Western Kentucky UniversityTopSCHOLAR®

Masters Theses & Specialist Projects Graduate School

5-1-2008

Longitudinal Variation in Fish Community Alongthe Upper Green River Below the Green RiverLakeJoshua WilseyWestern Kentucky University

Follow this and additional works at: http://digitalcommons.wku.edu/theses

Part of the Medical Sciences Commons

This Thesis is brought to you for free and open access by TopSCHOLAR®. It has been accepted for inclusion in Masters Theses & Specialist Projects byan authorized administrator of TopSCHOLAR®. For more information, please contact topscholar@wku.edu.

Recommended CitationWilsey, Joshua, "Longitudinal Variation in Fish Community Along the Upper Green River Below the Green River Lake" (2008).Masters Theses & Specialist Projects. Paper 387.http://digitalcommons.wku.edu/theses/387

LONGITUDINAL VARIATION IN FISH COMMUNITY ALONG THE UPPERGREEN RIVER BELOW THE GREEN RIVER LAKE.

A Thesis Presented toThe Faculty of the Department of Biology

Western Kentucky UniversityBowling Green, Kentucky

In Partial FulfillmentOf the Requirements for the Degree

Master of Biology

ByJoshua Delmar Wilsey

May 2008

LONGITUDINAL VARIATION IN FISH COMMUNITY ALONG THE UPPERGREEN RIVER BELOW THE GREEN RIVER LAKE.

Date Recommended

Director of Thesis

)ean, Graduate Studies and Research Date

Table of Contents

1) Abstract2) Introduction3) Study Area4) Habitat Characterization5) Temperature6) Fish Sampling7) Analytical Methods8) Results9) Discussion10) Appendix11) Literature Cited

pg. ii-iiipg. 3-5pg. 5-6pg. 6-7Pg-7Pg-7pg. 7-9pg. 9-13pg. 13-18pg. 19-39pg. 40-43

LONGITUDINAL VARIATION IN FISH COMMUNITY ALONG THEUPPER GREEN RIVER BELOW THE GREEN RIVER LAKE.

Joshua D. Wilsey May 2008 Pages 43

Directed By: Scott Grubbs, Philip Lienesch and Albert MeierDepartment of Biology Western KentuckyUniversity

The impoundment of flowing waters can cause decreased variation in flow and

temperature, reduce habitat availability, and decreased mean temperature below a

reservoir. These types of alterations can influence fish communities. Twelve sites were

sampled along the Upper Green river in order to address the potential influence of

impoundment on fish communities below the Green River Lake. The habitat features:

mean depth, flow, dominant substrate, embeddedness, percent riffle, run and pool were

measured. Temperature probes were placed at each site in order to measure temperature

over an extended period of time. Backpack electrofishing and seining was used to obtain

fishes. The lowest richness and Shannon diversity were from the two most upstream

sites. Principal components analysis of the wadeable fish data suggests possible

differences in species composition at the two most upstream sites when compared to the

other downstream sites. Mean daily temperature was significantly different between

sites (p = 0.05). Significant differences in seasonal daily means, minimums and

maximum temperatures suggests that there are three thermal sections along the Green

River within the study area: (1) a cooler section influenced by release patterns from the

Green River Dam, (2) a warmer section downstream of the Green River Dam receiving

surface inputs from the non-karst tributaries, and (3) the cooler section flowing through

the well-developed karst region. Linear regressions only showed significant relationships

n

between Etheostoma helium abundances and mean daily temperature (p = 0.002),

between Campostoma oligolepis abundances and mean daily temperature (p = 0.043) and

between Etheostoma zonale and % gravel. Finding only four significant linear

relationship between species abundances and environmental variable suggest other

environmental variable not measured may be important for determining species

composition in the Green River.

in

Introduction

The damming of streams is an anthropogenic disturbance is considered one of

major causes for the imperilment of native species in the southern United States because

of frequency of its occurrence across a wide range of stream sizes (Warren et. al. 2000).

Not only are impoundments widespread geographically but they influence a wide range

of communities including: macroinvertebrate communities (Vinson 2001; Lehmkuhl

1972), riparian communities within flood plains (Richter and Richter 2000) and fish

communities (Quist et. al. 2005; Eberle et. al. 2002). Dams have the potential to impact

streams by altering natural flow regime, habitat availability, and in-stream thermal

patterns (Allan 1995). Flow alteration is one of the most common and obvious changes

caused by impoundment. With the exception of hydroelectric dams, typically

impoundment causes decreases in the annual variation of water levels downstream

(Baxter 1977). The presence of a dam can lessen the impact of seasonal variation in water

level from increased water flow due to heavy rain events and decreased water levels that

occur during periods of drought. This buffering of hydrologic variability then leads to

decreased frequency of extreme high and low water events (Richter et al. 1996). These

kinds of changes to flow regime often negatively impact fish communities (Powers et al.

1988; Osmundson et al. 2002; Franssen et al. 2006; Freeman & Marcinek 2006).

Hydrologic changes below an impoundment can alter a stream's morphology. The

natural variation between high and low water events are important to the maintenance of

habitat features within a stream, including riffles, sand bars, and floodplain riparian zones

(Richter and Richter 2000). As water level increases so does its ability to pick up

sediments, organic material and other particles. Periodic high water events are often

necessary to prevent coarser substrates from being covered over by finer substrates. High

flow events causes a stream to transport these materials from further upstream and

deposit them in sand bars, stream banks and slack water areas once water levels recede

(Osmundson et al. 2002). Impoundments routinely disrupt this process by altering natural

flow regimes and by acting as a sink trapping finer sediments being transported from

upstream. This deposition is often necessary for the maintenance of those features

composed of finer substrates. Without a source of finer substrates from upstream these

features can erode away, leaving behind only courser substrates. When habitat structures

like sand bars, and riffles are lost it causes a loss of in-stream habitat heterogeneity. This

loss of habitat can influence fish communities (Schlosser 1982; Pearsons et al. 1992;

Bunn & Arthington 2002; Koel 2004), especially for those species that specialize in

utilizing these types of habitat.

Dams may alter downstream temperature regimes by decreasing thermal variation

both daily and seasonally (Jaske & Goebel 1967; Baxter 1977). Thermal alteration is

largely due to the decreased surface area to volume of reservoirs and thermal

stratification that typically decrease the influence of fluctuations in air temperature on

water temperature. In addition, if water is being released from the hypolimnion it is often

much cooler than the surface water of streams. These kinds of alterations in thermal

pattern can influence the distribution and diversity of fishes by causing reduced

abundances and absences of those species sensitive to this type of alteration (Baxter

1977; Paller& Saul 1996).

Many studies have previously studied from a temporal perspective the

downstream effects of impoundment on fish communities (Tieman & Gillette 2004;

Taylor et. al. 2001; Martinez et. al. 1994). Few studies have addressed the influence that a

disturbance gradient created by damming can have on the longitudinal distribution fish

communities. The purpose of this research was to address the influence of impoundment

on the longitudinal distribution of a riverine fish community in central Kentucky, U.S.A.

The key environmental variables that were the focus of this study were thermal

variability and habitat size. They were addressed in detail at several positions

downstream from the reservoir, by determining if the furthest upstream sites possessed

the lowest mean daily water temperatures, the least variation in water temperature, and

the least percent composition of fine substrates than sites further downstream, in order to

assess the possible disturbance gradient created by the reservoir. These key

environmental variables were compared to the fish communities at each site in order to

find if the changes in the environmental variables being caused by dam releases resulted

in decreases or absences of certain fish species or even changes to entire guilds of fishes.

These types of changes in the fish community would indicate that the presence of the

Green River Lake was impacting fish communities below it.

Study Area

All field work was conducted along the mainstem of the Green River in central

Kentucky, U.S.A. between the outflow of the Green River Lake and Mammoth Cave

National Park (Fig. 1). The Green River is a 6th-7lh order stream within the study area.

The Green River originates in Casey and Lincoln Counties and flows nearly 600 km west

to its confluence with the Ohio River in Henderson County.

Within the study area the Green River flows through three Level-IV Interior

Plateau Ecoregions (Woods et al. 2002). In longitudinal order, the Green River flows

through the Eastern Highland Rim, Western Pennroyal Karst Plain, and Crawford-

Mammoth Cave Uplands Ecoregions. The Eastern Highland Rim is underlain by

Mississippian-age sandstones. In contrast, the Western Pennyroyal Karst Plain is

characterized by Mississippian-age limestone with extensive networks of underground

fluvial systems. Surface stream density is very low. The Crawford-Mammoth Cave

Uplands is geologically characterized by Mississippian-age sandstone in upland

environments and Mississippian limestone in karst valleys.

Habitat Characterization

Twelve sites were sampled along the Green River within the study area (Fig. 1).

Habitat variables were quantified during summer baseflow conditions. Sampling reaches

at each site were defined as 15 times the wetted width of the stream or 1000 m,

whichever was smaller, and included both wadable and non-wadable portions. The

percent composition of riffle, run, and pool habitat features of the entire sampling reach

were visually estimated. Within the wadable section five transects were placed

perpendicular to flow and spaced every 30-40 m. Along each transect the dominant

substrate was estimated visually. Depth was measured to the nearest 0.5 cm and velocity

was quantified with a Marsh-McBirney Flo-Mate Model 2000 Portable Flowmeter.

Within the non-wadable portions of the sampling reach depth was measured at an

additional five transects spaced every 30-40 m. Large woody debris, root wads, undercut

banks and other habitat features along both sets of transects were noted. Percent

composition of each substrate class sampled, mean depth, mean velocity, and mean

discharge were calculated from the transect data collected. The percent composition of

riffle, run, and pool habitat features within each sampling reach were also visually

estimated.

Temperature

HOBO® Water Temp Pro v2 Logger temperature probes were used to measure

temperature over an extended period of time at each sampling site. The probes were

placed in-stream in May 2005 and removed in August 2006. Probes were redeployed in

August 2006 and retrieved in September 2007, yet because data could not be obtained

from a subset of three probes all thermal information for this time period was excluded

from analyses. Temperature was recorded every hour. Mean daily temperature, minimum,

maximum, standard deviation, and coefficient of variation were calculated per site for

each month, season, and the entire 15-mo period.

Fish Sampling

Fish were collected at each site during May - August 2007. Within wadable

sections of each sampling reach fish were sampled using 30-60 minutes of seining and 15

minutes of backpack electrofishing. Common and easily identifiable fish were identified

in the field. All remaining fish collected were fixed in 10% buffered formalin and taken

back to the lab for identification.

Analytical Methods

Habitat and temperature data were combined in one environmental matrix. The

mean daily, minimum, and maximum temperatures were compared between sites using

the analysis of variance (ANOVA). The mean daily, minimum and maximum

temperatures for each site were further partitioned by season and ANOVAs were

8

performed to explore seasonal trends in the data. Tukey's honestly significant difference

(Tukey's HSD) test was performed on all ANOVAs to detect statistical pair-wise site

differences. Bonferroni corrections were applied to minimize type I error. Principle

components analysis (PCA) was performed on the environmental data to explore

longitudinal relationships along the Green River.

Species richness, evenness and Shannon-Weaver diversity indices were calculated

from the fish data set. Rank abundance curves were plotted to explore changes in

abundance that occurs longitudinally along the Green River. In order to assess fish

community structure each species was assigned to guilds within four categories (trophic

guild, substrate preference, stream flow velocity preference, and reproductive guild;

Table 1). Life history information from multiple sources was used to assign species to

guilds (Frose & Pauly 2008; Natureserve 2008; Pflieger 1997; Etnier & Starnes 1993).

The fish data were ratified by removing those species that occurred only at one

site and were logio (x+1) transformed prior to analyses, were x = species abundance per

site. Principle components analysis was performed to explore longitudinal relationships

in species composition. Simple linear regressions were conducted to assess relationships

between species abundance data and environmental variables. Only those species and

variables that had strong PCA loadings were used. An objective mechanism was not

determined apriori when interpreting which variables possessed strong loadings. The

same loading values were not used between PCA axes, instead the first natural block of

the most extreme values positive or negative were selected. The number of those

variables selected was entirely based on the relative strength of their loading. Simple

linear regressions were also performed between species richness, Shannon-Weaver

diversity and fish guild data and environmental variable data. Bonferroni corrections

were applied to minimize Type I error. Indicator species analysis was performed on the

fish data sets in order to identify those taxa that were indicative of distinct groups among

the sites sampled. Ordination, calculation of indices and the indicator species analysis

were performed with PC-ORD Version 5 for Windows (MJM Software Design 1999)

while SPSS 12.0 (SPSS Inc. 2003) was used for the ANOVA's and linear regressions.

Results

Although wetted width varied among the upstream sites due to difference in

channel morphology, the downstream sites were markedly larger (Table 2). Mean depth

(50-101 cm) and discharge (1.0-5.8 m3/s) varied predictably, increasing in a downstream

direction (Table 2). Gravel was the primary substrate at most sites. With one exception

the two most upstream sites, GR 8.14 and GR 8.15, possessed the lowest percentage of

fine substrates.

Mean daily temperatures varied less between sites (16.3 - 17.6°C) than both

minimum (0.2 - 5.4°C) and maximum (28.4 - 32.4°C) temperatures. The lowest mean

temperatures occurred at the two most upstream sites, GR 8.14 and GR 8.15. Conversely,

these two sites also possessed the highest minimum daily temperatures and except for the

two most downstream sites (GR 8.0D and GR 8.1) exhibited the least variability (Figure

2; Table 2).

Over the whole 15-mo period, mean daily (Fdf=n = 1.79, p = 0.05) and daily

maximum (Fdf n = 2.09, p = 0.02) temperatures were significantly different between

sites, but daily minimums were not significant (Fdf=n = 1.40, p = 0.17). There were a

consistent series of within-season differences. Mean daily (Fdf=n = 6.99, p < 0.0005),

10

daily minimum (Fdf=n = 8.45, p < 0.0005) and maximum temperatures (Fdf=n = 5.43, p <

0.0005) were significantly different between sites for spring. Within summer, mean daily

(FdfHi = 44.89, p < 0.0005), daily minimum (Fdf=n = 40.105, p < 0.0005) and maximum

temperature (Fdf=i i = 43.92, p < 0.0005) were significantly different between sites.

Similarly, autumn mean daily (Fdf=n = 2.77, p = 0.001), daily minimums (Fdf=n =3.19,

p < 0.0005), and daily maximum temperatures were significantly different between sites

(Fdf=ii = 2.54, p = 0.004). Means for daily temperature (Fdf=n = 4.79, p < 0.0005), daily

minimum (F = 5.05, p < 0.0005) and daily maximum temperatures (Fdf=n = 4.72, p <

0.0005) were also significant during winter days (Table 3).

The first three PC A axes explained 74.3% of the variation of the environmental

data (Table 4). Strong positive scores were exhibited for sites GR 8.0D, GR 8.1, and GR

8.2 along PCA Axis 1. Strong negative scores were observed for GR 8.9 and GR 8.10.

GR 8.3 and GR 8.15 possessed strong positive scores along PCA Axis 2. Along PCA

Axis 3 there were strong positive scores for GR 8.14 and GR 8.15 and a strong negative

score for GR 8.3 (Table 4; Figure 3). Mean wetted width, percent pool, and both standard

deviation and coefficient of variation of the temperature data were strong contributors to

PCA Axis 1. Percent cobble, percent small boulder, and mean daily temperature

produced high loadings along PCA Axis 2 while percent gravel was the sole variable

important to PCA Axis 3.

There were 61 species total collected across the 12 sites. The four most common

species sampled were Etheostoma zonale, Pimephales notatus, Lepomis megalotis, and

Campostoma oligolepis, comprising 46.6 % of the total number of individuals collected.

Several species were obtained from all 12 sites, including C. oligolepis, E. blenniodes, E.

11

caeruleum, E. zonale, L. megalotis and P. notalus. Cottus carolinae was collected

mainly from the downstream sites. C.oligolepis, E. helium, E. zonale, L. megalotis and P.

notatus were most common in the middle sites. Pomoxis annularis was obtained only

from the two most upstream sites (GR 8.14 and GR 8.15) (Figure 4; Table 5).

Species richness ranged from 14 to 27 species per site. Shannon-Weaver diversity

ranged from 2.19 to 2.69, while evenness showed no obvious pattern between sites and

was between 0.95 and 0.96 (Table 6). The lowest richness and Shannon diversity were

from the two most upstream sites, GR 8.14 and GR 8.15. Benthic insectivores comprised

the largest trophic guild at most sites. Prominent species within the benthic insectivore

guild ordered from most abundant to least include: Etheostoma zonale, Cottus carolinea,

E. blennoides, and E. helium. Generalist insectivores composed the second largest trophic

guild (Figure 5). Lepomis megalotis and Notropis photogenis were two important species

within this group. Those species that had a preference for gravel substrates made up the

highest percentage of the fish community at all sites except GR 8.15. C. carolinae, N.

atherinoides, E. rafinesquei, and E. helium composed most of the individuals collected

that possessed a preference for gravel substrates. The next largest group was those

species preferring cobble substrates, followed closely by substrate generalists (Figure 6).

E. zonale, C. oligolepis, N. photogenis, and E. blenniodes were important cobble species

and Pimephales notatus and L. megalotis were the most abundant substrate generalists.

GR 8.15 possessed a greater proportion of silt and substrate generalist guilds than all

other sites. The presence of Pomoxis annularis and Lepomis macrochirus were primarily

responsible for the greater percentage of silt species. Percent composition within velocity

guilds revealed an even distribution between fast, moderate, slow, and no flow

12

preferences. Only a small percentage of species were flow velocity generalists (Figure 7).

GR 8.15 possessed greater percent composition of slow and generalist stream flow guilds

than other sites. Lepomis megalotis, Fundulus caienatus, and Gambusia affinis were

prominent species within the slow/no flow guild and Pimephales notatus the most

abundant species within the flow generalist guild. The two upstream sites, GR 8.14 and

GR 8.15, possessed higher percentages of nest spawners and substrate choosers than were

present at most other sites and fewer open water/substrate spawners (Figure 8). Lepomis

megalotis, and P. notatus were the most common nest spawner species collected. Within

the substrate chooser guild, E. zonale and E. blenniodes were important species.

Important species within the open water/substrate spawner group included: C. oligolepis,

Cottus carolinea, and N. photogenis.

The PC A of the fish data showed 60.9% of the variation was explained along the

first three axes (Table 4). Strong negative loadings were exhibited for sites GR 8.3, GR

8.14 and GR 8.15 along PCA Axis 1. Along PCA Axis 2, Sites GR 8.14 and GR 8.15

possessed strong positive loadings and GR 8.3 possessed a strong negative loading. A

plot of axes 1 and 2 isolated GR 8.3, GR 8.14 and GR 8.15 for the other sites sampled

(Figure 9). Campostoma oligolepis, Notropis atherinoides, N. photogenis, Phenacobius

uranops, Fundulus notatus, E. bellum, E. rafinesquei were strong contributors to PCA

axis 1. Notropis volucellus, Pylodictis olivaris, Gambusia affinis, Labdidesthes sicculus,

E. caeruleum, E. zonale produced strong loadings along PCA Axis 2. Increased

abundances of E. rafinesquei, G. affinis, and L.sicculus were important in determining

where GR 8.3 was located along PCA axes 1 and 2. Decreased abundances of E. bellum,

E. zonale and C. oligolepis were important in determining where sites GR 8.14 and

GR8.15 were located along PCA axes 1 and 2.

Linear regressions revealed three significant relationships between the abundance

of individual fish species and environmental variables. Both Etheostoma helium (b =

0.02, r2 = 0.62, p = 0.002) and Campostoma oligolepis (b = 0.01, r2 = 0.35, p = 0.04)

abundances were linearly related to mean daily temperature. The only remaining

significant linear relationship was between C. oligolepis and stream velocity (b = 4.11, r2

= 0.53, p = 0.007). Mean daily temperature was able to significantly explain values of

both species richness (b = 6.57, r2 = 0.54, p = 0.007) and Shannon-Weaver diversity (b =

0.35, r2 = 0.57, p = 0.005). Pertaining to guild data, there were significant positive

relationships between substrate choosers and percent riffle (b = 0.41, r2 = 0.43, p = 0.02)

and between nest spawners and percent cobble (b = 0.30, r = 0.38, p = 0.03). The

indicator species analysis revealed that the two upstream sites, GR 8.14 and GR 8.15,

were isolated from the remaining 10 sites (Figure 10). Etheostoma helium (p = 0.015) and

E. zonale (p = 0.015) were significant indicators of the ten more downstream sites while

P. annularis was the sole significant indicator of the two upstream sites (p = 0.015).

Discussion

This decrease in fine substrates and increased dominance of coarser substrates

frequently occurs directly below a dam outflow. The two most upstream sites, GR 8.14

and GR 8.15, possessed lower percentages of fine substrates than the other sites sampled

and as a result possessed greater percentages of coarser substrates. This results because

dams interrupt the transport of these fine substrates from further upstream and the

13

14

sediment-free waters leaving the dam outflow tend to pick up fine substrates mainly in a

downstream manner (Allan 1995).

Results from the 2005-2006 temperature data analyses at first glance appear to be

somewhat mixed. The Tukey's HSD revealed patterns in seasonal daily means,

minimums and maximums at GR 8.14 and GR 8.15 that were significantly different from

most of the other sites. The mean temperatures and maximums were significantly lower,

and minimums were significantly higher, indicating that the Green River Lake outflow

altered the thermal regime at these sites. This type of thermal pattern is common in the

tailwaters of streams, especially those with hypolimnetic release patterns (Jaske &

Goebel 1967; Baxter 1977). Thermal similarities between these two upstream sites and

the two most downstream sites (GR 8.0D and GR 8.1), however, appear to somewhat

contradict this conclusion unless the base geology of this study region is considered.

These two downstream sites are located in Crawford-Mammoth Cave Uplands Level-IV

Ecoregion, a well-developed karst region of Mississippian-age limestone. Surface stream

density is very low is this region. The thermal regime of the Green River flowing through

this region is heavily influenced by groundwater inputs. O'Driscoll and DeWalle (2006)

similarly found that stream sites that received significant groundwater inputs were less

responsive to changes in air temperature, often resulting in cooler water temperatures in

the summer and warmer temperatures in the winter. This type of a pattern is similar to

that of streams being influenced thermally by hypolimnetic release from reservoirs. This

suggests that there are three thermal sections along the Green River within the study area:

(1) a cooler section influenced by release patterns from the Green River Dam, (2) a

warmer section downstream of the Green River Dam receiving surface inputs from the

15

non-karst tributaries, and (3) the cooler section flowing through the well-developed karst

region.

The differences in species richness, Shannon-Weaver diversity, guild

composition, the PC A ordination of species abundances, and indicator species analysis all

indicate that the two upstream sites, GR 8.14 and GR 8.15, possess fish communities

distinct from each of the remaining study sites. Mean daily temperature possessed a

significant positive relationship to both richness and Shannon-Weaver diversity. Richness

and Shannon-Weaver diversity were lowest at GR 8.14 and GR 8.15. Of the species that

loaded strongly on PC A Axis 1, both Campostoma oligolepis and Etheostoma helium

abundances were significantly related in a positive direction to mean daily temperatures.

This makes it very possible that at least part of the variation explained along PC A Axis 1

was related to mean daily temperature. Decreases in diversity and changes to community

composition often occur within the tailwaters of a reservoir after the establishment of the

impoundment (Edwards 1978; Quinn & Kwak 2003).

Campostoma. oligolepis, one of the species possessing significant linear

relationships with environmental variables, is a stoneroller species found in drainages

within the Midwestern and southeastern United States (NatureServe 2008). They occur in

swift flowing rocky riffles within creeks, and small to medium streams (Fowler & Taber

1985). They graze gravel and rock substrates comsuming periphyton present on these

substrates (Fowler & Taber 1985). Lower mean daily temperatures could be negatively

influencing C. oligolepis abundances indirectly by limiting periphyton growth.

Rosemond (1994) found temperature to be a factor that positively influenced periphyton

biomass and abundance.

16

Etheostoma helium the other species possessing a significant linear relationship to

mean daily temperature, is endemic to the Green and Barren River Basins and inhabits

and spawns in fast-flowing reaches with coarse substrates (Fisher 1990; Page and Burr

1992). This species was absent only from the two upstream reaches, GR 8.14 and GR

8.15. Etheostoma bellum spawns between April and June when water temperatures are

between 20-25 °C. Although the upstream reaches possess the highest proportions of

coarse substrates, thermal release patterns from the Green River Lake produced stream

temperatures that are arguably too cool for effective spawning of this species. During

2006 the daily mean April - June temperature of GR 8.14 and GR 8.15 possessed only 37

and 25 days, respectively where mean daily temperature was between 20-25 °C, while the

remaining sites possessed between 51 and 63 days.

The significantly positive relationship between both E. bellum and C. oligolepis

and mean daily temperature both indicate that temperature is one factor influencing fish

communities in the Green River. This influence becomes clearer when the changes in

these two species are considered along with the results from the ANOVAs of the

temperature data that suggest three thermal regions. Abundances in both E. bellum and

C. oligolepis are highest in the middle sites where mean temperatures are warmer and

their abundances are lowest in both the most upstream two sites (GR 8.14, GR 8.15) and

the two most downstream sites (GR 8.0D, GR 8.1) where mean water temperatures are

the coolest. Why these two species possess lower abundances in sites with cooler mean

temperatures is unclear. The decreased abundances of these two species could be the

result of temperature influencing the larval development of these species. Darters have

been shown to possess an inverse-log relationship between temperature and incubation

17

time in days (Page 1983). Temperature has been shown to influence the larval growth

rates of other fish species (Claramunt & Wahl 2000; Lasker 1964). The influence of

temperature on growth rates is well known in marine and estuarine fishes (Klimogianni

et. al. 2004; Morehead & Hart 2003; McGurk 1984; Laurence & Rogers 1976). The

growth rate of larval fishes often determines the time of first feeding in these fish, which

is an important factor in determining their ability to avoid starvation (Rogers & Westin

1981; Laurence 1978; Houde 1974).

Other environmental parameters (e.g., substrate heterogeneity and hydrologic

variability) not specifically measured in this study may also be important factors

influencing fish community composition along the Green River. Several studies have

shown that hydrologic patterns can influence fish community structure (Poff & Allan

1997; Bain et al. 1988). Poff & Allan (1995) found that those sites that are hydrologically

variable have a greater composition of fishes that possess generalized feeding strategies,

are silt-tolerant, are slow stream velocity species and are associated with silt or are

substrate generalists. The greater percent composition of both species preferring silt and

substrate generalists and the greater percent composition of slow velocity preference

guilds at the most upstream site (GR 8.15) could suggest fishes at this site are

experiencing greater flow variability instead of more stable hydrologic conditions as

would be expected. However, direct measurement of hydrologic variability is necessary

to quantify such a relationship.

In conclusion this study was able to detect differences in key environmental

variables between upstream sites and downstream sites. Lower percent composition of

fine substrates, decreased mean daily temperatures, and minimum temperatures both

18

throughout the entire sampling period and within particular seasons within the upstream

sites were important changes found to be occurring as a result of the presence of the

Green River Lake. Being able to explain species richness, Shannon-Weaver diversity and

abundances of E. bellum and C. oligolepis, with mean daily temperature indicates that

releases from the Green River Lake are influencing the fish communities found below it

by causing decreased mean daily temperatures in the upstream sites.

\

Big Brush CreekLynn Camp Creek f Big Pitman Cieek

Mammoth CaveNational Park

Legend

O Sampling_SitesGreen River and Tributaries

Mammoth Cave National Park

Green River Lake

Upper Green River Subbasin

DDCD3Q.

5'

0 25 5 10 15 20

Figure 1. Map of study area along the upper Green River within central Kentucky, U.S.A.

20

Table 1. The four categories andthe guilds fish species wereassigned to within those categories.

ID Description

Trophic Guild123

45678

Hebivore-detritivoreOmnivoreGeneral InsectivoreSurface/water columninsectivoreBenthic InvertivorePiscivorePlanktivoreParasite

Substrate12345

CobbleGravelSandSiltGeneral

Current Velocity Preference1234

FastModerateSlow-NoneGeneral.

Reproductive Guild

123456

Open water/ substratumspawnerBrood HiderSubstratum ChooserNest SpawnerExternal BrooderInternal Bearer

Table 2. Environmental parameters sampled in 2007 from twelve sampling reaches along the Upper Green River. Samplingreaches arranged in order from downstream to upstream.

Par am

HabitatWidth

DepthVeloc

Disch

%Riff%Run%Pool%Fine

%Gr

%Cb

%SmB

Dist

Description

mean wetted width(m)mean depth (cm)mean velocity(m/s)mean totaldischarge (m3/s)% riffle habitat% run habitat% pool habitat% substrate as silt,sand and clay% substrate asgravel% substrate ascobble% substrate assmall boulderdistance from dam(km)

GR8.0D

47.8

930.20

5.8

35452038

62

0

0

155

GR8.1

47.5

760.35

5.0

20552520

80

0

0

130

GR8.2

45.3

1010.23

3.5

15701518

78

4

0

124

GR8.3

33.4

870.21

2.7

503020

6

28

30

36

96

GR8.7

34.5

690.16

1.8

10603018

64

6

12

64

GR8.9

29.9

670.33

1.7

204040

0

66

28

6

47

GR8.10

23.1

680.26

1.6

25403526

74

0

0

47

GR8.11

26.8

540.27

1.9

206020

8

72

20

0

43

GR8.12

36.3

500.25

2.0

30403034

66

0

0

39

GR8.13

34.9

600.19

1.7

30403034

62

0

4

24

GR8.14

32.3

500.20

1.0

354025

2

92

6

0

4

GR8.15

28.8

690.17

1.1

354025

2

74

22

2

2

to

Table 2. Continued

Par am Description GR GR GR GR GR GR GR GR GR GR GR GR8.0D 8.1 8.2 8.3 8.7 8.9 8.10 8.11 8.12 8.13 8.14 8.15

Temp.Min

Max

Mean

StDev

CV

minimum meandaily temperaturemaximum meandaily temperaturemean dailytemperaturestandard deviationof dailytemperaturescoefficient ofvariation of dailytemperatures

2.0

28.4

16.9

6.9

0.41

2.6 1.7 1.2 1.3 1.3 1.6 1.6 1.8 3.0 5.3 5.8

30.2 28.9 30.3 31.1 31.7 31.7 31.1 32.1 31.8 31.6 32.4

16.8 16.9 17.0 17.0 17.6 17.6 17.3 17.7 17.5 16.5 16.3

6.8 7.1 7.5 7.2 8.1 8.1 7.7 8.2 8.0 6.9 6.7

0.40 0.42 0.44 0.42 0.46 0.46 0.45 0.46 0.46 0.42 0.41

K)

23

OO

o

00

O

oc

O

Qo

o

o

oo

o

00

o

00

O

in

oo

a

00

a

en• i

00

O

00

o

Temperature °C

i/"i O ^"i O */"J O */"» Or*"t r*"( (N CNI •—« - H

Temperature °C

>/"i O V") O ^ J O «O Or^ f^i rs r-J -—i i—'

Temperature °C

V%

I P

\

V\

\.

Dat

e

1)

Q

•utaQ

ju13Q

Oo

c/33Of

1

oo

1co

o

0>

-pC3O

"3u

ratu

r

1

ean

d

•

ex

24

GR8.9A CV-

StDev

GR8.10A

•"""" GRS

°oPool y

A MeanGR8.12

A GR8.3

%SmB

%CB \ i o R l f f

13 /iir / \

^ / ^Veloc

A GR8.15

A G R 8 J > ^

( \

°o Fuie

^ °oGR

s Depth

GR8

S^Disch" Width

%Ruii

.OD •

AGR8.2

A

GR8.1

Axis 1

Figure 3. Plot of PCA Axes 1 and 2 of environmental data set

Table 3. Tukey's HSD results for mean daily, mean daily minimum, and mean daily maximum temperatures within each season.

SpringMeanTukey HSDDaily MinimumTukey HSDDaily MaximumTukey HSD

SummerMeanTukey HSDDaily MinimumTukey HSDDaily MaximumTukey HSD

GR8.0D

17.94b

17.70b

17.82abc

24.12ab

23.40bed

24.99a

GR8.1

18.07b

17.31b

18.86be

23.78ab

22.87abc

24.72a

GR8.2

17.87b

17.00b

18.80be

24.37be

23.49cd

25.28ab

GR8.3

17.68b

16.69b

18.73be

24.97cd

24.12d

25.83b

GR8.7

17.69b

16.84b

18.56be

24.47be

23.60cd

25.45ab

GR8.9

18.10b

17.07b

19.21c

26.33e

25.29e

27.44cd

GR8.10

17.93b

16.85b

19.09be

26.26e

25.23e

27.34cd

GR8.11

17.82b

16.71b

19.04be

25.27d

23.93d

26.72c

GR8.12

18.05b

16.80b

19.38c

26.57e

25.25e

27.91d

GR8.13

17.50b

16.17b

18.92be

26.16e

24.88e

27.52cd

GR8.14

15.61a

14.40a

17.29ab

23.94ab

22.50a

25.87b

GR8.15

15.25a

14.45a

16.33a

23.52a

22.72ab

24.78a

Table 3. Continued.

FallMeanTukey HSDDaily MinimumTukey HSDDaily MaximumTukey HSD

WinterMeanTukey HSDDaily MinimumTukey HSDDaily MaximumTukey HSD

GR8.0D

12.48a

12.04a

13.00ab

8.06c

7.68b

8.48b

GR8.1

12.47a

11.74a

13.35abc

8.01c

7.56be

8.49b

GR8.2

12.34a

11.69a

13.07abc

7.82be

7.29abc

8.40b

GR8.3

12.21a

11.64a

12.79a

7.63abc

7.02abc

8.31b

GR8.7

13.49ab

12.93ab

13.97abc

7.36abc

6.86abc

7.90be

GR8.9

13.03ab

12.28a

13.77abc

7.29abc

6.70a

7.89be

GR8.10

13.29ab

12.72ab

13.91abc

7.22abc

6.65a

7.80be

GR8.11

13.18ab

12.37ab

14.09abc

7.31abc

6.77ab

7.87be

GR8.12

13.25ab

12.45ab

14.13abc

7.18abc

6.59a

7.77be

GR8.13

13.65ab

12.93ab

14.39abc

7.06ab

6.43a

7.77be

GR8.14

14.97ab

14.26ab

15.74be

6.82a

6.45a

7.35a

GR8.15

15.39b

15.06b

15.90c

6.82a

6.63a

7.10a

O\

27

Table 4. Diversity indicies describing fish sampling data.

Site

GR8.0DGR8.1GR8.2GR8.3GR8.7GR8.9GR8.10GR8.11GR8.12GR8.13GR8.14GR8.15

Richness

182222241924222723221314

Evenness

0.9590.9380.9580.9630.9470.9510.9540.9590.9560.9570.9620.951

Shannon-WeaverDiversity ('H)

2.7732.9

2.963.0622.7893.0212.9473.1622.9992.9582.4692.509

Table 5. Abundances of individual fish species collected in 2007 from the twelve sampling sites along the Upper Green River.

Species

Ambloplites rupestrisCampostoma oligolepisCarassius auratusCottus carolinaeCyprinella spilopteraCyprinella venustaCyprinella whippleiErimystax dissimilisEtheostoma heliumEtheostoma blenniodesEtheostoma caeruleumEtheostoma flabellareEtheostoma maculatumEtheostoma raflnesqueiEtheostoma squamicepsEtheostoma stigmaeumEtheostoma zonaleFundulus catenatusFundulus notatusGambusia affinisHypentelium nigricansLabedesthes sicculusLepomis cyanellusLepomis macrochirus

Abbrev

AmbrupeCam_oligCar_auraCot_caroCypspilCyp_venuCyp_whipEri dissEth_bellEth_blenEth caerEth_flabEth macuEth_rafiEth_squaEthstigEth_zonaFun_cateFun notaGamaffiHyp_nigrLab_siccLepcyanLep macr

GR8.0D

060

17000333100200

2423

100000

GR8.1

543

1103

001414003102

17000

11027

GR8.2

240

17000422402102

331000600

GR8.3

4003007017908

3801

5568

371

2940

GR8.7

110

100000

1222

105201

560001010

GR8.9

06900000

1679101000

17100

208602

GR8.10

25400

13200

1412110015

2327

076000

GR8.11

257031001

13662070

1847

1082600

GR8.12

11009701544311100

211003600

GR8.13

8800001279320008

447001

1201

GR8.14

7300002003200600

100001000

GR8.15

210000000

1910070540200011

oo

Table 5. Continued

Species

Lepomis megalotisLepomis microlophusLuxilus chrysocephahisLythrurus faciolarisLythrurus fumeusMicropterus dolomieuMicropterus punctulatusMicropterus salmiodesMoxostoma sp.Notropis atherinoidesNotropis hoopsNotropis photogenisNotropis rubellusNotropis telescopusNotropis volucellusNoturus elegansNoturus eleutherusPercina caprodesPercina evidesPhenocobius mirabilisPhenocobius uranopsPimephales notatusPomoxis annularisPylodictus olivarus

Abbrev

Lep_megaLepmicrLux_chryLyt_faciLyt_fumeMicdoloMicjpuncMic_salmMox_sp.Not_atheNot_boopNot_photNot_rubeNotjeleNotjvoluNtu_elegNtu_eleuPer_caprPer evidPhe_miraPheuranPim_notaPomannuPyloliv

GR8.0D

80500210

321500000000000900

GR8.1

180012530000

42400000110100

GR8.2

1300101500

1704000000401100

GR8.3

2505000809000009000100

1001

GR8.7

500000009607006020100

2500

GR8.9

2500001318

380

19000001152

9000

GR8.10

350000020069

10000000101

5801

GR8.11

13000020009160

144100041

1700

GR8.12

90000000000

22004110002

2200

GR8.13

4000000200703067001000

8500

GR8.14

501200000000000000000

1510

GR8.15

1310000010000000000000420

too

30

Table 6. Scores of PCA axes 1,2 and 3 of environmental and biologicaldata sets for sites sampled along the Upper Green River

Site

GR 8.0DGR8.1GR8.2GR8.3GR8.7GR8.9GR8.10GR8.11GR8.12GR8.13GR8.14GR8.15

% var.explained

EnvironmentalAxis 1

3.2603.0803.350

-1.0800.530

-3.190-2.420-0.810-2.110-1.5600.3500.610

32.8

Axis 2

0.170-1.440-0.8505.000

-0.270-0.450-1.660-0.600-1.700-0.8100.5002.100

21.0

Axis 3

-2.250-0.220-0.170-1.9200.6000.280

-0.3301.280

-1.380-1.2202.7902.530

16.4

Axis 1

1.3800.4000.3002.5200.670

-4.450-3.430-4.060-1.530-2.1805.2805.100

20.9

BiologicalAxis 2

0.4003.6800.370

-7.8000.5701.920

-0.160-1.1700.310

-1.3201.3902.180

15.6

Axis 3

-0.800-4.860-2.220-2.6700.880

-2.4301.7602.4701.4402.1901.9902.260

12.8

31

Table 7. Loadings for PCA axes 1, 2 and3 of environmental data set for sitessampled along the Upper Green River.

WidthDepthVelocDisch%Riff%Run%Pool%Fine%Gr%Cb%SmBMeanStDevCV

Axis 1

0.7760.622

-0.1480.673

-0.1950.584

-0.7380.3880.198

-0.164-0.346-0.621-0.828-0.895

Axis 2

-0.1850.155

-0.568-0.3510.390

-0.404-0.207-0.523-0.4390.7440.629

-0.648-0.416-0.268

Axis 3

0.3250.4840.0000.5490.447

-0.446-0.1600.379

-0.819-0.2840.3450.3300.2620.204

32

Table 8. Loadings for PCA axes 1, 2 and 3 of biological data set for sites sampledalong the Upper Green River.

Amb_rupeCam_oligCot_caroCyp_spilCyp_whipEri_dissEth_bellEth_blenEth caerEth_flabEth_macuEth_rafiEth_stigEthe zonFun cateFun_notaGam_affiHyp_nigrLab_siccLep_cyanLep macr

Axis 1

-0.3600.880

-0.1080.473

-0.5630.5200.8330.114

-0.1440.601

-0.323-0.7370.2760.2560.398

-0.6660.1990.5650.122

-0.6000.126

Axis 2

-0.0830.2410.276

-0.169-0.4770.303

-0.348-0.201-0.723-0.341-0.282-0.389-0.200-0.673-0.502-0.551-0.6650.115

-0.673-0.2690.479

Axis 3

-0.2610.1250.525

-0.3100.1520.552

-0.114-0.324-0.190-0.5520.547

-0.103-0.4220.1460.2970.2010.4010.3370.1750.3530.457

Lep_megaLux_chryLyt_faciMic_doloMic_puncMic_salmMox_sp.Not_atheNot boopNot_elegNot_eleuNot_photNot_teleNotvoluPer_caprPer evidPhemiraPhe_uranPirn notaPomannuPyloliv

Axis 1

0.397-0.559-0.3780.212

-0.0320.0660.0330.6470.4620.3600.0470.7020.413

-0.0090.4930.0380.5810.6990.582

-0.548-0.101

Axis 2

-0.432-0.4440.5430.401

-0.3830.278

-0.364-0.075-0.236-0.0960.0160.264

-0.282-0.703-0.105-0.0210.090

-0.003-0.4100.408

-0.700

Axis 3

0.2790.3070.1330.5430.7940.0920.4940.253

-0.202-0.422-0.2640.258

-0.438-0.2830.1600.6220.2600.081

-0.277-0.4590.163

m Relative Abundance Relative Abundance Relative Abundance^3* C *t 0 s ! •""• ^3" f^l C**) r~^ ^f C^i C^~i r~~i

o o o o o o o o o o o o o o o

ai'D

rn

ao00Pi

oo

o06

o

oo

Pi

o

ooOiO

LT;

00

O

ooa:a

CN

00

O

om

oo

.Qi

oo

0CN -ig

0

0

0

en

3O

c-OC3X)03

oCN , as

—

Lynn Camp Creek B l * ? I I K h C r " k Big Pitman deek

8.2 ^ J H.9

8.01)

Cave

National Park

Legeuil

Herbivore/Detntivore

Omnivore

| Generalist Insectivore

9 SurfaceMfcter Column Insectivore

U BenthicInsectivore

| Rsovore

| Parasite

Green River and Tnbutanes

Mammoth Cave National Park

Green River Lake

Upper Green River Subbasm0 45 9

IWIometers18

Figure 5. Map depicting pie charts showing percent composition of trophic guilds communities at each site.

-u

Lynn Camp Ci eek BigBrush C reek B j g Famm Cnfk

8.7 8-9

801) GieeuRnei

Muiiiinntli CNational Pmk

iver8.12

Cobbled'avdSandSilr

| Substrate Genei'alistGieen River and TiibntaiiesISLuoinoth Cave National ParkGreen River L aleeUpper Green Ri\rei' Sub basin

Figure 6. Map depicting pie charts showing percent composition of substrate preference guilds within fishcommunities at each site.

Mammoth CaNational Piu k

Lyim Csunp Cr*dt B i r B l n s h C r ' f k B i 8 Pitman Creek

831 «T 8."

Legend

Velocity Guild

| Fast FlowModerate Flow

QB Slow/No FlowQFlow Generallst

— Green River and TributariesMammoth Cave National Park

g Green River Lake_ Upper Green River Subbasin 4.5 9

I Kilometers18

Figure 7. Map depicting pie charts showing percent composition of velocity preference guilds within fishcommunities at each site.

Mammoth CaveNational Park

Lynn Camp Creek f Big Pitman Creek

8-3A • J ' 8.7

8.0Di

Legend

Reproductive Guild

Open Water/Substrate Spawnerg Brood Hider3 Substrate Chooser~] Nest Spawner

0 Internal Bearer• Green River and TributariesMammoth Cave National Park

| Green River LakeUpper Green River Subbasin

I Kilometers18

Figure 8. Map depicting pie charts showing percent composition of reproductive guilds within fishcommunities at each site. U)

38

Cam olig

A GR8.9s^5:Eri_diss

Hyp_nigrl^^S.

A GR8.1

Lep macri

Mic d

\Phe m i r a \ ^ \ ^ v \

Phe uran—^^/^*GR8.12Notjrthe""^55^

A G R 8 _ 1 0 _ _ —Eth belT" ^^**

Eth flab—-^""^A-, ~, *GR8OR8.11 A /

Fun cate J//

Ethe_zon yLabsicc

^ ^ ' i ^

< <

13/0/

/ /

'/II 1im afT6

/Not_

V

\

\\lAWmm7/'1volu

Eth

Mic salm

/ " //

^GR8.7^^^**^ A

Lyt_faci^ Pom

^ ^

gR8.2 A GR80D

VV^Arat

Eth macuS

' \ \Cypwhip

caer

rupe

Lep cyan\ ~

^Eth ran\

Fun_nota

*GR8.15A

annu ^GR8.14

Axis 1

Figure 9. Plot of PCA axes 1 and 2 offish data.

4.3E-02

100

3.3E-01

75

Distance (Objective Function)

6.1E-01 9E-01

Information Remaining (%)

50

1.2E+00I

GP 8 ODGP 8 3 1GR8."7

GR8.11 ,GR8.13 1

OK&.li 'GR8.9 ,GR8.10 1GR8.1 ,GP 8 "• 'GR8.14

E. bellum

Group

1 •>

E. zonale

. P. annularisGR8.15 1

Figure 10. Dendrogram depicting results of indicator species analysis of the 2007 fish data.

Literature cited

Allan, J. David. 1995. Stream Ecology: Structure and function of running waters.Chapman & Hall. New York, NY. pg. 310.

American Public Health Association (APHA). 1998. Standard Methods for theExamination of Water and Wastewater. APHA. American Water WorksAssociation and Water Environment Federation. 20th edition. Washington D.C.

Bain, Mark B., John T. Finn, and Henry E. Booke. 1988. Streamflow regulation and fishcommunity structure. Ecology. 69: 382-392.

Baxter, R. M. 1977. Environmental effects of dams and impoundments. Annual Review ofEcology and Systematics 8: 255-283.

Bunn, Stuart E., and Angela H. Arthington. 2002. Basic principles and ecologicalconsequences of altered flow regimes for aquatic biodiversity. EnvironmentalManagement. 30: 492-507.

Claramunt, Randall M. and David H. Wahl. 2000. The Effects of Abiotic and BioticFactors in Determining Larval Fish Growth Rates: A Comparison Across Speciesand Reservoirs. Transactions of the American Fisheries Society. 129: 835-851.

Eberle, Mark E., Eric G. Hargett, Thomas L. Wenke, and Nicholas E. Mandrak. 2002.Changes in fish assemblages, Solomon River Basin, Kansas: habitat alterations,extirpations and introductions. Transactions of the Kansas Academy of Science.105: 178-192.

Edwards, Robert J. 1978. The Effect of Hypolimnion Reservoir Releases on FishDistribution and Species Diversity. Transactions of the American FisheriesSociety. 107: 71-77.

E.P.A. 1998. Guidance for Data Quality Assessment. EPA/600/R-96/084. U.S.Environmental Protection Agency; Office of Research and Development;Washington, DC.

Etnier, David A. & Wayne C. Starnes. 1993. The Fishes of Tennessee. The University ofTennessee Press/ Knoxville.

Fisher, William L. 1990. Life History and Ecology of the Orangefin Darter Etheostomahelium (Pisces: Percidae). American Midlands Naturalist. 123: 268-281.

Fowler, J.F. and C.A. Taber. 1983. Fowler, J.F. and C.A. Taber, 1985. Food habits andfeeding periodicity in two sympatric stonerollers (Cyprinidae). American MidlandNaturalist 113:217-224.

Froese, R. and D. Pauly (Eds). [Internet], [updated 2008 Feb]. Fishbase. World FishCenter, [accessed 2008 Feb]. Available at: http://fishbase.org

Franssen, N. R., K. B. Gido, C. S. Guy, J. A. Tripe, S. J. Shrank, T. R. Strakosh, K. N.Bertrand, C. M. Franssen, K. L. Pitts, and C. P. Paukert. 2006. Effects of floodson fish assemblages in an intermittent prairie stream. Freshwater Biology 51:2072-2086.

40

41

Freeman, M. C , and P. A. Marcinek. 2006. Fish assemblage responses to waterwithdrawals and water supply reservoirs in piedmont streams. EnvironmentalManagement 38: 435-450.

Grubbs, S. A., O. D. Meier, and A. J. Meier. 2007. Longitudinal patterns offishassemblage in small unregulated subbasins: evaluating reach- and watershed-scaleparameters. Hydrobiologia 592: 211-223.

Helsel, D.R. 1978. Less than obvious - statistical treatment of data below the detectionlimit. Environmental Science and Technology 24:1766-1774.

Houde, E. D. 1974. Effects of temperature and delayed feeding on growth and survival oflarvae of three species of subtropical marine fishes. Marine Biology. 26: 271-285.

Jaske, R. T., and J. B. Goebel. 1967. Effects of dam construction on temperatures of theColumbia River. Journal of the American Water Works Association 59: 935-942.

Kessler, Richard K., Andrew F. Casper, and Gordon K. Weddle. 1995. Temporalvariation in microhabitat use and spatial relations in the benthic fish communityof a stream. American Midlands Naturalist. 134: 361-370.

Kilar, J. A. and J. McLachlan. 1989. Effects of wave exposure on community structure ofa plant-dominated fringing-reef platform: Intermediate disturbance anddisturbance mediated competition. Marine Ecology Program Series 54: 265-276

Klimogianni, A., G. Koumoundouros, P. Kaspiris and M. Kentouri. 2004. Effect oftemperature on the egg and yolk-sac larval development of common pandora,Pagellus erythrinus. Marine Biology. 145: 1015-1022.

Koel, T. M. 2004. Spatial variation in fish species richness of the upper Mississippi riversystem. Transactions of the American Fisheries Society 133: 984-1003.

Lake, P. S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the NorthAmerican Benthological Society 19: 573-592.

Lasker, Reuben. 1964. An Experimental Study of the Effect of Temperature on theIncubation Time, Development, and Growth of Pacific Sardine Embryos andLarvae. Copeia. 1964: 399-405

Laurence, Geoffrey C. 1978. Comparative growth, respiration and delayed feedingabilities of larval cod {Gadus morhua) and Haddock {Melanogrammus aeglefinus)as influenced by temperature during laboratory studies. Marine Biology. 50: 1-7.

Laurence, Geoffrey C. and Carolyn A. Rogers. 1976. Effects of temperature and salinityon comparative embryo development and mortality of Atlantic cod {Gadusmorhua L.) and haddock {Melanogrammus aeglefinus (L.)). Journal du Conseil.36: 220-228.

Lehmkuhl, D. M. 1972. Changes in thermal regime as a cause of reduction of benthicfauna downstream of a reservoir. Journal of the Fisheries Research Board ofCanada. 29,1329-32.

Martinez, Patrick J., Thomas E. Chart, Melissa A. Trammell, John G. Wullschleger, andEric P. Bergersen. 1994. Fish species composition before and after construction of

42

a main stem reservoir on the White river, Colorado. Environmental Biology ofFishes. 40: 227-239.

McGurk, M. D. 1984. Effects of delayed feeding and temperature on the age ofirreversible starvation and on the rates of growth and mortality of Pacific herringlarvae. Marine Biology. 84: 13-26.

Morehead, D. T. and P. R. Hart. 2003. Effect of temperature on hatching success and sizeof striped trumpeter (Laths lineata) larvae. Aquaculture. 220: 595-606.

NatureServe. [updated 2008 Feb. 1]. NatureServe Explorer: An online encyclopedia oflife [web application]. Version 7.0. NatureServe, Arlington, Virginia. Availablehttp://www.natureserve.org/explorer. [Accessed: 2008 Feb.].

O'Driscoll, Michael A. and David R. DeWalle. 2006. Stream-air temperature relations toclassify stream-ground water interactions in a karst setting, central Pennsylvania,USA. Journal of Hydrology. 329: 140-153.

Osmundson, D. B., R. J. Ryel, V. L. Lamarra, and J. Pitlick. 2002. Flow-sediment-biotarelations: implications for river regulation effects on native fish abundance.Ecological Applications 12: 1719-1739.

Page, Lawrence M. 1983. Handbook of Darters. T.F.H. Publications Inc. Neptune City,New Jersey, pg. 169.

Page, Lawrence M., and B. M. Burr. 1992. A field guide to freshwater fishes: NorthAmerica north of Mexico. Houghton Mifflin Company, Boston, Massachusetts,pg. 432

Paller, M. H., and B. M. Saul. 1996. Effects of temperature gradients resulting fromreservoir discharge on Dorosoma cepedianum spawning in the Savannah River.Environmental Biology of Fishes 45: 151-160.

Pflieger, William L. 1997. The Fishes of Missouri. Missouri Department of Conservation.Jefferson City, MO.

Poff, N. LeRoy and J. David Allan. 1995. Functional organization of stream fishassemblages in relation to hydrological variability. Ecology. 76: 606-627.

Poff, N. LeRoy and J. David Allan. 1997. The natural flow regime. Bioscience. 47: 769.

Powers, M. E., R. J. Stout, C. E. Cushing, P. P. Harper, F. R. Hauer, W. J. Matthews, P.B. Moyle, B. Statzner, and I. R. Wais De Badgen. 1988. Biotic and abioticcontrols in river and stream communities. Journal of the North AmericanBenthological Society 7: 456-479.

Quinn, Jeffrey W. and Thomas J. Kwak. 2003. Fish Assemblage Changes in an OzarkRiver after Impoundment: A Long-Term Perspective. Transactions of theAmerican Fisheries Society. 132: 110-119.

Quist, Micheal C , Wayne A. Hubert, and Frank J. Rahel. 2005. Fish assemblagestructure following impoundment of a Great Plains river, Western NorthAmerican Naturalist. 65: 53-63.

43

Reynolds, C.S., J. Padisak and U. Sommer. 1993. Intermediate disturbance in the ecologyof phytoplankton and the maintenance of species diversity. Hydrobiologia 249:183-188.

Richter, Brian D., Jeffery V. Baumgartner, Jennifer Powell, and David P. Braun. Amethod for assessing hydrologic alteration within ecosystems. ConservationBiology. 10:1163-1174.

Rosemond, Amy D. 1994. Multiple Factors Limit Seasonal Variation in Periphyton in aForest Stream. Journal of the North American Benthological Society. 13: 333-344.

Richter B. D. and H. E. Richter. 2000. Prescribing flood regimes to sustain ecosystemsalong meandering rivers. Conservation Biology 14: 1467-1478.

Schlosser, I. J. 1982. Fish community structure and function along two habitat gradientsin a headwater stream. Ecological Monographs 52: 395-414.

Taylor, Christopher A., Jason H. Knouft, and Tim M. Hiland. 2001. Consequences ofstream impoundment in a small North American drainage. Regulated Rivers:Research and Management. 17: 687-698.

Tiemann, Jeremy S., and David P. Gillette. Effects of lowhead dams on riffle-dwellingfishes and macroinvertebrates in a Midwestern river. Transactions of theAmerican Fisheries Society. 133: 7.5-717.

Townsend, C.R. and A.G. Hildrew. 1993. Species traits in relation to a habitat templatefor river systems. Freshwater Biology 31: 265-275.

Vinson, Mark R. 2001. Long-term dynamics of an invertebrate assemblage downstreamfrom a large dam. Ecological Applications. 11: 711-730.

White, P. S. & S. T. A. Pickett, 1985. Natural disturbance and patch dynamics: Anintroduction. Pages 3-9 In S. T. A. Pickett and P. S. White (editors). The Ecologyof Natural Disturbance and Patch Dynamics. Academic Press, New York.