esopus respond to waters from the portal in a variety of ways. The most
important finding was that no adverse changes in fish assemblages could be attributed
directly to portal waters, and some of the noted effects might even be
considered beneficial. Though fish communities were altered at sites near and
downstream from the portal, many changes were positive and could be related to
changes in temperature or habitat quality and quantity. One companion study of
Figure 6. Non-metricMDS ordination plot of fish assemblages based on square-root-transformed
density of all species collected at 16–18 sites in the Upper Esopus Creek annually
from 2009–2011. The 5 symbols denote site membership in groups with significantly similar
communities (53–72% Bray-Curtis similarity, P ≤ 0.05) based on group-averaged cluster
analysis. Site locations are shown in Figure 1.
Figure 7 (following page). Density-bubble plots showing the distribution and density of
(A) Slimy Sculpin and (B) Cutlips Minnow populations across sites and years (labels on
bubbles) in the Upper Esopus Creek, 2009–2011. Site locations are shown in Figure 1.
Northeastern Naturalist
362
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
363
fatty acids and periphyton communities in the upper basin (S.D. George, US Geological
Survey, Troy, NY, unpubl. data) detected the only ecological impairment
that may be linked directly to the portal: a reduction in the standing crop of primary
producers at sites immediately downstream from the portal. Potentially adverse
effects were also noted in the density and biomass of juvenile life stages of Brown
Trout. Though such effects could be caused by impaired water quality, the portal
was not a major contributor to turbidity and suspended sediment loads in the upper
basin (McHale and Siemion 2014). Loads typically reflect short-duration
(high-flow) events and, thus, may not be as important to trout growth and survival
as long-term exposure to moderately elevated levels of suspended sediment and
turbidity. During water-years 2010 and 2011, median turbidity from 34 routine
grab samples collected on the same dates (every 2–4 weeks) upstream from the
portal at esop03a (7.5 NTRUs) and downstream from the portal at esop03b (9.4
NTRUs) (USGS 2014) did not differ significantly (Mann-Whitney U-Test: P =
0.1990), nor did the distributions of turbidity data from both sites (Kolmogorov-
Smirnov Test: P = 0.1601). The median turbidity for 24 mean-daily values (estimated
monthly from continuous data) at the portal during the same 2-year period
was 16.4 NTU (EPA 2014). Although these results indicate that waters from the
portal slightly reduced clarity at downstream sites, turbidity levels generally
remained below thresholds (approximately 40 NTU) found to impair feeding,
growth, and survival of several trout species (Newcombe 2003, Rowe et al. 2003,
Stuart-Smith et al. 2004, Sweka and Hartman 2001b), which suggests that these
small increases may not be biologically meaningful.
When compared to richness at other mainstem sites, the significantly larger
number of species at sites esop3a and esop3b, immediately upstream and downstream
of the portal, suggests that the portal’s connection to the Schoharie Basin
was the source of several fish species. Similarly, richness was high at the furthest
downstream site (espo6), near the Ashokan Reservoir where we encountered as
many as 15 species. The richness of fish species at sites in the nearby Beaverkill
Basin, with the same drainage area as esop6, ranged from 12 to 14, whereas comparable
sites in the adjacent Neversink Basin (which are close to its confluence
with the Delaware River) reach as high as 18 species (Baldigo, et al., in press).
Richness at sites in the Beaverkill with the same drainage area as esop3a and
esop3b ranged from 10 to12 species, whereas 7 species were collected at the 1
comparable site in the Neversink, which was upstream of a reservoir. Although
the small increases in richness at sites near the portal (and at esop6) did not cause
significant differences in diversity, the addition of species creates more complex
food webs, which could benefit local communities by making ecosystems less
susceptible to the effects of short- and long-term stresses or perturbation (Schaefer
et al. 2012). The increased richness at esop6 was likely related to its close
proximity to the reservoir and to its habitat, which is more suitable for these lentic
species than sites further upstream. Although increased species richness may be
beneficial, additional nonnative species could also be detrimental to resident fish
communities (Cucherousset and Olden 2011).
Northeastern Naturalist
364
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
Waters from the portal had a number of measureable effects on fish communities
at downstream mainstem sites that suggest the normal upstream–downstream
continuum was disrupted. Estimates of total density and biomass for fish communities
at individual mainstem sites revealed no significant differences that could
be directly attributed to the portal, yet total density and biomass at 4 mainstem
sites downstream were significantly lower than at the 3 mainstem sites upstream.
Conversely, the ANOSIM detected no significant impacts of the portal on fish
communities at mainstem sites downstream. More important, however, may be
the increased number of fish species at several sites (e.g., esop3a and esop3b) in the
Esopus. Fish assemblages (i.e., the distribution of individual species) in large undisturbed
basins normally follow a predictable succession (longitudinal zonation)
between small, cold, low-order headwater reaches, and large, warm, high-order
reaches reflecting a continuum of abiotic and biotic factors (Vannote et al. 1980).
Fish-species richness and diversity normally increase progressively with increasing
stream size and order (at low to mid-order reaches), often with additions to, rather
than replacement of, existing species (Hutchinson 1993, Whiteside and McNatt
1972, Zalewski et al. 1990). In general, physical and chemical factors including dissolved
oxygen, pH, alkalinity, suspended turbidity, and conductivity often change
with stream size; however, stream depth, size, order, and habitat heterogeneity
typically explain the largest amount of variability in richness and diversity of fish
communities (Hutchinson 1993; Schlosser 1987, 1991). In temperate rivers, abiotic
factors can limit the diversity of low-order fish communities to relatively few
species that are tolerant of lower temperatures, wide fluctuations in flow, small/
shallow channels with homogeneous habitat, and high water-velocities. High rates
of primary production in mid-order stream reaches can increase species richness,
biodiversity, density, and/or levels of biomass for secondary and tertiary consumers
(Ward and Stanford 1983). Other biotic factors, such as competition and predation,
typically control richness and diversity of fish communities at larger downstream
reaches where relatively stable flows, more heterogeneous stream channels and
habitat, and slower and warmer waters permit additional species to coexist (Vannote
et al. 1980). Mid-basin changes in richness and diversity indicate a small and
localized effect on the Esopus ecosystem, yet decreases in fish-community biomass
and density indicate that primary production, along with secondary and top consumers,
may have been unusually low at several sites downstream from the portal.
Altered composition of periphyton communities could be responsible for
decreases in primary production at mid-order sites in the Esopus. Periphyton is
responsible for most primary production and form the base of the food web in
low- to mid-order streams in temperate climates (Vannote et al. 1980). Chlorophyll
a (chl a) concentration and ash-free dry mass (AFDM) quantify the standing crop
of primary producers (assumed herein to be a surrogate for primary production) in
stream food webs (Rosemond et al. 1993). Although neither chl a nor AFDM differed
significantly between groups of upstream and downstream sites, these values
were significantly depressed at the first 3 downstream sites (esop3b, esop4, and
esop4a) compared to the 3 upstream sites (Smith et al. 2013). Large decreases in
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
365
primary production can cause cascading effects at higher trophic levels (e.g., primary
consumers, predators; Kurle and Cardinale 2011). Changes in the abundance
and/or biomass of certain preferred (or all) macroinvertebrate-prey species could
theoretically limit available food resources, and therefore, growth and survival
of some fish species, at one or more downstream sites. The associations between
standing crop, periphyton communities, and macroinvertebrate communities are
relatively strong in the Esopus (Smith et al. 2013) and provide additional evidence
that the food web is altered at several downstream sites. Because changes in primary
production can cause cascading effects throughout the food web (Henley
et al. 2000, Vannote et al. 1980), direct and indirect impacts to predator (fish and
macroinvertebrate) populations are possible.
The indirect effects of reduced primary production and the direct effects of
unusual thermal, hydrologic, and water-quality regimes on the 2 dominant trout
species are potentially important, but difficult to isolate in this system because of
their dissimilar life-histories. Although both types of effects may be attributed to
the portal, the disparate timing for spawning, incubation, and emergence (from
gravel beds) make the early life stages for each species differentially vulnerable to
storm flows occurring at different times of the year. Nevertheless, the median density
of trout at downstream sites was significantly lower than that at sites upstream
from the portal, and biomass differences were only nominal. Differences in median
Rainbow Trout metrics were generally minor; thus, changes in Brown Trout populations
drove differences in overall trout metrics. Median densities of all Brown
Trout and YOY Brown Trout were significantly lower at downstream sites than at
upstream sites, yet densities of mature Brown Trout did not differ between the 2
reaches. Ross (2012) determined that upstream reaches contained more habitat preferred
by juvenile Brown Trout than did downstream reaches, which agrees with the
significant differences we detected in densities of YOY Brown Trout between sites
from both reaches. Although the 37% decrease in biomass of Brown Trout populations
between upstream and downstream sites was not significant, the significantly
lower density of YOY Brown Trout suggests that their populations might have been
larger if not for their location downstream of the portal. Clearly, the abundance of
juvenile Brown Trout was reduced at many downstream Esopus sites.
Though the potential effects of the portal on trout provided the impetus for
this investigation, a number of related mechanisms could also adversely affect
productivity and fish assemblages downstream. This consideration is important
because observational studies can only describe the strength of the relationships
between or among factors; the findings cannot unequivocally attribute a specific
effect to a specific cause. Assuming that turbid waters directly reduce primary productivity,
then the portal would only play a nominal role because it has been a minor
source of turbidity and suspended sediments since the mid-1990s (CCE 2007). For
example, estimates from all samples collected during water-year 2010 for median
turbidity (NTRU) varied from 119 at Stony Clove, 26 at Broadstreet Hollow Brook,
41 at Woodland Valley Creek, and 47 at Beaverkill, to 77 at the last mainstem site
Northeastern Naturalist
366
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
(esop3a) upstream of the portal, and 20 at the first site (esop3b) downstream of
the portal (McHale and Siemion 2014, USGS 2014). The median turbidity for 12
mean-daily values (estimated monthly from continuous data) reported for waters
discharged from the portal during water-year 2010 was 15.3 NTU (EPA 2014),
which suggests that portal waters do not substantially alter natural turbidity levels
within the Esopus. Thus, turbidity from portal waters probably does not limit fishspecies
distributions and their assemblages at downstream Esopus sites. These data
indicate that several tributaries, not the portal, are the primary sources of turbidity
in the upper basin. In fact, Stony Clove Creek accounted for more of the total
suspended sediment load (30–57%) at the furthest downstream site, esop6, during
water-years 2010 and 2012 than did all other tributaries combined (McHale and
Siemion 2014).
The results from other investigations show that elevated turbidity levels can
strongly affect the behavior, growth, and condition of Brown Trout and other salmonids,
yet few have detected strong linkages in the wild (Henley et al. 2000).
During laboratory experiments, significant changes in behavior and activity levels
of Brook Trout only occurred when turbidity levels surpassed 7.1 NTU (Gradall and
Swenson 1982). Increasing turbidity levels from 0 to 45 NTU had no significant effect
on the rates at which Brook Trout consumed prey because they switched from
passive to more active searching and feeding behavior (Sweka and Hartman 2001a).
Although growth rates of Brook Trout decreased linearly with increasing turbidity,
they did not differ significantly from controls until turbidity levels reached
45 NTU, when they declined by 62% (Sweka and Hartman 2001a). Similarly, the
volume of stomach contents and prey diversity for Brown Trout sampled from Lake
Sorell, Tasmania, were greater in 1996 when turbidity averaged 26 NTU than they
were during 2001 when turbidity averaged 141 NTU (Stuart-Smith et al. 2004).
Repeated exposure to 0–6-d pulses of turbid (suspended sediment concentrations
of 700 mg/L) water also significantly affected growth of juvenile Rainbow Trout in
streamside (caged) experiments (Shaw and Richardson 2001). Model simulations
predicted that if Rainbow Trout were exposed to high turbidity (53 NTU), they
would occupy shallower- and slower-water habitats, switch from passive to active
prey-capture behavior, gain no net energy, and remain in poor condition (Harvey
and Railsback 2009). Despite these findings, Rainbow Trout had normal feeding
rates (White and Harvey 2007) and wild trout populations generally persisted
(Harvey and Railsback 2009) in streams with highly turbid regimes, purportedly
because trout can sense and capture prey using non-visual organs (e.g., the lateral
line) under highly turbid and low-light conditions (Rowe et al. 2003).
The supplemental flows from the portal could potentially benefit the health
of resident trout and their populations at downstream Esopus sites in several
ways. The intakes for the Shandaken tunnel are deep in the Schoharie Reservoir
and withdrawals usually originate in the hypolimnion. Thus, water temperatures
for these enhanced flows are typically lower at downstream Esopus sites
than they are at upstream sites during the warm months. The positive effect of
the supplemental flows on habitat quantity (area and volume) in the upper basin
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
367
is illustrated by adding flows at esop3a, upstream of the portal (USGS 2012a),
which averaged 97 ft3s-1, to those from the portal (USGS 2012b), which averaged
226 ft3s-1 during summer base-flow periods (July–September) between 1996 and
2011. Simple addition shows that Esopus flows immediately downstream from
the portal averaged at least 323 ft3s-1 during these same months, and would have
been about 77% lower without supplemental flows from the Schoharie Reservoir.
During the warmest months of June, July, and August 2011, water temperatures
averaged 11.4, 14.0, and 17.2 °C, respectively at the portal; 14.9, 17.9, and
17.8 °C, respectively at the upstream site esop3a; and 12.8, 15.6, and 17.3 °C,
respectively at the downstream site esop3b. Except for August 2010 (when the
Schoharie Reservoir reached unusually low levels and epilimnion withdrawals
replaced hypolimnion withdrawals), water temperatures during June–August in
2009–2011 averaged 2.7 °C lower at the portal than at esop3a, and temperatures
averaged 1.8 °C lower at esop3b than at esop3a. The large volumes of cold water
from the portal have an important effect on summer stream temperatures at Esopus
reaches downstream from the portal. Such thermal differences are crucial to
the viability of Brown Trout populations in reaches where water temperatures approach
and sometimes surpass thermal limits for growth and survival (Elliott and
Elliott 2010, Wehrly et al. 2007).
The findings by Ross (2012) help explain our conflicting results that identified
substantial effects on the densities of juveniles (YOY), yet only minor effects on
densities of mature Brown Trout at downstream sites. Ross (2012) reported that
biomarkers (serum chemistry, gill histology, and water content) and growth rates
indicated that Brown Trout were generally stressed at all reaches (upstream, downstream,
and further downstream), but that the portal had no significant impact on the
species at downstream reaches in the Esopus. In general, Ross (2012) determined
that (a) the health and growth rates of adult Brown Trout were poor across the Upper
Esopus, (b) the quality of habitat for adult trout was optimal at downstream
sites, and (c) the quality of habitat for juvenile trout was poor at downstream sites.
These findings partly explain why mature Brown Trout densities were generally
unaffected and YOY Brown Trout were less abundant at downstream sites. Possibly
more notable was the fact that growth of Brown Trout, during warm months, was
less negative in reaches immediately downstream from the portal than growth of
trout from upstream and further downstream reaches. The inference is that conditions
downstream from the portal were less stressful than conditions at upstream
and further downstream reaches during the warm months. These data suggest that
a stress-refuge may exist periodically at sites immediately downstream from the
portal, yet our results indicate that no broad benefit was conveyed to resident Brown
Trout or Rainbow Trout populations at downstream sites.
The results from annual surveys generated baseline information for an important
natural resource and also detected decreasing temporal trends in many fish-community
and population metrics. These data are important because they define the
current status of local fisheries and quantify the normal measures of error (natural
variability) that may occur in key metrics given the typical year-to-year variations in
Northeastern Naturalist
368
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
temperature. precipitation, and stream flows. As such, these results provide a baseline
dataset that defines the fishery status during the period 2009–2011. More important,
these data are now available to quantify the impacts of short-term (e.g., severe flooding
or contaminant spills) or long-term (e.g., altered thermal or hydrologic regimes
associated with climate change) disturbances on resident-fish assemblages. The
decline in many metrics between 2009 and 2011 may reflect natural responses to a
drought during summer 2010 and moderate floods during the 2011 water-year (before
we conducted our fish surveys). Although speculative, low flows and warmer than
normal temperatures during summer 2010 could have caused widespread stress and
mortality among resident-fish species, and flood flows during fall 2010 and spring
2011 could have negatively affected the survival of eggs and YOY Brown Trout that
spawned in the fall or Rainbow Trout that spawned in the spring. Thus, depressed fish
populations during 2010 could have further adversely affected the density and biomass
of species populations during 2011. Regardless of the reasons, site-to-site and
year-to-year variations in most fishery metrics are much larger and more notable than
any effects that the portal may have had on fish populations.
In summary, from 2009–2011, the fish assemblages at most sites in the Esopus
were comparable to those previously observed in other rivers of the region. Community
richness increased near the portal, and median densities and biomass of fish
communities generally declined significantly at the downstream sites. Such effects,
however, do not suggest a specific cause and could easily be related to a general
shift to larger and fewer individuals, some sampling bias, differences in sampling
efficiency, changes in habitat, or poor water quality at sites downstream from the
portal. A decrease in water quality would implicate tributaries because they were
the primary sources of increased turbidity and suspended sediments loads at all
mainstem reaches downstream of the portal. The only obvious deleterious effect of
the portal seemed to be a large decrease in the standing crop of primary producers
as indicated by changes in chl a and AFDM at sites immediately downstream from
the portal (Smith et al. 2013). Less conspicuous, the densities of YOY Brown Trout
were significantly reduced at several downstream sites, but associated impacts are
not transferred to their whole populations. The increased water volume and decreased
temperature at downstream reaches during summer months also increase
physical habitat and help moderate temperatures that would otherwise be harmful
to resident trout. Although such conditions may not truly be beneficial, the conditions
at mainstem sites immediately downstream from the portal are clearly less
harmful to trout (and their populations) than are conditions at upstream and further
downstream sites during warm months. Despite the notable decreases in primary
production and density of YOY Brown Trout at several sites, this study shows that
fish communities and trout populations across most reaches in the Esopus are generally
in good condition and unaffected by the portal.
Acknowledgments
The authors extend appreciation to David Munsey and Martyn Smith of the USGS;
Danyelle Davis and Mark Vian of the NYCDEP; Karen Stainbrook of the NYSDEC; and
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
369
Alissa Freligh, Justin Zimmerman, Luis Rodriguez, Kevin Hackett, Sam Hawspurg, Jesse
McCarthy, Katherine Keegan-Twombly, Bobby Watzka, Bradley Mclean, James Werner,
Evan Leahy, and Brandon Annabel of Ulster County Community College for technical and
field support. This research was funded by the NYSDEC, Cornell Cooperative Extension of
Ulster County, and the USGS. Anonymous reviewers made numerous helpful suggestions
during manuscript review.
Literature Cited
Baldigo, B.P., M.B. DeLucia, W.T. Keller, G.S. Schuler, and C.D. Apse. In press, Contrasting
fish assemblages in free-flowing and impounded tributaries to the Upper Delaware
River: Implications for conserving biodiversity. Advances in Environmental Research.
Bilotta, G.S., and R.E. Brazier. 2008. Understanding the influence of suspended solids on
water quality and aquatic biota. Water Research 42:2849–2861.
Clarke, K.R., and R.M. Warwick. 2001. Change in marine communities: An Approach to
Statistical Analysis and Interpretation, 2nd Edition. PRIMER-E Ltd, Plymout h, UK.
Cornell Cooperative Extension (CCE). 2007. Upper Esopus Creek management plan. Volume
I: Draft summary of findings and recommendations. Cornell Cooperative Extension
of Ulster County, Kingston, NY.
Cucherousset, J., and J.D. Olden. 2011. Ecological impacts of nonnative freshwater fishes.
Fisheries 36:215-30
Cumming, G., F. Fidler, and D.L. Vaux. 2007. Error bars in experimental biology. Journal
of Cell Biology 177:7–11.
Elliott, J.M., and J.A. Elliott. 2010. Temperature requirements of Atlantic Salmon, Salmo
salar; Brown Trout, Salmo trutta; and Arctic Charr, Salvelinus alpinus: Predicting the
effects of climate change. Journal of Fish Biology 77:1793–1817.
Environmental Protection Agency (EPA). 2014. Enforcement and complience history online
(ECHO) data for facility ID: 110028180304. Available online at https://echo.epa.
gov/. Accessed 28 October 2014.
Fitzpatrick, F.A., I.R. Waite, J. D’Arconte, M.R. Meador, M.A. Maupin, and M.E. Gurtz.
1998. Revised methods for characterizing stream habitat in the National Water-Quality
Assessment Program. US Geological Survey, Raleigh, NC.
Gradall, K.S., and W.A. Swenson. 1982. Responses of Brook Trout and Creek Chubs to
turbidity. Transactions of the American Fisheries Society 111:39–395.
Harvey, B.C., and S.F. Railsback. 2009. Exploring the persistence of stream-dwelling trout
populations under alternative real-world turbidity regimes with an individual-based
model. Transactions of the American Fisheries Society 138:348–360.
Hasnain, S.S., B.J. Shuter, and C.K. Minns. 2013. Phylogeny influences the relationships
linking key ecological thermal metrics for North American freshwater fish species. Canadian
Journal of Fisheries and Aquatic Sciences 70:964–972.
Henley, W.F., M.A. Patterson, R.J. Neves, and A.D. Lemly. 2000. Effects of sedimentation
and turbidity on lotic food webs: A concise review for natural resource managers. Reviews
in Fisheries Science 8:125–139.
Hutchinson, G.E. 1993. A Treatise on Limnology. Volume IV. The Zoobenthos. Pp. 944, In
Y.H. Edmonson (Ed.). John Wiley and Sons, New York, NY.
Kruskal, J.B. 1964. Multidimensional scaling by optimizing goodness of fit to a nonmetric
hypothesis. Psychometrika 29:1–27.
Kurle, C.M., and B.J. Cardinale. 2011. Ecological factors associated with the strength of
trophic cascades in streams. Oikos 120:1897–1908.
Northeastern Naturalist
370
B.P. Baldigo, S.D. George, and W.T. Keller
2015 Vol. 22, No. 2
McHale, M.R., and J. Siemion. 2014. Turbidity and suspended sediment in the upper Esopus
Creek watershed, Ulster County, New York. SIR 2014-5200. US Geological Survey,
Troy, NY. 42 pp.
Meador, M.R., J.P. McIntyre, and K.H. Pollack. 2003. Assessing the efficacy of single-pass
backpack electrofishing to characterize fish-community structure. Transactions of the
American Fisheries Society 132:39–46.
Newcombe, C.P. 2003. Impact-assessment model for clear-water fishes exposed to excessively
cloudy water. Journal of the American Water Resources Association 39:529–544.
Newcombe, C.P., and D.D. Macdonald. 1991. Effects of suspended sediments on aquatic
ecosystems. North American Journal of Fisheries Management 11:72–82.
Quinn, J., R. Davies-Colley, C. Hickey, M. Vickers, and P. Ryan. 1992. Effects of clay discharges
on streams. Hydrobiologia 248:235–247.
Redding, J.M. 1987. Physiological effects on Coho Salmon and Steelhead of exposure to
suspended solids. Transactions of the American Fisheries Society 116:737–744.
Rosemond, A.D., P.J. Mulholland, and J.W. Elwood. 1993. Top-down and bottom-up control
of stream periphyton: Effects of nutrients and herbivores. Ecology 74:1264–1280.
Ross, T.J. 2012. Effects of anthropogenic stream alteration on Brown Trout habitat, movement,
and physiology. M.Sc. Thesis. Cornell University, Ithaca, NY. 106 pp.
Rowe, D.K., T.L. Dean, E. Williams, and J.P. Smith. 2003. Effects of turbidity on the ability
of juvenile Rainbow Trout, Oncorhynchus mykiss, to feed on limnetic and benthic prey
in laboratory tanks. New Zealand Journal of Marine and Freshwater Research 37:45–52.
Schaefer, J.F., S.R. Clark, and M.L. Warren. 2012. Diversity and stability in Mississippi
stream-fish assemblages. Freshwater Science 31:882–894.
Schlosser, I.J. 1987. A conceptual framework for fish communities in small warmwater
streams. Pp. 17–24, In W.J. Matthews and D.C. Heins (Eds.). Community and Evolutionary
Ecology of North American Stream Fishes. Oklahoma University Press, Norman,
OK.
Schlosser, I.J. 1991. Stream fish ecology: A landscape perspective. Bioscience 41:704–712.
Shaw, E.A., and J.S. Richardson. 2001. Direct and indirect effects of sediment-pulse duration
on stream-invertebrate assemblages and Rainbow Trout (Oncorhynchus mykiss)
growth and survival. Canadian Journal of Fisheries and Aquatic Sciences 58:2213–2221.
Shepard, R.N. 1962. The analysis of proximities: Multidimensional scaling with an unknown
distance function. Psychometrika 27:125–140.
Sigler, J.W., T.C. Bjornn, and F.H. Everest. 1984. Effects of chronic turbidity on density
and growth of Steelheads and Coho Salmon. Transactions of the American Fisheries
Society 113:142–150.
Simonson, T.D., J. Lyons, and P.D. Kanehl. 1994. Quantifying fish habitat in streams:
Transect spacing, sample size, and a proposed framework. North American Journal of
Fisheries Management 14:607–615.
Simpson, E.H. 1949. Measurement of diversity. Nature 163:688.
Smith, A.J., B.T. Duffy, D.L. Heitzman, J. Lojpersberger, L.E. Abele, B.P. Baldigo, M.R.
McHale, S.G. George, J. Siemion, and M.A. Novak. 2013. Upper Esopus Creek: Biological
assessment, 2009–2010 Survey. New York State Department of Environmental
Conservation Troy, NY. 30 pp.
Stuart-Smith, R.D., A.M.M. Richardson, and R.W.G. White. 2004. Increasing turbidity
significantly alters the diet of Brown Trout: A multi-year longitudinal study. Journal of
Fish Biology 65:376–388.
Sweka, J.A., and K.J. Hartman. 2001a. Effects of turbidity on prey consumption and growth
in Brook Trout and implications for bioenergetics modeling. Canadian Journal of Fisheries
and Aquatic Sciences 58:386–391.
Northeastern Naturalist Vol. 22, No. 2
B.P. Baldigo, S.D. George, and W.T. Keller
2015
371
Sweka, J.A., and K.J. Hartman. 2001b. Influence of turbidity on Brook Trout reactive
distance and foraging success. Transactions of the American Fisheries Society
130:138–146.
US Geological Survey (USGS). 2012a. Water-resources data for the United States, water
year 2011: US Geological Survey water-data report WDR-US-2011, site 01362200.
Available online at http://wdr.water.usgs.gov/wy2011/pdfs/01362200.2011.pdf. Accessed
11 September 2014.
USGS. 2012b. Water-resources data for the United States, water year 2011: US Geological
Survey water-data report WDR-US-2011, site 01362370. Available online at http://wdr.
water.usgs.gov/wy2011/pdfs/01362370.2011.pdf. Accessed 11 September 2014.
USGS. 2014. USGS water data for the nation. Available online at http://waterdata.usgs.
gov/nwis/. Accessed 28 October 2014.
Van Deventer, J.S., and W.S. Platts. 1985. A computer-software system for entering, managing,
and analyzing fish-capture data from streams. US Forest Ser vice, Ogden, UT.
Vannote, R.L., G.W. Minshall, K.W. Cummins, J.R. Sedell, and C.E. Cushing. 1980.
The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences
37:130–137.
Wagener, S.M., and J.D. LaPerriere. 1985. Effects of placer mining on the invertebrate
communities of interior Alaska streams. Freshwater Invertebrate Biology 4:208–214.
Ward, J.V., and J.A. Stanford. 1983. The serial discontinuity concept of lotic ecosystems.
Pp. 29–42, In T.D. Fontaine and S.M. Bartell (Eds.). Dynamics of Lotic Ecosystems.
Ann Arbor Science, Ann Arbor, MI.
Wehrly, K.E., L.Z. Wang, and M. Mitro. 2007. Field-based estimates of thermal tolerance
limits for trout: Incorporating exposure time and temperature fluctuation. Transactions
of the American Fisheries Society 136:365–374.
White, J.L., and B.C. Harvey. 2007. Winter-feeding success of stream trout under different
streamflow and turbidity conditions. Transactions of the American Fisheries Society
136:1187–1192.
Whiteside, B.G., and R.M. McNatt. 1972. Fish-species diversity in relation to stream order
and physicochemical conditions in the Plum Creek drainage basin. American Midland
Naturalist 88:90–101.
Whittaker, R.H. 1975. Communities and Ecosystems, 2nd edition. MacMillan Publishing
Company, New York, NY. 352 pp.
Zalewski, M., B. Brewinska-Zaras, P. Frankiewicz, and S. Kalinowski. 1990. The potential
for biomanipulation using fry communities in a lowland reservoir: Concordance between
water quality and optimal recruitment. Hvdrobiologia 200/ 201:549–556.
Zippin, C. 1958. The removal method of population estimation. Journal of Wildlife Management
22:82–90.