t -

BEFORE THE ILLINOIS POLLUTION CONTROL BOARD

IN THE MATTER OF :

)

PROPOSED AMENDMENTS TO

)

R 04-25

DISSOLVED OXYGEN STANDARD )

35111. Adm. Code 302.206

)

NOTICE OF FILING

TO: See Attached Service List

PLEASE TAKE NOTICE that I have today filed with the Office of the Clerk of the

Pollution Control Board the following documents :

WRITTEN TESTIMONY OF DR JAMES E. GARVEY

a copy of which is served upon you .

Dated: October 4, 2006

Roy M . Harsch

GARDNER CARTON & DOUGLAS LLP

191 Wacker Drive - Suite 3700

Chicago, Illinois 60606

(312) 569-1000

ENC

of Its

ASSOCIATIO 0 WASTEWATER

S,

r

THIS FILING PRINTED ON RECYCLED PAPER

---

CERTIFICATE OF SERVICE

The undersigned certifies that a copy of the foregoing :

WRITTEN TESTIMONY OF DR. JAMES E. GARVEY

was filed electronically with the Clerk of the Illinois Pollution Control Board and served upon

the parties to whom said Notice is directed by electro

'

Ily on Wednesday, October 04, 2006

.

41

---

Fred L. Hubbard

16 West Madison

P

.O. Box 12

Danville, IL 61834-0012

Bernard Sawyer

Metropolitan Water Reclamation District

6001 W

. Pershing Rd .

Cicero, IL 60650-4112

Claire A. Manning

Posegate & Denes, P.C.

I i I N. Sixth Street

Springfield, IL 62705

Deborah J . Williams

Stefanie N . Diers, Assistant Counsel

Illinois EPA

1021 North Grand Avenue

P.O. Box 19276

Springfield, IL 62794-9276

Dorothy M. Gunn

Illinois Pollution Control Board

100 W. Randolph Street - Suite 11-500

Chicago, IL 60601

Frederick D . Keady

Vermilion Coal

1979 Johns Drive

Glenview, IL 60025

James T. Harrington

Ross & Hardies

150 North Michigan Avenue - Suite 2500

Chicago, IL 60601-7567

John Donahue

City of Geneva

22 South First Street

Geneva, IL 60134-2203

Service List

R2004-025

Alex Messina

Illinois Environmental Regulatory Group

3150 Roland Avenue

Springfield, IL 62703

Charles W

. Wesselhoft

Ross & Hardies

150 North Michigan Avenue - Suite 2500

Chicago, IL 60601-7567

Connie L. Tonsor

Illinois EPA

1021 North Grand Avenue

P

.O

. Box 19276

Springfield, IL 62794-9276

Dennis L Duffield

City of Joliet

Department of Public Works and Utilities

921 E. Washington Street

Joliet, IL 60431

Erika K

. Powers

Barnes & Thornburg

I N. Wacker - Suite 4400

Chicago, IL 60606

James L

. Daugherty

Thom Creek Basin Sanitary District

700 West End Avenue

Chicago Heights, IL 60411

Joel J. Sternstein

Office of the Attorney General

188 West Randolph Street - 20th Floor

Chicago, IL 60601

Stanley Yonkauski

Illinois Department of Natural Resources

One Natural Resources Way

Springfield, IL 62702-1271

---

Katherine D . Hodge

Hodge Dwyer Zeman

3150 Roland Avenue

P.O. Box 5776

Springfield, IL 62705-5776

Lisa Frede

Chemical Industry Council of Illinois

2250 E. Devon Avenue - Suite 239

Des Plaines, IL 60018-4509

Matthew J. Dunn

Office of the Attorney General

188 West Randolph - 20a' Floor

Chicago, IL 60601

Mike Callahan

Bloomington Normal Water Reclamation Dist.

PO Box 3307

Bloomington, IL 61702-3307

Richard McGill

Illinois Pollution Control Board

100 W. Randolph Street

-

Suite 11-500

Chicago, IL 60601

Stephanie N . Diers

IEPA

1021 North Grand Avenue East

P.O. Box 19276

Springfield, IL 62794-9276

Susan M . Franzetti

10 South LaSalle Street

- Suite 3600

Chicago, IL 60603

V icky McKinley

Evanston Environment Board

233 Grey Avenue

Evanston, IL 60202

Edward Hammer

Larry Cox

Downers Grove Sanitary District

2710 Curtiss Street

Downers Grove, IL 60515

Margaret P

. Howard

2601 South Fifth Street

Springfield, IL 62703

Michael G. Rosenberg, Esq .

Metropolitan Water Reclamation District

100 East Erie Street

Chicago, IL 60611

Richard Lanyon

Metropolitan Water Reclamation District

100 East Erie Street

Chicago, IL 60611

Sanjay K . Sofat

Illinois EPA

1021 North Grand Avenue East

P.O. Box 19276

Springfield, IL 62794-9276

Sue Schultz

Illinois American Water Company

300 North Water Works Drive

P.O. Box 24040

Belleville, IL 62223-9040

Tom Muth

Fox Metro Water Reclamation District

682 State Route 31

Oswego, IL 60543

W

.C

. Blanton

Blackwell Sanders Peper Martin LLP

2300 Main Street -

Suite 1000

Kansas City, MO 64108

Albert Ettinger

---

U.S. Environmental Protection Agency

WQ-16J

77 West Jackson Boulevard

Chicago, IL 60604

Todd Main

Director of Policy and Planning

Friends of the Chicago River

407 S. Dearborn - Suite 1580

Chicago, IL 60605

N

. LaDonner Driver

Illinois Environmental Regulatory Group

3150 Roland Avenue

Springfield, IL 62703

Marc Miller, Senior Policy Advisor

Michael J. Fischer, Policy Advisor

Office of Lt . Governor Pat Quinn

Room 214 State House

Springfield, IL 62706

CHOU 12490125 .1

Senior Staff Attorney

Environmental & Law Policy Center

35 E. Wacker-Suite 1300

Chicago, IL 60601

Irwin Polls

Ecological Monitoring and Assessment

3206 Maple Leaf Drive

Glenview, IL 60025

Tracy Elzemeyer

General Counsel

American Water Company

727 Craig Road

St. Louis, MO 63141

Dr. Thomas J . Murphy

2325 N. Clifton Street

Chicago, IL 60614

---

BEFORE THE ILLINOIS POLLUTION CONTROL BOARD

IN THE MATTER OF :

)

PROPOSED AMENDMENTS TO

)

R 04-25

DISSOLVED OXYGEN STANDARD

)

35 111. Adm. Code 302.206

)

Written Testimony of James E

. Garvey

I thank the Illinois Pollution Control Board (Board) for allowing me to present my

testimony. My name is Dr. James E

. Garvey, Associate Professor of Zoology and

Associate Director of the Fisheries and Illinois Aquaculture Center at Southern Illinois

University Carbondale (SIUC)

. I also hold several other appointments such as Chair of

the American Fisheries Society (AFS) Farm Bill Advisory Task Force, Executive Officer

of the Illinois Chapter of the AFS, Member of the US Army Corps of Engineers

Environmental Management Program - Project Sequencing Team, and North Central

Representative of the Early Life History Section of the AFS . As you know, I am an

aquatic ecologist with an active research program that revolves around environmental and

human-induced factors influencing the abundance and distribution of fishes in lakes and

rivers. I have published well over forty publications that are widely cited in the discipline

of fisheries, aquatic ecology, and general ecology .

I also have an active graduate training

program . My graduate students often join natural resource agencies such as the Illinois

EPA, the US Fish and Wildlife Service, and the Missouri Department of Conservation .

My participation in this process began over two years ago when the Illinois

Association of Wastewater Agencies (IAWA) asked Dr. Matt Whiles and me to evaluate

the current dissolved oxygen standard in Illinois . After an extensive literature review,

we

generated a report that stated that the current standard is too simplistic for the diverse

I

---

waters of Illinois . We supported many of the recommendations that were developed in

the US EPA National Criteria Document (NCD) for dissolved oxygen .

Review

Over the course of two years, much data collection, literature review, and

discourse among stakeholders have occurred

. I have attended all the stakeholder

meetings and hearings before the Board

; I have had the opportunity to review all the

technical information and data presented in this rulemaking process thanks to the

cooperation of the stakeholders . The end result of this process is that the

recommendations that Dr . Whiles and I set forth largely have been supported

. I have

appeared before the Board on several occasions to present my findings . Recall, we

recommended that a two-season standard be adopted throughout the state.

During March

through June, when the majority of early life stages of many fishes and other aquatic

organisms are produced, we recommended a standard dissolved oxygen concentration be

met that provides sufficient oxygen to support the metabolic needs of eggs and larvae .

During this time of year, streams are typically flowing, primary productivity is

accelerating but not peaking, and temperatures are cool to moderate . Thus, high

dissolved oxygen concentrations are expected to be available to young aquatic organisms ;

this expectation has been well supported by my findings described in previous testimony

.

The literature and growing state-wide oxygen data set demonstrate that, for warm-water,

low gradient systems common in Illinois, concentrations should not decline below 5

mg/L and weekly averages should not decline below 6 mg/L . We also suggested a 30-

day running average of 5.5mg/L, which has little biological support in my view, but is

recommended in the NCD.

2

---

As temperatures increase during summer, increased biological activity and

water's reduced oxygen capacity should reduce dissolved oxygen concentrations,

particularly during night. Evidence is mounting that the majority of reproduction of

aquatic organisms in Illinois either occurs before July 1 (see Csoboth 2006 thesis, SIUC

;

Exhibit 1) or late-spawning organisms have early life stages that are tolerant to low

dissolved oxygen concentrations (e.g.,

freshwater mussels) . Thus, we recommended that

during July through February, Illinois adopt a daily acute minimum of 3

.5 mg/L and a

seven-day average of daily minima of 4 mg/L . In previous testimony before the Board, I

have demonstrated that streams that meet these dissolved oxygen conditions appear to

contain diverse, robust biological assemblages . Those that do not are typically impaired

.

During the past year, the Illinois Department of Natural Resources (IDNR) and

the Illinois Environmental Protection Agency (IEPA) have proposed an alternative, two-

tier oxygen standard for the state and have expended much energy to develop it .

The

"general use" tier is very similar to the IAWA state-wide recommendation with slightly

higher concentrations . Also, the criteria for early life stages are extended through July .

In addition, the agencies recommended an "enhanced oxygen" tier for streams that

contain fishes and invertebrates that were found by Ohio Environmental Protection

Agency to occur in Ohio waters with high average oxygen concentrations

. My concern

about this approach is that the selection of streams based solely on associations between

aquatic organisms and average oxygen concentrations ignores other potential causal

factors such as habitat quality, gradient, and temperature .

Thus, coining these organisms

as "oxygen sensitive" and then using them to select enhanced tier waters may by

completely spurious

. Only through experiments that establish causality between oxygen

3

---

tolerance and fish life processes can tolerance be assessed . Again, these issues have been

addressed in previous testimony when I described the research by Smale and Rabeni

published in the Transactions ofthe American Fisheries Society .

Recall, these

investigators used a combination of lab assays and surveys to develop an index of oxygen

sensitivity in Missouri streams .

Overview of Testimony

I present results that continue to support the recommendations in the Garvey and Whiles

report . First, I review the results of recent peer-reviewed papers that show that dissolved

oxygen concentrations in Illinois streams are difficult to predict and largely influenced by

characteristics of stream habitat and morphology. I then explore the implications of the

two-tier oxygen standard for Illinois using data that were collected both by the IDNR and

JEPA as well as data that were collected by IAWA members. In my view, the most

compelling results derive from stream segments slated for enhanced dissolved oxygen

protection by the proposed IDNR/IEPA two-tier approach . As I analyzed these data, it

became apparent that many of these segments likely violate both the IDNR/IEPA and

perhaps the IAWA proposed standards, even though "enhanced oxygen" taxa are present

in the streams

. Further, daily discharge (i

.e., volume of water moving per second through

the stream) explained as much as 50% of the variation in daily median and minimum

dissolved oxygen concentrations in several of these systems . Thus, the physical

characteristics of streams interacting with flow largely drove much of the oxygen

dynamics. In my view, this further complicates any attempts to fit a single standard to

any stream in the State and renews the urgent need to develop tiered, habitat-based

criteria that incorporate how discharge affects aquatic communities and water quality .

4

---

Literature Review

Several papers that were generated by Dr

. Mark David and colleagues at the

University of Illinois Urbana-Champaign through support by the C-FAR program

recently have been published (Exhibit 2)

. Although the general expectation was for

dissolved oxygen dynamics in their research streams in Illinois to be affected by nutrient

loading, they found that stream physical characteristics, primarily basin shape and its

propensity to hold organic matter and intercept light, were more important in influencing

oxygen concentrations

. As I have argued throughout this process and in the original

IAWA-sponsored report, these results indicate that stream physical characteristics

trump

water quality and need to be the primary focus of standard development

.

Analysis of Historical Grab Data and 2004-2005 Continuous Data

Illinois DNR/EPA provided me with "grab" dissolved oxygen data collected

during 1994 through 2003 in streams that have fully met their aquatic use designation

. In

addition, they provided data from 2004 and 2005 collected with semi-continuous data

logging probes in streams that have been tapped for inclusion in the "enhanced oxygen"

tier. I sent the results I present below to Mr

. Matt Short and Mr

. Joel Cross for their

review

. As of the date I am drafting this testimony, they have not responded

. The grab

data demonstrate that median dissolved oxygen concentration declines during June

through August relative to other months (Exhibit 3)

. Concentrations did decline below a

benchmark of 5 mg/L during the summer months, although rarely

. Given that these grabs

were typically taken during the day, it is not surprising that relatively low dissolved

oxygen concentrations were not frequently encountered.

5

---

The continuous data demonstrated that dissolved oxygen in "enhanced" segments

more frequently declined below 5 mg/L and even occasionally below 3

.5 mg/L (Exhibit

3). These low concentrations which often exceeded both the IAWA and DNR/EPA

proposed standards typically occurred during the night through dawn

. Interestingly, these

enhanced-tier segments more frequently (up to 20% of observations) exceeded the

DNR/EPA minimum of 5 mg/L during July than the IAWA proposed standard of 3 .5

mg/L during that month (Exhibit 3)

. The streams that contained "oxygen sensitive"

species failed to meet the standard set for them by the IDNR/EPA proposal .

On 24 April 2006, Mr

. Toby Frevert sent a letter to Mr. Dennis Streicher

including several disclaimers about the above data set .

He indicated that the grab data

were a worst case scenario, including only data collected in the early morning hours. On

the contrary, the data set I received from the agencies and recently sent back to them for

confirmation included grab data that were collected during morning through afternoon

(median collection time was 1100 hours, with times as late as 1700 ; Exhibit 3) . Thus, it

appears to me that the data represent the range of daily conditions that affect oxygen

concentrations. Time of day was positively related to DO concentration in this data set,

but explained less than 1 % of the variation . Although the continuous data show that the

enhanced streams cannot meet the IDNR/IEPA expected standard, Mr

. Frevert noted that

these data included results from 2005 when a drought gripped much of the state

. Because

these results were collected under extreme conditions, he argued they should be

discounted. I respectfully disagree

.

Few laws exist in the tangled and complex discipline of ecology. However, one

of the most commonly agreed tenets in our discipline is Liebig's Law of the Minimum,

6

---

taught in every general ecology course, including my own at SIUC

. Liebig aptly noted

that the distribution of all living organisms will not be dictated by the average conditions,

but rather the availability of the most limiting condi ion . This condition does not always

have to be limiting, but only when organisms are ex eriencing some critical period such

as reproduction or growth

. The condition could be an occasionally limited nutrient, or in

our case, oxygen. In other words, the occasional "worst case" scenario which limits the

oxygen available to the local fauna will determine the species composition and

abundance present at all times . Only by identifying the limiting conditions, in other

words the acute minimum oxygen concentration, can we determine what should be

present through time. The extreme drought conditions in the enhanced streams likely

provided the worst case scenario and thereby insight into what the acute minimum should

be to support a diverse aquatic assemblage . The proposed minimum standard of 3 .5

mg/L was rarely exceeded in these streams (Exhibit 3) and likely is near that extreme

lower limit.

Illinois Water Survey Data

Illinois DNR via Ms . Ann Holtrop provided me with grab dissolved oxygen data from

various studies compiled through the Illinois State Water Survey . These data extend

from the early 1970s through the 1990s . After reviewing the reports from which these

data were collected (see Exhibit 4), it was clear that the 20,101 individual observations

that I analyzed were collected in many ways. Even given this caveat, I thought it might

be interesting to determine whether average dissolved oxygen concentrations "improved"

in Illinois surface waters through time as nutrient loading abated during the past thirty

years as a function of the Clean Water Act . I was rather surprised to find that no real

7

---

pattern occurred through the decades, with concentrations varying widely among sites

and years for which data were available

. As per the results emerging from Dr

. David's

laboratory as well as the results I will present below, it appears that oxygen

concentrations in streams are likely influenced by habitat and its interactions with many

other factors, of which nutrient loading is but one component

.

IAWA 2005 and 2006 Semi-Continuous Monitoring

Several IAWA members have installed semi-continuous dissolved oxygen loggers (15-60

minute intervals depending on the source) in streams that are in segments slated for

enhanced tier standards by the agencies

. Segments for which I have received data are on

the Fox, DuPage, Kickapoo, Rock, and Vermilion Rivers (Exhibit 5

; 24,575 individual

observations)

. With the exception of the Fox River where the data derive from 2005, the

remainder of the data derives from summer 2006

. I also procured USGS daily

monitoring data for discharge from gauging stations near the river segments to test the

hypothesis that discharge drives much of the variation in dissolved oxygen concentrations

in low-gradient Illinois streams . The IAWA members who collected the data have

reviewed these summary results.

Dynamics of dissolved oxygen vary widely among the enhanced tier stream

segments (Exhibit 5), from daily concentrations varying widely in the Fox River to less

so in the Vermilion River

. Both median and minimum daily dissolved oxygen

concentrations typically declined as the summer progressed in the Fox, DuPage,

and

Kickapoo Rivers, but not the others (Exhibit 5)

. Probably the most compelling result is

the linear or log-linear relationship between daily discharge and median and minimum

daily dissolved oxygen concentrations in the streams (Exhibit 5)

. In 2005 for the Fox

8

---

River, dissolved oxygen concentrations declined sharply with declining daily discharge

(Exhibit 5)

. Conversely, in the other streams during 2006, dissolved oxygen

concentrations were either unrelated to discharge or negatively related (Exhibit 5)

. I

could speculate broadly about the underlying mechanisms including flow-related

biochemical oxygen demand, hypoxic groundwater intrusion, and changes in water

quality due to run-off

. Regardless of the underlying causes, given that discharge can

explain up to 50% of the variation in dissolved oxygen concentrations during both severe

drought (2005) and non-drought years, this issue needs to be incorporated into standard

development and interpretation

.

I applied both the enhanced tier standard and the proposed IAWA standard to the

semi-continuous data

. Typically, both standards demonstrate that several of the stream

segments including those in the DuPage, Fox, and Kickapoo Rivers fail to meet the

season-dependent acute minima, even given the proposed enhanced status of these

systems (Exhibit 6) . This is not surprising given that some portions of the DuPage and

Fox Rivers are currently listed with low dissolved oxygen as a probable cause for

impairment (see map in Exhibit 5)

. However, the Rock River which is listed as impaired

due to low oxygen did not fail to meet any of the minimum criteria (Exhibit 6)

.

Seven-day means ending in July for IAWA and August for the IDNR/IEPA

proposals were generally insensitive (Exhibit 6)

.

Interestingly, the IAWA proposed 7-

day minimum standard of 4 mg/L which applies during July through February generated

more violations than the IDNR/IEPA 7-day mean minimum of 4

.5 mg/L which starts in

August (Exhibit 6)

. Although I did not expect this to occur, apparently applying the

mean minimum criterion during July as per the IAWA proposal is more sensitive

.

9

---

Because the daily variation in dissolved oxygen concentrations differs more than the

daily average (i .e., it is the variation not the mean that is sensitive), it appears that the

mean-minimum criterion is more sensitive to frequent declines in oxygen during the

summer. In my view, it appears that many of these streams, particularly the Fox River,

fail to provide adequate oxygen for aquatic life during part of the summer

. This causes

me to question the linkage between the aquatic assemblages used to select the sites for

enhanced status and oxygen needs of the resident organisms

.

Summary

One of the major conclusions of the Garvey and Whiles report was that we have

much to learn about associations between aquatic organisms and spatial and temporal

heterogeneity in dissolved concentrations of warm surface waters in the US .

Since that

report was completed, I have had the privilege of exploring this issue in depth and

receiving some unprecedented (and fun) data sets

. As Liebig stated generally for all

ecology, it is clear that oxygen can become a limiting dissolved gas for aquatic organisms

and, below some threshold concentration, we should expect to see deleterious effects

and

reductions in species composition and abundance. To this date, all the data I have

reviewed suggest that a threshold does exist and that it occurs during the summer when

concentrations are less than or equal to 3 mg/L as stated in the NCD and the Garvey and

Whiles report . If a stream remains consistently above this level (i .e., never violates a 3 .5

mg/L minimum), oxygen is no longer limiting for life and some other factor then limits

organisms . . . .probably habitat

. All of the stream data and the literature (see Dr

. David's

research) support this view .

10

---

I favor scrapping dissolved oxygen as a standard altogether

. Although under

extreme conditions it can become limiting

(e.g.,

in the Gulf of Mexico hypoxic zone),

variable or low concentrations are largely a symptom of habitat problems and their

interactions with other factors such as chemical and biological pollutants

. .

. and, as this

testimony suggests, discharge

. However, given that this is not currently a possibility,

it

appears that the set of standards proposed in the Garvey and Whiles report stand the test

of the data and should be adopted in the interim

. I do urge the stakeholders to move

rapidly toward a habitat-based tier designation where oxygen is but one of a suite of

physical and chemical parameters used to diagnose root causes and develop sound

solutions .

CH02/ 22464211

.1

II

---

EARLY LIFE HISTORY OF FISHES IN RESTORED AND UNRESTORED

BACKWATERS

by

Laura A

. Csoboth

B.S ., University of Delaware, 2002

A Thesis

Submitted in Partial Fulfillment of the Requirements for the

Master of Science Degree

Department of Zoology

in the Graduate School

Southern Illinois University Carbondale

December 2006

---

THESIS APPROVAL

EARLY LIFE HISTORY OF FISHES IN RESTORED AND UNRESTORED

BACKWATERS

By

Laura A

. Csoboth

A Thesis Submitted in Partial

Fulfillment of the Requirements

for the Degree of

Master of Science Degree

in the field of Zoology

Approved by :

Dr James E

. Garvey, Chair

Dr

. Frank M. Wilhelm

Dr. Matthew R. Whiles

Graduate School

Southern Illinois University Carbondale

August 25, 2006

---

AN ABSTRACT OF THE THESIS OF

LAURA A

. CSOBOTH, for the Master of Science degree in ZOOLOGY, presented on

AUGUST 25, 2006, at Southern Illinois University Carbondale

.

TITLE : EARLY LIFE HISTORY OF FISHES IN RESTORED AND UNRESTORED

BACKWATERS

MAJOR PROFESSOR : Dr. James E. Garvey

River modifications have degraded backwaters, reducing critical habitat for larval

fishes

. Restoration projects, such as on the Illinois River's Swan Lake backwater, may

be improving fish spawning and nursery areas

. During 2004 and 2005,1 (1) assessed

restored and unrestored backwaters as fish nurseries by quantifying the density of larval

fishes and their zooplankton prey and (2) investigated movement of larval fish between

the river and Swan Lake by setting drift nets on a diel and seasonal basis .

A flood pulse occurred during June 2004, but no spring or summer flood event

during 2005

. When restored and unrestored backwater sites were compared to the

Illinois River, families of fish changed between years, with fluvial-dependent taxa

present during the flood pulse of 2004 and relatively absent during the drought of 2005

.

Higher fish densities and larger fish larvae were collected in backwaters as compared to

the Illinois River during both years, probably due to abundant zooplankton and warmer

temperatures

. Larval drift was positively related to water velocity during 2004, and an

estimated 32 .3 million larvae drifted into Swan Lake

. No other relationships between

drift direction and abiotic variables occurred

. In 2005, drift was related to larval density,

probably due to the lack of a flood pulse.

I

---

Backwater-river connectivity in the Swan Lake HREP was preserved through the

construction of water control stop-log structures, which allowed continual access by all

fishes

. The backwater management plans at Swan Lake enhanced habitat diversity and

realized beneficial effects by creating a diverse and abundant age-O fish community

.

ii

---

ACKNOWLEDGMENTS

I thank my advisor, Dr

. Jim Garvey, for his continual support and guidance

.

Funding was provided by the U .S

. Army Corps of Engineers, St . Louis District .

Technical and logistical support was provided by the Two Rivers National Wildlife

Refuge, Chad Dolan of the Illinois Natural History Survey, and Neil Booth of the Illinois

Department of Natural Resources . I thank countless individuals within the Department of

Zoology, Southern Illinois University Carbondale, particularly my committee members

Dr. Matt Whiles and Dr

. Frank Wilhelm

. I want to recognize students, faculty, and staff

of the Fisheries and Illinois Aquaculture Center and particularly

: R. Brooks, K.

DeGrandchamp, and D

. Schultz for their assistance throughout all phases of this project

;

Dr. Burr, D. Knuth, A . Lohmeyer, and S

. Tripp for their help with larval fish

identification; R. Colombo, R. Lane, and Q

. Phelps for their assistance with the analysis

and review of my manuscripts

. I am grateful for the diligent patience of many

undergraduate students and, particularly two extra help employees, Michielle Lyman and

Kristal Derr, who processed countless larval fish and zooplankton samples

. Finally,

many grateful thanks go to my family for their endless support and encouragement and

my friends, both new and old .

iii

---

Chapter

ABSTRACT

i

ACKNOWLEDGMENTS

iii

LIST OF TABLES

v

LIST OF FIGURES

vii

CHAPTER I

-

OVERVIEW

I

CHAPTER2 -°L'

V

CHAPTERS

OLEOFBA'CKWATERS . ...a 27

CHAPTER4- NOIN

TABLE OF CONTENTS

iv

Page

88

---

LIST OF TABLES

TABLE

PAGE

Table 1 .

Fish families commonly found in the Illinois River basin grouped

into one of three life history classes (e.g.,

Nelson 1994, Dettmers et

al

. 2001, and Galat and Zweimuller 2001), fluvial specialists

(FS),

fluvial dependents (FD),

and macrohabitat generalists (MG) xx

Table 2

.

Monthly means (± SE) for abiotic variables (temperature

[°C],

dissolved oxygen [DO, mg/L], water depth

[m], secchi depth [cm],

wind speed [km/h], channel velocity [m/s]) collected at sites

Table 3

.

during 2004 in the Illinois River system

Monthly means (± SE) for abiotic variables (temperature

[°C],

dissolved oxygen [DO, mg/L], water depth

[m], secchi depth [cm],

average wind speed [km/h], channel velocity [m/s]) collected at

Table 4.

tow sites during 2005 in the Illinois River system

Site comparisons tested for overall differences in larval densities

during 2004 and 2005 among Calhoun Point (CP),

the Illinois

River (ILR), Lower Swan

(LS), and Middle Swan (MS), with

Table 5

.

adjusted P-values

Comparisons of stratified transects within each site for 2004 and

2005 in the lower Illinois River system, where backwater systems

had inshore and offshore transects (Calhoun Point

[CP], Lower

Swan [LS], and Middle Swan

[MS]), and the Illinois River (ILR)

Table 6

. Fish families

had upstream

grouped

and

into

downstream

one of

transects

three early

at

life

CP and

history

LS classes

(Galat and Zweimuller 2001) with percent of total catch during

2004 and 2005 for the Illinois River (ILR), Lower Swan

(LS),

Table 7

. Length

Middle

analysis

Swanof (MS),larval

and

fish

Calhoun

caught

Point

in

(CP)Lower Swan(LS),Middle

Swan (MS),

Calhoun Point (CP),

and the Illinois River (ILR) using

Kolmogorov-Smimov pairwise comparisons, with results listed for

sites or years that collected larger-sized

fish. MS

was drawn down

before the end of the 2004 sampling season

. Therefore, pairwise

comparisons with this site were conducted on truncated data sets

Table 8.

that included only samples taken on or before 23 July 2004

Mean zooplankton density (#/L) f1 SE by taxa described as a

percent of the total annual density for each site in 2004 and 2005

v

---

Table 9.

Drift of fish during 2004 and 2005 in Swan Lake, Illinois River,

with net sets averaged by time period (standard error represents ±1

Table 10 .

of mean catch rate)

Fish families grouped into one of three early life history classes

(Galat and Zweimuller 2001) with percent of total catch during

2004 and 2005 by gear

. Drift net total catch for 2004 : N= 5,756

;

for 2005 : N= 1,506. Tow net total catch for 2004

: N= 56,476; for

2005: N= 60,509

. No fluvial specialists were caught in either drift

Table 11

. Kolmogorov-Smirnov

or tow nets tests

for 2004 and 2005 pooled diet drift net

data and Lower Swan Lake (LS) and the Illinois River (ILR) larval

tow data

. Test results are listed for treatments which collected

significantly larger-sized fish

vi

---

LIST OF FIGURES

FIGURE

PAGE

Figure 1 .

Study site of the lower Illinois River and two major backwaters,

Figure 2 .

Swan Lake and Calhoun Point

Mean daily temperature of Illinois River and backwater sites

during 2004 through 2005, depicted

as solid gray lines

. Mean

daily depth data are depicted as broken black lines for the Illinois

River and Lower Swan

. Channel depth of Lower Swan was

recorded using a submerged device at the stop-log structure

. River

data were collected at Hardin, Illinois, USA (river kilometer 34

.4).

Depth was not recorded for Middle Swan and Calhoun Point

.

Figure 3 .

Principle

Shaded regions

component

represent

analysis

sampling

(PCA) of

periodsabiotic

variables

for the

lower Illinois River (D) and backwater sites, Lower Swan

(A),

Middle Swan (Y), and Calhoun Point

(•). Weekly means of

abiotic variables were included in the analysis : temperature

(°C),

dissolved oxygen (DO, mg/L), depth (m), secchi depth

(cm), and

Figure 4 .

wind (km/h)

Mean density (#/m

3) per day of fish caught in the lower Illinois

River during 2004 and 2005

. Error bars represent =1:1 standard

error of the mean transect density

. Calhoun Point was not sampled

Figure 5 .

during 2005 because of low water levels

Fluvial specialist (FS), fluvial dependent

(FD), and macrohabitat

generalist (MG) mean densities by site and year in the lower

Illinois River system

. Error bars represent ±1 standard error of the

Figure 6.

mean of dates sampled during that year

Length frequency distributions of all larval and juvenile fish

caught in each site in of the lower Illinois River system during

2004 and 2005 . Sample size is an

average of transects within that

site summed over the sampling season

. Length frequencies are

expressed as a percentage of average catch

. Middle Swan samples

were not collected after 23 July 2004

. During 2005, lengths were

collected in Middle Swan samples, and Calhoun Point was too

shallow to be sampled

. Descriptive statistics were calculated on

Figure 7.

raw length data

Mean density (#/L) of cladocerans, copepods, nauplii, and other

zooplankters per day in each site of the lower Illinois River system

through 2004 and 2005

. Error bars represent ±1 standard error of

vii

---

the mean of transects sampled that day

. Middle Swan samples

were not collected after 23 July 2004

. During 2005, zooplankton

data were not identified for Middle Swan and no zooplankton

Figure 8 .

samples were taken in Calhoun Point

Mean density (#/L) of rotifers per day in each site of the lower

Illinois River through 2004 and 2005

. Error bars represent

f 1

standard error of the mean of transects sampled that day

. Middle

Swan samples were not collected after 23 July 2004

. In 2005,

zooplankton data were not identified for Middle Swan and no

Figure 9.

zooplankton samples were taken in Calhoun Point

Fixed drift net site during 2004 and 2005 on the lower Illinois

River (top inset)

. Tandem nets were floated on the Lower Swan

(LS) side of the stop-log water control structure

. Schematic

depicts nets sampling larval fish drifting out of LS

. Bottom inset

portrays the drift net frame positioned to sample larval fish drifting

Figure 10.

into LS

Mean daily temperature of the Illinois River and Lower Swan for

2004 and 2005, depicted as solid gray lines

. Mean daily depth data

are depicted as broken black lines

. Channel depth of Lower Swan

was recorded at the stop-log structure between the backwater and

the river

. River stage and temperature data were collected at

Hardin, Illinois, USA (river kilometer 34

.4)

. Shaded regions

Figure 11 .

represent sampling periods

Mean daily depth (top) and differential velocity (bottom)

measurements in the channel between Swan Lake and the Illinois

River

. Differential velocity was calculated as the sum of mean

channel velocities per day

(i.e.,

the sum of 48 values) . Positive

velocity values represent net inflow of water into Swan Lake and

negative values correspond to net outflow to the river

. Continuous

data were recorded using a submerged device anchored at the stop-

log structure

. No depth data were collected from September

through November 2004, and no velocity data were collected from

October 2004 through April 2005

. Shaded regions represent the

Figure 12 .

2004

Drift

and

of ichthyoplankton

2005 sampling

intoperiods(•)

and

out (0) of Swan Lake in

2004 and 2005

. Symbols represent mean catch (4/minute) per net

set and line data represent continuous depth data recorded at the

stop-log structure

. Error bars represent ±1 standard error for the

mean of each net set

---

Figure 13 .

Mean density (#/m3)

per day of fish caught in the Illinois River and

Swan Lake during 2004 and 2005

. Error bars represent ±1

Figure 14 .

standard error of the mean transect density

Mean catch rate (#/minute) of fish in three early life histo7 classes

drifting into and out of Swan Lake and mean density (#/m

) of

classes in the Illinois River and Swan Lake during 2004 and 2005

.

All families were included and grouped into one of three early life

history classes as defined by Galat and Zweimuller (2001)

: fluvial

specialists (FS), fluvial dependents

(FD), and macrohabitat

generalists (MG).

Error bars represent ±1 standard error of the

Figure 15 .

mean

Length frequency distributions of all larval and juvenile fish

caught in the Illinois River, Lower Swan, and in the drift during

2004 and 2005

. Sample size is an average of transects or nets

within that site summed over the sampling season for larval tow

data or drift data, respectively

. Length frequencies are expressed

as a percentage of the average caught

. Descriptive statistics were

Figure 16 .

calculated on raw length data

Length frequency distributions of all fish caught drifting during

each time of day for 2004 and 2005

. Data of fish drifting into and

out of Lower Swan were pooled by time of day (i.e.,

dawn, day,

dusk, night)

. Length frequencies are expressed as a percentage of

the mean number per net set

(i.e., mean of three nets) summed over

the sampling season

. Descriptive statistics were calculated on raw

length data

ix

---

CHAPTER 1

OVERVIEW

River floodplain habitats are an endangered landscape (Tockner and Stanford

2002), which is a concern because they function as a critical component in the early life

history of fishes by providing rich food resources and structural refuge (Junk et al

. 1989) .

In their natural state, rivers are dynamic systems where flood pulses cue biological

processes, increase productivity, and maintain diversity (Junk et al

. 1989) . These flood

pulses also alter the river's course, where meanders become backwaters, and backwaters

become marshland or low-land forests (Lusk et al

. 2003) . Continued meandering of the

river continually creates and alters new backwaters habitats

. The degree of connectivity

between a river and backwater affects the backwater's ability to function as a spawning

and nursery habitat (Grift et al

. 2003) . The higher the connectivity, the more available

the backwater is for fish use

.

In backwater habitats, congregations of larval fishes vary in species composition

and abundance through time (Holland 1986)

. To explain this, I adopted a generalized life

history classification based on fishes' relative dependence to flowing water (adapted

from Galat and Zweimuller 2001)

. The first early life history class is classified as fluvial

specialists which spend most of their life in flowing water and rarely use backwater

habitats . Examples of these fishes are sturgeon Acipenseridae spp

. and perches

Stizostedion spp

. The second category, fluvial dependents, have only a portion of its life

history

I

---

2

requiring flowing water (Galat and Zweimuller 2001)

. These fishes, like temperate

basses Morone spp.,

paddlefish Polyodon spathula,

and freshwater drum Aplodinotus

grunniens,

either spawn in the river and their larvae move laterally into backwaters or

spawn in the backwaters and their larvae move into rivers at some point during their first

year (Dettmers et al

. 2001, Galat and Zweimuller 2001)

. The extent of this larval fish

movement between rivers and adjacent, connected backwaters is currently

undocumented . Finally, macrohabitat generalists, such as sunfishes

Lepomis spp., shads

Dorosoma spp.,

and gars Lepisosteus spp.,

conduct most of their life in backwater areas .

When macrohabitat generalists are found in the river, they typically use it as a corridor to

access other backwater areas (Junk et al

. 1989). Based on these life history classes, many

riverine fishes rely on semi- or permanently-connected backwaters

.

Backwater habitat and river connectivity have been reduced because of high

sedimentation rates (Havera and Bellrose 1985)

. In a natural floodplain system,

sedimentation of some backwaters would be counteracted by the creation of other

backwater areas (Miranda 2005)

. Unfortunately, because rivers have become

immobilized due to levee construction and floodplain settlement, existing backwaters are

becoming terrestrialized without new backwaters being created

(, Lusk et al

. 2003,

Miranda 2005) .

Spurred by backwater research in the 1980s, the Habitat Restoration and

Enhancement Program (FIREP) was initiated as a partnership between state and federal

agencies to fund large-scale restoration projects in the Upper Mississippi River (USACE

1993)

. The continuing goal of the HREPs is to improve habitat for waterfowl and fish by

reducing sedimentation and enhancing backwater-river connectivity

. Few studies have

---

3

documented the effects of these HREPs and their subsequent management regimes on

river fishes

. Fewer still have documented the effect of these improvements on spawning

and nursery habitats .

This study was conducted on the lower Illinois River (ILR) and contained two

backwater complexes, the HREP-managed Swan Lake (river kilometer [RKM] 8

.0) and

natural Calhoun Point (CP, RKM

1 .6),

and two adjacent river segments, RKM 1

.6 and

RKM 8.0

(measuring from its confluence with the Mississippi River

; Figure 1) .

Historically, Swan Lake was only connected to the river at its downstream end where a

0

.5 km-wide opening to the river existed

. To reduce sediment loading, this opening was

restricted to the width of a stop-log water control structure during restoration, about 5 m

wide

. Additionally, the backwater was compartmentalized by constructing cross-levees,

creating a lower compartment (Lower Swan

[LS]),

which was managed to be

continuously connected to the river, and a middle compartment (Middle Swan

[MS]),

which was drawn down each year during this study to consolidate sediments and promote

moist-soil vegetation

. The CP backwater complex (466 ha), at the confluence of the

Illinois and Mississippi Rivers, had a low-lying levee, isolating the backwater and

maintaining water levels during non-flood periods (Figure 1)

. Thus, LS was a restored,

unmanipulated backwater, MS was a restored and manipulated counterpart, and CP was

an unrestored off-channel backwater

.

Therefore the objectives of this study were two-fold

. First, I assessed the benefit

of a restored HREP-supported backwater system relative to an unrestored, manipulated

backwater by quantifying the response of larval fish communities

. Secondly, I

---

4

investigated the interplay of life history strategies with lateral drift dynamics on a diel

and seasonal basis within a restored backwater system

. I tested the hypotheses that :

i) Backwater sites, whether restored or unrestored, would be used as

spawning and nursery habitat in higher densities than river segments

sampled (Holland 1986, Junk et al . 1989).

ii) Seasonal abundances of larval fishes would be different among sites based

on their river connectivity, with the site being the most connected

(i.e.,

LS) to the river yielding the highest larval abundance

.

iii) Family composition, described by the life history classification, would

also vary among sites, where sites with the lowest river connectivity

would contain relatively higher densities of macrohabitat generalists

(Miranda 2005)

.

iv) Larval exchange would occur between the restored backwater, LS, and the

lower ILR, such that an

influx of larvae into the backwater would be

related to rising water temperatures and river stage (Junk et al

. 1989).

v) Because Swan Lake is a major backwater of the lower ILR, ingress or

egress of larvae between LS and the ILR would result in density

differences between river segments upstream and downstream of the

backwater-river confluence (Sheaffer and Nickum 1986)

. However,

restriction of the LS-ILR confluence may affect the magnitude of larval

movement between the river and backwater

.

---

5

vi) Diel periodicity in larval drift patterns may occur where drift rates and

sizes of larvae differ among sampling times (Gale and Mohr 1978,

Gadomski and Barfoot 1998)

.

These are ecological and management-oriented questions regarding backwater

function and the influence backwater restoration may have in the early life history of

fishes

. Understanding larval dynamics and production within backwater systems,

especially within restored habitats, will enable researchers to determine the quality of

backwaters as larval fish nurseries and to address the influence of habitat alterations on

larval fish assemblages among backwater lakes and river reaches

.

---

CHAPTER

2

THE ROLE OF BACKWATER RESTORATION IN LARVAL FISH ECOLOGY

ABSTRACT

Large river modifications have widely degraded backwaters, reducing critical

habitat for larval fishes

. During 2004 and 2005, 1 assessed how river backwaters function

as fish nurseries by quantifying the response of larval fish communities to restoration of

Swan Lake, a major (>1,000 ha) backwater complex on the lower Illinois River

. The

response was compared to a nearby unrestored backwater complex and to adjacent river

segments

. The densities of zooplankton prey in each site also were quantified

. Families

of fish changed between years, with fluvial-dependent taxa present during the flood pulse

of 2004 and relatively absent during the drought of 2005

. During both years, tenfold

greater larval densities were produced in backwaters than the river

. Larvae were larger in

backwaters and during the non-flood year, probably due to abundant zooplankton

. All

backwaters produced similar larval densities regardless of restoration

. Growth was

highest in the most isolated and regulated backwater portion of Swan Lake

. Predictable

flood pulses coupled with habitat heterogeneity in the backwaters may be important for

larval abundance, assemblage composition, and recruitment

. Current site-specific

restoration efforts constructed with the built-in flexibility to adapt to other management

regimes will likely be most beneficial to all early life history strategies of fishes, while

also providing widespread benefits and supporting the move towards system-wide

management programs on large rivers

.

6

---

7

INTRODUCTION

Natural river ecosystems support abundant and diverse species assemblages due

to high habitat diversity and physical complexity (Junk et al

. 1989, Dettmers et al . 2001).

Off-channel aquatic habitats, such as backwater lakes and sloughs, provide productive

habitat and a lentic-lotic gradient whereby fishes find structural refuge, food resources,

and spawning and nursery grounds (Junk et al

. 1989)

. For many riverine fish species,

floodplains are optimal spawning and nursery habitats and are actively sought out via

lateral spring spawning migrations of adults (Molls 1999)

. However, river regulation and

degradation have reduced backwater habitat quality and river connectivity, jeopardizing

larval fish abundance and success (Havera and Bellrose 1985, Tockner and Standford

2002)

. Larval fish diversity and recruitment are likely related to river-backwater

connectivity, such that reduced access could alter riverine fish communities (Pezold

1988, Turner et al . 1994)

. Although this association with the river is critical to backwater

health (Gore and Shields 1995), few studies have related connectivity to early life history

requirements and larval fish communities (Miranda 2005) .

Despite the important links between backwaters and larval fish abundance,

floodplain habitat has become severely degraded and is among the most endangered

landscape in the world (Tockner and Stanford 2002)

. Dam and levee construction

coupled with high sediment loads from agricultural lands extensively modified large river

hydrology, altered the flood-pulse, reduced backwater habitat quality and river

connectivity, and decreased system productivity (Havera and Bellrose 1985)

. In the

1980s, the loss of backwater habitat along the Illinois River spurred research and the

development of projects aimed to restore essential fish and waterfowl habitat (Sheehan et

---

8

al. 1990)

. A Habitat Restoration and Enhancement Project (HREP), through the federally

supported Environmental Management Program (EMP), was initiated on Swan Lake, a

major backwater of the Illinois and Mississippi Rivers

. One goal of the Swan Lake

HREP, to improve spawning and nursery habitat for fishes, was addressed by controlling

river connectivity of the backwater and increasing habitat heterogeneity (USACE 1991)

.

HREP techniques must be evaluated because current paradigms that influence

river management and serve as the basis for these restorations will continue to be

implemented in future programs

. However, a paucity of information about life histories

and habitat needs of larval and juvenile fishes in large river ecosystems, which may

hinder restoration efforts and effectual progress (Galat and Zweimuller 2001)

. All

riverine fishes exhibit certain life history strategies based on their relative dependence on

flowing water

; some are specialized for riverine environments, while others require

flowing water for only a portion of their life history

. Lastly, generalist fishes reside

mostly in lentic backwater areas, especially during their first year of life

. Therefore,

understanding larval dynamics and abundance within restored habitats will enable

researchers and managers to determine the influence of habitat alterations on larval fish

assemblages, where alterations that increase connectivity may produce more diverse and

abundant assemblages

. I quantified the response of larval fish communities within a

restored HREP-supported backwater system relative to an

unmanipulated of-channel

complex to assess the benefit of this HREP restoration program

. I tested the hypotheses

that (1) fish and zooplankton densities differed among sites and between years and that

(2) life history classes (i

.e.,

family composition) varied among sites or between years

. I

---

also evaluated how flood regimes and backwater habitat heterogeneity affected larval

abundance and composition within managed and natural backwater habitats

.

METHODS

STUDY AREA

The study area was located on the lower Illinois River (ILR) and contained two

backwater complexes, HREP-managed Swan Lake (river kilometer [RKM] 8

.0) and

natural Calhoun Point (CP, RKM

1 .6), and two adjacent river segments, RKM 1.6

and

RKM 8.0

(measuring from its confluence with the Mississippi River ; Figure 1)

.

Historically, Swan Lake was only connected to the river at its downstream end

where a 0.5 km-wide opening to the river existed

. During restoration, a cross-levee was

constructed to create a lower compartment (Lower Swan [LS],

567 ha) and middle

compartment (Middle Swan

[MS], 485 ha; Figure 1), and the historic LS connection was

restricted to the width of a stop-log water control structure, about 5 m wide

. An

additional river connection was created in MS via a stop-log structure

. During normal

pool stage, the stop-log water control structures at LS and MS were the only avenue

through which backwater-river movement of larval fish could occur

. The cross-levee and

stop-log connections allowed the compartments to be managed independently (Figure 1)

.

Lower Swan was managed to be continuously connected to the Illinois River, while the

MS water control structure was opened in early winter, allowing spawning fish to access

the lake before it was disconnected from the river in early spring and pumped a meter or

two below pool level each June

. Moderate flooding could top the MS stop-logs in the

water control structure and the cross-levee, which would serve to connect MS to the river

9

---

and LS during additional times of the year

. Thus, LS was a restored, unmanipulated

backwater, while MS was a restored and manipulated counterpart

.

The CP backwater complex (466 ha), at the confluence of the Illinois and

Mississippi Rivers, was unrestored and regularly connected to the rivers during floods via

low-lying levees (Figure 1)

. These levees provided limited flood control, and functioned

more to isolate the backwater from the river and maintain water levels during non-flood

periods

. This backwater system was dendritic and contained many vegetated islands

.

However CP was also shallow and had an

unconsolidated lakebed . This system

represented an unrestored, unmanipulated backwater

.

LARVAL ABUNDANCE

To understand drift patterns and compare the outcome of management practices

on study sites, I quantified seasonal larval abundance using paired, bow-mounted

ichthyoplankton nets (0

.5 m diameter x 2 m long, 500-µm mesh)

. River and backwater

habitats were sampled during late March through September 2004 and 2005

. Each week,

four stratified transects were randomly chosen within each Swan Lake compartment and

two transects were randomly chosen in the ILR (RKM

8 .0).

Every two weeks, I sampled

CP and its adjacent segment of the ILR (RKM

1 .6).

1 stratified backwater transects into

inshore and offshore tows, with two inshore plus two offshore transects per Swan Lake

compartment, and one inshore plus one offshore tow within CP (N = 10 backwater

transects)

. River tows were conducted within one km upstream and one km downstream

of each backwater-river confluence (N = four river transects)

. In MS, sampling ended on

22 July 2004, and on 27 June 2005, because the summer drawdown made the lake too

10

---

11

shallow to sample (i.e., water level management program)

. Calhoun Point was not

sampled during 2005 due to low water levels

.

At each transect, tows were conducted at the surface for five minutes at a constant

speed, with a calibrated mechanical flow meter (Model 2030R, General Oceanics, Inc

.,

Miami, Florida, USA) mounted in the mouth of one net to estimate volume sampled and

standardize samples

. Inshore backwater tows followed the shoreline, offshore backwater

tows were straight transects, and river tows were straight transects conducted

perpendicular to flow direction. River tows started at the main channel border and

continued across to the opposite main channel border

. If five minutes had not passed by

the time the opposite side was reached, the direction was reversed with nets still in the

water, and sampling continued until five minutes had elapsed

.

Upon completion of each transect, net contents were flushed into the cod end and

preserved in 95% ethanol

. Samples were split to approximately 200 fish using a Folsom

plankton splitter (Aquatic Research Instruments, Hope, Idaho, USA)

. All age-0 fish in

the subsample were counted, identified to the lowest possible taxon, typically genus, and

classified to a developmental stage

(i.e.,

yolk-sac, larval, juvenile) using descriptions and

keys in Soin and Sukhanova (1972), Auer (1982), Murty et al

. (1986), Tweb et al. (1990)

and voucher specimens from Southern Illinois University's Fluid Vertebrate Collection

(Brooks Burr, Carbondale, Illinois, USA) and Colorado State University's Larval Fish

Laboratory (Darrel Snyder, Fort Collins, Colorado, USA)

. A subsample of fish from

each taxon and stage identified was measured (total length

[TL] ;

mm) using Scion

Image® software or metric calipers (N = 10 per net)

. During 2005, MS samples were

counted, but not identified .

---

12

At the start of each transect, I sampled water chemistry parameters that could

affect the density of larval fish, including temperature

(°C), dissolved oxygen (mg/L [YSI

Model 52 Dissolved Oxygen Meter

; Yellow Springs Instruments, Yellow Springs, Ohio,

USA]), water depth, secchi depth, and average wind speed (km/h [Kestrel 1000, NFS

-

Radiation Protection Systems, Inc

., Groton, Connecticut, USA]) . Surface water velocity

(cm/s) was measured in the main channel at river transects with an electronic (Flo-Mate

Model 2000, Marsh McBirney, Inc

., Frederick, Maryland, USA) or mechanical flow

meter (Model 2030R Flowmeter, General Oceanics, Inc

., Miami, Florida, USA)

.

Continuous monitoring temperature loggers (8-bit Minilog-TR, Vemco Ltd

., Nova

Scotia, Canada) were located in each backwater and the river to supplement the water

temperature data taken at each transect

. River stage data were recorded at Hardin,

Illinois by the St

. Louis District U.S

. Army Corps of Engineers (river kilometer 34.4)

.

Swan Lake channel depth data were recorded using a Doppler unit (Model 6526-51

Starflow Ultrasonic Doppler Flow Meter

; Unidata America, Lake Oswega, Oregon,

USA)

. This unit was anchored to the bottom of the water control structure and

continuously recorded temperature (°C)

and depth

(mm) .

Depth data were not available

for MS and CP .

LIFE HISTORY CLASSIFICATION

Larval and juvenile fish collected were grouped by family into one of three

generalized classes to better explain trends between years and treatment groups (Galat

and Zweimuller 2001 ; Table I)

. The classes were

: fluvial specialists, fluvial dependent,

and macrohabitat generalists

. Fluvial specialists inhabit streams and rivers throughout

---

13

their entire life and rarely enter floodplain habitats (Galat and Zweimuller 2001)

. Fluvial

dependent species regularly use lentic backwater or reservoir habitats, but certain life

stages depend on lotic environments (Galat and Zweimuller 2001)

. These species are

typically broadcast spawners, where developing eggs and larvae are semi-buoyant and

passively drift in wind-induced or downstream currents (Holland 1986)

. Adult fluvial

dependent fishes also may make lateral migrations into slow-flowing lentic areas to

spawn-activities likely corresponding with increasing temperatures and rising water

levels (Junk et al . 1989)

. Macrohabitat generalists include species commonly found in

reservoirs and off-channel habitats that do not depend on lotic systems (Galat and

Zweimuller 2001)

. When these fishes use the river, it is as a corridor to move among

backwaters (Junk et al

. 1989, Dettmers et al . 2001)

. Spawning usually occurs in off-

channel habitats and offspring generally do not leave this habitat until the juvenile stage

(Holland 1986)

. I grouped families based on Galat and Zweimuller (2001

; taxonomy

from Nelson [1994]), and the only deviation from their groupings was Sciaenidae, which

I classified as fluvial dependent based on life history descriptions from Dettmers et al

.

(2001) and Koel and Sparks (2002) .

ZOOPLANKTON DENSITY

Because zooplankton abundance and composition affect growth and survival of

exogenous feeding larvae (Miller et al . 1988, Miller et al

. 1990), zooplankton were

sampled in each backwater and river segment

. Similar to larval tows, I randomly

stratified samples between inshore/offshore and upstream/downstream habitats

.

Backwaters were sampled before a corresponding inshore and offshore larval tow (N=

---

14

two samples/backwater), and river sites were sampled in the thalweg at each upstream

and downstream transect (N= two samples/river segment)

. Samples consisted of four, 1-

m vertical hauls from the boat using a conical net (0

.5 m x 2 m, 64 µm mesh)

. If depth

was less than 1 m, the entire water column was sampled four times, with the depth noted

to adjust volume sampled

. During 2005, neither MS nor CP were sampled

.

After each haul, net contents were flushed into a removable collecting bucket and

rinsed into a 64 µm sieve

. All four hauls from one site were preserved with 10%

buffered sugar-formalin in a single container (Haney and Hall 1973), and returned to the

laboratory for processing

. Copepods including cyclopoids, calanoids, and nauplii,

cladocerans including Bosmina spp.,

Chydorus spp ., Daphnia spp

., Diaphanosoma spp .,

and Moina spp.,

and other zooplankters including ostracods were identified and counted

.

Rotifers were dyed with Rose Bengal and counted, but not identified

. Using a Henson-

Stempel Pipette, each sample was processed until two taxa reached counts of 200 or until

10% of the sample volume had been processed

. Due to the high density of rotifers, they

were counted until approximately 100 individuals had been processed

. Densities were

calculated by dividing number of taxa in a subsample by the fraction of subsample

counted, then dividing that amount by the total volume of water filtered in the field

.

Zooplankton were grouped as rotifers and macro-crustaceans (i

.e.,

copepods,

cladocerans, nauplii, etc

.) for data analysis due to large differences in density

.

DATA ANALYSIS

To examine environmental relationships among sites and between years, weekly

means of abiotic variables were analyzed using principle component analysis (PCA)

. The

---

15

abiotic matrix, containing temperature

(°C),

dissolved oxygen (mg/L), depth (m), secchi

(cm), and wind (km/h) data, was analyzed using PC-ORD with the correlation option to

center and standardize parameters (McCune and Medford 1999)

.

To standardize samples by volume, densities for larval tow and zooplankton data

were calculated as fish/m 3

and zooplankton/L

. Total larval abundance at each site was

conservatively calculated as the sum of weekly densities

. For the abundance estimate,

when a site was not sampled or processed during a week, the larval density from the

previous weeks was substituted for the missing value

. All data were log-transformed to

meet assumptions of normality

. Two-way repeated measures ANOVA (proc MIXED,

SAS Institute 1999) was used to test for differences among sites and between years that

were sampled over time (Hurlbert 1984)

. For tow and zooplankton data, mean densities

among sites were compared over time

.

Fish length data were analyzed to determine whether sizes differed among tow

and drift samples

. Proportions of fish per I-mm length group were calculated and used to

make pairwise comparisons

. Kolmogorov-Smirnov tests were used to compare size

structure of fish collected from tows and between year differences for LS and ILR tow

data

. Size structure among tow sites was compared using Bonferroni adjusted a-values

.

Because MS was drawn down before the end of the 2004 sampling season, pairwise

comparisons within this site only included samples collected on or before 23 July

.

I determined overall trends within and between years for larval and zooplankton

densities

. A one-way ANOVA design tested density differences between years for larvae

and zooplankton

. I investigated larval and zooplankton density differences within each

---

16

year using a two-way ANOVA randomized block design

. The randomized block design

test assessed hypotheses about overall lake effects

: (1) stratified habitats differed in

density (i .e., inshore v

. offshore), (2) sites differed in density (i.e.,

LS v. CP), and (3) fish

families and zooplankton groups varied among sites

. To control for experimentwise error

rates, Tukey-Kramer post-hoc (Sokal and Rohlf 1995)

. Pearson correlation coefficients

were calculated for untransformed larval and zooplankton densities to reveal any

association between larvae and their food source

.

RESULTS

ENVIRONMENTAL FACTORS

A moderate flood pulse occurred in 2004, during which water levels in the lower

Illinois River were above flood stage for approximately five weeks during June (Figure

2)

. In contrast, water levels remained at or below normal pool level of 128 m during the

2005 sampling season .

Water temperatures in the river gradually rose and fell during 2004, peaking in

late July at 28 °C,

and varied in 2005, exceeding 30 °C

twice (Figure 2) . Lower Swan

conditions varied in a manner similar to the ILR, although water temperatures were

higher and more variable and depths fluctuated less (Figure 2)

. The 2004 flood pulse

topped the water control structure at MS and the low-lying levee at CP, causing water

levels in these backwaters to rise rapidly and connecting all the backwaters to the river

.

Middle Swan and CP remained isolated during 2005

.

From the PCA output, two axes, which had broken-stick eigenvalues less than the

actual eigenvalues, were used to graphically represent the data (Jackson 1993)

. River

---

17

and backwater sites were spatially separated by water and secchi depths, while water

temperatures and other abiotic parameters were similar among all sites (Figure 2 & 3,

Table 2 & 3)

.

LARVAL ABUNDANCE

During both years, fish larvae in the backwaters first appeared in low densities

during late March and early April, whereas larvae were not collected in river sites until

late April (Figure 4)

. Backwaters consistently produced higher densities of larvae than

the ILR (P < 0.01,

Table 4), with 2004 and 2005 ILR densities peaking at means of 11

and 14 fish/m3

(Figure 4), and seasonal abundance was estimated at 41 and 31 fish/m

3,

respectively

. A peak of larvae occurred in all backwaters and the river during June 2004,

the period of floodwater inundation (Figure 4)

. During that time, Lower Swan and CP

experienced a larval pulse of similar size (Figure 4, Table 4), and had similar total

seasonal abundance estimates of 435 and 409 fish/m 3,

respectively. Middle Swan larvae

peaked at the highest density of 400 larvae/m 3,

though they were not statistically higher

than LS (Table 4), and total seasonal abundance was estimated at 1,276 fish/m

3.

These

among site differences (two-way repeated measures ANOVA

; 2004:

F3,1o

= 55 .37, P <

0.01) differed across time (2004

:

F21,133

= 55

.06, P < 0

.01) and also interacted (2004 :

F43,133 = 8.07,

P < 0 .01)

. During 2005, sites also differed (two-way repeated measures

ANOVA : F2,9 = 53

.37, P < 0 .01) across time (2005

:

F22,115

= 15.05, P < 0.01)

and

interacted (2005 :

F31,11s =

5 .00,P<0

.01).

Larval pulses did not occur in synchrony

among sites, but occurred in MS during May, LS during late May, and the ILR during

June 2005 (Figure 4)

. My estimation of MS larval abundance during 2005, at 516

---

18

fish/m3,

was less than LS' estimated total larval abundance

. However, during the same

time interval, of late-March through 23 July 2005, MS had a higher estimated larval

abundance than LS

. Seasonal abundance in LS increased between years to 531 fish/m

3

during 2005 .

Each site had homogeneous larval distributions, with no differences between

inshore and offshore or upstream and downstream stratified transects (P > 0

.05, Table 5) .

During 2004, offshore transects in Swan Lake had slightly higher densities than inshore

transects, although they were not significantly higher (Table 5)

.

The larval fish assemblage in the backwaters was comprised predominantly of

macrohabitat generalists during both years, but changed between years in the river

(Figure 5)

. In the ILR, fluvial dependent taxa, consisting mostly of sciaenidae,

catostomidae, and cyprinidae, occurred at higher densities during 2004, but were

relatively absent during the low water year of 2005 (Figure 5)

. Macrohabitat generalists

dominated the ILR during 2005 (Figure 5)

. MS contained higher densities of fluvial

dependents (i.e.,

catostomids, cyprinids) compared to other backwaters during 2004

(Table 6, Figure 5)

. Clupeids, the most abundant family, drove macrohabitat generalist

patterns in all sites and heavily influenced system-wide trends in total density (Table 6)

.

Despite being continuously connected to the ILR, LS had substantially higher densities of

every family sampled except sciaenidae and moronidae during 2004

. During 2005,

higher densities of these families were collected in LS compared to the ILR, though these

larval densities in LS were lower than during 2004

.

Larval sizes were larger in backwater systems than the ILR, regardless of year

(Figure 6)

. During 2004, Calhoun Point had larger fish than MS and LS, possibly

---

19

indicating this backwater provided better fish nursery habitat (P < 0

.001, Table 7, Figure

6)

. Progressively smaller sized fish occurred in MS, LS, and the smallest were collected

in the ILR (Table 7)

. The size structure of fishes in LS and ILR were significantly larger

during 2005, though larger fish again occurred in the backwater (all P < 0.001,

Table 7).

ZOOPLANKTON DENSITY

Patterns in zooplankton and rotifer density differed by site and exhibited temporal

patterns

. During 2004, macro-crustaceans, consisting mostly of cladocerans, copepods,

and nauplii, peaked during late May and June in the ILR, LS, and CP, while MS densities

steadily declined during April through June (Figure 7)

. The opposite occurred for rotifer

densities, with the ILR, LS, and CP showing declining densities during late May and a

pulse of rotifers occurring in MS during 2004 (Figure 8)

.

Although rotifers dominated the zooplankton assemblage during both years, 2005

rotifer densities appeared greater while macro-crustaceans occurred in lower densities

than 2004

. In the ILR, while there was no between-year difference in rotifer density

(F1 ,34 = 1 .28, P = 0 .27

; Figure 8), higher densities of macro-crustaceans occurred during

2004 (F1,34 = 12.43, P < 0.01

; Figure 7)

. In LS, the opposite occurred, with higher rotifer

densities during 2005 (F1,34

= 8.46,

P < 0 .01

; Figure 8) and no difference of macro-

crustacean densities between years (F1,34

= 1

.43, P = 0 .24

; Figure 7) . The repeated

measures analyses for 2004 revealed temporal variation in density, but no difference

among sites for rotifers (site :

F3,63 = 0.80,

P = 0 .50; week :

F15,63 = 2.61,

P < 0.01

;

lake*week :

F32,63 = 1

.20, P = 0.27)

or macro-crustaceans (site

:

F3,64 = 1

.63, P = 0.19 ;

week

:

F15,64

= 11 .92, P < 0.01 ; lake*week

:

F3,64 = 5

.37, P = 0.01)

. During 2005, all

---

20

effects were significant for both rotifer densities (site

: F,,2 = 91 .56, P = 0.01

; week:

F21,33

= 2 .74, P < 0 .01 ; lake*week

:

F,6,33 = 3

.70, P < 0

.01) and macro-crustaceans (site

: F 1,2 =

93.56, P = 0.01 ; week : F21

,33 = 3.69, P < 0.01 ; lake*week

:

F16,33 = 2.89,

P < 0.01).

Macro-crustaceans in LS and the ILR were positively correlated to larval fish

density during 2004 (LS

: r = 0.71, P < 0.01 ; ILR

: r = 0.63, P = 0.01)

. Only rotifer

densities in MS correlated with larval densities (r = 0

.80, P < 0.01)

. No other patterns

occurred during 2004, and no correlations occurred during 2005 for either rotifers or

macro-crustacean densities and fish larvae

.

DISCUSSION

Connectivity is a critical feature of floodplain habitats, and likely influenced

larval abundance patterns in the lower ILR and its backwaters

. Larval abundance during

the flood year was synchronized among all sites

; well-timed to the period of inundation

where all sites were connected

. More families in greater evenness also occurred during

the flood year

. Although typically isolated from the river, a large variety of families

occurred in MS during 2004, most notably a large proportion of catostomids and

cyprinids that likely were entrained in the backwater when floodwaters topped its levee

.

This restored and manipulated site also produced more larvae during both years than any

other site in the same time frame, which hints towards the benefits garnered from regular

periods of connectivity and increased aquatic vegetation

.

During the non-flood year I saw a lack of fluvial dependent taxa in all sites, which

was more pronounced in the riverine habitat than the backwaters

. Moreover,

---

21

macrohabitat generalists tended to have higher densities during the non-flood year,

apparently capitalizing on the stable water levels and low flows

. Similarly, Brown and

Coon (1994) showed a decreased number of taxa during a non-flood year, with lower

densities of fluvial dependent species (i.e., goldeye

Hiodon alosoides, buffalo

Ictiobus

spp., and carpsuckers Carpiodes spp.)

and an increased abundance of macrohabitat

generalists (i.e.,

centrarchids) in most tributaries

. Changes in family composition

between years emphasize the importance of a predictable flood pulse in larval fish

ecology where rising waters cue spawning activity and permit access to floodplain habitat

(Junket al . 1989).

In accordance with other studies that have found lower or no reproduction in

rivers during non-flood years (Brown and Coon 1994, Agostinho et al

. 2004), I expected

significantly lower densities and larval abundance estimates during 2005

. However,

seasonal abundance of macrohabitat generalists in LS was higher than the year before,

and only MS abundance and peak densities were markedly lower

. Moreover, the ILR,

which functioned nearly exclusively as spawning habitat during 2004, as indicated by the

larval lengths, had a larger size structure during 2005, suggesting these segments of the

river provided relatively better nursery habitat for larvae during the non-flood year

.

Lower velocities during summer 2005, often below 0

.1 m/s, transformed the ILR into

habitat suitable for larvae

. This compensatory pattern in larval abundance occurred due

to certain species, mainly gizzard shad

Dorosoma cepedianum,

mosquitofish Gambusia

affinis, brook silversides Labidesthes sicculus,

and emerald shiners Notropis

atherinoides,

exploiting the low flow conditions, a concept dubbed the `low flow

recruitment hypothesis' (Humphries et al

. 1999, King 2004). The hypothesis postulates

---

22

that during low flow periods in the river, appropriately sized prey is concentrated, and

under these conditions, some species spawn and can easily make the transition from

endogenous to exogenous feeding, thereby having high recruitment (Humphries et al

.

1999)

. I do not know whether recruitment differences occurred between years, but the

significantly higher rotifer densities in LS during the low-flow period likely supported

the increased rates of larval abundance (Aoyagui and Bonecker 2004)

.

Limited information exists on the interaction between larval fishes and

zooplankton densities in large rivers

. The positive correlations between plankton and

larvae may be due to mutually favorable abiotic conditions, such as warm temperatures

(Wetzel 2001)

. However, it has been speculated that biotic factors, such as top-down

effects of larvae, may play an

important role in riverine foodwebs (Jack and Thorp 2002)

.

During the flood year, plankton and fish densities were positively related

; ILR and LS

zooplankton to fish densities in those sites, and MS larvae to rotifer densities

. The

positive relationship may be due to autochthonous inputs that increased seasonal

abundance in backwaters during inundation, allowing for large pulses of larvae and

plankton (Junk et al . 1989)

. The drastic summer decline in zooplankton densities may be

due to increased foraging from larger larvae or simply summer declines often observed in

large rivers (Gosselain et al . 1998)

. Whether temperature related or not, these patterns of

summer declines remain unexplained (Gosselain et al

. 1998). Although significant

patterns were not found in either CP during 2004, or in LS and the ILR during 2005, total

plankton densities increased with larval abundance and decreased after mid-summer,

indicating the use of plankton as a prey source by larval fishes

.

---

23

Promoting vegetation growth in backwaters through annual drawdowns may have

positive effects for these restored systems

. Vegetation could have provided food and

cover for larvae, possibly promoting invertebrate populations in densely vegetated areas

and affording protection from predation (Dewey et al

. 1997, Flinn et al . 2005)

. However,

fish survival in MS was likely low due to the drawdowns

. High mortality often occurs

through bird and fish predation (Crowder et al

. 1997) and anoxic conditions in shallower

waters

. The true potential of MS for larval abundance may only be realized if the

compartment is managed in a rotating fashion, whereby drawdowns would be conducted

every few years to maintain vegetation growth and during other years, it would be

continuously connected to the river .

Floodwater stability (gradual rise and fall) coupled with habitat heterogeneity in

the backwaters may be important

. The unrestored backwater, CP, had the largest larval

sizes during 2004, which may have been due to more stable water levels where larvae

were not stranded on lake banks and fish nests were not desiccated due to rapidly

receding waters (Adams et al

. 1999, Brown and Coon 1994)

. In many species, larval

survival and eventual recruitment of the adult population has been show to directly relate

to larval (Miller et al . 1988)

. Progressively smaller larvae were collected in MS and LS,

a pattern which corresponds to the degree of emergent structure available in each site,

with CP having the most emergent vegetation and LS having none

.

To enhance the complex biotic and abiotic interactions so valued in riverine-

floodplain habitats, restoration projects must have a built-in flexibility to either create or

sustain different habitat types

. This approach will likely be most beneficial to all life

history strategies by offering a wide variety of habitat characteristics to meet specific

---

24

early life history environmental and habitat requirements (Grift et al

. 2003)

. Restoration

projects similar to the Swan Lake HREP offer the added benefit of flexibility

. The

compartmentalization allows managers the option of rotating management regimes

between the two lower compartments to promote habitat diversity while maintaining river

connectivity

. One or both compartments may be left open to the river at any one time,

ensuring backwater access to riverine fishes

. Given the widespread benefits of these

management practices and the move towards system-wide management programs on

large rivers (Theiling 1995, Flinn et al

. 2005), current site-specific restoration projects

should be constructed with the flexibility to adapt to other management regimes

.

---

25

CHAPTER 3

LATERAL EXCHANGE OF LARVAL FISH BETWEEN A RESTORED

BACKWATER AND A LARGE RIVER IN

THE EAST-CENTRAL U.S.

ABSTRACT

The lateral exchange of larval fish between a river's main channel and its

floodplain waters may be compromised by the widespread degradation of backwaters

.

During 2004 and 2005, drift nets were set bi-directionally within a constructed channel

between the Illinois River and an adjacent, restored backwater, Swan Lake, to investigate

movement between these sites on a diel and seasonal basis

. Larval density and

composition data within the river and backwater were also collected

. Drift was positively

related with water velocity during 2004, and an estimated 32

.3 million larvae drifted into

Swan Lake that season

. No other relationships with drift direction and abiotic variables

occurred

. Lateral drift patterns were strongly related to the flood pulse during 2004, but

in the absence of a flood, as during 2005, the ambient biotic assemblage influenced drift

timing, magnitude, and composition

. Swan Lake's restoration appears to have

successfully altered the backwater for multiple-use management while maintaining river

connectivity and allowing exchange between the backwater and river to occur

.

---

26

INTRODUCTION

Propensity to drift in streams and rivers is an adaptive response in aquatic

organisms (Eckblad et al . 1984, Kennedy and Vinyard 1997)

. In fishes,

drift is usually restricted to early life stages, such as seasonal larval fish drift in lotic

systems and largely influences spatiotemporal patterns in larval density (Holland 1986)

.

Drift in riverine larval fishes is typically downstream, but can also be lateral, where larval

fishes may be exchanged between river and slackwater habitats (Humphries et al

. 1999).

This exchange has been hypothesized due to ichthyoplankton (hereafter termed larvae)

density differences among slackwater and river habitats (Sheaffer and Nickum 1986,

Brown and Coon 1994, King 2004) and due to downstream larval drift from tributaries to

main channel areas (Eckblad et al

. 1984, Muth and Schmulbach 1984)

. However, the

extent of lateral larval drift into and out of backwaters relative to downstream drift in

rivers is currently undocumented and would be useful to determine origins and

destinations of larval fishes .

I defined lateral drift as the movement of organisms between a river's main

channel and its adjacent floodplain waters

. When connected to the river, these

backwaters provide a lentic-lotic gradient along which fish and other aquatic organisms

find spawning grounds, structural refuge, food resources, and overwinter habitat (Junk et

al. 1989)

. Slackwater areas have been recognized as a critical component in the early life

history of fishes (Holland 1986)

. These areas may be a productive source of age-0 fishes

given the higher larval densities downstream of backwater outflow (Sheaffer and Nickum

1986) and large migrations of juveniles to the main channel from backwaters (Molls

1999)

. However, the contribution of river larvae to backwaters may also be significant

---

27

and a vital process bringing riverine spawned larvae to productive nurseries (de Graaf et

al. 1999)

. Due to the nature of these backwater-river confluences, with periods of inflow

to the backwater, outflow to the river, and stagnant waters following high water events,

larval exchange is likely complex (Brown and Coon 1994)

. These flow patterns at the

confluence are apt to drive the ability of backwaters to function as nursery habitat for

larval and juvenile fishes, potentially affecting the diversity and abundance of fishes

(Brown and Coon 1994)

.

Most investigators have focused on patterns of downstream drift in rivers (Gale

and Mohr 1978, Muth and Schmulbach 1984, Johnston et al

. 1995), but not in the context

of how timing and behavior may shape trends in lateral exchange

. Fish spawning

behavior dictates temporal patterns in larval drift, where drift duration increases with the

duration of spawning (Reichard et al

. 2001), and rising water temperatures and spring

flood events that cue spawning often result in peak larval drift densities (Carter et al

.

1986, de Graaf et al .1999, Auer and Baker 2002)

. Because feeding may influence the

presence of absence of larvae in the drift, larval metamorphosis into exogenously feeding

fish can either initiate or cease drift behavior (Carter et al

. 1986, Dudley and Platania

2000), or have no influence on the species' presence or absence in the drift (Auer and

Baker 2002), depending on the species

. Downstream drift probably allows access to

nursery habitats with amenable growing conditions, and scatters the cohort, which may

reduce competition for food and space as well as reduce conspicuousness to predators

(Bardonnet 2001)

. Thus, the interplay of abiotic and biotic factors determines the timing,

duration, and taxa of fish drifting, but their influence on lateral exchange is unknown

.

---

28

Larvae have diurnal, nocturnal, or crepuscular drift patterns that depend on

species developmental stage (Gale and Mohr 1978, Muth and Schmulbach 1984,

Gadomski and Barfoot 1998)

. These photokinetic responses may change in some species

as larvae age (Bulkowski and Meade 1983), possibly a result of ontogenetic changes in

diet or susceptibility to or avoidance of predation

. However, diet patterns in drift are

contradictory, where fish species of the same life stage exhibit significant nocturnal drift

in some water bodies and diurnal drift in others (Muth and Schmulbach 1984, Smith and

Brown 2002) .

Therefore, management decisions to improve nursery and spawning habitats and

survival of early life stages of fish may be ineffective due to gaps in our understanding of

the timing and behavior of lateral larval drift

. All riverine fishes exhibit certain life

history strategies based on their relative dependence on flowing water

. Some are

specialized for riverine environments and adapted for downstream drift, others require

flowing water for a portion of their life history where habitat changes likely entail lateral

movement

; and finally generalists reside mostly in lentic areas and should be less prone

to enter the drift

. I investigated the interplay of life history strategies with lateral drift

dynamics on a diel and seasonal basis within a restored system, which will assess the

success of the restoration and the backwater's use as a productive fish nursery

.

Quantifying lateral movement will enable researchers to determine the role of backwaters

as larval fish nurseries in large rivers and to address the influence of habitat alterations on

larval fish assemblages among backwater lakes and river reaches

. I hypothesized lateral

movement at the restored connection between Swan Lake and the lower Illinois River to

be influenced by abiotic factors that cue spawning, such as temperature and river stage,

---

and that drift composition would be linked to life history strategy, where taxa requiring

flowing water would comprise a majority of the drift

.

METHODS

STUDY AREA

Swan Lake, a 1,100-ha Illinois River (ILR) backwater located between river

kilometer 8 and 21, is vitally important for fishes of the Illinois and Mississippi Rivers

(USACE 1991)

. A Habitat Restoration and Enhancement Project (HREP) through the

federal Environmental Management Program (EMP) was initiated in the 1980s

. One goal

of the Swan Lake HREP was to improve spawning and nursery habitat for fishes by

improving river connectivity of the backwater and increasing habitat heterogeneity

(USACE 1991). The downstream portion of the backwater (Lower Swan

[LS], 567 ha)

was managed to be continuously connected to the ILR through a water control structure

.

Historically, Swan Lake was connected to the ILR through a 0

.5-km wide opening .

Restoration of the backwater complex restricted the river connection to the width of a

stop-log water control structure, about 5-m wide, which was nested within a rip-rap lined

channel (c . 50 m long x 12 m wide)

. During normal pool stage, the stop-log structure at

LS is the only avenue through which larval fish movement between the backwater and

river occurs, making the stop-log structure the focal point of this study site

.

FISH SAMPLING

To quantify bidirectional larval drift between LS and the ILR, three conical drift

nets (0.5

m x 2 m, 500-pm mesh) were attached to a floating, anchored PVC frame and

29

---

fished during late March through July 2004, and March through August 2005 (Figure 9)

.

I sampled larvae for 15 minutes at the surface (approximately one-third channel depth)

every two weeks on the lake-side of the LS stop-log structure (Figure 9)

. Two directional

net sets, one sampling larvae potentially moving into LS and one set sampling larvae

leaving LS, were conducted at dawn, mid-day, dusk, and mid-night within 36 hours

.

During flood events, sampling frequency was increased to weekly, but fewer night sets

occurred

.

At each sampling time (e.g.,

dawn, mid-day, dusk, mid-night), surface water

temperature (°C)

and dissolved oxygen (mg/L [YSI Model 52 Dissolved Oxygen Meter

;

Yellow Springs Instruments, Yellow Springs, Ohio, USA]), secchi depth

(cm), surface

water velocity (cm/s [Flo-Mate Model 2000, Marsh McBimey, Inc

., Frederick, Maryland,

USA]), and average wind speed and direction (km/h [Kestrel 1000, NFS

- Radiation

Protection Systems, Inc

., Groton, Connecticut, USA]) were quantified

. A Doppler bi-

directional velocity meter (Model 6526-51 Starflow

; Unidata America, Lake Oswega,

Oregon, USA), anchored to the bottom of the water control structure, recorded

temperature ( °C), depth (mm),

and mean velocity (mm/s) twice an hour

. River stage data

from Hardin, Illinois (river kilometer 34

.4) were obtained from the U .S

. Army Corps of

Engineers

. No Swan Lake depth data were collected during September through

November 2004, and no velocity data were collected during October 2004 through April

2005.

Upon completion of each drift net set, contents were flushed into the cod end and

preserved in 95% ethanol

. Each sample was split to approximately 200 fish using a

Folsom plankton splitter (Aquatic Research Instruments, Hope, Idaho, USA)

. All age-0

30

---

fish in the subsample were counted, identified to the lowest possible taxon, typically

genus, and classified to a developmental stage

(i.e.,

yolk-sac, larval, juvenile) using

descriptions and keys in Soin and Sukhanova (1972), Auer (1982), Murty et al

. (1986),

Tweb et al

. (1990) and voucher specimens from Southern Illinois University's Fluid

Vertebrate Collection (B

. Burr, Carbondale, Illinois, USA) and Colorado State

University's Larval Fish Laboratory (D

. Snyder, Fort Collins, Colorado, USA)

. A

subsample of fish from each taxon and stage identified was measured (total length

[TL];

mm) using Scion Image® software, which was calibrated to 0

.5 mm, or metric calipers

(N = 10 per net)

.

Larval samples also were collected in LS and the ILR to characterize the taxa and

abundance within each site and compare larval densities to drift composition at the water

control structure

. I used paired, bow-mounted ichthyoplankton nets (0

.5 m diameter x 2

m long, 500-µm mesh) to sample these habitats on the same dates as drift nets were set

.

Four stratified transects randomly chosen within LS and two in the ILR (river kilometer

7.0 to 9

.0) were sampled weekly

. I stratified backwater transects into two inshore and

two offshore tows, while tows in the ILR were conducted within 1 km up- and

downstream of the LS-ILR confluence .

At each transect, nets were towed at the surface for 5 minutes at a constant speed,

with a calibrated mechanical flow meter (Model 2030R, General Oceanics, Inc

., Miami,

Florida, USA) mounted in the mouth of one net to standardize sampling effort (i

.e.,

volume sampled)

. Inshore backwater tows followed the shoreline

; offshore backwater

tows were straight transects

; and river tows were straight transects taken perpendicular to

flow direction

. River tows started at the main channel border and continued to the main

31

---

32

channel border on the opposite side of the river

. If 5 minutes had not passed during river

transects, the direction was reversed with nets still in the water, and sampling continued

until 5 minutes was reached

. All tow samples were preserved, processed, and identified

as with drift net samples .

LIFE HISTORY CLASSIFICATION

The collected fish were grouped by family into one of three generalized classes to

better explain trends observed between years and treatment groups (e.g.,

Galat and

Zweimuller 2001)

. The classes were as follows

: fluvial specialists, fluvial dependent,

and macrohabitat generalists (Table 1)

. Fluvial specialists are fish that inhabit streams

and rivers throughout their entire life and rarely enter floodplain habitats (Galat and

Zweimuller 2001)

. Fluvial dependent species regularly use lentic backwater or reservoir

habitats, but certain life history traits depend on lotic environments (Galat and

Zweimuller 2001)

. These species are typically broadcast spawners, where developing

eggs and larvae are semi-boyant and passively drift in wind-induced or downstream

currents (Holland 1986)

. Adult fluvial dependent fishes also may make lateral migrations

into slow-flowing lentic areas to spawn-activities likely corresponding with increasing

temperatures and rising water levels (Junk et al

. 1989)

. Macrohabitat generalists include

species commonly found in reservoirs and off-channel habitats that do not depend on

lotic systems (Galat and Zweimuller 2001)

. When these fishes use the river, it is either as

a corridor to move among backwaters (Junk et al

. 1989, Dettmers et al . 2001)

. Spawning

usually occurs in off-channel habitats and offspring generally do not leave this habitat

until the juvenile stage (Holland 1986)

. 1 grouped families based on Galat and

---

Zweimuller (2001

; taxonomy from Nelson [1994]), and the only deviation from their

'

groupings was Sciaenidae, which I classified as fluvial dependent (Dettmers et al

. 2001,

Koel and Sparks 2002) .

DATA ANALYSIS

Larval tow data at each site were standardized as

fish per m 3. Larval drift

represented larval exchange between the backwater and river and was calculated

as

number of fish per minute

. Some fish were caught in nets set opposite to the direction of

flowing water when velocities were 0

.1 m/s or greater (e.g., frame positioned to catch

fish drifting out of LS while water flowed into LS at 0

.11 m/s; Scheidegger and Bain

1995)

. This was not considered drift and thus was removed from data sets before

analyzing

. Total larval abundance from tow data was calculated as the sum of weekly

densities during each sampling season

. All data were logio(x+l) transformed to meet

assumptions of normality

.

Two-way repeated measures ANOVA (proc MIXED, SAS Institute 1999) was

used to test for differences among treatments sampled over time (Hurlbert 1984)

: (1) bi-

weekly drift catch rates into and out of the backwater, with catch rates as the response

variable and week and direction as

predictors, (2) bi-weekly drift catch rates at dawn,

day, dusk, and night, where week and time of day are predictor variables, and (3) weekly

mean densities of tow data between sites were compared, with density as a response

variable and week and site as predictors .

I estimated total number of larvae drifting for each season and determined overall

trends within and between years for larval densities

. Non-zero catch rates of drift data

33

---

for each direction and year were regressed (Type I regression, proc REG, SAS Institute

1999, Sokal and Rohlf 1995) as a function of environmental parameters collected from

the Doppler device corresponding to the time and date of drift

(e.g., depth, temperature,

velocity)

. For significant relationships, these environmental data were used to estimate

the total amount of larval drift for each season

. A one-way ANOVA design tested how

larval density and catch rates differed between years

. To control for experimentwise

error rates, I used Tukey-Kramer post-hoc (Sokal and Rohlf 1995)

. Relationships

between larval densities and lateral drift were examined using linear regression on

untransformed data

.

Proportions of fish per 1-mm total length group were used to compare sizes of

fish between tow sites, drift direction, year, and among diel time points

. Kolmogorov-

Smirnov tests were used to compare (1) size distribution of fish drifting into versus out of

LS in each year, (2) total size structure (into and out combined) of fish drifting during

2004 and 2005, (3) size structure of fish collected from tows, and (4) between-year

differences for LS and ILR tow data

. Multiple pairwise comparisons were used to

examine size structure differences among the times of day sampled (e

.g., dawn vs . night,

dawn vs . day, dawn vs . dusk, etc

.), with Bonferroni adjusted a-values (a"= 0

.05/6).

RESULTS

ABIOTIC PATTERNS

A moderate flood pulse occurred in 2004, during which water levels in the lower

Illinois River were above flood stage for approximately five weeks during June (Figure

34

---

35

10). In contrast, water levels remained at or below normal pool level

(i.e.,

128 m) during

the entire 2005 sampling season

. Water temperatures in the river gradually rose and fell

during 2004, peaking in late July at 28

°C (Figure 10)

. However, during 2005, water

temperatures were more variable and rose over 30

°C twice. In general, water

temperatures in LS varied in a manner similar to those in the ILR, where the amplitude

and daily variation was larger because of the shallower water depths

. Lower Swan depth

fluctuations were slightly dampened compared to the river (Figure 10)

.

Movement of water through the LS water control structure was bi-directional,

often changing direction multiple times a day likely due to river boat traffic and wind,

and velocities varied greatly

. Aside from the 2004 flood pulse which increased water

depth in the channel to 3

.98 m, channel depth remained above 2 m for most of 2004 and

averaged 2

.04 m after May 2005 (Figure 11)

. The bi-directional movement of water into

and out of Swan Lake typically occurred daily, but on average water was moving into LS,

with 86% of differential velocity values being positive (Figure 11)

. Mean channel

velocities were typically less than I m/s during the two sampling seasons, and only three

percent of data points equaled or exceeded this amount

. In 2004, velocities flowing into

LS peaked at 2

.2 m/s during the rise in flood waters, dropped to near zero at flood crest,

and flowed out at peak velocities of 1

.5 m/s during the rapid recession of flood waters .

Near-zero velocities occurred during low water periods, particularly those occurring

during summer 2005, where differential velocity values were clustered around zero

(Figure 11) .

SEASONAL PATTERNS OF LARVAL FISHES

---

36

Larval drift patterns varied between years, while larval fish densities and timing

in LS and the ILR were similar between years

. Fish recruited to drift nets during 1 May

through mid-July 2004, with a major pulse of fish drifting into LS during late May 2004,

timed with the rising flood waters (Figure 12)

. During 2005, a smaller, discrete peak of

larvae was exchanged between LS and the ILR

; three weeks later and at less than half the

magnitude of catch rates (Figure 12)

. Larval fish appeared (garnered from tow data) in

the ILR during mid-April and during late-March 2004 in LS, and there was a

synchronous pulse of larvae in tows during June 2004, the period of floodwater

inundation (Figure 13)

. During 2005, larval fish appeared in both LS and the ILR during

mid-April, and the larval pulses were asynchronous, peaking two weeks later in the ILR

than in LS (Figure 13) .

Higher rates (#/minute) of fish drifted into and not out of LS during 2004,

although differences over time and an interaction between drift direction and time

occurred (Figure 12

; all P < 0 .01 ; direction : F1,6 = 18

.53 ; week : F,1,6o = 7.67

;

direction*week

: F11,60

= 8

.2)

. Ingress and egress of ichthyoplankton were similar in 2005

(Figure 12 ; F1,6 = 3 .8, P = 0 .10)

. However, larval densities in LS were 10 times higher

than the ILR during both years (Figure 13

; two-way repeated measures ANOVA

: P <

0.01

; 2004: F1,6 = 66.97; 2005: F,,6

= 109.78),

although they differed over time (P < 0

.01 ;

2004: F21,91 = 23

.27; 2005 :

F22,76

= 15

.11) and also interacted (P < 0 .01

; 2004:

F20,91 =

4.43 ; 2005: F,8,76 = 3.74)

. Total annual larval abundance estimates were similar between

years in the ILR, 42 and 31 larvae/m3, and LS, 435 and 531 larvae/m3, during 2004 and

2005, respectively

. Mean daily drift rates of larvae by direction did not differ between

years (two-way ANOVA :

F3,,32 = 1

.6, P = 0.21)

despite an almost eight-fold difference

---

37

in rates of larvae drifting into Swan Lake during 2004 and 2005, with means of 4

.2

fish/minute and 0

.5 fish/minute, respectively

. The non-significant between-year result

was due to high variance within year and similar rates of larvae drifting out of the

backwater between years, at 0 .4 fish/minute each year (Figure 12)

.

Both LS and the ILR had homogeneous larval distributions, with no differences in

LS between inshore and offshore (2004 : tj,,9 = -2.08,

P = 0 .05 ; 2005: t,,,s

= -0.66,

P =

0

.52) or in the ILR between upstream and downstream stratified transects (2004

: 11,21 _

0.94, P = 0 .36; 2005 :

11,22 = -0.27, P = 0.79),

though there was a propensity for higher

densities to occur downstream of LS during 2004 .

No diel patterns in catch rates occurred during 2004 (Table 9

;

F3,4 = 0

.1, P >

0

.05), though a diel drift pattern occurred during 2005, where more fish larvae drifted at

night than during other times of the day (F3,4

=

15 .9, P = 0 .01).

Grouping families by life history class revealed that drift composition varied

between years (Table 10, Figure 14)

. Fluvial dependents were most abundant drifting

into Lower Swan during 2004, where sciaenids and cyprinids represented 60% of the fish

exchange (Table 10, Figure 14)

. Macrohabitat generalists were collected in the drift

during 2004 and 2005 without much directional difference, making these taxa some of

the only that drifted out of LS (Figure 14)

.

Temporal drift patterns varied by life history class

. Drift of fluvial dependent

taxa was initiated and peaked in a discrete pulse during rising flood waters

. Macrohabitat

generalists, comprised mostly of clupeids, also drifted in peak rates during the rising

flood waters, beginning in May before the pulse and continuing for two months

. During

---

38

2005, neither timing nor peak drift between fluvial dependents and macrohabitat

generalists varied .

Macrohabitat generalists dominated the ambient larval composition during both

years in tows conducted in LS, but not the ILR

. During 2004, fluvial dependent taxa

dominated the composition in the ILR, and were five times higher than densities of those

taxa in LS

. Regardless of year, fluvial dependent taxa were relatively absent in LS

compared to the densities of macrohabitat generalists (Table 10, Figure 14)

. As with drift

rates during the non-flood year, densities of fluvial dependents were lower in the ILR

(Table 10, Figure 14)

. No fluvial specialists were sampled during either year

. There

were no differences in family composition in tows conducted upstream or downstream of

LS.

Drift rates at LS were related to abiotic and biotic factors

. Non-zero catch rates of

fish drifting into the backwater during 2004 was positively related to velocity, but not

temperature or channel depth (multiple regression model

: adj . ? = 0.92,

F3,13 = 61 .83, P

< 0.01 ; Temperature

: t = 0.64, P = 0.53

; Velocity: t = 13.19, P < 0.01

; Depth: t = 1.23,

P

= 0.24)

. Using the regression model and continuous velocity data, I extrapolated catch

rates to the seasonal duration of drift

(i .e.,

1 May through 15 July) and channel volume

.

An estimated 32

.3 million fish drifted into LS during the 2004 sampling season (mean

catch = [33 .86 * velocity]

- 2 .91)

. Drift catch rates into LS were weakly but positively

correlated with ILR tow-derived densities during 2004 (P = 0

.02, r 2 = 0.36)

. During

2005, although drift was unrelated to abiotic variables, it was related to ambient larval

densities

. Drift into LS was positively related to ILR larval densities (P < 0

.

01

, ? =

0

.78), while LS larval tow-derived densities likely influenced catch rates of drift to the

---

river (P < 0 .01, r2 = 0.94)

. Therefore, abiotic factors influenced 2004 larval drift and

density affected 2005 drift

.

SIZE STRUCTURE OF LARVAL FISHES

Larval fish sizes differed between years and among sites

. Larger fish drifted out

of LS during 2004, while during 2005, larger fish drifted into LS (Table 11, Figure 15)

.

However, there were no differences in the size distribution of larvae caught in tows

upstream and downstream of LS in either year

. Higher river and channel velocities

during the flood pulse (2004) did not entrain larger fish into LS, and, regardless of net

direction (i.e.,

in versus out), larger fish occurred in drift nets during 2005 (KSa = 8

.01, P

< 0.01)

. Despite this change in size structure, larval sizes remained larger in the

backwater than the ILR regardless of year (Table 11, Figure 15)

. Furthermore, both

backwater and river yielded larger larval sizes during 2005 (LS

: KSa = 2 .53, P < 0.01 ;

ILR

: KSa = 10.55, P < 0.01) than 2004.

Lengths of fish drifting at the LS water control structure also varied among times

of day sampled (i.e.,

dawn, day, dusk, night) . Multiple pair-wise Kolmogorov-Smirnov

comparisons revealed a propensity for larger fish to drift at night than at other times

(Table 11)

. In 2004, fish were similarly sized during dawn, day, and dusk (Table 11,

Figure 16)

. Stronger diel patterns occurred during 2005, with night catch collecting the

largest fish and dusk, dawn, and day catches each sampling progressively smaller fish

(Table 11)

. Regardless of year, ichthyoplankton drifting at dawn, day, and dusk were

predominantly less than 8 mm TL, with cumulative percent frequencies between 74

-

80% for each time period in 2004, and between 56

- 82% in 2005

. Almost 70% of larvae

39

---

caught at night were 8 mm or greater (Figure 16)

. Therefore, during 2004 and 2005,

larger fish drifted at night between LS and the ILR

.

DISCUSSION

Seasonal lateral drift occurred at the restored connection between LS and the

lower ILR, with patterns shaped by the annual flood pulse and the fishes' life history

strategies

. The flood pulse concept specifies the importance of coupling increasing water

levels and temperatures to cue spawning and yield high recruitment of riverine fishes

(Junk et al

. 1989, Harvey 1987) . In this study, fish abundance in LS was similar

throughout two physically contrasting seasons, a flood and a non-flood year

. Fish

abundance estimates for the ILR were also similar between years

. Yet these abiotic

conditions induced a change in ambient family densities

(i.e.,

life history classes)

between 2004 and 2005, which may have influenced between-year variation in

magnitude, direction, and composition of lateral larval exchange

. During 2004, the large

pulse of larvae drifting into LS occurred with the rising flood waters, and exchange

between the river and backwater was temporally isolated to the flood event

. Similarly,

peak richness and lateral drift was synchronized, but not correlated, with peak densities

in the river and backwater during rising flood waters

. In contrast, drift had no net

directionality during 2005, and peak drift rates were less than half that of 2004

. Only

during 2005 were ambient larval densities positively correlated to drift rates,

demonstrating a strong biotic influence in the absence of the spring flood

.

The change in life history classes also conveys the significance of abiotic cues to

initiate and influence larval drift

. The propensity for fluvial dependents to drift into

40

---

41

backwaters during the flood year was likely related to the coupling of temperature and

flooding, while the relative absence of these species in the drift and the larval assemblage

was related to their decoupling during the non-flood year

. A lack of a spring flood pulse,

as seen during 2005, may have resulted in poor reproductive or recruitment of fluvial

dependents, where many of those taxa either did not reproduce or their eggs did not hatch

or develop successfully (Humphries and Lake 2000)

. Some between-year differences in

life history classes may be related to the strong reliance of fluvial dependents on annual

flood pulses.

Aside from the influence of abiotic cues on floodplain habitat use by fluvial

dependents, the ecological role of the restored LS, as it contributes to fish reproduction in

the river-floodplain, needs to be identified and evaluated

. In other large river systems,

slackwater areas contribute larvae and juveniles to the river such that densities

downstream of the backwater-river confluence become higher (Muth and Schmulbach

1984, Sheaffer and Nickum 1986)

. However, I did not find any difference in larval fish

densities in the river upstream or downstream of the backwater

. Substantial movement

between the backwater and river occurred, even though Swan Lake seemed to neither

function as

a major sink nor source of larvae

. Despite this, I believe LS to be a vital

component of the lower ILR, with potential benefits to the Mississippi River due to its

close proximity (USACE 1991) .

The low drift rates of macrohabitat generalists denote a behavioral component to

lateral drift, thereby discounting suspicions of LS acting solely as a sink

. The dominant

taxa in the larval assemblage were found in the drift (Jurajda 1995, Reichard et al

. 2001);

yet clupeids were less abundant in the 2004 drift than their ambient densities would have

---

42

warranted

. During 2005, macrohabitat generalists drifted in lower rates despite their

higher densities within the river and LS

. This under-representation of macrohabitat

generalists in the drift may be indicative of drift avoidance (Brown and Armstrong 1985,

Reichard et al. 2001)

. Ultimately, these findings imply a purposeful lateral drift pattern,

where some families avoid or are less prone to drift

.

Additional behavioral components of larval drift were reflected by diel and size

structure patterns

. Nocturnal lateral movement of larger individuals occurred during both

years, while higher catch rates at night only occurred during 2005

. Larvae, particularly

larger fish, may innately drift at the bottom during the day and move to the surface at

night to feed or evade predation (Gale and Mohr 1978, Muth and Schmulbach 1984,

Johnston et al. 1995). Carter et al

. (1986) captured larvae drifting in densities in the

Colorado River almost four times higher at night than during the day

. Other abiotic

factors, such as water clarity, likely drove interannual diel variations, where catch rates

showed no diel patterns during the flood year and were more apparent during low flow

when water transparency was likely higher (Reichard et al

. 2001, Araujo-Lima et al .

2001)

. Therefore, the diel patterns quantified in LS probably resulted from phototaxic

responses and changes in water clarity between years

.

Larger fish drifted out of LS during 2004 than 2005, which supports conventional

larval drift hypotheses, where backwaters function as nursery habitat, provide ideal

conditions for growth, and later become a source for age-0 fish (Sheaffer and Nickum

1986)

. Lateral drift during 2005 was not necessary as a life history strategy given that

the ILR offered amenable habitat similar to backwaters, with slow flows, warm habitat,

and high plankton densities (see Chapter 3)

. Furthermore, additional research on the

---

43

Swan Lake backwater complex showed that a large abundance of juvenile clupeids

emigrated from LS into the ILR during summer 2004, and juvenile sciaenid and moronid

fish during fall 2004 (Schultz 2006)

. Thus, it is my thought that larvae entering the

backwater or spawned in the backwater were able to feed and grow throughout the

season, eventually making an ontogenetic habitat shift by exiting the backwater sometime

that fall and recruiting to the river fishery (King 2004)

.

In restored systems, continuous connectivity should be maintained where

possible, as species use the floodplain throughout the spring and summer

. Limiting

backwater access or reducing river access could impair the recruitment potential of

certain species, such as fluvial dependents, eventually leading to a less diverse riverine

fish community (Turner et al

. 1994)

. Historically, the LS connection to the ILR was

about 99% wider than its post-restoration state, and likely permitted gradual changes in

water level, direction, and velocity

. Currently, water levels lag behind the river, and

typically flow directions and velocities change multiple times per day

. Velocities are

most likely higher through the restricted channel, especially during flood pulses

. It is

possible that the narrowed connection has negatively altered lateral movement of fishes

through the stop-log structure, and that changes in water patterns since the stop-log

structure construction have altered exchange patterns

.

Although I have no larval exchange information from natural systems, the

restricted channel connection between LS and the ILR did not seem to hinder fish

movement (Schultz 2006)

. Instead, one effect of the narrowed channel connection

between LS and the ILR on larval fishes may be to concentrate larvae as they are

transported

; however it is unknown whether this concentration of fishes would have a

---

44

compensatory or depensatory effect on survival, if any

. Few studies have evaluated

restoration techniques and their influence on fish early life stages

. A marine study

demonstrated the efficacy of artificial channels in permitting larval exchange between

estuarine and ocean habitats, showing a similar number of species and concentration

between an artificial and natural estuary-ocean channel, despite a large difference in the

topography and volume of water passing through the channels (Young and Potter 2003)

.

Further research efforts should focus on the effects of interannual variability on drift

patterns (Johnston et al

. 1995), as well as investigate exchange between natural

backwaters and riverine habitat

. Although potentially tedious, studies investigating

lateral movement of ichthyoplankton to backwaters or tributaries should be conducted to

improve energy exchange estimates of river-floodplain systems and enable us to estimate

contribution and productivity of backwater systems to the river

.

This study was unique, as no other study has quantified lateral larval drift in a

river-floodplain system

. Lateral drift is an important component of fishes' life history in

lotic systems, but these strategies and ontogenetic habitat changes may not withstand the

anthropogenic disturbances in our streams and rivers, such as levees and impoundments

.

Altered hydrology may affect larval drift ecology and influence recruitment of fishes

.

This study has shown that lateral drift patterns may be related to the flood pulse, but in

the absence of a flood, the ambient biotic assemblage could influence drift timing,

magnitude, and composition

. Generally, Swan Lake's restoration appears to have

successfully altered the backwater for multiple-use management while maintaining river

connectivity and allowing exchange between the backwater and river to occur

.

Ultimately, future conservation efforts aimed at restoring hydrology should not focus on

---

one particular habitat, but should equally consider main channel, floodplain, and tributary

habitats (Galat and Zweimuller 2001)

.

45

---

CHAPTER 4

MANAGEMENT IMPLICATIONS

I completed two objectives in this study

: (1) to assess the benefit of a restored

backwater system relative to an unrestored, manipulated backwater by quantifying the

response of larval fish communities, and (2) to investigate the interplay of life history

strategies with lateral drift dynamics on a diet and seasonal basis within a restored

backwater system

. The following evaluations were made on tested hypotheses

:

i) Backwater sites, whether restored or unrestored, were used as spawning

and nursery habitats in higher densities during both years than river

segments sampled

. During 2004, the Illinois River (ILR) was used almost

exclusively as spawning habitat, based on the larval size structures

.

However, during 2005, the non-flood year, size structures in the ILR

indicated those river segments were being used as both spawning and

nursery habitat .

ii) Seasonal abundances of larval fishes varied among sites, but this variation

was not solely due to backwater-river connectivity

. Middle Swan (MS)

experienced the highest larval fish densities during both years, but had

limited river connectivity

. I believe high larval fish densities can be

attributed to the annual drawdowns that promoted aquatic vegetation as

well as the brief periods of connectivity

.

iii) Family composition, described by the life history classification, varied

among sites and between years

. During 2004, the ILR and

46

---

47

MS had high densities of fluvial dependent larval fishes compared to

Lower Swan (LS) and Calhoun Point

(CP), which were dominated by

macrohabitat generalists

. Despite LS having the highest river

connectivity, macrohabitat generalists dominated the assemblage during

both years

. Fluvial dependent taxa were relatively absent in all sites

during 2005, the non-flood year

.

iv) Larval exchange occurred between the restored backwater, LS, and the

lower ILR

. However, the influx of larvae into the backwater occurred

only during 2004 and was more related to velocity than to water

temperatures or river stage

. Although drift rates were not different

between years, timing and composition seemed more related to ambient

larval densities in the ILR and LS than abiotic conditions

(i .e.,

water

temperature, water depth, velocity) .

v) Although Swan Lake is a major backwater of the lower ILR, density

differences between river segments upstream and downstream of the

backwater-river confluence did not occur

. It is unclear whether restriction

of the LS-ILR confluence affected the magnitude of larval movement

between the river and backwater

. It may have reduced larval movement or

may have concentrated larvae moving through the connection

.

vi) Diel periodicity in larval drift patterns occurred at the LS-ILR confluence

.

During both years, larval fish drifting at night were larger than those

drifting at other times of day

. During the non-flood year, larval fishes also

drifted in higher rates at night

.

---

48

The results from this study were intended to provide researchers and managers

with information on the role restored backwaters play in the life history strategies of

larval fishes and to help guide the development of future backwater management and

restoration programs

.

In restored systems, continuous connectivity should be maintained where

possible, as species use the floodplain throughout the spring and summer

. Limiting

backwater access could impair the survival of some families, such as fluvial dependents

that require multiple habitat types to complete their life history, eventually leading to a

less diverse river fish community

. Connectivity in the Swan Lake HREP was preserved

through the construction of water control stop-log structures

. These fixed structures

either created new river connections or severely restricted original river confluences

.

The effect of the narrowed channels on larval movement between the river and

backwaters cannot be assessed due to the lack of information on natural systems

. Despite

this, the Swan Lake HREP appears to have successfully restored the backwater while still

maintaining its river connectivity.

In addition to reconnecting the floodplain to the river, natural flood pulses serve

an important biological role in river ecosystems

.

The flood pulse concept specifies

the importance of coupling increasing water levels with rising temperatures during spring

to cue spawning and yield high recruitment of fishes

. In this study, fish production in LS

and the ILR was similar throughout two physically contrasting seasons, a flood and a

non-flood year

. These abiotic conditions induced a change in ambient family densities

between 2004 and 2005, which may have influenced between-year variation in

magnitude, direction and composition of lateral larval exchange

. A lack of a spring flood

---

49

pulse, as seen during 2004, may have reduced reproduction or recruitment of fluvial

dependents.

Although river connectivity is an important attribute of backwaters and should not

be overlooked, maintaining this connectivity might not be enough to ensure successful

habitat restoration

. Increasing the habitat heterogeneity through system-wide

management regimes can positively influence larval fishes and other target organisms by

directly impacting organic matter and macroinvertebrate and zooplankton densities (Flinn

et al. 2005)

. Often, natural backwaters go through an annual succession from lotic to

lentic conditions, thereby accommodating several life history strategies and promoting

species richness (Aarts et al . 2004)

. Restored systems that are drawn down to consolidate

sediments and promote vegetation growth also have greater water clarity and will likely

be productive areas for aquatic vegetation and zooplankton (Flinn et al

. 2005)

. By

increasing food resources and offering protective refuge, river restoration schemes aimed

at enhancing habitat diversity in off-channel areas can have similar beneficial effects by

creating a stable, diverse, and abundant age-0 fish community (Pezold 1998, Langler and

Smith 2001) .

However, research conducted to establish the role of off-channel habitat in fish

recruitment (Langler and Smith 2001) and to determine the function of connectivity in

recruitment models is still necessary

. Schultz (2006) showed that because LS was

continuously connected to the ILR, mass emigrations of juvenile fishes could occur

during the late summer and fall

. Additionally, restoration projects should not be viewed

as sustainable once completed

. Natural floodplain succession is stifled by an

immobilized main channel, so restored and unrestored backwater systems must be

---

50

actively managed to prevent sedimentation from degrading the habitat (Lusk et al

. 2003,

Aarts et al

. 2004) .

Therefore, human restoration and management of floodplain habitat can

potentially create valuable spawning and nursery areas for many fish species, and offer

the necessary habitat diversity lost since the degradation of natural floodplain regions

(Grift et al

. 2003). These alterations can restore fragmented areas, providing

macrohabitats in proximity, thereby meeting specific requirements of riverine species

early life stages (e.g.,

fluvial specialists, fluvial dependents

; Aarts et al . 2004). Also, the

important interplay among hydrologic patterns, habitat quality and availability, and the

ontogeny of larval fish in river-floodplain systems should be noted, where flood regimes

influence family composition and production improves with habitat diversity

. Our

continued attention to river connectivity is paramount, but the importance of habitat

heterogeneity in larval production and assemblage structure should not be neglected

.

---

Table 1

. Fish families commonly found in the Illinois River basin grouped into one of

three life history classes (e.g.,

Nelson 1994, Dettmers et al

. 2001, and Galat and

Zweimuller 2001), fluvial specialists

(FS), fluvial dependents (FD),

and macrohabitat

generalists (MG) .

51

Life History

Life History

Family

Class

Family

Class

Acipenseridae

FS

Hiodontidae

FD

Amiidae

MG

Ictaluridae

MG

Anguillidae

MG

Lepisosteidae

MG

Aphredoderidae

MG

Moronidae

FD

Atherinidae

MG

Percidae

MG

Catostomidae

FD

Percopsidae

MG

Centrarchidae

MG

Petromyzontidae

FD

Clupeidae

MG

Poeciliidae

MG

Cyprinidae

FD

Polyodontidae

FD

Esocidae

MG

Sciaenidae

FD

Fundulidae

MG

Umbridae

MG

Gadidae

MG

---

52

Table 2. Monthly means (± SE) for abiotic variables (temperature [°C],

dissolved oxygen [DO, mg/L], water depth [m], secchi depth

[cm], wind speed [km/h], channel velocity [m/s]) collected at sites during 2004 in the Illinois River system

.

Site

Temp

DO

Depth

Secchi

Wind

Velocity

Illinois River

March-04

7.81 (0.56)

7.58

(0.98)

22.6 (0.6)

7.0 (2.2)

April-04

11.22 (1 .10)

7.86 (0.24)

6.25 (1 .75)

27.5 (5.5)

9.2 (1 .3)

0.46

(0.07)

May-04

18.62 (1 .07) 5.03 (0.70)

6.07

(0.54) 25.2 (4.2)

6.4 (1 .3)

0.54

(0.15)

June-04

23 .28 (0.44)

4.49 (0.80)

7.89

(0.14) 23.6 (3.7)

8.3 (2.2)

0.63

(0.11)

July-04

25 .91 (0.54)

3 .61 (0.15)

6.62 (0.34) 30.3 (3.1)

7.0

(0.8)

0.55

(0.13)

August-04

26.87 (0.87)

3 .85 (0.13)

6.44 (0.69)

29.1 (0.7)

8.4 (3 .3)

0.19

(0.07)

September-04

23 .77

8 .18 (0.35)

6.81 (0.57)

25.6 (2.4)

7.7 (1 .0)

0.08

Lower

Swan

March-04

13.90 (1.75)

7.26 (1 .00)

0.87 (0.25)

15.3 (1 .9)

13 .6 (2.7)

April-04

16.42 (2.30)

8.02 (0.55)

0.77 (0.03)

23.2 (0.5)

9.0 (1 .6)

May-04

22.03 (1.31)

6.31 (0.55)

1 .01 (0.07)

19.5 (1 .2)

8 .2 (1

.8)

June-04

25.05 (1 .08)

8.61 (1 .53)

2.09 (0.10)

32.7 (1 .5)

6.8 (1 .3)

July-04

27.58 (0.77)

6.30 (0.65)

0.96

(0.20)

22.7 (3 .5)

4.7 (0.6)

August-04

24 .02 (0.64)

3 .78 (0.27)

0.70 (0.14)

14.3 (0 .8)

6.3 (3 .0)

September-04

20 .24 (0.64)

5.08 (1 .03)

0.77 (0.03)

14.1 (1 .4)

6.4 (0.6)

Middle Swan

March-04

7.64 (0 .95)

1 .33

(0.05)

25 .6 (7.6)

4.0 (0.4)

April-04

14.40 (0.76)

9.53 (0.65)

1 .26 (0.06)

27.8

(1 .0)

10.0 (2.1)

May-04

22.29 (2.01)

6.50 (0.54)

1 .13

(0.10)

24.1

(3 .5)

7.8 (2.4)

June-04

25 .18 (0.96)

8.08 (1 .73)

2.03

(0.17)

34.5 (2.2)

9.6 (1 .4)

July-04

28.41 (1 .35)

7.46 (0.86)

1 .25 (0.26)

33 .6 (6.2)

7.6 (1 .1)

Calhoun Point

March-04

7.12 (0.42)

0.68 (0.33)

28.0 (1 .0)

9.3 (5 .4)

April-04

6.13 (0.05)

0 .80 (0.00)

19.0 (1 .0)

12.9 (5 .7)

May-04

25.27

4.97 (1 .27)

0 .82

(0.08)

19.6 (0.6)

12.6 (0.3)

June-04

24.85 (1 .80)

6.00 (3.31)

1 .45 (0.20)

44.1 (18 .4)

4.7

(3

.5)

July-04

29 .34 (2.44)

7.10 (0.96)

0.79

(0.11)

29 .1 (3.6)

6.0 (1 .9)

August-04

26.63 (0.82)

5.15 (0.70)

0.77 (0.03)

18 .3 (0.8)

8.9 (2.3)

September-04

21 .34

4.45 (0.05)

0.83

(0.02)

19 .0 (1 .0)

14.5 (0.6)

---

53

Table 3 . Monthly means (± SE) for abiotic variables (temperature [°C],

dissolved oxygen [DO, mg/L], water depth

[m], secchi depth

[cm],

average wind speed [km/h], channel velocity [m/s]) collected at tow sites during 2005 in the Illinois River system

.

Site

Temp

DO

Depth

Secchi

Wind

Velocity

Illinois River

March-05

6.57

5.25 (0.75)

27.0

(3.0)

14.5 (0.4)

0.44

April-05

15 .16 (1 .37)

6.66 (1 .77)

6.91 (0.41)

28.9 (6.6)

9.9 (1.9)

0.50 (0.06)

May-05

18 .68 (1 .72)

9.66 (2.19)

6.51 (0.22)

18 .1 (4.3)

10.1 (1.1)

0.37 (0.03)

June-05

26.64 (1 .10)

9.69 (1 .15)

6.36 (0.22)

25.6 (1.3)

5.8 (1 .3)

0.16 (0.03)

July-05

29.60 (0.36)

7.04 (0.51)

6.00 (0.14)

24.9 (1 .9)

4.6 (0.7)

0.15 (0.04)

August-05

28.44 (0.53)

6.78 (1 .24)

6.50

(0.23)

25 .9 (3 .2)

6.1 (2.1)

0.12 (0.04)

September-05

26.65

8.94 (0.31)

6.13 (1 .38)

25 .1 (0.6)

3 .4 (1 .1)

0.15

Lower Swan

March-05

8.86 (0.33)

0.78 (0.03)

12.5

(0.2)

7 .7 (2.7)

April-05

16 .15 (2.18)

8.31 (1 .42)

2.11 (1.37)

15.3 (1 .8)

13 .2 (0.7)

May-05

20 .94

(2.35)

9.25 (1 .32)

0.75 (0.03)

14.5 (1 .8)

8 .9

(1 .7)

June-05

27 .09 (0.62)

10 .02 (1 .61)

0.73 (0.04)

14.4 (0.6)

5 .0 (3 .2)

July-05

29

.47 (1 .34)

9 .54 (1 .27)

0.68 (0.06)

14.5 (1 .2)

7.1 (1 .2)

August-05

27 .53 (1 .31)

5 .76 (1 .18)

0.73 (0.04)

13.5 (1 .0)

5 .6 (1 .3)

September-05

26.62

12 .72 (1 .65)

0.70 (0.09)

13 .3 (0.5)

4.5

(1

.4)

Middle Swan

March-05

7.74

1 .08 (0.21)

23 .3 (1 .2)

9.1

(2.5)

April-05

16 .81 (1 .89)

8.31 (1 .40)

0.95 (0.07)

20.9 (1 .6) 12.3

(2.4)

May-05

21 .45 (2.34)

9.25 (1.13)

0.89 (0.07)

21 .5 (0.5)

8.6 (2.0)

June-05

27.54 (0.91)

7.45 (1.26)

0.97 (0.03)

19 .1 (2.4)

7.6 (2.0)

---

54

Table 4

. Site comparisons tested for overall differences in larval densities during 2004

and 2005 among Calhoun Point (CP),

the Illinois River (ILR), Lower Swan (LS), and

Middle Swan (MS), with adjusted P-values .

Site Comparison

t

df

P

2004

CP vs. ILR

1 .61

48.1

0.38

CP vs. LS

-1

.50

48.3

0.45

CP vs. MS

-2.51

48.0

0.07

ILR vs. LS

-4.04

46.6

< 0.01

ILR vs. MS

-4.85

47.7

< 0.01

LS vs. MS

-1

.32

47.7

0.55

2005

ILR vs. LS

-4

.43

35 .3

< 0.01

ILR vs. MS

-7

.11

36 .6

< 0.01

LS

vs. MS

-3.07

37.5

0.01

---

Table 5

. Comparisons of stratified transects within each site for 2004 and 2005

in the

lower Illinois River system, where backwater systems had inshore and offshore transects

(Calhoun Point [CP],

Lower Swan [LS],

and Middle Swan [MS]), and the Illinois River

(ILR) had upstream and downstream transects at CP and LS

.

55

Within-Site Comparison

t

df

P

Inshore vs. offshore

CP 2004

0.72

11 .0

0.49

LS 2004

-2.08

18.9

0.05

LS 2005

-0.66

18.0

0.52

MS 2004

-1 .96

13.8

0.07

MS 2005

-1 .49

12.2

0.16

Upstream vs. downstream

ILR at CP 2004

-0.04

10.0

0.97

ILR at LS 2004

0.94

22.0

0.36

ILR at LS 2005

-0.27

22.0

0.79

---

56

Table 6

. Fish families grouped into one of three life history classes (e .g.,

Galat and

Zweimuller 2001) with percent of total catch during 2004 and 2005 for the Illinois River

(ILR), Lower Swan (LS), Middle Swan

(MS), and Calhoun Point (CP).

2004 % Catch

2005 % Catch

Family

ILR

LS

MS

CP

ILR

LS

Fluvial Specialist

0.0

0.0

0.0

0.0

0.0

0.0

Fluvial Dependent

Catostomidae

1 .1

0.2

24.7

0.0

0.3

0.2

Cyprinidae

5.7

2.2

5 .2

4.2

7.8

3 .1

Hiodontidae

0.0

0.0

< 0.1

0.0

0.0

0.0

Moronidae

1 .1

< 0.1

0.0

< 0.1

0.1

< 0.1

Sciaenidae

34.0

0.2

< 0.1

0.0

0.6

0.7

Macrohabitat Generalist

Atherinidae

< 0.1

< 0.1

0.0

0.5

0.2

0.2

Centrarchidae

1

.5

1 .9

2.3

8.5

1 .3

0.6

Clupeidae

56.4

95 .4

67.5

86.8

89.6

94.9

Gasterosteidae

< 0.1

0.0

0.0

0.0

0.0

0.0

Ictaluridae

0.0

0.0

0.0

0.0

0.0

< 0.1

Lepisosteidae

< 0.1

< 0.1

0.2

< 0.1

0.0

0.0

Percidae

0.0

< 0.1

0.0

0.0

0.0

0.0

Poeciliidae

< 0.1

< 0.1

< 0.1

0.1

< 0.1

0.4

Total N

8,086

48,390 105,906 11,500

4,476 56,033

---

57

Table 7. Length analysis of larval fish caught in Lower Swan

(LS), Middle Swan

(MS),

Calhoun Point

(CP), and the Illinois River (ILR) using Kolmogorov-Smirnov pairwise

comparisons, with results listed for sites or years that collected larger-sized

fish. MS was

drawn down before the end of the 2004 sampling season

. Therefore, pairwise

comparisons with this site were conducted on truncated data sets that included only

samples taken on before 23 July 2004

.

* Indicates significant differences between treatments, where P

< a or a".

Site Comparison

KSa

P

Results

2004

a"-value = 0.008, Bonferroni adjusted

LS

vs. MS

4.67

< 0.001 *

MS

LS

vs. CP

3 .49

< 0.001 *

CP

LS

vs.

ILR

8.83

< 0.001

*

LS

MS vs. CP

1.74

0.005*

CP

MS vs.

ILR

12.79

< 0.001 *

MS

CP

vs. ILR

11 .54

< 0.001 *

CP

2005

ce-value = 0.05, not adjusted

LS

vs. ILR

5 .37

<0 .01*

LS

2004

vs. 2005

LS

2.53

< 0.01 *

2005

ILR

10.55

<0 .01*

2005

---

58

Table 8

. Mean zooplankton density (#/L) ±1

SE by taxa described as a percent of the

total annual density for each site in 2004 and 2005

.

Cladoceran

Copepod

Nauplii

Rotifer

Site

Density

(SE)

o

/o

Density

(SE)

o/o

Density

(SE)

o o

Density

(SE)

o o

2004

Illinois

0.20

< 0.1

1 .53

< 0.1

28.79

0'6

River

(0

.08)

(0 .45)

(7 .38)

(2753)

4868

99.4

LowerSwan

(0

0

.04)

.15

< 0.1

(0

1.14

.35)

0.1

35

(7.87)

.00 1.8

1897

(365)

98.1

MiddleSwan

(0

1 .10

.54)

0.1

(3

8

.90)

.47

0.4

105

(36.22)

.54

4.8

2101

(542)

94.8

CalhounPoint

(0

0

.09)

.32

< 0.1

(0

0

.38)

.98

< 0.1

(13

40

.00)

.50

1 .6

2443

(745)

98.3

2005

Illinois

0.87

0.69

11 .04

1888

River

(0 .31)

0 ' 1

(0.23)

<0.1

(3 .29) 0 '6

(241)

99.3

Lower

0.31

< 0.1

1 .74

0.1

61 .97

1 .6

3774 98.3

Swan

(0 .08)

(0

.56)

(23.65)

(618)

---

N

Catch (#/minute)

59

Table 9

. Drift of fish during 2004 and 2005 in Swan Lake, Illinois River, with net sets

averaged by time period (standard error represents ±1 of mean catch rate)

.

Note

: Night net sets were not conducted during rising flood waters during 2004, which

reduced net set count and likely mean night catch rate

.

Year

Time of day

Net sets

Fish

Mean ± SE

2004 Dawn (0450-0711)

24

1,956

1 .80± 1 .76

Day

(1215-1500)

24

2,551

2 .38±2.00

Dusk (1815-2050)

22

1,137

0.96 ± 0.94

Night (0000-0110)

20

112

0.12 ± 0.08

2005 Dawn (0450 - 0640)

24

490

0.45 ± 0.42

Day (1145 -1435)

23

77

0.07 ± 0.05

Dusk

(1815-2110)

26

70

0.06±0.04

Night (2340-0110)

20

869

0.93+0.79

---

Table 10

. Fish families grouped into one of three life history classes (Galat and

Zweimuller 2001) with percent of total catch during 2004 and 2005 by gear

. Drift net

total catch for 2004 : N= 5,756; for 2005 : N= 1,506

. Tow net total catch for 2004 : N=

56,476; for 2005 : N= 60,509

. No fluvial specialists were caught in either drift or tow

nets .

60

2004

2005

Family

Drift Nets Tow Nets

Drift Nets Tow Nets

Fluvial Specialist

0.0%

0.0%

0.0%

0.0%

Fluvial Dependent

Catostomidae

8 .4%

0 .3%

1 .6%

0.2%

Cyprinidae

28.8%

2 .7%

1 .0%

4.9%

Moronidae

0.2%

0.2%

0.0%

<0.1%

Sciaenidae

31

.0%

5.0%

2.2%

0.4%

Macrohabitat Generalist

Atherinidae

0.0%

< 0.1%

0.0%

0.2%

Centrarchidae

0.3%

1 .8%

0.7%

1 .1%

Clupeidae

30.6%

89.8%

92.6%

92.5%

Gasterosteidae

0.0%

<0.1%

0.1%

0 .0%

Ictaluridae

0.0%

0.0%

0.5%

< 0.1%

Lepisosteidae

0.1%

< 0.1%

0.0%

0.0%

Percidae

< 0.1%

< 0.1%

0.1%

0.0%

Poeciliidae

< 0.1%

< 0.1%

1 .2%

0.6%

---

61

Table 11 . Kolmogorov-Smimov tests for 2004 and 2005 pooled diel drift net data and

Lower Swan Lake (LS) and the Illinois River (ILR) larval tow data . Test results are

listed for treatments which collected significantly larger-sized fish .

* Indicates significant differences between treatments, where

P :5 a or a"

.

Treatment

2004

2005

KSa

P

Result

KSa

P

Result

IN vs

. OUT

10.91 <0.01*

Out

4.41

0.01*

In

LS vs. ILR

8 .83 <0.01*

LS

5.37 < 0.01*

LS

NIGHT vs

. DAWN 9.48 < 0.001* Night

7.78 < 0.001* Night

NIGHT vs. DAY

9.57 < 0.001

Night

11 .99 < 0.001* Night

NIGHT vs. DUSK 10.93 <0.001* Night

2.55

< 0.001

*

Night

DAWN vs. DAY

0.56 0.914

9.70 < 0.001* Dawn

DAWN vs

. DUSK

2.03 < 0.001 Dusk

6.06 < 0.001* Dusk

DAY vs . DUSK

1 .48 0.026

4.74

< 0.001* Dusk

---

2 km

Mississippi River

62

Figure 1

. Study site of the lower Illinois River and two major backwaters, Swan Lake

and Calhoun Point.

---

Oct-04 Feb-05 Jun-05

Oct-05

Time (d)

Figure 2 . Mean daily temperature of Illinois River and backwater sites during 2004

through 2005, depicted as solid gray lines

. Mean daily depth data are depicted as broken

black lines for the Illinois River and Lower Swan

. Channel depth of Lower Swan was

recorded using a submerged device at the stop-log structure

. River data were collected at

Hardin, Illinois, USA (river kilometer 34

.4)

. Depth was not recorded for Middle Swan

and Calhoun Point

. Shaded regions represent sampling periods

.

63

40

131

E

30

130

wa7

20

129

a)

10

128

i

0

127

---

Low temperature 4

High DO

High temperature

Low DO

2

-2 -

-4

-4

O Illinois River

A Lower Swan

V Middle Swan

A

•

Calhoun Point

0

:A

0

4V

AZ 0

0o% it ct

, I

0 L,

VAA+--

A4

0

El

f

00

V

0

0 0

High depth

High secchi

Low wind

-2

0

Axis 1

2

4

Low depth

Low secchi

High wind

64

Figure 3

. Principle component analysis (PCA) of abiotic variables for the lower Illinois

River (0) and backwater sites, Lower Swan (A),

Middle Swan (V), and Calhoun Point

( •). Weekly means of abiotic variables were included in the analysis

: temperature

(°C),

dissolved oxygen (DO, mg/L), depth (m),

secchi depth (cm),

and wind (km/h) .

---

500

400

300

200

100

0

Apr-04

Jun-04 Aug-04 Oct-04

65

Time (d)

Figure 4 . Mean density (#/m3 )

per day of fish caught in the lower Illinois River during

2004 and 2005

. Error bars represent ±1 standard error of the mean transect density

.

Calhoun Point was not sampled during 2005 because of low water levels

.

20041

Illinois River

22005J

Illinois River

2004

Lower Swan

2005

Lower Swan

5*

4 }

2004

2005

Middle Swan

Middle Swan

4

2004

Calhoun Point

Apr-05 Jun-05 Aug-05 Oct-05

400

20

10

0

400

300

M

200

100

1'

0

N

dJ

C

400

C

300

200

100

0

---

20

M1

E

5

0

FS FD MG

FS

Habitat Use Class

FD MG

66

Figure 5. Fluvial specialist (FS), fluvial dependent (FD),

and macrohabitat generalist

(MG) mean densities by site and year in the lower Illinois River system

. Error bars

represent t1 standard error of the mean of dates sampled during that year

.

Illinois River

Middle Swan

2004

15

∎

2005

10

Lower Swan

Calhoun Point

---

IL,

LL

CC4f7

a

0

40

30

20

10

0

30

2004

Illinois River

Avg . N = 2.757

Mean = 9.6 m m

Median = 5.5

mm

S .O . = 9.0 MM

k

A

2005

Illinois River

Avg . N = 3,112

Mean= 10.9

mm

Median = 7.2 mm

S .D . = 10.6 mm

2004

Lower Swan

Avg . N = 23,432

Mean = 18.2 mm

Median = 13.4

mm

S .D . = 15.7 mm

2005

Lower Swan

Avg . N = 21,421

Mean = 19.2 mm

Median= 12

.7 mm

S .D . = 15.3 mm

20 40

2004

Calhoun Point

Avg. N = 6,588

Mean = 13.8 mm

Median = 9.7

mm

S .D . = 11 .5 mm

60

80

100

67

Length group (mm)

Figure 6 .

Length frequency distributions of all larval and juvenile fish caught in each site

in of the lower Illinois River system during 2004 and 2005

. Sample size is an average of

transects within that site summed over the sampling season

. Length frequencies are

expressed as a percentage of average catch

. Middle Swan samples were not collected

after 23 July 2004

. During 2005, lengths were collected in Middle Swan samples, and

Calhoun Point was too shallow to be sampled

. Descriptive statistics were calculated on

raw length data .

2004

20

40

60

80

100

Middle Swan

Avg

. N = 36,991

Mean = 15 .8 mm /

Median = 13

.6 mm

S .D . = 10 .9 mm

---

200

150

100

50

0

150

100 -

200

0

150 -

100

50 -

Time (d)

68

Figure 7

. Mean density (#/L) of cladocerans, copepods, nauplii, and other zooplankters

per day in each site of the lower Illinois River system through 2004 and 2005

. Error bars

represent ±1 standard error of the mean of transects sampled that day

. Middle Swan

samples were not collected after 23 July 2004

. During 2005, zooplankton data were not

identified for Middle Swan and no zooplankton samples were taken in Calhoun Point

.

-

Illinois

2004

River -

Illinois

•

2005

River

2004

2005

Lower Swan

f

Lower Swan

f,**,* 1 }+

to

.+•+

+ •

f

f+

2004

Apr-05 Jun-05 Aug-05

Oct-05

j

Middle Swan

~j

T

.-.

j

2004

j 1

Calhoun Point

7

4 ••}

Apr-04 Jun-04 Au g04 Oct-04

---

12000

10000

8000

6000

4000

2000

8000

6000

4000

2000

0

0

9

00

0

00

0

Apr-04

Jun-04

Aug-04

Oct-04

Time (d)

Figure 8

. Mean density (#/L) of rotifers per day in each site of the lower Illinois River

through 2004 and 2005

. Error bars represent tl standard error of the mean of transects

sampled that day

. Middle Swan samples were not collected after 23 July 2004

. In 2005,

zooplankton data were not identified for Middle Swan and no zooplankton samples were

taken in Calhoun Point

.

69

2004

Illinois River

2005

Illinois River

I

0

0

0

4

060

0

4

41

0

00 000

00

2004

.

Lower

Swan

2005

Lower wan

4,

00

f

o

~

00

o

0

~~

J o

O 0

2004

Apr-05 Jun-05 Aug-05

Oct-05

0

Middle Swan

0000

$§

0

2004

Calhoun Point

0

10000

8000

6000

J

40002000

Yy

0

CCD 10000

O

8000

6000

4000

2000

0

10000

---

Access road

H5m

Illinois

River

(Flow direction)

70

Figure 9

. Fixed drift net site during 2004 and 2005 on the lower Illinois River (top inset)

.

Tandem nets were floated on the Lower Swan (LS) side of the stop-log water control

structure

. Schematic depicts nets sampling larval fish drifting out of LS

.

Bottom inset

portrays the drift net frame positioned to sample larval fish drifting into LS

.

---

oU

20

a)

N

10

CL

)

1

0

127

I-a)

I

5 ^

30

4 ta

20

3 a

2

C

10

1C

N

00

0

Feb-04

Jun-04

Oct-04

Feb-05 Jun-05

Oct-05

71

^

Time (d)

Figure 10 . Mean daily temperature of the Illinois

River and Lower Swan for 2004 and

2005, depicted as solid gray lines

. Mean daily depth data are depicted as broken black

lines

. Channel depth of Lower Swan was recorded

at the stop-log structure between the

backwater and the river . River stage and temperature

data were collected at Hardin,

Illinois, USA (river kilometer 34

.4).

Shaded regions represent sampling periods .

---

Time (d)

Figure 11

. Mean daily depth (top) and differential velocity (bottom) measurements in the

channel between Swan Lake and the Illinois River

. Differential velocity was calculated

as the sum of mean channel velocities per day

(i.e., the sum of 48 values)

. Positive

velocity values represent net inflow of water into Swan Lake and negative values

correspond to net outflow to the river

. Continuous data were recorded using a submerged

device anchored at the stop-log structure

. No depth data were collected from September

through November 2004, and no velocity data were collected from October 2004 through

April 2005 . Shaded regions represent the 2004 and 2005 sampling periods

.

72

---

3

G)

2-

*

1

C

1-

0

0

75 -

2005

4

s

20 -

3

V

15-

5-

1

0

0

March April

May

June

July August

Time (d)

73

Figure 12 . Drift of ichthyoplankton into

(•) and out (0) of Swan Lake in

2004 and

2005 . Symbols represent

mean catch (#/minute) per

net set and line data represent

continuous depth data recorded at the stop-log structure

. Error bars represent ±1

standard

error for the mean of each net set.

---

200

150 -

•

Illinois River

O Swan Lake

0

2004

30 -

20 -

10-

0

March April May

June

July

August

Time (d)

Figure 13

. Mean density (#/m3)

per day of fish caught in the Illinois River and Swan

Lake during 2004 and 2005

. Error bars represent ±1 standard error of the mean transect

density .

74

---

.~

C

E

M

8 .0

6 .0

0 .6

0 .4

0 .2

6 .0

0 .6

0 .4

0 .2

0 .0

75

FS

FD

MG

FS

Habitat Use Classes

FD MG

Figure 14

. Mean catch rate (#/minute) of fish in three early life history classes drifting

into and out of Swan Lake and mean density (#/m3)

of classes in the Illinois River and

Swan Lake during 2004 and 2005

. All families were included and grouped into one of

three early life history classes as defined by Galat and Zweimuller (2001)

: fluvial

specialists (FS),

fluvial dependents (FD),

and macrohabitat generalists (MG) .

Error bars

represent f 1 standard error of the mean

.

Drift Into Swan

2004

2005

6.0

0 .2

Drift Out of Swan

~

0.1

Illinois River

Swan Lake

6.0

4.0-

2.0-

---

UC0)0Q

N

U ..

0

40

30

20

10

0

30

10

5

30

20

10

0

30

20

10

0

0

20

40

60

0

20

Length group (mm)

40

60

80

76

Figure 15

. Length frequency distributions of all larval and juvenile fish caught in the

Illinois River, Lower Swan, and in the drift during 2004 and 2005

. Sample size is an

average of transects or nets within that site summed over the sampling season for larval

tow data or drift data, respectively

. Length frequencies are expressed as a percentage of

the average caught

. Descriptive statistics were calculated on raw length data

.

Illinois

2004

River

Avg . N=2 .757

Mean =9.6 mm

Median 5.5 mm

S .D . = 9.0 mm

2005

Illinois River

Avg

. N=3,112

Mean = 10.9 mm

Median 7 .2 mm

S .D . = 10.6 mm

2004

Lower Swan

Avg . N = 23,432

Mean = 18,2 MM

Median 13 .4 mm

S .D . = 15.7

mm

Lower Swan

2005

Avg

. N = 21,421

Mean= 19,2 MM

Median 12 .7 mm

S .D . = 15.3

mm

2004

Drift Into

Avg . N = 2,213

Mean = 7.9 mm

Median 7.4 mm

S .D . = 5.4

mm

2005

Drift Into

Avg . N

= 410

Mean = 8.3 m m

Median 6.9 mm

S .D . = 4.1 mm

2004

2005

Drift Out

Drift Out

Avg . N =

185

Avg . N = 296

Mean =8.0 mm

Mean = 9 .2 mm

Median 7.6 mm

Median 7.8 mm

S .D . =4.6 mm

S.D . =4 .8 mm

---

10 20

30

40

0

10

20

Length group (mm)

40 50

77

Figure 16

. Length frequency distributions of all fish caught drifting during each time of

day for 2004 and 2005

. Data of fish drifting into and out of Lower Swan were pooled by

time of day (i.e., dawn, day, dusk, night)

. Length frequencies are expressed as a

percentage of the mean number per net set (i

.e., mean of three nets) summed over the

sampling season

. Descriptive statistics were calculated on raw length data

.

2004-Dawn

Avg . N= 693

Mew =7.7mm

Median=7.5mrr

S .D. =6.2mm

.

2005-Dawn

Avg. N= 217

Mean=9.7rrm

Medan=7.2nm

S .D. =4.8rrm

2004-Day

Avg. N=1,243,

Mean= 7.2rrr

Median=7.2mrr

S .D. =2.8mrr

2005 - Day

Avg. N=41

Mean= 6 .5nm

Median= 6 .4nm

S .D.=1 .4mn

2004 - Dusk

Avg . N= 406 .

Mean= 8 .5mrr

Median= 7

.2mrr

k~~_

10_63mw

.

. .S.D.=5.6mn

2005-Dusk

Avg. N= 39

Mean=7

.7mn

Medan=6 .3mm

k

1l

1

2004-Night

2005-Night

Avg. N= 57

Avg . N= 409

Mean= 9.3mrr

Mew= 10 .0rrm

Median =9.Omrr

Medan=9 .5mn

L

S .D. = 2.7 mrr

A

S .D. =4 .1am

40

30

20

10

0

30

20

V

C 10

N

0I4)

.

o

L

0

2030

10

30

20

10

0

---

LITERATURE CITED

Adams, S . R., T

. M. Keevin, K . J. Killgore, and J

. J. Hoover. 1999 . Stranding potential

of young fishes subjected to simulated vessel-induced drawdown

. Transactions of

the American Fisheries Society 128 :1230-1234 .

Auer, N. A

. 1982

. Identification of larval and juvenile fishes of the Great Lakes Basin

with emphasis on the Lake Michigan drainage

. Great Lakes Fishery Commission,

Special Publication 82-3, Ann Arbor, Michigan .

Auer, N . A., and E . A. Baker. 2002

. Duration and drift of larval lake sturgeon in the

Sturgeon River, Michigan

. Journal of Applied Ichthyology 18 :557-564 .

Agostinho, A . A., L. C. Genies, S

. Verissimo, and E . K. Okada

. 2004. Flood regime,

dam regulation, and fish in the Upper Parana River

: effects on assemblage

attributes, reproduction and recruitment

. Reviews in Fish Biology and Fisheries

14(1): 11-19.

Aoyagui, A. S ., and C . C. Bonecker. 2004

. Rotifers in different environments of the

Upper Parana River floodplain (Brazil)

: richness, abundance and the relationship

with connectivity

. Hydrobiologia. 522(1-3) : 281-290 .

Araujo-Lima, C . A. R

. M., V. V. daSilva, P . Petry, E . C

. Oliveira, and S . M. L. Moura.

2001

. Diel variation of larval fish abundance in the Amazon and Rio Negro

.

Brazilian Journal of Biology 61 :357-362 .

Bardonnet, A . 2001

. Spawning in swift water currents : implications for eggs and larvae

.

Archiv fir Hydrobiologie (Supplementband) 135

:271-291 .

Brown, A . V ., and M. L

. Armstrong . 1985

. Propensity to drift downstream among

various species of fish

. Journal of Freshwater Ecology 3(1)

:3-17.

78

---

79

Brown, D . J., and T . G. Coon. 1994

. Abundance and assemblage structure of fish larvae

in the lower Missouri River and its tributaries

. Transactions of the American

Fisheries Society 123 :718-732

.

Bulkowski, L., and J. W

. Meade. 1983

. Changes in phototaxis during early development

of walleye

. Transactions of the American Fisheries Society 112

:445-447 .

Carter, J . G., V. A

. Lamarra, and R. J. Ryel . 1986

. Drift of larval fishes in the Upper

Colorado River

. Journal of Freshwater Ecology 3 :567-577 .

Crowder, L . B.,

D. D. Squires, and J . A. Rice. 1997

. Nonadditive effects of terrestrial

and aquatic predators on juvenile estuarine fish

. Ecology 78(6)

: 1796-1804 .

de Graaf, G. J., A. F

. Born, A . M. K. Uddin, and S. Huda

. 1999. Larval fish movement

in the River Lohajang, Tangail, Bangladesh

. Fisheries Management and Ecology

6:109-120

.

Dettmers, J . M., D. H

. Wahl, D . A. Soluk, and S. Gutreuter

. 2001 . Life in the fast lane

:

fish and foodweb structure in the main channel of large rivers

. Journal of the

North American Benthological Society 20

:255-265 .

Dewey, M. R., W. B. Richardson, and S

. J. Zigler . 1997

. Patterns of foraging and

distribution of bluegill sunfish in a Mississippi River backwater

: influence of

macrophytes and predation

. Ecology of Freshwater Fish 6

: 8-15 .

Dudley, R. K., and S

. P. Platania . 2000

. Downstream transport rates of passively drifting

particles and larval Colorado pikeminnow in the San Juan River in 1999

.

Division of Fishes, Museum of Southwestern Biology.

---

Flinn, M . B ., M. R. Whites, S . R. Adams, and J. E. Garvey

. 2005 . Macroinvertebrate

and zooplankton responses to emergent plant production in upper Mississippi

River floodplain wetlands . Archiv fir Hydrobiologia 162(2)

: 187-210 .

Eckblad, J . W., C. S

. Volden, and L . S. Weilgart . 1984

. Allochthonous drift from

backwaters to the main channel of the Mississippi River . American Midland

Naturalist 111(1) :16-22.

Gadomski, D. M., and C . A. Barfoot

. 1998. Diel and distributional abundance patterns

of fish embryos and larvae in the lower Columbia and Deschutes rivers

.

Environmental Biology of Fishes 51

:353-368 .

Galat, D. L., and J. Zweimuller

. 2001 . Conserving large-river fishes: is the highway

analogy an appropriate paradigm? Journal of the North American Benthological

Society 20 :266-279 .

Gale, W . F., and H . W. Mohr, Jr. 1978

. Larval fish drift in a large river with a

comparison of sampling methods

. Transactions of the American Fisheries

Society 107 :46-55 .

Grift, R . E., A. D. Buijse, W . L. T. Van Densen, M . A . M

. Machiels, J . Kranenbarg, J . G.

P

. Klein Breteler, and J. J. G. M. Backx. 2003

. Suitable habitats for 0-group fish

in rehabilitated floodplains along the lower River Rhine

. River Research and

Applications 19

: 353-374 .

Gore, J . A., and F . D

. Shields, Jr . 1995

. Can large rivers be restored? BioScience

45(3):142-152 .

80

---

81

Gosselain, V ., J. Descy, L

. Viroux, C. Joaquim-Justo, A

. Hammer, A . Metens, and S .

Schweitzer. 1998

. Grazing by large river zooplankton

: a key to summer

potamoplankton decline? The case of the Meuse and Moselle rivers in 1994 and

1995. Hydrobiologia 367/370

: 199-216 .

Haney, J . F ., and D . J. Hall

. 1973 . Sugar-coated Daphnia :

A preservation technique for

Cladocera

. Limnology and Oceanography 18 :331-333 .

Harvey, B . C. 1987

. Susceptibility of young-of-the-year fishes to downstream

displacement by flooding

. Transactions of the American Fisheries Society

116:851-855 .

Havera, S. P.,

and F . C. Bellrose . 1985

. The Illinois River: A lesson to be learned

.

Wetlands 4:29-41 .

Holland, L. E. 1986

. Distribution of early life history stages of fishes in selected pools

of the Upper Mississippi River

. Hydrobiologia 136 :121-130 .

Humphries, P

., A . J. King, and J . D. Koehn. 1999

. Fish, flows and floodplains

: links

between freshwater fishes and their environment in the Murray-Darling River

system, Australia

. Environmental Biology of Fishes 56 :129-151

.

Humphries, P ., and P . S. Lake

. 2000

. Fish larvae and the management of regulated

rivers. Regulated Rivers

: Research and Management 16

:421-432 .

Hurlbert, S

. H. 1984

. Pseudoreplication and the design of ecological field experiments

.

Ecological Monographs 54(2)

:187-211 .

Jack, J. D., and J. H. Thorp

. 2002

. Impacts of fish predation on an Ohio River

zooplankton community

. Journal of Plankton Research 24(2)

:119-127 .

---

82

Jackson, D . A. 1993 . Stopping rules in principal components analysis

: a comparison of

heuristical and statistical approaches

. Ecology 74: 2204-2214 .

Johnston, T

. A., M. N. Gaboury, R . A

. Janusz, and L . R. Janusz . 1995

. Larval fish drift

in the Valley River, Manitoba

: influence of abiotic and biotic factors, and

relationships with future year-class strengths

. Canadian Journal of Fisheries and

Aquatic Sciences 52 :2423-243 1 .

Junk, W . J., P. B . Bayley, and R . E. Sparks. 1989

. The flood pulse concept in river-

floodplain systems, p . 110-127

. In D. P. Dodge (ed.) Proceedings of the

International Large River Symposium

. Canadian Special Publication of Fisheries

and Aquatic Sciences 106 .

Jurajda, P. 1995

. Effect of channelization and regulation on fish recruitment in a

floodplain river. Regulated River

: Research and Management 10 :207-215 .

Kennedy, T. B., and G

. L . Vinyard . 1997

. Drift ecology of western catostomid larvae

with emphasis on Warner suckers,

Catostomus warnerensis

(Teleostei).

Environmental Biology of Fishes 49

:187-195 .

King, A. J . 2004

. Ontogenetic patterns of habitat use by fishes within the main channel

of an Australian floodplain river

. Journal of Fish Biology 65

:1582-1603 .

Koel, T. M., and R. E. Sparks

. 2002. Historical patterns of river stage and fish

communities as criteria for operations of dams on the Illinois River

. River

Research and Applications 18 :3-19

.

McCune, B., and M. J. Mefford . 1999

. Multivariate analysis of ecological data, version

4

. MjM Software Design, Gleneden Beach, Oregon, USA

.

---

83

Miller, T . J., L . B. Crowder, J. A. Rice, and E . A. Marshall . 1988

. Larval size and

recruitment mechanisms in fishes: Towards a conceptual framework . Canadian

Journal of Fisheries and Aquatic Sciences 45(9)

: 1657-1670 .

Miller, T. J., L. B . Crowder, and F. P. Binkowski . 1990. Effects of changes in the

zooplankton assemblage of growth of bloater and implications for recruitment

success

. Transactions of the American Fisheries Society 119(3) :483-491 .

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.

Molls, F. 1999. New insights into the migration and habitat use by bream and white

bream in the floodplain of the River Rhine

. Journal of Fish Biology 55 :1187-

1200.

Murty, D. S., K . K

. Sukmaran, P. V. G. K. Reddy, and R. K. Dey . 1986. Observations

on the life history of silver carp Hypophthalmichthys molitrix

(Valenciennes) .

Journal of the Inland Fisheries Society of India 18 :4-14 .

Muth, R. T., and J . C . Schmulbach. 1984

. Downstream transport of fish larvae in a

shallow prairie river. Transactions of the American Fisheries Society 113

:224-

230.

Nelson, J. S. 1994. Fishes of the world

. John Wiley and Sons, New York, N .Y.

Pezold, F . 1998

. Fish diversity in an isolated artificial wetland

. Journal of Freshwater

Ecology 13 :171-179 .

---

84

Reichard, M ., P. Jurajda, and R

. Vaclavik. 2001 . Drift of larval and juvenile fishes

: a

comparison between small and large lowland rivers

. Archiv fur Hydrobiologie

(Supplementband) 135 :373-389 .

SAS Institute Inc . 1999. SAS Version 8 .1

. Cary, North Carolina, 27513 .

Sheehan, R . J., W. M

. Lewis, L . R. Bodensteiner, D . King, D . E

. Logsdon, S . D. Scherck,

and P. S. Wills. 1990

. Winter habitat requirements and overwintering of riverine

fishes

. Southern Illinois University, Federal Aid in Sport Fish Restoration,

Project F-79-R, Project Completion Report, Carbondale

.

Scheidegger, K . J., and M

. B. Bain . 1995

. Larval fish distribution and microhabitat use

in free-flowing and regulated rivers

. Copeia :125-135 .

Schultz, D . W. 2006

. Quantifying fish movement between the Illinois River and an

associated backwater

. Master's thesis. Southern Illinois University, Carbondale,

Illinois .

Sheaffer, W . A. and J . G

. Nickum . 1986 . Backwater areas as nursery habitats for fishes

in Pool 13 of the Upper Mississippi River

. Hydrobologia 136 :131-140 .

Smith, K . A. and M. L . Brown. 2002

. Seasonal composition and abundance of

ichthyoplankton entrained through Big Bend Dam, South Dakota

. Journal of

Freshwater Ecology 17(2) :199-207 .

Soin, S . G., and A . I. Sukhanova

. 1972. Comparative morphological analysis of the

development of the grass carp, the black carp, the silver carp and the bighead

(Cyprinidae) . Journal of Ichthyology 12(1)

:61-71 .

Sokal, R

. R. and F. J. Rohlf. 1995. Biometry

: the principles and practice of statistics in

biological research. W. H

. Freeman and Co ., NY.

---

85

Theiling, C

. H

. 1995. Habitat rehabilitation on the Upper Mississippi River. Regulated

Rivers Research and Management . 11 : 227-238.

Tockner, K ., and J. A. Stanford . 2002

. Riverine flood plains : present state and future

trends. Environmental Conservation 29(3) : 308-330 .

Turner, T . F., J. C . Trexler, G . L. Miller, and K . E. Toyer. 1994

. Temporal and spatial

dynamics of larval and juvenile fish abundance in a temperate floodplain river .

Copeia (1) :174-183 .

Tweb, A ., A. Ahmed, P . Baroi, and G. Mustafa. 1990

. Studies on the development of

the silver carp,

Hypophthalmichthys molitrix (Val .). Bangladesh Journal of

Zoology 18(2):139-145

.

Young, G . C., and I. C. Potter. 2003

. Do the characteristics of the ichthyoplankton in an

artificial and a natural entrance channel of a large estuary differ? Estuarine,

Coastal, and Shelf Science 56 :765-779 .

U.S

. Army Corps of Engineers . 1993. Upper Mississippi River system

Environmental Management Program definite project report (SL-5) with

integrated environmental assessment

. Swan Lake rehabilitation and

enhancement : main report. Pool 26, Illinois River, USACE, St

. Louis

District, St . Louis, Missouri .

Wetzel, R . G. 2001 . Limnology

: lake and river ecosystems . Third edition . Academic

Press. San Diego, C .A.

---

VITA

Graduate School

Southern Illinois University

Laura A . Csoboth

Date of Birth

: November 6, 1979

215 Arizona Rd ., Carbondale, Illinois 62903

University of Delaware

Bachelor of Science, Natural Resource Management, May 2002

Special Honors and Awards :

Travel award to the Illinois Chapter of the American Fisheries Society Meeting

Travel award to the Annual Larval Fish Conference

Thesis Title :

Early Life History of Fishes in Restored and Unrestored Systems

Major Professor : Dr

. James E. Garvey

Publications :

Duke, J. M . and L. A. Csoboth. 2003

. Increased Scientific Capacity and Endangered

Species Management : Lessons from the Red Wolf Conflict

. Drake Journal of

Agricultural Law 8 :539-590 .

86

---

Timing of Riverine Export of Nitrate

and Phosphorus from Agricultural

Watersheds in Illinois : Implications

for Reducing Nutrient Loading to the

Mississippi River

MARKTODD BV

.

.

ROYER,'DAVID,* .tAND

LOWELL E . GENTRY*

School of Public

and Environmental Affairs, Indiana

University, 1315 E. Tenth

Street, Bloomington, Indiana 47405,

and Department of Natural Resources and Environmental

Sciences, University of Illinois

at Urbana-Champaign,

1102 S. Goodwin Avenue, Urbana, Illinois 61801

Agricultural watersheds in the upper Midwest are the

major source of nutrients to the Mississippi River and Gulf

of Mexico, but temporal patterns in nutrient export and

the role of hydrology in controlling export remain unclear .

Here we report on N03-N, dissolved reactive phosphorus

(DRP), and total P export from three watersheds in Illinois

during the past 8-12 yea rs

. Our program of intensive, long-

term monitoring allowed us to assess how nutrient export

was distributed across the range of discharge that

occurred ateach site andto examine mechanistic differences

between N03 -N and DRP export from the watersheds .

Last, we used simple simulations to evaluate how nutrient

load reductions might affect N03 -N and P export to

the Mississippi River from the Illinois watersheds . Artificial

drainage through under-field tiles was the primary

mechanism for N03 - - N export from the watersheds . Tile

drainage and overland flow contributed to DRP export,

whereas export of particulate P was almost exclusively from

overland flow . The analyses revealed that nearly all

nutrient export occurred when discharge was a median

discharge, and extreme discharges (>_ 90th percentile) were

responsible for >50% of the N03-

-N export and >80%

of the P export

. Additionally, the export occurred annually

during a period beginning in mid-January and continuing

through June

. These patterns characterized all sites, which

spanned a 4-fold range in watershed area . The simulations

showed that reducing in-stream nutrient loads by as

much as 50% during periods of low discharge would not

affect annual nutrient export from the watersheds .

Introduction

Nitrogen and P enrichment from nonpoint sources and

resulting eutrophication is a main cause of poorwater quality

and biotic impairment in many streams and rivers in the

United States (1) .

To address nutrient inputs from nonpoint

sources, states currently are developing nutrient criteria,

email'Correspondingauthorphone

: troyer@indiana.edu .

: (812)855-0563;fax(812)855-7802 ;

1 Indiana University .

t University of Illinois at Urbana-Champaign .

10 .1021/es052573n CCC : $X150

Published on Web 05/24/2006

a xxxx American Chemical Society

PAGE EST : 5.6

numeric standards, and total maximum daily loads (I'MDLs)

for nutrient-impaired streams and rivers. ATMDLrepresents

the maximum load (kg d-1 ) of a nutrient that a stream can

receive and still maintain water quality sufficient to meet its

designated uses

. Determination of the TMDL is based on the

assimilative capacity of the stream for that nutrient, natural

or background sources, point and nonpoint inputs, and a

specified margin of safety (2) . Because the assimilative

capacity, discharge, and magnitude of inputs can vary

throughout a year, the approach allows for seasonal variation

in the TMDL. In streams of the midwestern United States,

the effects of eutrophication are most pronounced during

periods of low discharge and warm water temperatures in

summer and autumn

. Nutrient TMDLs tend to focus on

critical periods of summer low discharge, while allowing

increased nutrient loads during times of high discharge . For

example, the Stillwater River drains an agricultural watershed

in western Ohio and the approved TMDL for N02 + NO3

increases from 3122 kg N d - ' in October and November to

load

>6700

is

kg

needed

Nd_Ito for

accommodate

December through

the increased

June(3)discharge

. The higherand

nonpoint

spring.

source runoff that occur from late winter through

In addition to degrading local water quality, nutrient

enrichment of midwestem streams has increased N and P

loading to the Mississippi River and Gulf of Mexico (4-6) .

Rivers draining agricultural regions of the upper Midwest

(i .e., the cornbelt) export large quantities of P and N

(predominantly as N03 -

-N)

as a result of extensive fertiliza-

hectares

tion and artificial

of cropland

drainage

in the

(5,

Mississippi

7,11)

. Approximately

River basin

20 million(MRB)

are artificially drained by under-field (tile) systems, particu-

larly in intensively farmed and fertilized areas such as Iowa,

Illinois, Indiana, and Ohio (5, 9) . Tile drainage provides a

bypass

mechanism

groundwater

by which water

flow paths

and dissolved

and move

nutrientsrapidly

fromcan

fertilized cropland to streams and rivers

(10, 11) . Because of

channel and hydrological modifications, streams in agri-

cultural watersheds are not efficient at nutrient removal by

processes such as denitrification, and a large fraction of the

nutrient load in such streams is transported to downstream

water bodies (12, 13) .

Nutrient export to the Mississippi River and eutrophi-

cation of midwestem streams both result from nutrient inputs

to surface waters but are associated with different hydrologi-

cal conditions . Eutrophication of streams is a primarily

biological process driven by nutrient uptake when local

conditions favor rapid growth of nuisance algae, such

as

during extended periods of low discharge . Conversely,

nutrient export from the Midwest to the Mississippi River is

a primarily hydrological process driven by precipitation and

drainage of the agricultural landscape (14) . To address water

quality in the Midwest, it is critical to gain a mechanistic

understanding of how N and P enter, and are exported from,

midwestem streams . There is a particular need for long-

term data to address (i) the temporal patterns in nutrient

loads and export, (ii) the role of hydrology in controlling

export, and (iii) the implications of these patterns for efforts,

such asTMDLs, to reduce eutrophication and nutrient export

from the agricultural Midwest .

A robust analysis of these issues requires long-term data

on NO3-N and P concentrations and river discharge, and

a sampling scheme that targets periods of high discharge

and rapidly changing nutrient loads . In this paper we use

long-term, intensive monitoring data from three agricultural

watersheds in Illinois to examine the above issues in relation

VOL . xx, NO . xx, xxxx / ENVIRON. SCI. & TECHNOL. • A

---

Analysis'TABLE

1 . Location, Watershed Characteristics . Period of Record, and Discharge Statistics for the River Sites Used in the

a Discharge statistics based on average daily discharge values from the period of record .

to nutrient management scenarios

. Usingsimple simulations

of

might

nutrient

affect

load

N03

-reductions,

- N and DRP

we

export

evaluate

to the

how

Mississippi

such reductionsRiver

from areas such as Illinois . Last, we discuss the implications

of our results for designing programs to reduce riverine export

of nutrients from agricultural regions of the upper Midwest .

Site Descriptions . We used long-term data on discharge

and nutrient concentrations from sites in the Embarras,

Kaskaskia,

Illinois

. The

and

sites

Sangamon

range in

river

drainage

systems

area and

inpeak

east-centraldischarge

but have similar land use dominated by row-crop agriculture,

mainly corn and soybean (Table 1) . East-central Illinois soils

are poorly drained Moflisols and the landscape was mostly

wetland

landscape

and

is

mesic

now tile-drained,

prairie prior

with

to

tile

settlementdensities

. Much

of 3-5

of thekm

km' (15),

and headwater streams have been extensively

channelized and dredged to accommodate high discharges

(16)

.

Nitrogen losses from the watersheds typically range

from 20 to 50 kg N ha- ' yr ', depending on precipitation,

and

streams,

The land

are among

use

inorganic

and

the

drainage

highest

N loads

modifications

in

are

the

90%

MRB

or

(5,

more

at

7,

the

NO17)sites

3-

.

-N

In

(7)arethe.

representative of tile-drained and intensively farmed areas

of the upper Midwest, and patterns in these streams likely

characterize much of the combelt region.

Methods

Stream discharge was monitored at the Embarras River and

Kaskaskia River sites by the U .S . Geological Survey (stations

Sangamon

03343400

Riversite

and 05590800,

was monitored

respectively)by

the Illinois

.

Discharge

State

at

Waterthe

determined

Water

Survey

Survey

(station

from

provided

106)hourly

. For

precipitation

or

each

15-min

site,

readingsmean

data

daily

.

(Champaign

The

discharge

Illinois Statesta-was

tion) and NO3 - - N concentrations for the Sangamon River

site for 1994-1999. All other nutrient concentrations were

determined from sampleswecollected approximatelyweekly,

either manually or with automated samplers

. Streams and

small rivers in the Midwest have flashy hydrology, and

discharge during floods represents a significant fraction of

annual discharge (18) . To account for this, we collected

additional samples when discharge was changing rapidly .

During most floods samples were collected daily, but in some

cases 2-4 samples were collected in a 24-hr period . In total,

our analysis is based on >4000 nutrient concentrations

determined from >2000 individual water samples collected

from October 1993 through September 2005 .

Nitrate was determined, after filtration through a 0 .45µm

membrane, on an ion chromatograph (Dionex, Inc . model

DX- 120 or model 2000i) . Dissolved reactive phosphorus (DRP)

was determined colorimetrically on filtered samples using a

spectrophotometer or (after 2001) a Lachat QuikChem8000

flow injection analyzer

. Total P was determined as described

for DRP except that samples were unfiltered and digested

with sulfuric acid and ammonium persulfate prior to analysis

.

B

• ENVIRON

. SCI, & TECHNOL . / VOL. xx . NO, xx, xxxx

Particulate P is defined as the difference between total P and

DRP . Internal and external standards for each nutrient were

analyzed routinely throughout the study as part of a quality

assurance plan . Daily in-stream nutrient loads were deter-

mined

nutrient

by

concentration

multiplying

(kg

mean

m-daily

') . Linear

discharge

interpolation

(m3din

-')SASby

(19) was used to estimate nutrient concentrations between

sampling dates . Annual and total nutrient export were

determined by summing the daily nutrient loads for each

kmwater

2 agricultural

year or the period

watershed

of record,

in Illinois,respectivelyweekly

. For

NO3-Na

1406

error

sampling

of

resulted

<4%

(20) .

in

We

a load

examined

estimate

smaller

with

watersheds

a root-mean-square(Table

1), but with a greater sampling frequency, and believe our

load estimates have similarly small error .

For the period of record at each site, we ranked the daily

discharge values and calculated the fraction of the

total

nutrient export attributable to each day, and the fraction

that occurred between various discharge percentiles . This

allowed us to assess how NO3-N, DRP, and total P export

were distributed across the range of discharge that occurred

at each site . The mass of nutrients exported each year from

the MRR is controlled to a large extent by precipitation (14) .

To examine patterns across years and account for inter-

annual differences in precipitation and export, we con-

smarted graphs of cumulative NO3 - - N and DRP export based

on percentage of the total for each water year .

To examine how reduced NO 3 - - N and DRP loads during

different hydrological conditions might affect export of these

nutrients to downstream water bodies, we performed three

simple simulations for each site . The first simulation reduced

in-stream N03-- N and DRP loads by 50% on all days with

discharge <50th percentile ; this simulation represented the

effect of focusing nutrient reductions on periods of low

discharge only . The second simulation reduced in-stream

loads

this represented

by 25% on all

a smaller

days with

reduction

discharge

in nutrient

<75th percentileloads

but

;

year)applied

. The

to a

final

larger

simulation

range of discharge

focused on

(ihigh

.e .,

more

flow

days

periods

of

andthe

reduced in-stream loads of NO3-N and DRP by 25% on all

days with discharge 2:75th percentile. For all simulations,

NO,- -N and DRP export during the period of record was

recalculated based on the adjusted loads and expressed as

a

simulations

percent of

were

the

selected

original exportto

evaluate

. The

the

reductions

interaction

used

betweenin

the

hydrology and nutrient export and do not necessarily reflect

management goals, although N load reductions of 20-30%

in the Mississippi River will be required to reduce hypoxia

in the Gulf of Mexico (21) .

Results

The volume of water and mass of N03- -N, DRP, and total

P exported from the watersheds varied considerably during

the study (Table 2), mainly as a consequence of variable

precipitation among years . In wet years, such as 1998 and

systemriver

site

coordinates

watershedarea

(km')

raw-crop

(%

agricultureland

cover)

period of

(water

recordyears)

maximum

discharge(m's

')

median

discharge(m's')

discharge(m'3mean-1)

EmbarrasSangamonKaskaskia

88'19'35"W39°50'09"N,40°16'06"N,39°47'29"N,88°11'08"W88°29'18"W

386101481

918691

1998-20051994-20031994-2005

10819871

011.0.6.3

035.0.6.8

---

Watersheds

TABLE 2

. Annual

during

Water

the

and

StudyNutrient

Export from the Three

the

2002

N

2003

yields

sites

(82

(117

cm

there

ranged

and

of

108

precipitation)

was

from

cm

a clear

of

45

precipitation,

to

pattern

55

yields

kg

of

hanutrient

were

1 .

respectively),

In

<

dry

15

export

years,

kgha'occurring

.

such

N0Across3-as-

predominantly

1)

. During the

at

period

the high

of record,

end of the

days

discharge

with discharge?

range (Figuremedian

discharge accounted for 97-98% of the NO3-N export and

98-99% of the DRP export (Table 3)

. Extremes in discharge

total

(>90th

NO,percentile)

-

-N export

accounted

and 84%

for

of the

an average

DRP exportof

56%

.

of the

The temporal distribution of N0

3 - - N export indicated

that the majority of the annual export occurred during a 5 .5

month

and years

period

(Figure

from

2)mid-January

. Within a water

through

year,

June

the

across

first 3 .5

all

monthssites

(October-mid-January) and the last 3 months Uuly-

September) together typically accounted for <30% of the

annual N03-N export. Within the January-June period,

N03- -N export often occurred in discrete events, as evi-

denced by the abrupt increases in the slopes of the lines in

Figure 2

. The watersheds spanned a 4-fold range in size, but

consistent

the pattern

across

in temporal

sites

.

distribution

For P, we focus

of N03on

DRP

-

-N

rather

export

thanwas

total P because we have longer records for DRP and it

represents the immediately available P . As with NO3-N,

DRP export occurred mainly from January to June in most

years, although substantial export occasionalyoccurred later

with

in the

individual

summer (Figure

floods

3)and

. Export

in several

of DRP

years

was often

40-80%

associatedof

the

annual DRP export occurred during a period of < 1 month

.

Mechanistically, NO3-N, DRP, and total P responded

differently to the occurrence of overland flowandwe illustrate

these differences with the 2002 and 2003 water years (wet

K

80

100

1

100

EmWnastie

--- Kaskasklasite

- --- sangamon sae

H

i

/

100

80

60

40

i

20

010ti410

.

I

101'

10°

10'

102

10'

Discharge (m' s')

FIGURE 1 . Cumulative nutrient export as a function of discharge

during the period of record for each site .

TABLEduring

Occurred

3

the

. Percentage

at

Periods

or above

of

of

Various

Recordthe

Total

DischargeNutrient(Q)

Export

Percentilesthat

% of export

river

a >- 98th

a >- 75th

system

a >-5M

percentile

percentile

percentile

and dry years, respectively) at the Embarras River site (Figure

4) . In 2002 a series of precipitation events from February

through mid-April initiated flow through agricultural tile

drains, increased discharge, and resulted in a steady export

of NO3 -N, with that period accounting for approximately

55% of the total 2002 N0

3 - - N export . During that same

of

period,

2002

DRP

annual

and total

exportP

.

exports

From April

were 33

8 through

and 17%,

May

respectively,17,

2002,

total precipitation was 21 cm and much of that water entered

the Embarras River as overland flow and caused sustained

flooding

. There also were several periods of high discharge

through tile drains throughout the Embarras River watershed .

During these floods and high tile discharge periods, DRIP

and total P export increased substantially from the previous

months, whereas NO, - -N export continued at approximately

the

for

same

N03-rate

- N export

(Figure

from

4) .

the

Overland

watershed,

flow was

but

not

during

important2002

overland flow was important for export of DRP and total P

(with particulate P accounting for a large fraction of total P)

.

We cannot assign precise values to the fraction of the

DRP and total P export originating from tile drainage versus

V OL. xx. N O. xx, xxxx

/ ENVIRON. SCI. & TECHNOL a C

water

year

discharge

410° m')

Mg

kg ha - '

N03-N DRP total P ND3-N DRP total P

200519992004200220011998200020031995199619971994

228290186180137132235125170539252

139615062728137222191599199911401081962490426

Embarras2421352429314321121154

10216515197 56312029284623104133228.5.0.7.3.9.5.1.0.5.2.7.3

000000000000.1.2.5.4.2.1.7.6.6.5.4.9

020101.1.1.2.3.1.2

1998 179

2223

Kaskaskia

20001999

9646

1389514

2138

3657.0.6

00.2.5

2001

100

1164

10

6

13 .3

0 .1 0.2

2002

187

2129 13

14 30 .2

0 .3 0.4

2003

31

292 2

38 55.2

0.3 1 .0

4

7 .6

2004 130

1382 11

<

0.1 0.1

2005 150

1288 22

45 35 .8

0.3

1 .2

40

33 .4

0.6 1 .1

1994

38

262

Sangamon

2001200019991998199519971996

3629313519198

21631833736932819391

4

2631362132339.6.5.0.3.4.5.0

2002

34

473

3

7

19.1

0.4 0.7

2003

16

146

1

8 46.8

0.3

0.8

3

14 .5

0 .1 0 .3

SangamonEmbarrasKaskaskiamean

57545856

N03-N

848281

989797

DRP

82

97

SangamonEmbarrasKaskaskiamean

86848580

94969495

99989898

---

0

Oct Nov Dec Jan FebMar Apr May Jun Jut Aug Sep

Water Year

FIGURE2the

period

.

of

Cumulative

record for

NO,--N

each

export

site

.

during each wateryear during

during

overland

peak

flow,

tile

because

flows

DRP

but

export

then

through

decreases

tiles

rapidly

is greatestand

This

drainageP

and

floods

two

a

disproportionately

drainage

contrasts

discharge

flow

and

dry

and

flow

total

total

pattern

year

event

total

and

.

approximately

events,

and

with

P

P

not

with

P

both

inputs

in

is

tile

export

associated

small

evident

May,

NO3--N

but

increased

flow

as

to

N

these

increased

but

declined

and

in

tile

the

in

export

proportion

with

then

the

P

river

were

along

discharge

export

dry

surface

(Figure

discretely

diverged

which

much

appeared

with

2003

(Table

to

smaller

4)declines

continues

NO3--N

runoffwater

tile

.

2),

from

Dissolved

during

to

discharge

.

year,

nearly

than

in

be

Therefore,

(11)N03--N

the

from

the

via

the

when

reactiveall

.

firstothertile(10)tile2002ThisasDRPDRPin

.

N

<median

and

The

DRP

simulations

discharge)

loads during

showed

would

periods

that

reduce

a

of

509

total

low

reduction

export

discharge

of

in

theseNO3--(i.e.,

an

nutrients

in

discharge

(Table

reduction

nonetheless

Conversely,

of DRP

average

discharge

('Fable

4)

.

in

(>_75th

by

decline

disproportional,

total

reducing

(<75th

<2%

4) .

in

export

percentile)

Reducing

of

percentile)

the

13loads

.0%

of

case

loads

N03

reduction

for

25%

of

gave

-N

N03--N

resulted

25%

during

N03

(20a

across

nearly

in

.7%)andDRP(23-N

and

total

and

periods

in

4a

a

proportional<

.2%

wider

export,

larger,

1%

of

for

in

rangehighcase.6%)withDRPbut

.

Discussion

In midwestern agricultural watersheds, in-stream nutrient

concentrations

time

because

in the

(22,

streams

of

23)the . This

we

increase

are

examined

pattern

greatest

in

(7,

runoff

has

during

12,

been

that

24)

winter

previously

and

occurs

we

through

believe

during

documentedspringthatthat

D • ENVIRON

. SO

. & TECNNOL

. / VOL. xx, NO . xx, xxxx

m 100

m 80

60

r 40

20

0- 0

100

80

60

40

20

0

100

80

60

40

20

Embarras , -

(

I

(r

(

( I(

I I

(r-l

~ rTi

l

11

..1

Y1==~

Kaskaskia

Sangamon

I

rJ I

r

rI

I

r

1

_J

I

Oct Nov Dec Jan FebMar Apr May Jun Jul Aug Sep

Water Year

FIGURE

the period

3 .

of

Cumulative

record for

DRP

each

export

site .

during each water year during

Oct Dec Feb Apr Jun Aug Od Dec Feb Apr Jun Aug Od

2002 -

2003 --

Water Year

discharge

2003

FIGURE4water

.

(lower)

Cumulative

years.

from

nutrient

the Embarras

export

River

(upper)

site

and

during

precipitation

the 2002 andand

watersheds

ourstudysites,

>-

First,

important

from

Illinois,

median

the

nearlyall

agricultural

are

characteristics

discharge,

of

representative

although

the

nutrient

upper

Midwest

and

limited

Midwestexport

regarding

extreme

of

geographically

intensively

during

occurred

. Our

discharges

the

the

analyses

export

when

drained

past

to

discharge

of

revealed

east-central8-12

(Q

and

nutrients>_

yearsfarmed90thwastwo

.

N

percentile)

and

mid-January

exchange

and

As

>80%

discharge

P from

of

of

the

were

nutrients

through

the

increases,

P

responsible

exportwatersheds

Junebetween

. Second,

there

.

for

occurred

is

the

the

>

less

50%

water

annual

of

consistently

opportunity

the

column

export

NO3-N

of

and

for

exportNO3--fromthethe

---