Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations

2012 Status, distribution, and habitat associations of Topeka Shiners in west-central Iowa Bryan David Bakevich Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/etd Part of the Aquaculture and Fisheries Commons

Recommended Citation Bakevich, Bryan David, "Status, distribution, and habitat associations of Topeka Shiners in west-central Iowa" (2012). Graduate Theses and Dissertations. 12269. https://lib.dr.iastate.edu/etd/12269

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected].

Status, distribution, and habitat associations of Topeka shiners in west-central Iowa

by

Bryan David Bakevich

A thesis submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Major: Fisheries Biology

Program of Study Committee: Clay L. Pierce, Co-Major Professor Michael C. Quist, Co-Major Professor Philip Dixon

Iowa State University

Ames, Iowa

2012

Copyright © Bryan David Bakevich, 2012. All rights reserved. ii

TABLE OF CONTENTS

LIST OF TABLES iii

LIST OF FIGURES iv

ACKNOWLEDGEMENTS vi

CHAPTER 1. GENERAL INTRODUCTION 1

CHAPTER 2. STATUS, DISTRIBUTION, AND HABITAT ASSOCIATIONS OF

TOPEKA SHINERS IN WEST-CENTRAL IOWA 4

CHAPTER 3. GENERAL CONCLUSIONS 51

APPENDIX A. TOTAL NUMBER OF TOPEKA SHINER DETECTED AT SITES IN

WEST-CENTRAL IOWA IN 2010 TO 2011 54

APPENDIX B. PERCENT OCCURRENCE AND CATCH PER UNIT EFFORT (CPUE)

OF FISH SPECIES COLLECTED IN STREAM AND OFF-CHANNEL SITES 57

iii

LIST OF TABLES

CHAPTER 2

Table 1. Historic (IAGFA 2005), Menzel and Clark (2002) study (1997-2000), and current

(2010-2011) Topeka shiner detections in ten-digit (HUC10) and eight-digit (HUC8)

hydrological units. Topeka shiner status for each HUC10 was determined to be stable (detected

during 1997 to 2000 and during 2010 to 2011), at risk (detected during 1997 to 2000 and not

detected during 2010 to 2011), or possible extirpated (not detected during 1997 to 2000 or during

2010 to 2011). Percent decline is the proportion of the number HUC10s where Topeka shiners

were not detected to the total number of HUC10s in their historic range. 41

Table 2. Habitat and biotic variables measured at stream sites (n = 67), off-channel sites

(n = 27) in west-central Iowa. Means and standard deviations (SD) were calculated for

variables in stream and off-channel sites separately and combined (n=94). 42

Table 3. Confidence models selected (ΔAICc less than 2) from the combined, stream and off-

channel candidate set of a priori logistic regression models as determined by Akaike’s

information criterion for small sample size (AICc) ranking. Also included are the number

of parameters in each model (k) and the Akaike’s weight (w). 43

Table 4. Model averaged coefficient estimates, standard error, 95% confidence intervals,

and relative weights for the combined, instream, and off-channel models. 44

iv

LIST OF FIGURES

CHAPTER 2

Figure 1. Location of study HUC8 basins within the Des Moines Lobe subecoregion

of Iowa. 45

Figure 2. Frequency of the stream and off-channel sites where Topeka shiner were

and were not detected. Numbers of sites where Topeka shiners were and were not

detected are shown above the bars. 46

Figure 3. Topeka shiner occurrence documented in 1997 to 2000 by Menzel and

Clark (2002) in (from left to right) the North Raccoon, middle Des Moines, Boone,

and upper Iowa river basins in central Iowa (A). Topeka shiner occurrence documented

in 2010 to 2011 (B). 47

Figure 4. NMDS ordination of fish assemblages in stream (circles) and off-channel

sites (triangles) combined. Grey symbols represent sites where Topeka shiner were

not detected and black sites represent those where Topeka shiner were present.

Isobars represent the differing levels of species richness among all sites. 48

Figure 5. NMDS ordination of fish assemblages in stream sites. Grey symbols represent

sites where Topeka shiner were not detected and black sites represent those where

Topeka shiner were present. 49

v

Figure 6. NMDS ordination of fish assemblages in off-channel sites. Grey symbols

represent sites where Topeka shiner were not detected and black sites represent those

where Topeka shiner were present. 50

vi

ACKNOWLEDGEMENTS

First, I would like to thank my major professors Dr. Clay Pierce and Dr. Michael

Quist for their guidance and instruction. Their expertise in scientific research, writing, and communication was invaluable to this project and to my growth as a student and fisheries professional. Much thanks also to Dr. Philip Dixon, who served on my committee and provided valuable insight. This project would not have been possible without the faculty, staff, and students of Iowa State University, the Department of Natural Resources, and the

Iowa Cooperative Fish and Wildlife Research Unit. I would like to thank fellow graduate

students Michael Colvin, Jesse Fischer, Timothy Parks, Tony Sindt, for their contributions to

this project and my overall experience in Iowa. Without them, I would have never realized

the awesomeness of Program R, shot my first goose, played drums in a rock band, or

discovered the joy of sitting silently in a tree stand, respectively. I would next like to thank my family for their unwavering support and encouragement. I would especially like to thank my wife, Jennifer, for leaving mountains of Montana to follow her future husband to the

flatlands of Iowa. I would never have made it through this experience without her love,

understanding, and friendship.

Lastly, I would like to dedicate this thesis to the memory of my grandfather, George

Bakevich Sr., who took me fishing before I learned how to walk. Because of him, I grew up to appreciate the beauty, complexity, and value of watery places.

1

CHAPTER 1. GENERAL INTRODUCTION

When European settlers began colonizing North America nearly four hundred years

ago, they were met with an abundance of natural resources that were used to build houses

and cities, grow crops, and provide food and clothing to a growing population. In the

Midwest, the most valued resource lay beneath the great expanse of native grasslands: a deep

layer of fertile soil. In Iowa, approximately 78% of land once dominated by tallgrass prairie,

wetlands, and forested areas was rapidly transformed into cropland and pasture for livestock

grazing (Gallant et al. 2011). Taking advantage of the abundant and rich soils in Iowa has

been crucial to the prosperity of the United States, but it has had some undesirable

consequences. Such costs of agriculture are readily apparent in aquatic systems throughout

the state. Practices associated with row crop agriculture and livestock grazing have

negatively impacted water quality, riparian and instream habitat conditions, and altered flow

regimes in Iowa (Bulkley 1975; Menzel 1983; Skaggs 1994). In turn, physical changes in

aquatic systems can impact the many species that live in or near streams, rivers, and lakes.

However, not all species respond to changes in their environment in a similar way. A change that has little effect on one species may cause the decline of another. Thus, the fundamental biology and life history of imperiled species must be understood if the goal is to prevent their continued decline.

This research focused on Topeka shiner, a federally endangered prairie stream fish that has declined throughout its historic range in the Midwest. In Iowa, historic records indicate Topeka shiners were once widely distributed, but are now restricted to only a few streams in central and northwest part of the state (IAGFA 2005). One of our goals was to 2

determine the current distribution of Topeka shiners in Iowa and if that distribution has changed since they were last investigated (Clark 2000). A second goal was to identify abiotic and biotic factors associated with Topeka shiner occurrence. Both of the objectives of this study will provide valuable information to assist managers with the recovery and conservation of this imperiled fish.

Thesis Organization

This thesis contains two additional chapters. The second chapter is a manuscript that will be submitted for publication in the North American Journal of Fisheries Management. This

manuscript provides an abstract, introduction, methods, results, discussion,

acknowledgement, and references section. All tables and figures are included at the end of

the text. The third chapter provides a general conclusion and synthesis. Appendices provide

further information and analyses.

References

Bulkley, R. V. 1975. Inventory of major stream alterations in Iowa. Completion report: a

study of the effects of stream channelization and bank stabilization on warm water

sport fish in Iowa. U.S. Fish and Wildlife Service, Subproject No. 1. Contract No. 14-

16-008-745, Ames, Iowa.

Clark, S. J. 2000. Relationship of Topeka shiner distribution to geographic features of the

Des Moines Lobe in Iowa. M.S. Thesis, Iowa State University, Ames, Iowa. 3

Gallant, A. L., W. Sadinski, M. F. Roth, and C. A. Rewa. 2011. Changes in historical Iowa

land cover as context for assessing the environmental benefits of current and future

conservation efforts on agricultural lands. Journal of Soil and Water Conservation

66:67-77.

IAGFA (Iowa Aquatic Gap Fish Atlas). 2005. Iowa Aquatic Gap Fish Atlas, Iowa Rivers

Information System. (http://maps.gis.iastate.edu/iris/fishatlas/), accessed February

2010.

Menzel, B. W. 1983. Agricultural management practices and the integrity of in-stream

biological habitat. Pages 305-329 in F. W. Schaller and G. W. Bailey, editors.

Agricultural management and water quality. Iowa State University Press, Ames.

Skaggs, R. W., M. A. Breve, and J. W. Gilliam. 1994. Hydrologic and water quality impacts

of agricultural drainage. Critical Reviews in Environmental Science and Technology

24:1-32. 4

CHAPTER 2. STATUS, DISTRIBUTION, AND HABITAT ASSOCIATIONS OF

TOPEKA SHINERS IN WEST-CENTRAL IOWA

A manuscript to be submitted for publication in the North American Journal of Fisheries

Management.

Bryan D. Bakevich1,2, Clay L. Pierce3, Michael C. Quist4

Abstract

The distribution of Topeka shiner Notropis topeka has declined across its historic

range and was listed as endangered under the Endangered Species Act in 1998. In Iowa, the

habitat associations of this imperiled fish are not well understood. Our goals were to

understand the current distribution and identify abiotic and biotic factors associated with the

occurrence of Topeka shiners in stream and off-channel habitats of west-central Iowa. Fish

assemblages and habitat characteristics were sampled in 67 stream and 27 off-channel sites

during 2010 – 2011. Topeka shiners were found in 52% off-channel sites, but only 9% of

stream sites, supporting the hypothesis that off-channel habitats are an important component

of their life history. When compared to prior distributions, our results indicated a recent reduction in the distribution of Topeka shiners in Iowa. Fish assemblages in stream sites

1Corresponding author: [email protected] 2Department of Natural Resource Ecology and Management, Iowa State University 3U.S. Geological Survey, Iowa Cooperative Fish and Wildlife Research Unit, Department of Natural Resource Ecology and Management, Iowa State University 4U.S. Geological Survey, Idaho Cooperative Fish and Wildlife Research Unit, Department of Fish and Wildlife Resources, University of Idaho 5

differed significantly from off-channel sites and had higher species richness. Fish assemblages containing Topeka shiner were different from those that did not contain Topeka shiner in off-channel sites, but not in stream sites. Results from logistic models suggested that Topeka shiner presence was associated with increased submerged vegetation and abundance of fathead minnow Pimephales promelas. Contrary to the findings of other studies, the abundance of large piscivorous fishes was not associated with the occurrence of

Topeka shiner. Our results provide new information about the biology and life history of

Topeka shiners in west-central Iowa that will guide restoration and other recovery efforts.

Introduction

Biodiversity in freshwater systems is in decline throughout the world (Dudgeon et al.

2006). Human-induced changes in biotic (e.g., species invasions, overgrazing) and abiotic conditions (e.g., flow modification, water pollution) are the primary causes of species declines in lotic systems (Allan and Flecker 1993). Declining populations of fishes in North

America have led to the listing of 149 fishes as threatened or endangered under the

Endangered Species Act (USFWS 2012). Most threatened and endangered fish are non- game species that, prior to federal listing, received little attention. As such, our understanding of the biology and ecology of many imperiled fishes is insufficient.

Information on the habitat requirements, symbioses, and physiological tolerances of these imperiled species is needed to develop science-based recovery plans that will aid in their conservation and guide restoration. 6

The landscape in the United States experienced drastic change since the arrival of

European settlers, but few areas were impacted so strongly by agriculture than the land in

Iowa. In the mid 1800s, the landscape of Iowa was primarily covered with tallgrass prairie, forests, and wetlands. By 2001, approximately 78% of Iowa’s land had been converted to cropland or pasture (Gallant et al. 2011). Agricultural practices have changed the physical, chemical, and hydrological characteristics of streams in Iowa (Bishop 1981; Menzel et al.

1984; Shilling and Hemlers 2008). Such changes in stream habitats can have a negative influence on fish and other aquatic species (Roth et al. 1996; Wang et al. 1997; Poff and

Zimmerman 2010; Rowe et al. 2009). In Iowa, 68 native fish species are listed as species of

greatest conservation need due to the loss or degradation of aquatic habitats (Zohrer 2005), and a recent study suggests that some of those species are declining (Sindt et al. 2012).

The Topeka shiner Notropis topeka is a small minnow native to streams of Iowa,

Kansas, Minnesota, Missouri, Nebraska, and South Dakota (Lee et al. 1980). When surveys indicated an approximately 80% reduction of its historic distribution, it was listed as endangered under the Endangered Species Act in 1998 (Tabor 1998). Decline of Topeka shiners in Iowa has been attributed to hydrologic changes, agricultural impacts on water quality, and increased predation (Tabor 1998), but the specific factors associated with

Topeka shiner occurrence are poorly understood. One recent study investigated how landscape-scale factors (e.g., land cover type, stream slope) affected Topeka shiner occurrence (Menzel and Clark 2002), yet the habitat associations of many fishes can occur at a finer scale (Pont et al. 2005). If fact, Wall and Berry (2006) found that Topeka shiners occurrence in South Dakota was associated with factors at multiple scales. Investigating 7

reach-scale habitat associations will provide novel information about Topeka shiner in Iowa,

but further study of landscape-scale factors may also be beneficial.

The importance of understanding factors associated with Topeka shiner ecology has

been highlighted as scientists have made major discoveries regarding their habitat use.

Recently, Topeka shiners have been documented in off-channel habitats such as oxbows and livestock watering ponds (Menzel and Clark 2002; Thomson and Berry 2009); however, the

role these habitats play in the life history of Topeka shiners is unknown. If off-channel habitats represent a significant portion of their total habitat use, efforts to recover Topeka shiner may need to be directed toward these habitats. Traditionally, stream restoration has focused on improving habitat within the stream channel and reducing nutrient or other inputs from the landscape. These actions can be beneficial to a suite of native fishes, but may not improve habitat for Topeka shiners if they are primarily using off-channel habitats.

Recognizing this, the U.S. Fish and Wildlife Service (USFWS) has restored over 25 off- channel habitats in Iowa in hopes of creating suitable habitat for Topeka shiners (USFWS

2009). During restoration, accumulated sediment is removed to increase depth and reconnect the off-channel habitat to groundwater sources. Connections to the stream are also dug from the off-channel habitat to facilitate fish movement between the two habitats. Since the relationship that Topeka shiners have with off-channel habitats in not well understood, there is little information to guide further restoration efforts.

Recovery efforts, such as critical habitat designation and habitat restoration, can only be effective if the distribution and habitat associations of Topeka shiners in west-central Iowa are understood. The goal of our study was to determine factors associated with the occurrence of Topeka shiners in west-central Iowa. Since Topeka shiners inhabit streams 8

that are not often sampled, this research will also provide information about the current

distribution of Topeka shiners in west-central Iowa.

Methods

Study area

The study area was confined to the North Raccoon, Boone, upper Des Moines, and

upper Iowa river basins located in the Des Moines Lobe (Griffith et al. 1994) landform of central Iowa (Figure 1). This landscape is characterized by gently rolling terrain and is

dominated by row crop agriculture. Although Topeka shiners have historically occurred in

all of these watersheds (IAGFA 2005), recent surveys indicate that significant populations

only remain in the North Raccoon and Boone river watersheds (Clark 2000). These two

watersheds contain the only known populations of Topeka shiner in Iowa that are within the

Mississippi River catchment. Populations in west-central Iowa were chosen because they

may have a unique evolutionary history compared to populations that exist in different

landforms, climates, hydrological regimes, and across a wide spectrum of biotic

communities.

Study sites

Based on our current knowledge of Topeka shiner habitat use, we chose to sample

stream and off-channel sites. Stream sites were typical of those on the Des Moines Lobe

with low gradients and riparian areas of grasses, row crops, or pasture. Many streams were

channelized and had low habitat complexity. Off-channel sites were pond-like water bodies

within the stream floodplain that remained disconnected from the stream channel during 9

normal flow conditions. Off-channel habitats were characterized by silt substrate, aquatic macrophytes, and moderate turbidity. Several off-channel sites were used to store water for livestock resulting in trampled areas within and around the site. Many of the sites were natural oxbows, but several restored oxbows occurred in the study area. Restored oxbows had been dredged to create deeper, more permanent off-channel habitats that could frequently connect with the main stream channel. Restoration often increases groundwater inflow by removing sediment and exposing coarse substrates from the former stream bed. Since the two site types differed physically, they were sampled using slightly different protocols.

Because Topeka shiners are rare in Iowa, sample sites that had an increased likelihood of Topeka shiner occurrence were chosen for this study. We used three criteria to select sample sites. First, we selected sites where Topeka shiners were predicted to occur based on two occurrence models. One model was developed by Menzel and Clark (2002) and the other was the Iowa Aquatic GAP model (Loan-Wilsey et al. 2005). Both models used landscape-scale variables (e.g., land cover type, stream gradient) to predict Topeka shiner occurrence. Second, we selected sites where Topeka shiners have been previously documented (IAGFA 2005). Third, we selected off-channel sites that could be identified from aerial photographs taken during 2009 and 2010 since these habitats were rare throughout our study area.

Sampling

Stream sites were sampled following standard Iowa DNR protocols (IDNR 2001), but with some modifications to increase the likelihood of Topeka shiner detection. Each stream site was at least 100 meters in length and was divided into macrohabitat units defined as a 10

run, riffle, or pool (Bisson et al. 1982). The end of the reach was determined by the end of

the last macrohabitat unit that exceeded the 100 meter minimum reach length. We used two

gear types to sample fishes. First, the site was sampled by upstream single-pass pulsed-DC

electrofishing. For small streams, a battery-powered backpack LR-20 electrofishing unit

(Smith Root Inc., Vancouver, WA USA) was used. For larger streams, a generator-powered, barge mounted VVP-15B electrofishing unit (Smith-Root Inc., Vancouver, WA, USA) was used. After the site was sampled with electrofishing, it was then sampled with bag seines

(6.0 Ø 1.5m, 6-mm mesh). Since we were interested in sampling the entire fish assemblage, using both methods likely ensured a more accurate account of the species present (Onorato et al. 1998). A high level of effort also increased the likelihood of detecting Topeka shiners if they were present at a given site. All fish were identified to species, enumerated, and released. Total length (mm) of all piscivores (e.g., largemouth bass Micropterus salmoides, channel catfish Ictalurus punctatus) was recorded prior to their release.

Off-channel sites were considered to be a single macrohabitat unit and were sampled using bag seines (6.0 Ø 1.5m, 6-mm mesh). Standard sampling protocols are not available for

these habitats, but our methods were similar to those of other studies of fish in small off-

channel habitats (e.g., Thomson and Berry 2009). Because electrofishing requires sufficient

water clarity to see stunned fish, electrofishing would not have been an effective sampling

method for these silty, shallow, and often turbid habitats. The act of wading in off-channel

habitats caused more sediment to be suspended, further limiting the effectiveness of

electrofishing. Bag seines, on the other hand, were well suited for sampling off-channel

sites. Sites were often physically homogenous (e.g., free of snags, undercut banks) and shallow enough for seines to thoroughly sample the wetted area of the site. All fish were 11

identified to species and released. Total length of all piscivores was recorded prior to their

release.

For all sampling (i.e., stream and off-channel sites), catch per unit effort (CPUE) for

each species was calculated as the number of individuals per 100 m2. Composite variables were created by summing the relative abundance of two or more fishes. For example,

Sunfish CPUE was equal to the sum of CPUEs for green sunfish Lepomis cyanellus and orangespotted sunfish Lepomis humilis. Piscivore CPUE was equal to the sum of CPUEs for largemouth bass, smallmouth bass Micropterus dolomieu, northern pike Esox Lucius, channel catfish, and flathead catfish Pylodictis olivaris. Only piscivores with total lengths large enough to feed primarily on fish (Mittelbach and Persson 1998) were included in this group.

Habitat characteristics were sampled separately within each macrohabitat. Wetted width of each macrohabitat unit was measured at 25%, 50%, and 75% of the macrohabitat length (Bisson et al. 1982). Transects perpendicular to the thalweg were established at 25%,

50%, and 75% of the macrohabitat length. Water depth, substrate type, stream velocity, canopy cover, and bank characteristics at 20%, 40%, 50%, 60%, and 80% of the stream width were measured at each transect. Substrate was classified as boulder (>256 mm), cobble (64-256 mm), coarse gravel (16-64 mm), gravel (2-16 mm), sand (0.062-0.2 mm), silt

(0.039-0.062), clay (<0.0390 mm), bedrock, hardpan, detritus, wood, soil, vegetation (e.g., submerged grass), or artificial. Average stream velocity was measured at 60% of water depth with a Marsh McBirney Flo-Mate Portable Velocity Meter (Model 2000; Marsh-McBirney

Inc., Frederick, MD, USA). Canopy cover was measured with a spherical densiometer facing each bank, and upstream and downstream from the center of each transect. Bank characteristics (e.g., percent woody vegetation, non-woody vegetation, eroding, rip-rap, 12

roots, bare ground) were visually estimated for both streambanks. All units of stream cover were classified (e.g., woody debris, macrophyte, terrestrial vegetation, small brush, overhanging vegetation, undercut bank, rip-rap, artificial structure) and measured by taking one length, three width, and three depth measurements. Average width, depth, and velocity were calculated for each macrohabitat. Average percent canopy cover and average percent of bank characteristic types were also calculated for each macrohabitat. The coefficient of variation for each characteristic was calculated as 100 times the standard deviation divided by the mean for each habitat variable. All averaged values were then weighted by the proportion of the total site area represented by each macrohabitat type. The percent of each stream cover type at a site was calculated by dividing the area of the cover unit by the wetted area of the entire site.

Model validation

If a species occurrence model performs well, it can identify areas suitable for conservation (Williams and Araujo 2000), restoration (Wenger et al. 2009), or reintroduction

(Evans and Oliver 1995). Menzel and Clark (2002) developed an occurrence model for

Topeka shiners in west-central Iowa to aid in their recovery. The Menzel and Clark model

(MCM) used landscape-scale variables (e.g., adjacent land cover, soil type) to predict the presence or absence of Topeka shiner in central Iowa streams. As is the case with many other occurrence models (Manel et al. 2001), the MCM has not been validated.

Traditionally, model performance is often evaluated using the same data used to build the model. However, this method usually overestimates correct classification rate (Efron 1986). 13

Therefore, the best method for testing model performance is using an independent data set.

Our study provided 94 independent sites that were used to test the MCM.

We used several methods to evaluate the predictive power of the MCM. The principal statistic used to measure model performance was Cohen’s kappa (κ) which compares the correct classification of observations to those expected by random chance

(Cohen 1960). Values between 0.0 and 0.4 signify “slight to fair” model performance, values

between 0.4 and 0.6 “moderate” performance, values between 0.6 and 0.8 “substantial”

performance, and values between 0.8 and 1.0 indicate near “perfect” performance (Landis

and Koch 1977). We created a confusion matrix to determine if observed data were in

agreement with model predictions (Fielding and Bell 1997). Each site was classified as a

true presence, false presence, true absence, or false absence. This information was then used

to calculate the percentage of sites correctly classified (PCC), model sensitivity (percent of

presences correctly classified), and specificity (percent of absences correctly classified).

Data analysis

Changes in the distribution of a species can help determine whether a population is

declining, stable, or increasing and can then be used to prioritize conservation efforts (Moyle

and Nichols 1974; Piller et al. 2004; Sindt et al. 2012). For example, areas of decline could

be potential restoration sites while areas of stable or increasing distributions could be

identified for protection. To determine the status of Topeka shiner in central Iowa, we

compared its current distribution to its prior distribution. We chose to compare our

distribution data to those collected by Menzel and Clark (2002) during 1997 to 2000 because

it was the most recent survey targeting Topeka shiners in Iowa. Topeka shiner status for all 14

HUC 10 watersheds within its historic range was classified as increasing, stable, at risk, or possibly extirpated. If the watershed was not occupied by Topeka shiners during 1997 to

2000, but detected during 2010 to 2011, it was considered increasing. If the watershed was occupied by Topeka shiners during 1997 to 2000 and during 2010 to 2011, it was considered stable. If it was occupied by Topeka shiners from 1997 to 2000, but not detected from 2010 to 2011, it was considered to be at risk. Lastly, if the watershed was within the historic distribution of Topeka shiners but they were not detected from 1997–2011, it was determined that Topeka shiners were likely extirpated.

We used two approaches to better understand the abiotic and biotic factors associated with the occurrence of Topeka shiners in our study area. First, we examined fish assemblage data from each site to evaluate the association of Topeka shiners with other members of the fish assemblage. We then developed multiple logistic regression models to identify reach- scale factors associated with Topeka shiner occurrence. Each method of analysis was applied to all sites and to stream and off-channel sites separately, thereby allowing us to identify important factors associated with Topeka shiners overall, as well as those that may only exist in stream or off-channel habitats.

Nonmetric multi-dimensional scaling (NMDS) ordination was used to visualize the different fish assemblages in all sites and those of stream and off-channel sites. Ordinations were created from distance matrices based on the relative abundance (number of individuals per 100 m2) of fishes using the Bray-Curtis distance measure with standardization for site total (Faith et al. 1987). Significant habitat variables (i.e., mean depth, canopy cover) were fit onto ordinations as vectors using the ENVFIT function in the vegan library (Oksanen et al.

2011) for Program R. Vectors were added to ordination if its r2 value was greater than the 15

95th percentile of 1,000 randomly permuted correlations. We tested for differences in fish

assemblages by using an analysis of variance using distance matrices (ADONIS) in the vegan package of Program R (R Development Core Team 2011).

Logistic regression is a common technique used to identify factors associated with occurrence of fishes (Harig and Fausch 2002; Rich et al. 2003; Quist et al. 2005; Fischer and

Paukert 2008). We used an information theoretic approach to select a set of candidate models that best explained the occurrence of Topeka shiners (Burnham and Anderson 2002).

These a priori candidate models were generated using factors known to be of biological importance to Topeka shiners. The number of variables included in each candidate model was limited to 10% of the sample size to prevent overfitting. The most parsimonious candidate models were included in the confidence model sets. Sets of the best performing models were evaluated using Akaike’s information criterion corrected for small sample size

(AICc). The AICc value reflects parsimony of the model while penalizing the inclusion of

additional variables. Only candidate models with a ΔAICc ≤ 2 were included in the

confidence set to ensure that the confidence set contained models that were nearly as

parsimonious as one another (Burnham and Anderson 2002). Model averaged coefficients

and 95% confidence intervals were then calculated from the confidence sets of competing

models to determine which factors significantly contributed to the prediction of Topeka

shiner occurrence. Model fit was evaluated using McFadden’s (1974) pseudo r2. Three

models were constructed to determine habitat and biological associations of Topeka shiners: a combined model (using both stream and off-channel sites), a stream model, and an off- channel model. 16

A combined model of associations among both site types (stream and off-channel)

was developed because some associations could exist that were independent of habitat type.

There is evidence that Topeka shiners are associated with Lepomis spp. in both stream

(Minckley and Cross 1950; Stark et al. 2002) and off-channel habitats (Thomson and Berry

2009). Similarly, fathead minnow Pimephales promelas are often associated with Topeka

shiners in streams (Minckley and Cross 1959; Winston 2002) and off-channel habitats

(Thomson and Berry 2009), and presence of piscivorous fishes is thought to have a negative influence on Topeka shiner populations (Schrank et al. 2001; Mammoliti 2002). Juvenile

Topeka shiners use submerged vegetation (Kerns and Bonneau 2002) which was present in streams and off-channel habitats in our study area. Thus, fathead minnow CPUE, sunfish

CPUE, piscivore CPUE, and percent submerged vegetation were used to create a set of

candidate models for the combined model.

Because stream and off-channel habitats differ greatly in physical characteristics

(e.g., substrate composition, water velocity, channel morphology), we included additional

variables that may be associated with Topeka shiner occurrence in those different habitats.

In stream habitats, Topeka shiner occurrence is often associated with habitat characteristics

such as coarse substrates (Wall and Berry 2006), banks with substantial vegetation (Bayless

2003), and stream vegetation cover (Kerns and Bonneau 2002). In off-channel habitats in

South Dakota, Thomson and Berry (2009) suggested that water depth influences the

occurrence and abundance of Topeka shiners. We included submerged vegetation as well as

mean depth in the off-channel model. All biotic variables included in the combined model

were also included in the separate stream and off-channel models.

17

Results

A total of 94 sites, representing 67 stream and 27 off-channel sites, was sampled in

2010 and 2011. We encountered 59 fish species and identified 68,177 individual fish.

Topeka shiners were detected in 6 stream and 14 off-channel sites (Figure 2) and ranked 24th

in abundance with a total of 790 individual fish sampled. Topeka shiners were sampled in 6

out of 22 HUC 10 watersheds where they occurred historically. East Buttrick, West Buttrick,

Hardin, Cedar, and Purgatory creeks in the North Raccoon River basin, and Eagle Creek in

the Boone River basin had at least one site where Topeka shiners were detected. Topeka

shiners were not detected in any watershed where they have not been detected within the last

twenty years. We failed to detect Topeka shiners in many of the watersheds where they had

been previously documented, as reflected by their “at risk” status (Table 1). Topeka shiners in our study were only detected in watersheds where they were detected in 1997 to 2000.

The spatial difference in Topeka shiner detections between the Menzel and Clark study

(2002) and our study suggests a recent decline in its distribution (Figure 3).

The MCM did not predict Topeka shiner occurrence accurately for sites in our study.

The Cohen’s kappa value for the Menzel and Clark (2002) model was 0.19, indicating relatively poor model performance. The model correctly classified 59% of the sites in our study, and was more successful at predicting presences (sensitivity = 0.8) than absences

(specificity = 0.51).

Fish assemblages and Topeka shiner occurrence in stream and off-channel sites were characterized using NMDS ordination (Figure 4). Assemblage structure is shown in two dimensions with a stress value of 0.18, indicating a fair match between the pairwise 18

assemblage distances and those distances in the ordination space (r2 = 0.86). Although a 3-

dimensional ordination had a stress of 0.13, the general patterns did not differ from the more

interpretable 2-dimensional representation. Significant differences were detected in fish assemblages between stream and off-channel sites (ADONIS: P < 0.001). No differences in

fish assemblages between sites where Topeka shiner were present and absent were detected

after adjusting for site type (ADONIS: P > 0.49). A contour surface indicating species

richness isobars was added to the ordination, indicating that differences in assemblages were

partially attributed to the number of species at each site. Mean species richness was

significantly lower in off-channel sites than in stream sites (two sample t test: t = 5.89 , df =

62.89 , P < 0.001).

Another NMDS ordination (stress = 0.17, r2 = 0.90) was created to characterize fish

assemblages at stream sites (Figure 5). Fish assemblages in stream sites that contained

Topeka shiners were not significantly different from those that did not contain Topeka

shiners (ADONIS: P = 0.75). Several habitat variables (Table 1) were significantly

correlated with the NMDS scores and indicated habitat gradients, but Topeka shiner

occurrence did not differ along those gradients. Fish assemblages at off-channel sites were

also characterized using NMDS ordination (Figure 6). This ordination reflects the true

pairwise distance between assemblages relatively well (stress = 0.14, r2 = 0.94). Fish

assemblages at sites with Topeka shiners differed significantly from sites without Topeka

shiners (ADONIS: P = 0.03). Assemblages that included Topeka shiners also contained

more lentic species (e.g., fathead minnow, largemouth bass, common carp Cyprinus carpio)

than assemblages without Topeka shiners. The only habitat vectors (Table 1) that were

significantly correlated with NMDS scores were mean canopy cover, proportion of site with 19

no visible disturbance, and percent coarse gravel substrate. Decreasing scores on the y-axis

indicated an increase in forested area and reduced land use disturbance. Topeka shiner sites

tended to be in the less forested areas that had some level of disturbance (i.e., pasture, row

crop, road). Topeka shiners also tended to occur in off-channel sites with less coarse

substrate that sites without Topeka shiners.

Important habitat and biotic variables (Table 2) were used to create sets of candidate

logistic regression models. Confidence model sets for combined, stream, and off-channel

models contained one, seven, and four candidate models, respectively (Table 3). The

combined model (i.e., stream and off-channel sites) contained only one model in its

confidence model set. All competing models in the combined model had a ΔAICc >2

indicating that no candidate model was nearly as parsimonious as the top model. Fathead

minnow CPUE appeared in the top five candidate models, while percent submerged

vegetation and sunfish CPUE only appeared in three of the five. The stream model had

seven candidate models in the confidence model set. Each confidence model was similarly

parsimonious and no single variable was common to all. The off-channel model had four

candidate models in its confidence set (Table 3). The top model containing only fathead

minnow CPUE had a larger Akaike weight than the other three models. Fathead minnow

CPUE was in all of the confidence models for off-channel sites.

Model averaged parameter estimates, standard errors, 95% confidence intervals, and

relative weights were calculated from each confidence model set (Table 4). The combined model (pseudo r2 = 0.21) contained two variables with parameter estimates significantly

different from zero. Both submerged vegetation and fathead minnow CPUE parameters were

greater than zero, although the size of the coefficients was relatively small. Sunfish CPUE 20

was not a significant predictor of Topeka shiner occurrence in the combined model. The

stream model (pseudo r2 = 0.001) contained no parameter estimates significantly different from zero. Similarly, the off-channel model (pseudo r2 = 0.13) contained no parameter

estimates significantly different from zero.

Discussion

The pattern of Topeka shiner distribution in our study indicates that their abundance

may be declining in several watersheds where they had been documented only a decade

earlier by Menzel and Clark (2002). However, Topeka shiners typically occur in low

densities and it is possible that we failed to detect them at some sites where they were

present. Ommission error could lead to an underestimation of abundance, but we assumed that the Menzel and Clark study (2002) had similar error. It should also be noted that

Menzel and Clark (2002) sampled more sites than our study, which could have increased

their overall probability of detecting Topeka shiners in a watershed. However, our sampling

efficiency was likely higher since our stream sites were electrofished and seined, while

Menzel and Clark only seined. Topeka shiners in our study were detected using both gear

types, indicating that the previous study could have failed to detect Topeka shiners by using

only one gear type. Past Topeka shiner watersheds were also sampled more intensely than

other watersheds to improve our confidence that a decline was likely in areas determined to

be at risk. For example, Indian Creek and Lake Creek watersheds in the western part of the

North Raccoon River basin historically contained Topeka shiners. We sampled previous

Topeka shiner sites and elsewhere within those watersheds and concluded that these areas 21

could be experiencing declines in abundance or distribution. In the Boone River basin,

Topeka shiners were only detected in the Eagle Creek watershed. This smaller fragment of the population in the Boone River basin could have a higher risk of local extirpation due to its size and isolation from the North Raccoon River basin (Lawton 1993). Topeka shiner populations found in only one watershed within a basin, such Eagle Creek in the Boone River basin, may be more susceptible to fish kills caused by point source pollution. From 1995 to

2011, 202 major fish kills (>1000 fish) occurred in streams throughout Iowa (IDNR 2012).

Increasing the abundance and expanding the distribution of Topeka shiners to other watersheds in the Boone River basin could ensure their persistence in that area.

Although we found no expansion of Topeka shiner range at the HUC10 level, we detected them farther upstream than they have been previously documented in Iowa. Topeka shiners were detected in small headwater streams of Cedar Creek and Hardin Creek in the

North Raccoon River basin. In the second year of sampling, these sites were completely dry, suggesting that they are somewhat ephemeral habitats. This provides further evidence that

Topeka shiners are adapted to drought conditions and capable of recolonizing small streams once flows return. Others have shown that Topeka shiner range has become more restricted to headwaters of the watersheds they were once widely distributed in (Winston 2002;

Thomson et al. 2005), but we detected Topeka shiners in both the upper and lower reaches of streams in Iowa.

The habitat associations identified in this study provide novel information about the biology and life history of Topeka shiner which can be used to guide restoration, reintroduction, and other recovery efforts. Our formal modeling identified some possible habitat associations between Topeka shiners and their biotic and abiotic environment. An 22

increase in fathead minnow CPUE was associated with occurrence of Topeka shiners in the

combined model. Fathead minnow CPUE was also present in many of the top candidate

models in all three (combined, stream, and off-channel) models. Others have shown that

fathead minnows commonly occur with Topeka shiners in other states (Minckley et al. 1959;

Thomson and Berry 2009; Winston 2002), but none have documented this in west-central

Iowa. One explanation for this association could be that fathead minnows and Topeka

shiners have similar physiological tolerances. For example, fathead minnows and Topeka

shiners can survive in drought conditions (Minckley et al. 1959) that other species cannot

tolerate. Fathead minnows might also act as a predation buffer to Topeka shiner. When a prey species becomes rare, predators may seek prey species that are more abundant

(Murdoch 1969). In warm, oxygen-limited habitats, as found during dry years and in off-

channel habitats, fathead minnows are one of the few species that could provide a predation

buffer for Topeka shiners. It is known that Topeka shiners are nest associates of

orangespotted sunfish and green sunfish (Pflieger 1997; Stark 2002), but little is known about

symbioses with fathead minnows. One study documented Topeka shiners establishing

territories on the periphery of fathead minnow nests (Stark et al. 2002), suggesting Topeka

shiners may be “nest associates” of fathead minnows in addition to green sunfish and orangespotted sunfish. They also observed groups of Topeka shiners overwhelming nest-

guarding male fathead minnows and feeding, presumably, on fathead minnow eggs. Since

Topeka shiner spawn slightly later than fathead minnows, feeding on nutrient rich eggs could

enhance improve female Topeka shiner condition prior to spawning (Belles-Isles and

Fitzgerald 1993). Our study identified an association between fathead minnows and Topeka 23

shiners, but to identify the mechanism underlying the association would require further research.

The previously documented positive associations between Topeka shiner and green sunfish and orangespotted sunfish was not apparent in our study. All sites (stream and off- channel) that supported Topeka shiner also included either green sunfish or orangespotted sunfish. Green sunfish and orangespotted sunfish CPUE was not a significant predictor of

Topeka shiner occurrence, likely because of their ubiquity throughout the study area. On two occasions we sampled spawning Topeka shiners near sunfish in stream sites. Although green sunfish grow larger and are more aggressive than other sunfishes (Werner and Hall 1977), these differences seem to have little effect on Topeka shiner reproduction or persistence. In off-channel sites, we documented Topeka shiner reproduction in habitats primarily dominated by orangespotted sunfish, but others included green sunfish. Topeka shiners may be nest associates of green sunfish and orangespotted sunfish in Iowa, but our study does not provide definitive evidence supporting this.

Negative association between Topeka shiner and piscivorous fishes (e.g., largemouth bass) has been documented in the Flint Hills of northeastern Kansas, due mainly to the occurrence of stream impoundments that create ideal habitats for piscivores (Schrank et al.

2001; Mammoliti 2002). In contrast, largemouth bass and other piscivorous fishes were often sympatric with Topeka shiners in our study. This discrepancy could be explained by differences in habitat between the Flint Hills of Kansas and the Des Moines Lobe of Iowa.

Streams in the Flint Hills are surrounded by native grasslands, are less eroded, and are relatively clear. In the Des Moines Lobe, streams and off-channel habitats are surrounded by cropland or pasture, are considerably eroded, and often highly turbid. It is known that 24

turbidity can affect foraging efficiency of largemouth bass (Shoup and Wahl 2009) and other

piscivores (Turesson and Bronmark 2007). It is quite possible that visual predators are much

more efficient at capturing Topeka shiners in the clear streams and impoundments of Kansas,

and less efficient in turbid conditions that characterize habitats in Iowa.

Biotic associations can occur with individual species, such as fathead minnows, but

also with fish assemblages. Since Topeka shiners were more common in off-channel

habitats, their fish assemblages could be associated with Topeka shiner occurrence. Off-

channel habitats in our study were inhabited more commonly with lentic species such as

common carp Cyprinus carpio, largemouth bass, black bullhead Ameiurus melas, green

sunfish, and orangespotted sunfish. Other species such as fathead minnow, brassy minnow

Hybognathus hankinsoni, and Topeka shiner also occurred at higher densities in off-channel habitats. These three species are tolerant of high water temperatures and low dissolved oxygen, conditions which are characteristic of disconnected off-channel habitats (Brungs

1971a; Brungs 1971b; Copes 1975; Koehle and Adelman 2007). Many lotic species, such as bigmouth shiners Notropis dorsalis and central stonerollers Campostoma anomalum (Pflieger

1997), were represented in off-channel habitats but at lower abundance than in streams.

Fishes only have the opportunity to enter these off-channel habitats during flood events that offer a connection with the main channel of the stream. Thus, the presence and abundance of lotic specialists in off-channel habitats could be a function of time since a flooding event occurred. Others have shown that varying levels of connectivity with the main channel can structure fish assemblages in large oxbows (Miranda 2005; Dembkowski and Miranda 2011;

Zeug et al. 2005). Similar findings in small off-channel habitats of wadeable streams are 25

lacking. Further research is needed to understand the mechanisms that underlie differences

in fish assemblage structure between off-channel habitats and smaller stream systems.

Within off-channel sites, Topeka shiner occurrence was associated with different fish

assemblages. Not only did lentic species tend to occur more often in off-channel sites, but

they also occurred more often with Topeka shiner than without them. On the other hand,

Topeka shiners were less often detected with fish assemblages containing more lotic species

(e.g., bigmouth shiner, sand shiner Notropis stramineus, common shiner Luxilus cornutus, highfin carpsucker Carpoides velifer). Lotic specialists that require flowing water and higher dissolved oxygen could enter an oxbow during a flood event. After flood waters recede, however, these species may perish as conditions become more lentic (Halyk and Balon

1983). Priority effects are known to structure fish assemblages in other aquatic systems

(Almany 2003) and could explain Topeka shiner persistence in recently flooded off-channel habitats. Since flooding events were not identified, initial fish assemblages could not be investigated in our study.

Topeka shiner occurrence is not only related to the occurrence of other fishes, but also to habitat features. Our results indicate a positive relationship between Topeka shiner occurrence and submerged vegetation. Juvenile fish of many species are known to use submerged vegetation as nursery habitat (Lobb and Orth 1991; Venugopal and Winfield

1993). In fact, Kerns and Bonneau (2002) observed juvenile Topeka shiners congregating in areas with submerged vegetation in the shallow margins of pool habitats. The type of

vegetation found in streams was slightly different than that of off-channel sites in our study.

In streams, most of the submerged vegetation consisted of submerged terrestrial bank

vegetation and a lesser amount of aquatic macrophytes. Submerged vegetation in off- 26

channel habitats consisted primarily of aquatic macrophytes but with small contributions of flooded terrestrial vegetation. Since livestock grazing is very common along wadeable streams in Iowa, measures to reduce overgrazing along stream banks could improve Topeka shiner habitat, though Wall and others (2004) suggest that Topeka shiners can survive in streams that experience “moderate” grazing. Increasing or maintaining aquatic vegetation in streams and off-channel habitats may not only be beneficial to Topeka shiners, but to other fish species that may use those areas as rearing habitats.

Our results suggest a positive association between Topeka shiner and off-channel habitats. Although this phenomenon has been noted by others (Minckley and Cross 1959;

Hatch 2001), there are no clear hypotheses as to why Topeka shiners were more common in off-channel habitats of west-central Iowa. One possibility is that off-channel habitats represent a considerable proportion of their total habitat use and, thus, could be considered a floodplain-exploitative species (Ross and Baker 1983). Topeka shiners can also tolerate low levels of dissolved oxygen and high water temperatures (Koehle and Adelman 2007), which are conditions typical of shallow, unconnected off-channel habitats. Topeka shiners are also known to persist during droughts when streams are reduced to a series of deep pools with little or no surface flow (Minckley and Cross 1959). Since off-channel habitats are similar to pool habitats during droughts, Topeka shiners are likely adapted to conditions typical of off- channel habitats. Topeka shiner can not only survive in off-channel habitats, but reproduction was documented at two off-channel sites during our study and by others

(Thomson et al. 2005). This research provides further evidence that off-channel habitats are an important component of Topeka shiner life history. 27

Higher sampling efficiency might also explain why Topeka shiners were more often

detected in off-channel habitats than in streams. Many of the stream sites had undercut

banks, dense bank vegetation, and stream cover that likely made sampling less efficient than

in the physically homogenous off-channel habitats. Topeka shiners are typically rare in sites

where they are detected so it is likely that this species was present but not detected in some

stream sites. Since this species is considered a prairie stream fish (Pflieger 1997), sampling

efforts targeting Topeka shiners has generally occurred solely in streams (Bayless et al.

2003). Topeka shiners often use (Thomson and Berry 2009) and are easily detected in off-

channel habitats, so future research should incorporate these habitats whenever possible.

Although others have suggested that off-channel habitat depth is associated with

Topeka shiner occurrence and abundance (Thomson and Berry 2009), we did not find this in our study. Eight of the fourteen off-channel habitats that contained Topeka shiners were relatively shallow (mean depth < 0.5 m). Juvenile Topeka shiners were also abundant in one shallow off-channel habitat, suggesting that reproduction is possible is such habitats. Robb

(2002) suggested that fathead minnow could persist by seeking shallow areas in a pond

where predators with lower physiological tolerances could not survive. Since Topeka shiners

are similarly tolerant to low dissolved oxygen and high water temperatures, they too could

physiologically exclude predators and even competitors.

Shallow off-channel habitats may be suitable for Topeka shiners, but they can also be ephemeral. During dry years, shallow habitats containing Topeka shiners could dry completely and be sinks to the overall population. Similarly, shallow off-channel habitats are more likely to freeze solid during the winter than deeper habitats, again becoming sinks

(Pulliam 1988) to the Topeka shiner population. The only way successful reproduction in 28

shallow off-channel habitats can contribute to the overall population is through late season flooding events that allow juvenile Topeka shiners to disperse. Restored or naturally deep off-channel habitats may be able to support more predators and competitors, but they are also less susceptible to drying during drought years. Deep off-channel habitats are less likely to completely freeze, allowing Topeka shiners to survive through winter. Although Thomson and others (2005) suggest that deep and frequently flooded habitats may be ideal for Topeka shiners, the tradeoffs between varying depths and flood frequencies should be further investigated to better guide off-channel habitat restoration.

Given their affinity for off-channel habitats, availability of these habitats could be a factor that limits the Topeka shiner population in Iowa. Only 5% of Iowa’s original wetlands currently exist (Bishop 1981), including off-channel habitats such as natural oxbows. In eastern South Dakota there are nearly 56,000 livestock ponds constructed near or in streams

(Johnson and Higgins 1997). This large number of off-channel habitats could be one reason why Topeka shiner populations are relatively stable in South Dakota. One reason why these habitats are so rare in Iowa is because they were drained and converted to row crop production or pasture (Best et al. 1978). Natural formation of off-channel habitats has also decreased due to the channelization and bank armoring that limits lateral movement of stream channels. Channel incision due to channelization lowers the channel bed (Shields et al. 1994), preventing floodwaters from reaching the floodplain where suitable off-channel habitats may exist. Since these habitats have become increasingly rare, off-channel habitat restoration could play an important role in the recovery of Topeka shiners in Iowa. In fact, several of the restored off-channel habitats in our study contained Topeka shiners and two contained indications of successful reproduction. This evidence suggests that continued 29

restoration of off-channel habitats may increase the likelihood of Topeka shiner recovery in

Iowa.

Although the distribution and abundance of Topeka shiners does not appear to be

increasing, we are beginning to understand some of the physical and biological needs of this imperiled fish. Without such knowledge, costly efforts aimed at Topeka shiner conservation

could provide little return. However, identifying associations should only be the first step in

understanding Topeka shiner biology and life history. Identifying the underlying

mechanisms behind biotic and abiotic associations could provide us with valuable

information that can be applied to Topeka shiner conservation. For example, identifying the

nature of the relationship of Topeka shiners and fathead minnows could help guide recovery

efforts. For example, fathead minnow presence might only identify suitable habitats for

Topeka shiner reintroduction. However, if there is a symbiotic relationship between the two

species, off-channel habitat restorations could include the stocking of known symbionts.

Restoration of off-channel habitats may be an important part of Topeka shiner

recovery in Iowa, but little is known about how they and other fishes use them throughout the

season or from year to year. Depth, flood frequency, and habitat factors such as submerged vegetation should be further investigated to determine the ideal conditions for Topeka shiner in off-channel habitats. Managers must often make decisions about imperiled species recovery based on limited or incomplete information. Since little is known about the role of off-channel habitats in Topeka shiner populations, there is a possibility that some restorations could have limited positive or even negative effects on Topeka shiners. However, inaction is a management decision that would likely lead to the continued decline of this species in

Iowa. Although there is some evidence indicating their importance to Topeka shiners in 30

Iowa, all off-channel habitat restorations should be monitored and evaluated to guide future

restoration work. Such an adaptive management technique would improve the chances of

Topeka shiner recovery and improve our understanding of these understudied aquatic

systems.

Acknowledgements

We thank Savanna Bice, Jacob Miller, Cole Harty, Jared Brashears, Brett Meyers,

Grant Scholten and Michael Sundberg for their assistance in the field. We also thank Daryl

Howell and Aleshia Kenney for their assistance and cooperation. This project was supported

in part by the Department of Natural Resource Ecology at Iowa State University, Iowa

Cooperative Fish and Wildlife Research Unit, the U.S. Fish and Wildlife Service, and the

Iowa Department of Natural Resources. Use of trade names does not imply endorsement by

the U.S. Government.

References

Adelman, I. R., L. I. Kusilek, J. Koehle, and J. Hess. 2009. Acute and chronic toxicity of

ammonia, nitrite, and nitrate to the endangered Topeka shiner (Notropis topeka) and

fathead minnows (Pimephales promelas). Environmental Toxicology and Chemistry

28:2216-2223.

Allan, J. D., and A. S. Flecker. 1993. Biodiversity conservation in running waters.

BioScience 43:32-43. 31

Almany, G. R. 2003. Priority effects in coral reef fish communities. Ecology 84:1920-1935.

Bayless, M. A., M. G. McManus, and J. F. Fairchild. 2003. Geomorphic, water quality and

fish community patterns associated with the distribution of Notropis topeka in a

central Missouri watershed. American Midland Naturalist 150:58-72.

Belles-Isles, J. C., and G. J. Fitzgerald. 1993. A fitness advantage of cannibalism in female

sticklebacks (Gasterosteus aculeatus L.). Ethology, Ecology, and Evolution 5:187-

191.

Best, L. B., K. L. Varland, and R. B. Dahlgren. 1978. Effects of stream channelization on

land-use practices in Iowa. Iowa State Journal of Research 52:411-423.

Bishop, R. A. 1981. Iowa’s wetlands. Proceedings of the Iowa Academy of Science 88:11-

16.

Bisson, P. A., J. L. Nielson, R. A. Palmason, and L. E. Grove. 1982. A system of naming

habitat types in small streams, with examples of habitat utilization by salmonids

during low stream flow. Pages 63-73 in N. B. Armantrout, editor. Acquisition and

utilization of aquatic habitat information. American Fisheries Society, Bethesda,

Maryland.

Brungs, W. A. 1971a. Chronic effects of constant elevated temperature on the fathead

minnow (Primephales promelas Rafinesque). Transactions of the American Fisheries

Society 100:659-664.

Brungs, W. A. 1971b. Chronic effects of low dissolved oxygen concentrations on the fathead

minnow (Pimephales promelas). Journal of the Fisheries Research Board of Canada

28:1119-1123. 32

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multi-model inference: a

practical information theoretic approach, second edition. Springer, New York.

Clark, S. J. 2000. Relationship of Topeka shiner distribution to geographic features of the

Des Moines Lobe in Iowa. M.S. Thesis, Iowa State University, Ames, Iowa.

Cohen, J. 1960. A coefficient of agreement for nominal scales. Educational and

Psychological Measurements 20:37-46.

Copes, F. A. 1975. Ecology of the brassy minnow, Hybognathus hankinsoni (Cyprinidae).

Reports on the fauna and flora of Wisconsin. Museum of Natural History, University

of Stevens Point, Stevens Point, Wisconsin.

Dembkowski, D. J., and L. E. Miranda. 2011. Comparison of fish assemblages in two

disjoined segments of an oxbow lake in relation to connectivity. Transactions of the

American Fisheries Society 140:1060-1069.

Dudgeon, D., A. H. Arthington, M. O. Gessner, Z. Kawabata, D. J. Knowler, C. Leveque, R

J. Naiman, A Prieur-Richard, D. Soto, M. L. J. Siassny, and C. A. Sullivan. 2006.

Freshwater biodiversity: importance, threats, status and conservation challenges.

Biological Reviews 81:163-182.

Efron, B. 1986. How biased is the apparent error rate of a prediction rule? Journal of the

American Statistical Association 81:461-470.

Evans, D. O., and C. H. Oliver. 1995. Introduction of lake trout (Salvelinus namaycush) to

inland lake of Ontario, Canada: factors contributing to successful colonization.

Journal of Great Lake Research 21(Supplement 1):30-53.

Faith, D. P., P. R. Minchin, and L. Belbin. 1987. Compositional dissimilarity as a robust

measure of ecological distance. Plant Ecology 69:57-68. 33

Fielding, A. H., and J. F. Bell. 1997. A review of methods for assessment of prediction errors

in conservation presence absence models. Environmental Conservation 24:38-49.

Fischer, J. R., and C. P. Paukert. 2008. Historical and current environmental influences on an

endemic Great Plains fish. The American Midland Naturalist 159:364-377.

Gallant, A. L., W. Sadinski, M. F. Roth, and C. A. Rewa. 2011. Changes in historical Iowa

land cover as context for assessing the environmental benefits of current and future

conservation efforts on agricultural lands. Journal of Soil and Water Conservation

66:67-77.

Griffith, G. E., J. M. Omernik, T. F. Wilton, and S. M. Pierson. 1994. Ecoregions and

subregions of Iowa: a framework for water quality assessment and management.

Journal of the Iowa Academy of Science 101:5-13.

Halyk, L. C., and E. K. Balon. 1983. Structure and ecological production of the fish taxocene

of a small floodplain system. Canadian Journal of Zoology 61:2446-2464.

Harig, A. L., and K. D. Fausch. 2002. Minimum habitat requirements for establishing

translocated cutthroat trout populations. Ecological Applications 12:535-551.

Hatch, J. T., and S. Besaw. 2001. Food use in Minnesota populations of the Topeka shiner

(Notropis topeka). Journal of Freshwater Ecology 16:229-233.

IAGFA (Iowa Aquatic Gap Fish Atlas). 2005. Iowa Aquatic Gap Fish Atlas, Iowa Rivers

Information System. (http://maps.gis.iastate.edu/iris/fishatlas/), accessed February

2010.

Iowa Department of Natural Resoureces (IDNR). 2012. Iowa DNR fish kill database.

(http://programs.iowadnr.gov/fishkill/search.aspx?std=mod), accessed February 2011. 34

Iowa Department of Natural Resources (IDNR). 2001. Biological sampling procedures for

wadeable streams of Iowa. Iowa Department of Natural Resources, Des Moines,

Iowa.

Johnson, R. R., and Higgins, K. F. 1997. Wetland resources of eastern South Dakota. South

Dakota State University, Brookings, South Dakota, USA.

Kerns H. A., and J. L. Bonneau. 2002. Aspects of the life history and feeding habits of the

Topeka shiner (Notropis topeka) in Kansas. Transactions of the Kansas Academy of

Science 105:125-142.

Koehl, J. J., and I. R. Adelman. 2007. The effects of temperature, dissolved oxygen, and

asian tapeworm infection on growth and survival of the Topeka shiner. Transactions

of the American Fisheries Society 136:1607-1613.

Landis, R. J., and G. G. Kock. 1977. The measurement of observer agreement for categorical

data. Biometrics 33:159-174.

Lawton, J. H. 1993. Range, population abundance and conservation. Trends in Ecology and

Evolution 8:409-413.

Lee, D. S., C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister, and J. R. Stauffer.

1980. Atlas of North American freshwater fishes. North Carolina Museum of Natural

History, Raleigh.

Loan-Wilsey, A. K., C. L. Pierce, K. L. Kane, P. D. Brown, and R. L. McNeely. 2005. The

Iowa Aquatic Gap Analysis Project. Final report, Iowa Cooperative Fish and Wildlife

Research Unit, Iowa State University, Ames, Iowa.

Lobb, M. D., and D. J. Orth. 1991. Habitat use by an assemblage of fish in a large warmwater

stream. Transactions of the American Fisheries Society 120:65-78. 35

Mammoliti, C. S. 2002. The effects of small watershed impoundments on native stream

fishes: a focus on the Topeka shiner and hornyhead chub. Transactions of the Kansas

Academy of Science 105:219-231.

Manel, S., H. C. Williams, and S. J. Ormerod. 2001. Evaluating presence-absence models in

ecology: the need to account for prevalence. Journal of Applied Ecology 38:921-931.

McFadden, D. 1974. Conditional logit analysis of qualitative choice behavior. Pages 105 –

142 in P. Zarembka, editor. Frontiers in Econometrics. Academic Press, New York.

Menzel, B. W., J. B. Barnum, and L. M. Antosch. 1984. Ecological alterations of Iowa

prairie-agricultural stream. Iowa State Journal of Research 59:5-30.

Menzel, B. W., and S. J. Clark. 2002. Final report to Iowa Department of Natural Resources

on agreement for consultant services for refinements to a habitat model of the Topeka

shiner, Ames, Iowa.

Minckley, W. L., and F. B. Cross. 1959. Distribution, habitat, and abundance of Topeka

shiner, Notropis topeka (Gilbert), in Kansas. American Midland Naturalist 61:210-

217.

Miranda, L. E. 2005. Fish assemblages in oxbow lakes relative to connectivity with the

Mississippi River. Transactions of the American Fisheries Society 134:1480-1489.

Mittelbach, G. G., and L. Persson. 1998. The ontogeny of piscivory and its ecological

consequences. Canadian Journal of Fisheries and Aquatic Sciences 55:1454-65.

Moyle, P. B., and R. D. Nichols. 1974. Decline of the native fish fauna of the Sierra Nevada

foothills, central California. American Midland Naturalist 92:72-83.

Murdoch, M. W. 1969. Switching in general predators: experiments on predator specificity

and stability of prey populations. Ecological Monographs 39:335-354. 36

Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. B. O’Hara, G. L.

Simpson, P. Solymos, M. Henry, H. Stevens, and H. Wagner. 2011. Vegan:

community ecology package. R package version 2.0-1. http://CRAN.R-

project.org/package=vegan.

Onorato D. P., R. A. Angus, and K. R. Marion. 1998. Comparison of a small-mesh seine and

a backpack electroshocker for evaluating fish populations in a north-central Alabama

stream. North American Journal of Fisheries Management 18:361-373.

Pflieger, W. L. 1997. The fishes of Missouri, revised edition, Jefferson City: Missouri

Department of Conservation.

Piller, K. R., H. L. Bart, Jr., and J. A. Tipton. 2004. Decline of the frecklebelly madtom in

the Pearl River based on contemporary and historical surveys. Transactions of the

American Fisheries Society 16:391-409.

Poff, N. L., and J. K. H. Zimmerman. 2010. Ecological responses to altered flow regimes: a

literature review to inform the science and management of environmental flows.

Freshwater Biology 55:194-205.

Pont, D., B. Hugeuny, and T. Oberdorff. 2005. Modelling habitat requirement of European

fishes: do species have similar responses to local and regional environmental

constraints? Canadian Journal of Fisheries and Aquatic Sciences 62:163-173.

Pulliam, H. R. 1988. Sources, sinks, and population regulation. The American Naturalist 132:

652-661.

Quist, M. C., F. J. Rahel, and W. A. Hubert. 2005. Hierarchical faunal filters: an approach to

assessing effects of habitat and nonnative species on native fish. Ecology of

Freshwater Fish 14:24-39. 37

R Development Core Team. 2011. R: a language and environment for statistical computing.

R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,

http://www.R-project.org/.

Rich, C. F., and T. E. McMahon. 2003. Local-habitat, watershed, and biotic features

associated with bull trout occurrence in Montana streams. Transactions of the

American Fisheries Society 132:1053-1064.

Robb, T., and M. V. Abrahams. 2002. The influence of hypoxia on risk of predation and

habitat choice by the fathead minnow, Pimephales promelas. Behavioral Ecology and

Sociobiology 52:25-30.

Ross, S. T., and J. A. Baker. 1983. The response of fishes to periodic spring floods in a

Southern stream. The American Midland Naturalist 109:1-14.

Roth, N. E., J. D. Allan, and D. L. Erickson. 1996. Landscape influences on stream biotic

integrity assessed at multiple spatial scales. Lanscape Ecology 11:141-156.

Rowe D. C., C. L. Pierce, T. F. Wilton. 2009. Fish assemblage relationships with physical

habitat in wadeable Iowa streams. North American Journal of Fisheries Management

29:1314-1332.

Schrank, S. J., C. S. Guy, M. R. Whiles, and B. L. Brock. 2001. Influence of Stream and

landscape-level factors on the distribution of Topeka shiners (Notropis topeka) in

Kansas streams. Copeia 2001:413-421.

Shields, F.D., S. S. Knight, and C. M. Cooper. 1994. Effects of channel incision on base flow

stream habitats and fishes. Environmental Management 18:43-57. 38

Shilling, K. E., and M. Helmers. 2008. Effects of subsurface drainage tiles on streamflow in

Iowa agricultural watershed: exploratory hydrograph analysis. Hydrological

Processes 22:4497-4506.

Shoup, D. E., and D. H. Wahl. 2009. The effects of turbidity on prey selection by piscivorous

largemouth bass. Transactions of the American Fisheries Society 138:1018-1027.

Sindt, A. R., C. L. Pierce, and M. C. Quist. 2012. Fish species of greatest conservation need

in wadeable Iowa streams: current status and effectiveness of aquatic gap program

distribution models. North American Journal of Fisheries Management 32:135-146.

Stark, W. J., J. S. Luginbill, and M. E. Eberle. 2002. Natural history of a relict population of

Topeka shiner (Notropis topeka) in northwestern Kansas. Transactions of the Kansas

Academy of Science 105:143-152.

Tabor, V. M. 1998. Final Rule to list the Topeka shiner as endangered. Federal Register 63(7

December 1998):69008-69021.

Thomson, S. K., and C. R. Berry. 2009. Stream fishes inhabit livestock watering ponds

(dugouts) near Six Mile Creek, Brookings County, South Dakota. Proceedings of the

South Dakota Academy of Science 88:127-138.

Thomson, S. K., C. R., Berry, C. A. Niehus, and S. S. Wall. 2005. Constructed

impoundments in the floodplain: a source or sink for native prairie fishes, in

particular the endangered Topeka shiner (Notropis topeka)? Pages 1337-134 in G. E.

Moglen, editor. Managing watersheds for human and natural impacts: engineering,

ecological, and economic challenges. Proceedings of the 2005 Watershed

Management Conference. 39

Turesson, H., and C. Bronmark. 2007. Predator-prey encounter rates in freshwater piscivores:

effects of prey density and water transparency. Oecologia 153:281-290.

U.S. Fish and Wildlife Service (USFWS). 2012. U.S. Fish and Wildlife Service endangered

species program.(www.fws.goc/endangered/species/us.species.html), accessed March

2012.

U.S. Fish and Wildlife Service (USFWS). 2009. Topeka shiner 5-year review: summary and

evaluation. Kansas Ecological Services Field Station, Manhatten, Kansas.

Venugopal, M. N., and N. J. Winfield. 1993. The distribution of juvenile fishes in a

hypereutrophic pond: can macrophytes potentially offer a refuge for zooplankton?

Journal of Freshwater Ecology 8:389-396.

Wall, S. S., C. R. Berry, C. M. Blausey, J. A. Jenks, and C. J. Kopplin. 2004. Fish-habitat

modeling for gap analysis to conserve the endangered Topeka shiner (Notropis

topeka). Canadian Journal of Fisheries and Aquatic Sciences 61:954-973.

Wall, S. S., and C. R. Berry. 2006. The importance of Multiscale habitat relations and biotic

associations to the conservation of an endangered fish species, the Topeka shiner.

American Fisheries Society Symposium 48:305-322.

Wang, L., J. Lyons, P. Kanehl, and R. Gatti. 1997. Influences of watershed land use on

habitat quality and biotic integrity in Wisconsin streams. Fisheries 22:6-12.

Wenger, S. J., M. M. Hagler, and B. J. Breeman. 2009. Prioritizing areas of the Conasauga

River sub-basin in Georgia and Tennessee for preservation and restoration.

Proceedings of the Southeastern Fishes Council 51:31-38.

Werner, E. E., and D. J. Hall. 1977. Competition and habitat shift in two sunfishes

(Centrarchidae). Ecology 58:869-876. 40

Williams, P. H., and M. B. Araujo. 2000. Using probability of persistence to identify

important areas for biodiversity conservation. Proceedings of the Royal Society of

London 267:1959-1966

Winston, R. W. 2002. Spatial and temporal species associations with the Topeka shiner

(Notropis topeka) in Missouri. Journal of Freshwater Ecology 17:249-261.

Zeug, S. C., K. O. Winemiller, and S. Tarim. 2005. Response of Brazos river oxbow fish

assemblages to patterns of hydrologic connectivity and environmental variability.

Transactions of the American Fisheries Society 134:1389-1399.

Zohrer, J. J. 2005. Securing the future for fish and wildlife: a conservation legacy for Iowans.

Iowa Department of Natural Resources, Iowa Wildlife Action Plan Report, Des

Moines.

41

Tables

Table 1. Historic (IAGFA 2005), Menzel and Clark (2002) study (1997-2000), and current (2010- 2011) Topeka shiner detections in ten-digit (HUC10) and eight-digit (HUC8) hydrological units. Topeka shiner status for each HUC10 was determined to be stable (detected during 1997 to 2000 and during 2010 to 2011), at risk (detected during 1997 to 2000 and not detected during 2010 to 2011), or possibly extirpated (not detected during 1997 to 2000 or during 2010 to 2011). Percent decline is the proportion of the number HUC10s where Topeka shiners were not detected to the total number of HUC10s in their historic range. HUC10 HUC8 Topeka shiners detected Status 2010- Historic 1997-2000 2011 Headwaters North Raccoon Yes No No Possibly extirpated Cedar Creek - Upper North Raccoon Yes No No Possibly extirpated Camp Creek Yes Yes No At risk Indian Creek Yes Yes No At risk Upper North Raccoon River Yes Yes No At risk Lake Creek Yes Yes No At risk Purgatory Creek Yes Yes Yesa Stable Cedar Creek - Middle North Raccoon Yes Yes Yes Stable Middle North Raccoon River Yes Yes No At risk Hardin Creek Yes Yes Yesa Stable Buttrick Creek Yes Yes Yesa Stable East Buttrick Creek Yes Yes Yes Stable Lower North Raccoon River Yes Yes No At risk Lower Boone River Boone Yes Yes No At risk Middle Boone River Yes Yes No At risk White Fox Creek Yes No No Possibly extirpated Eagle Creek Yes Yes Yesa Stable Otter Creek Yes No No Possibly extirpated Prairie Creek Yes Yes No At risk Upper Des Bluff Creek Moines Yes No No Possibly extirpated Brushy Creek Yes Yes No At risk East Branch Iowa River Upper Iowa Yes No No Possibly extirpated

Percent decline 27% 73% aEvidence of reproduction noted during sampling – presence of young-of-year Topeka shiner Table 2. Habitat and biotic variables measured at stream sites (n = 67), off-channel sites (n = 27) in west-central Iowa. Means and standard deviations (SD) were calculated for variables in stream and off-channel sites separately and combined (n=94). Variable Description Stream Off-channel Combined Mean SD Mean SD Mean SD

Habitat variables

Canopy Mean proportion of canopy cover 25.67 23.20 17.88 23.90 23.43 23.54 Coarse Proportion of substrate > 0.2 cm (%) 36.21 21.52 2.47 7.43 26.52 24.07 CoarseGravel Proportion of coarse gravel substrate (1.6 – 6.4 cm; %) 10.72 11.09 1.23 4.54 8.00 10.57 Depth Mean depth (m) 0.33 0.30 0.59 0.30 0.41 0.25 DistAbsent Proportion of banks with no disturbance within view (%) 31.71 43.57 17.28 31.85 27.60 40.92 Eroding Proportion of bank eroded (%) 23.66 12.86 0.37 1.92 16.97 15.19 Gravel Proportion of gravel substrate (0.2 – 1.6 cm; %) 16.01 14.06 0.99 3.56 11.70 13.80 Sand Proportion of sand substrate (0.006 – 0.2 cm; %) 41.07 23.33 1.35 3.94 29.66 26.77 Silt Proportion of silt substrate (0.004 – 0.006 cm; %) 14.70 19.66 86.16 19.37 35.22 37.89 42 VegBank Proportion of bank covered by woody vegetation and non-woody 60.32 16.09 84.11 15.39 67.15 19.16 vegetation (%) VegCover Proportion of wetted area covered by terrestrial vegetation, aquatic 7.19 10.48 20.81 23.03 11.1 16.37 macrophytes or overhanging vegetation (%) Width Mean wetted width (m) 6.95 6.36 16.03 7.87 9.56 7.95 WidthtoDepth Ratio of mean wetted width (m) to mean depth (m) 23.21 15.85 32.11 16.90 25.76 16.57

Biotic variables

FHMinnow Number of fathead minnows per 100m2 2.71 6.99 34.75 89.38 11.91 49.80 Piscivore Number of piscivorous fishes (largemouth bass, smallmouth bass, 0.12 0.21 0.21 0.37 0.14 0.27 channel catfish, and northern pike) per 100m2 Sunfish Number of orangespotted sunfish and green sunfish per 100m2 0.99 2.02 12.34 25.51 4.25 14.54

43

Table 3. Confidence models selected (ΔAICc less than 2) from the combined, stream and off-channel candidate set of a priori logistic regression models as determined by Akaike’s information criterion for small sample size (AICc) ranking. Also included are the number of parameters in each model (k) and the Akaike’s weight (wi).

Confidence models k AICc ΔAICc wi Combined Model

VegCover, FHMinnow, Sunfish 4 84.9 0.00 0.46

Stream model

Coarse, VegCover 3 45.67 0.00 0.17 FHMinnow, Sunfish 3 46.24 0.58 0.13 VegCover, Piscivore 3 46.31 0.64 0.12 VegCover, Sunfish 3 46.31 0.64 0.12 VegCover, FHMinnow 3 46.40 0.73 0.12 Coarse, FHMinnow, Sunfish 4 47.35 1.68 0.07 Coarse, VegCover, VegBank 4 47.36 1.69 0.07

Off-channel model

FHMinnow 2 37.64 0.00 0.20 Depth, FHMinnow 3 38.83 1.19 0.11 FHMinnow, Pisc 3 38.90 1.25 0.11 VegCover, FHMinnow 3 39.50 1.85 0.08

44

Table 4. Model averaged coefficient estimates, standard error, 95% confidence intervals, and relative weights for the combined, instream, and off-channel models. Model parameters Estimate SE 95%CI Relative weight Combined Model VegCovera 0.031 0.015 0.002 .059 1.00 FHMinnowa 0.041 0.012 0.003 0.080 1.00 Sunfish 0.049 0.036 -0.075 0.172 1.00

Stream model

Coarse 0.018 0.021 -0.024 0.061 0.39 VegBank 0.016 0.029 -0.043 0.075 0.09 VegCover 0.0001 0.043 -0.085 0.086 0.75 FHMinnow 0.018 0.065 -0.112 0.149 0.39 Piscivore -0.690 2.337 -5.358 3.978 0.15 Sunfish -0.093 0.286 -0.664 0.178 0.40

Off-channel model

Depth -1.990 1.887 -5.884 1.905 0.22 VegCover 0.014 0.018 -0.022 0.051 0.16 FHMinnow 0.030 0.023 -0.017 0.077 1.00 Sunfish 0.023 0.025 -0.029 0.075 0.22 a Coefficients significantly different from zero

45

Figures

Figure 1. Location of study HUC8 basins within the Des Moines Lobe subecoregion of Iowa.

46

0.7 61 Absent 0.6 Present

0.5

0.4

0.3

0.2 Percent of Sites Percent 13 14

0.1 6

0.0 Stream Off-channel

Figure 2. Frequency of the stream and off-channel sites where Topeka shiner were and were not detected. Numbers of sites where Topeka shiners were and were not detected are shown above the bars.

47

Figuure 3. Topeka shiner occurrence documented in west-central Iowa watersheds in 1997 to 2000 (A) and in 2010 to 2011 (B).

48

Stream TS Absent Stream TS Present Off-channel TS Absent Off-channel TS Present Species richness isobar

1

8 6 1

0 2

8 2 2

4 1

2 1

Combined NMDS2 Combined

0 1

8

-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 8

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

Combined NMDS1

Figure 4. NMDS ordination of fish assemblages in stream (circles) and off-channel sites (triangles) combined. Grey symbols represent sites where Topeka shiner were not detected and black sites represent those where Topeka shiner were present. Isobars

represent the differing levels of species richness among all sites.

49

Sand Gravel Eroding WidthtoDepth Coarse

Width

Depth

Silt Stream NMDS2 Stream VegCover

TS Present TS Absent -1.5 -1.0 -0.5 0.0 0.5

-1.0 -0.5 0.0 0.5 1.0

Stream NMDS1

Figure 5. NMDS ordination of fish assemblages in stream sites. Grey symbols represent sites where Topeka shiner were not detected and black sites represent those where Topeka shiner were present.

50

TS Present TS Absent

CoarseGravel Off-channel NMDS2 Off-channel

Canopy

DistAbsent -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

-1 0 1 2

Off-channel NMDS1

Figure 6. NMDS ordination of fish assemblages in off-channel sites. Grey symbols represent sites where Topeka shiner were not detected and black sites represent those where Topeka shiner were present.

51

CHAPTER 3. GENERAL CONCLUSIONS

The preceding chapter provided identified distributional trends and habitat

associations of Topeka shiner in west-central Iowa. Novel information about Topeka shiner

biology and life history can be used to guide recovery efforts, but specific management

suggestions were not included in Chapter 2 due to their more speculative nature. This

chapter includes suggestions and recommendations that may be useful to managers involved

with Topeka shiner recovery and off-channel habitat restoration in Iowa.

Encouraging the growth of aquatic macrophytes in off-channel habitats may improve

the chances of Topeka shiner persistence in off-channel restorations. Submerged vegetation

cannot become established if an off-channel habitat is excessively turbid. Restored habitats

that receive water from drainage tiles might provide a better environment for macrophyte

growth and possibly improve overwinter survival of Topeka shiners. Water from drainage tiles is typically less turbid than water entering an off-channel habitat from overland flow.

This could increase the depth of the photic zone and encourage macrophyte growth. Water

from drainage tiles could buffer water temperatures in off-channel habitats. Although

Topeka shiners can survive very warm water temperatures in the summer, tile water could prevent the off-channel habitat from completely freezing during the winter.

Another way to encourage macrophyte growth would be to include shallow areas in off-channel restorations. Maintaining shallow areas may be challenging since water levels can frequently change in these dynamic habitats. However, gradual slopes along some part of the restoration could provide quality macrophyte habitat in high and low water conditions. 52

Deep areas may be important for overwinter survival, but shallow areas should also be

considered when designing off-channel restorations.

Off-channel habitat restoration is costly, so managers should target areas for

restoration that have the best return on investment. Restoration or conservation area

placement has been debated for years in the academic community (e.g., single large or several small [SLOSS] debate). One approach to prioritization would be to restore several

habitats that are within close proximity to each other, creating areas of higher Topeka shiner

density. Increasing the density of Topeka shiners in Cedar Creek, for example, could

decrease the likelihood of local extirpations by providing a significant source for the recolonization of nearby stream and off-channel habitats. On the other hand, targeting areas that are at higher risk of extirpation could be beneficial to Topeka shiners. If Topeka shiners are limited to only a few areas of high concentration, a change in local conditions (point source pollution, localized drought) could negatively impact the entire population. Buttrick and Cedar Creeks seem to have relatively stable levels of Topeka shiner, but Purgatory and

Eagle Creek could be areas that are most likely to lose Topeka shiners first. Restoration

efforts should be considered in these areas as well. It is important to note that Hardin Creek

(located between Buttrick and Cedar Creeks) has a substantial amount of Topeka shiners, but

no restoration has occurred in this watershed. As natural oxbows disappear in this area, off-

channel habitat restorations could become increasingly important here.

As suggested in Chapter 2, off-channel habitats and their associated streams should

be monitored as often as possible. Off-channel habitat restorations have only recently been

used to improve fish habitat, so there is little information to guide these projects. However,

wetland restorations to improve water quality and waterfowl habitat are quite common. 53

Perhaps wetland restoration protocols could be used to guide off-channel habitat restorations in Iowa. Nevertheless, post-restoration monitoring will be the only way to evaluate restoration effectiveness. Although monitoring requires the use of scarce monetary resources, it will provide information that can be used to increase the success of future

Topeka shiner recovery efforts.

54

APPENDIX A. TOTAL NUMBER OF TOPEKA SHINER DETECTED AT SITES IN WEST- CENTRAL IOWA IN 2010 TO 2011. DATE OF THE SAMPLE, STREAM NAME, SITE TYPE (STREAM AND OFF-CHANNEL) AND THE PRESENCE OF OFF-CHANNEL RESTORATION Date Stream name County Site Type # Topeka shiners Restoration

5/17/10 Hardin Creek Greene Off-channel 6 5/17/10 Hardin Creek Greene Off-channel 8 Buena 5/24/10 North Raccoon River Vista Off-channel 5/25/10 North Raccoon River Sac Off-channel Y 5/27/10 East Buttrick Creek Greene Off-channel 5/27/10 East Buttrick Creek Greene Stream 5/28/10 West Buttrick Creek Greene Off-channel 139 Y 6/1/10 Lake Creek Calhoun Off-channel 6/1/10 Lake Creek Tributary Calhoun Stream 6/4/10 Cedar Creek Calhoun Off-channel 37 Y 6/9/10 Lake Creek Calhoun Off-channel Y 6/9/10 Lake Creek Calhoun Off-channel Y 6/11/10 Purgatory Creek Calhoun Off-channel 11 7/2/10 East Buttrick Creek Webster Stream 7/6/10 Cedar Creek Calhoun Off-channel 26 7/8/10 East Cedar Creek Calhoun Off-channel 3 7/8/10 East Cedar Creek Calhoun Off-channel 26 7/8/10 East Cedar Creek Tributary Calhoun Stream 7/9/10 Lost Branch Creek Greene Stream 7/9/10 Cedar Creek Greene Off-channel Y 7/12/10 Boone River tributary Hamilton Stream 7/13/10 Pauper's Gulch Greene Stream 7/14/10 East Hardin Creek Greene Stream 5 7/15/10 Hardin Creek Greene Off-channel 7/15/10 Prairie Creek Calhoun Stream 7/21/10 North Raccoon River Dallas Off-channel 7/27/10 Hardin Creek Greene Off-channel 3 7/27/10 Hardin Creek Greene Off-channel 7/28/10 West Buttrick Creek Greene Off-channel 14 Y 7/29/10 Reading Creek Greene Stream 8/3/10 Eagle Creek Hamilton Off-channel 4 8/10/10 Prairie Creek Humboldt Off-channel 8/16/10 Boone Tributary Wright Stream 8/17/10 White Fox Creek tributary Wright Stream 8/18/10 Eagle Creek Wright Off-channel 14 8/19/10 West Buttrick Creek Webster Off-channel 143 8/20/10 Eagle Creek Wright Stream 8/27/10 Hardin Creek Webster Stream 2 55

Appendix A. continued

5/12/11 North Raccon tributary Greene Stream 5/13/11 Skillet Creek Webster Stream 5/16/11 Elm Branch Dallas Stream 5/17/11 North Raccoon tributary Greene Stream 5/18/11 Lyon Creek tributary Hamilton Stream 5/19/11 Short Creek Greene Stream 5/23/11 Purgatory Creek Calhoun Stream 5/24/11 Buck Run Calhoun Stream 5/26/11 Brushy Creek Outlet Webster Off-channel 5/27/11 Camp Creek tributary Calhoun Stream 5/31/11 Snake Creek Greene Stream 6/1/11 North Raccoon tributary Sac Stream 6/3/11 Lost Grove Creek Webster Stream 1 6/7/11 East Buttrick Creek Greene Stream 6/8/11 North Raccoon tributary Carroll Stream 6/8/11 Doe Brook Carroll Stream 6/9/11 Purgatory Creek Carroll Stream 6/13/11 Cedar Creek Calhoun Stream 9 6/14/11 Boone River Hamilton Off-channel 6/15/11 Boone River tributary Hamilton Stream 6/20/11 Brewers Creek Hamilton Stream 6/22/11 Eagle Creek tributary Wright Stream 1 6/24/11 White Fox Creek tributary Wright Stream 7/8/11 Eagle Creek tributary Wright Stream 1 7/11/11 Caton Branch Boone Stream 7/12/11 Prarie Creek Kossuth Stream 7/13/11 North Racoon tributary Dallas Stream 7/15/11 Boone River tributary Sac Stream 7/15/11 Black Hawk Outlet Sac Stream Black Hawk Lake Outlet 7/15/11 Creek Sac Stream 7/24/11 Prairie Creek Kossuth Stream 7/27/11 West Fork Camp Creek Calhoun Stream 7/27/11 Camp Creek Calhoun Stream 7/28/11 Camp Creek Calhoun Stream 7/28/11 Boone River tributary Wright Stream 8/1/11 White Fox Creek Wright Stream 8/3/11 Prairie Creek Calhoun Stream 8/3/11 West Buttrick Creek Greene Off-channel 194 Y 8/4/11 White Fox Creek Hamilton Stream 8/4/11 Iowa River Hancock Stream 8/8/11 Snake Creek Greene Stream 8/9/11 Boone River tributary Hamilton Stream 56

Appendix A. continued

8/9/11 Buck Creek Hamilton Stream 8/10/11 Lake Creek Calhoun Stream 8/10/11 West Fork Camp Creek Calhoun Stream 8/10/11 Camp Creek Calhoun Stream 8/11/11 Indian Creek Sac Stream 8/11/11 North Raccoon tributary Sac Stream 8/11/11 North Raccoon River Sac Stream 8/11/11 Cedar Creek Sac Stream 8/11/11 Elk Run Carroll Stream 8/12/11 Purgatory Creek Carroll Stream 8/12/11 North Raccoon River Calhoun Stream 8/12/11 North Raccoon River Carroll Stream 8/15/11 North Raccoon River Greene Stream 8/18/11 Otter Creek Wright Stream

APPENDIX B. PERCENT OCCURRENCE AND CATCH PER UNIT EFFORT (CPUE) OF FISH SPECIES COLLECTED IN CENTRAL IOWA STREAM AND OFF-CHANNEL SITES. DATA ARE PRESENTED SEPARATELY FOR SITES WHERE TOPEKA SHINER WERE PRESENT AND ALL SITES SAMPLED. SPECIES ARE LISTED IN DESCENDING ORDER OF % OCCURRENCE IN ALL STREAM SITES % Occurrence CPUE (number per 100 m2) TS Present All TS Present All Off- Off- Off- Off- Common name Scientific name Stream channel Stream channel Stream channel Stream channel Bluntnose minnow Pimephales notatus 100 28.57 95.52 37.04 54.48 4.56 427.48 115.98 Creek chub Semotilus atromaculatus 100 42.86 91.04 40.74 46.17 2.73 735.25 9.52 Johnny darter Etheostoma nigrum 83.33 21.43 85.07 18.52 10.64 0.99 147.75 3.47 Bigmouth shiner Notropis dorsalis 83.33 21.43 85.07 14.81 62.22 9.86 782.74 32.85 Blacknose dace Rhinichthyes atratulus 83.33 0 83.58 0 28.75 0 719.33 0 Common shiner Luxilus cornutus 100 35.71 82.09 33.33 6.98 21.76 343.72 28.47 Central stoneroller Campostoma anomalum 100 0 80.6 0 13.28 0 300.31 0 57 White sucker Catostomus commersoni 100 50 77.61 44.44 18.3 14.04 206.51 32.85 Fathead minnow Pimephales promelas 66.67 100 76.12 88.89 17.65 385.51 181.51 495.48 Spotfin shiner Cyprinella spiloptera 83.33 21.43 71.64 29.63 15.49 2.12 213.6 47.9 Sand shiner Notropis stramineus 83.33 28.57 67.16 37.04 34.25 0.43 225.12 266.2 Green sunfish Lepomis cyanellus 66.67 92.86 64.18 81.48 4.06 55.14 62.66 83.35 Golden redhorse Moxostoma erythurum 83.33 7.14 46.27 7.41 1.92 0.29 25.15 0.41 Blackside darter Percina maculata 33.33 0 46.27 0 2.44 0 26.57 0 Northern hog sucker Hypentelium nigricans 50 0 44.78 0 0.63 0 10.96 0 Brassy minnow Hybognathus hankinsoni 33.33 71.43 41.79 44.44 1.37 10.12 68.65 10.19 Moxostoma Shorthead redhorse macrolepidotum 33.33 0 34.33 11.11 0.25 0 10.33 0.37 Brook stickleback Eucalia inconstans 50 35.71 31.34 33.33 3.56 32.09 28.81 33.97 Black bullhead Ameiurus melas 16.67 64.29 28.36 62.96 0.43 184.38 10.78 293.95 Yellow bullhead Ameiurus natalis 33.33 14.29 26.87 11.11 2.01 0.6 10.33 0.63 Smallmouth bass Micropterus dolomieu 16.67 0 26.87 3.7 0.02 0 4.79 0.12 Fantail darter Etheostoma flabellare 33.33 0 26.87 0 5.83 0 53 0 Common carp Cyprinus carpio 16.67 78.57 25.37 70.37 0.23 60.55 47.69 79.67 Appendix B. continued

Rosyface shiner Notropis rubellus 33.33 0 23.88 7.41 5.36 0 56.47 0.15 Hornyhead chub Nocomis biguttatus 50 0 23.88 0 2.68 0 38.44 0 Orangespotted sunfish Lepomis humilis 33.33 57.14 22.39 62.96 0.56 136.88 3.84 187.1 Channel catfish Ictalurus punctatus 0 0 22.39 7.41 0 0 5.01 0.5 Largemouth bass Micropterus salmoides 16.67 50 20.9 48.15 0.38 4.99 5.6 11.54 Stonecat Noturus flavus 16.67 0 20.9 0 0.05 0 3.94 0 Suckermouth minnow Phenacobius mirabilis 16.67 0 19.4 0 0.07 0 7.22 0 Bluegill Lepomis macrochirus 0 14.29 17.91 29.63 0 0.48 22.2 25.29 River carpsucker Carpoides carpio 0 0 17.91 18.52 0 0 2.43 3.35 Gizzard shad Dorosoma cepedianum 0 0 14.93 3.7 0 0 11.76 0.25 Northern rock bass Ambloplites rupestris 33.33 7.14 14.93 3.7 0.79 0.06 4.36 0.06 58 Quillback carpsucker Carpiodes cyprinus 0 0 14.93 0 0 0 1.64 0 Highfin carpsucker Carpiodes velifer 0 0 10.45 3.7 0 0 1.07 0.41 Silver redhorse Moxostoma anisurum 16.67 0 10.45 0 0.15 0 0.99 0 Topeka shiner Notropis topeka 100 100 8.96 51.85 2.27 100.18 2.27 100.18 Banded darter Etheostoma zonale 16.67 0 7.46 0 0.4 0 1.14 0 Slender madtom Noturus exilis 0 0 7.46 0 0 0 0.43 0 Slenderhead darter Percina phoxocephala 16.67 0 7.46 0 0.02 0 1.21 0 Walleye Sander vitreus 0 0 7.46 0 0 0 0.13 0 White bass Morone chrysops 0 0 7.46 0 0 0 1.13 0 Southern redbelly dace Phoxinus erythrogaster 0 0 5.97 0 0 0 3.39 0 Yellow perch Perca flavescens 0 7.14 4.48 14.81 0 0.26 1.68 2.3 Northern pike Esox lucius 0 7.14 4.48 11.11 0 0.1 0.24 1.36 Freshwater drum Aplodinotus grunniens 0 0 4.48 3.7 0 0 0.07 0.05 Smallmouth buffalo Ictiobus bubalus 0 0 4.48 0 0 0 0.26 0 Flathead catfish Polyodictis olivaris 0 0 2.99 0 0 0 0.04 0 Appendix B. continued

Iowa darter Etheostoma exile 0 0 2.99 0 0 0 7.44 0 Yellow bass Morone mississippiensis 0 0 2.99 0 0 0 0.1 0 Bigmouth buffalo Ictiobus cyprinellus 0 0 1.49 22.22 0 0 1.15 4.22 White crappie Pomoxis annularis 0 7.14 1.49 22.22 0 1.34 0.21 2.16 Black crappie Pomoxis nigromaculatus 0 0 1.49 11.11 0 0 0.23 48.54 Golden shiner Notemigonus crysoleucus 0 0 1.49 7.41 0 0 0.21 0.06 Emerald shiner Notropis atherinoides 0 0 1.49 3.7 0 0 0.1 0.05 Red shiner Notropis lutrensis 0 0 1.49 0 0 0 0.1 0 Bullhead minnow Pimephales vigilax 0 0 0 3.7 0 0 0 0.33 Longnose gar Lepisosteus osseus 0 0 0 3.7 0 0 0 0.02

59