   

3,33% WIWINUHWWW“WWW”WWII

io>.'»_"."4?"~

(2/5?th 2

This is to certify that the
thesis entitled

QUANTIFYING EFFECTS OF RESIDENTIAL LAKESHORE
DEVELOPMENT ON LITTORAL FISHES AND HABITAT:
TOWARD A FRAMEWORK FOR LAKE ECOSYSTEM
CONSERVATION

presented by

Aaron Kenneth Jubar

has been accepted towards fulﬁllment
of the requirements for the

 

 

 

 

MS. degree in Fisheries and Wildlife
”ff/WW 7 EMA/(MC; W
ajor Professor’s Signature
/ 2 1/57“ {/0921
Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

.UBRARY
Michigan State
University

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue,p65-p.15

QUANTIFYING EFFECTS OF RESIDENTIAL LAKESHORE DEVELOPMENT
ON LITTORAL FISHES AND HABITAT:
TOWARD A FRAMEWORK FOR LAKE ECOSYSTEM CONSERVATION
By

Aaron Kenneth J ubar

A THESIS
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Fisheries and Wildlife

2004

ABSTRACT
QUANTIFYING EFFECTS OF RESIDENTIAL LAKESHORE DEVELOPMENT
ON LITTORAL FISHES AND HABITAT:
TOWARD A FRAMEWORK FOR LAKE ECOSYSTEM CONSERVATION
By

Aaron Kenneth Jubar

Extensive alterations to north temperate lakes due to residential lakeshore
deve10pment (LSD) and associated activities have the potential to negatively affect
habitat features in the littoral zones of lakes. To quantify the effects of residential LSD, I
surveyed littoral habitat features of Six Michigan lakes that varied primarily in their
degree of LSD measured in dwellings per kilometer. Undeveloped Sites had signiﬁcantly
greater abundance of coarse woody material and submersed macrophyte cover compared
to developed sites. Substrate particle size was Signiﬁcantly larger at retaining wall sites
compared to undeveloped and maintained sites. In order to assess the indirect effects of
LSD on lakewide response variables, I examined whole-lake macrophyte cover, water
chemistry, and growth rates of bluegill (Lepomis macrochirus) and largemouth bass
(Micropterus salmoides) in 15 Michigan lakes representing a gradient of LSD. Whole-
lake littoral ﬂoating macrophyte cover decreased with increasing amount of LSD, but
other macrophyte growth forms and water chemistry showed no signiﬁcant response to
LSD. Bluegill growth increased while largemouth bass growth Showed a marginally
Significant decrease with increasing LSD. The effects of LSD on fish growth also
depended upon growth year (for bluegill), and ﬁsh Size (for both Species). Collectively,
these ﬁndings demonstrate the importance of investigating both local and lakewide (i.e.,

cumulative) effects of LSD on lake ecosystems.

ACKNOWLEDGMENTS

I owe a great deal of gratitude to my major advisor, Dr. Mary Bremigan, for all
her insight, support, patience and friendship during the course of this research. Dr.
Bremigan was always available and willing to assist with any aspect of my research. I am
truly grateful for the opportunity to pursue graduate studies under her tutelage. I would
also like to thank my graduate committee, Dr. Patricia Soranno and Dr. Graham Hickling,
for providing thoughtful comments and constructive criticism on this work. Their
feedback has proved to be an integral part of this thesis. Thanks to Dr. Alan Tessier for
his contribution during the early stages of my research. I would also like to thank Dr.
James Bence, Dr. Daniel Hayes, and Tyler Wagner for their statistical advice and
support. I owe a great deal of thanks to the numerous graduate students, undergraduate
students, and others that provided field and laboratory assistance, especially R. Brim, C.
Carman, K. Cheruvelil, B. Goodwin, P. Hrodey, K. Jubar, N. Longbucco, S. Martin, C.
Meier, D. Myers, J. Roasa, A. Shultz, T. Wagner, and S. Wills. Special thanks to Dr.
James Dragun for his guidance and advice, and for encouraging me to pursue my
graduate degree. Heartfelt thanks to Jill, my loving wife, for all of her support over the
past couple years. Funding for this research was provided by the Michigan Department of

Natural Resources.

iii

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. v
LIST OF FIGURES .......................................................................................................... Vii
INTRODUCTION .............................................................................................................. 1
CHAPTER 1: LOCAL EFFECTS OF RESIDENTIAL LAKESHORE DEVELOPMENT
ON LITTORAL HABITAT ................................................................................................ 3
Introduction ............................................................................................................. 3
Methods ................................................................................................................... 7
Results ................................................................................................................... 10
Discussion ............................................................................................................. 15

CHAPTER 2: LAKEWIDE EFFECTS OF RESIDENTIAL LAKESHORE
DEVELOPMENT ON WATER CHEMISTRY, MACROPHYTE COVER, AND

LITTORAL FISH GROWTH ........................................................................................... 20
Introduction ........................................................................................................... 20
Methods ................................................................................................................. 24
Results ................................................................................................................... 29
Discussion ............................................................................................................. 32
Conclusion ............................................................................................................ 36

APPENDIX ....................................................................................................................... 38

LITERATURE CITED ..................................................................................................... 58

LIST OF TABLES

Table 1. Physical characteristics, lakeshore development (LSD), water chemistry and
land cover of the six study lakes .................................................................. 39

Table 2. Percentages of expected and observed shoreline for each modiﬁcation type
(retaining wall, maintained, and undeveloped) in each of the six study lakes. Percent (%)
expected indicates the amount of shoreline in a lake for each combination of natural
features (hi gh/low wind, and high/low slope). Percent (%) observed indicates the amount
of shoreline within a particular combination of natural features that is made up of each
modiﬁcation type. Even distribution of modiﬁcation types relative to available natural
features would be indicated by observed percentages equal to expected percentages. P-
values are derived from a Chi-square analysis. Bold indicates the higher value for
comparisons that were signiﬁcantly different. Lakes are ordered according to increasing
lakeshore development .............................................................................. 40

Table 3. Within—lake variable P-values and associated degrees of freedom (in
parentheses) derived from ANCOVA for sampling transects conducted along 0.5 m and l
m depth contours. Main effects in the model include shoreline modiﬁcation type
(undeveloped, maintained, and retaining wall) and lakeshore development (LSD;
#dwellings/km). Modiﬁcation*LSD refers to the interaction between modiﬁcation type
and lakeshore development. Signiﬁcant P-values are in bold ................................. 41

Table 4. Physical characteristics, lakeshore development (LSD), water chemistry and
land cover of the 15 study lakes ................................................................... 42

Table 5. Macrophyte assemblage characteristics of the 15 study lakes in southeast
Michigan. Cover refers to the percentage of sampled points at which plants were present.
Total plant cover indicates the percent occurrence of plants sampled over the entire lake,
whereas littoral plant cover is the percentage of sampled points at which plants were
present in the littoral zone. Littoral zone is deﬁned as the area from shore to the depth at
which plants consistently occurred. Dense cover indicates the percent occurrence of sites
with macrophyte cover scores of 2 or 3 for any macrophyte category. Plant data were
collected during August and September 2003 ................................................... 43

Table 6. Back-calculated lengths (as determined by the body-proportional hypothesis) for
bluegill from all study lakes by year-of-capture (2003-2004) and year class. Lee's
phenomenon would be indicated when the estimated size at a particular age declines
within a year class with increasing year of capture or within a year of capture with
decreasing year class. Because Lee's phenomenon appears evident, I only used back-
calculated lengths for bluegill from 2001 and 2002 year classes ............................. 44

Table 7. Back-calculated lengths (as determined by body-proportional hypothesis
method) for largemouth bass from all study lakes by year-of-capture (2003-2004) and
year class. Lee's phenomenon would be indicated when the estimated size at a particular
age declines within a year class with increasing year of capture or within a year of

capture with decreasing year class. Because Lee’s phenomenon appears evident, I only
used back-calculated lengths of largemouth bass from 1998 to 2001 year classes. . ....45

Vi

LIST OF FIGURES

Figure 1. Map illustrating the location of the six study lakes. The six study lakes are
located within the Huron River Watershed, southeast Michigan ............................. 46

Figure 2. Percentages of the four combinations of natural features along the shorelines of
the six study lakes. Lakes are presented from lowest LSD (7.76) to highest LSD (22.31).
Note: wind exposure is not evenly distributed due to differences in lake

morphometry ...................................................................................................................... 47

Figure 3. Percentages of the three different modiﬁcation types along the shorelines of the
six study lakes. Lakes are presented from lowest LSD (7.76) to highest LSD (22.31). . ..48

Figure 4. Relationship between lakeshore development (#dwellingS/km) and coarse
woody material abundance (mean intersections per meter, a.), and submersed macrophyte
cover (cover score, b.), i1 SE, along the 0.5 m depth contour for the three shoreline
modiﬁcation types: undeveloped, maintained, and retaining wall. Note difference in y-
axes between coarse woody material abundance and submersed cover ..................... 49

Figure 5. Relationship between lakeshore development (#dwellingS/km) and ﬂoating
macrophyte cover (cover score, a.), and emergent macrophyte cover (cover score, b.),
:tISE, along the 0.5 m depth contour for the three shoreline modiﬁcation types:
undeveloped, maintained, and retaining wall. Note difference in y—axes between ﬂoating
macrophyte and emergent macrophyte cover ................................................... 50

Figure 6. Relationship between lakeshore development (#dwellings/km) and mean
substrate particle size, iISE, along the 0.5 m depth contour for the three shoreline
modiﬁcation types: undeveloped, maintained, and retaining wall ........................... 51

Figure 7. Map illustrating the location of the ﬁfteen study lakes. The ﬁfteen study lakes
are located within the Huron River Watershed, southeast Michigan ......................... 52

Figure 8. Whole-lake macrophyte response to LSD for littoral ﬂoating cover (A), whole-
lake macrophyte cover (B), littoral emergent cover (C), and littoral submersed cover (D).
Data presented are actual values. Statistical values are based upon transformed data
(when appropriate) ................................................................................... 53

Figure 9. Annual growth increment (mm) relative to ﬁsh size (back-calculated total
length, mm) of bluegill collected in 2003 and 2004 in the 15 study lakes. Data points on
far left indicate size at hatching (3mm, Carlander 1977) ....................................... 54

Figure 10. Annual growth increment (mm) relative to ﬁsh size (back-calculated total

length, mm) of largemouth bass collected in 2003 and 2004 in the 15 study lakes. Data
points on far left indicate size at hatching (3mm, Carlander 1977) ........................... 55

Vii

Figure 11. Annual growth increment for bluegill (growth years 2001 and 2002) as a
function of back-calculated total length. Low development lakes (<13 dwellings-km")
are represented by the solid line, medium development lakes (13-20 dwellings0km") are
represented by the wide-dashed line, and high development lakes (>20 dwellingS°km'l)
are represented by the short-dashed line ......................................................... 56

Figure 12. Annual growth increment for largemouth bass (growth years 1998 and 2001)
as a function of back-calculated total length. Low development lakes (<13 dwellings-km")
are represented by the solid line, medium development lakes (13-20 dwellings-km") are
represented by the wide-dashed line, and high development lakes (>20 dwellings-km’l)
are represented by the Short-dashed line ......................................................... 57

viii

INTRODUCTION

As the demand for waterfront property increases, lakeshore owners are
extensively developing riparian areas adjacent to north temperate lakes. Human
alterations do not cease at the land-water interface, but often extend into the nearshore
littoral areas of the lake. There is a growing body of literature examining the effects of
residential lakeshore development (LSD) on nearshore habitat (Christensen et al. 1996,
Radomski and Goeman 2001, Jennings et al. 1996, Jennings et al. 2003, Hatzenbeler et
al. 2004) and ﬁsh (Jennings et al. 1999, Schindler et al. 2000, Scheuerell and Schindler
2004). Many of these studies, however, consider lakes from northern regions of
Minnesota and Wisconsin, USA, which may have different limnology and food web
ecology compared to lakes from other north temperate regions, and thus differ in their
response to LSD. To date, no research has examined the effects of LSD on aquatic
ecoSystems in southern Michigan. Riparian lake owners in southern Michigan have
caused extensive development in recent decades, as many seasonal cottages have been
built, and subsequently converted to or replaced with year-round residences.
Additionally, rapid suburban development in this area over the past few decades has
forced many residences on these lakes to convert from septic to sewer waste disposal
systems.

My research examined lakes located in one major river watershed (Huron River)
in southeast Michigan. This study area provided me with a unique opportunity to examine
habitat and ﬁsh responses to LSD for lakes ranging from low development (located in
some of the many recreation areas) to extremely high development (due to the relatively

close proximity of these lakes to the urban areas of Ann Arbor and Detroit, Michigan).

AS the human population increases in southeast Michigan, boaters and anglers have
increasingly used lakes in this area for recreational purposes. Lakeshore development and
associated activities have the potential to negatively affect lake littoral habitat, which
may, in turn, affect ﬁsh populations with close linkages to the littoral environment of
lakes.

I studied multiple attributes of lake ecosystems to identify the response of various
lake ecosystem components to LSD at local and lakewide scales. This research will
contribute to our understanding of how lake ecosystems respond to changes caused by
human perturbation, particularly in southeast Michigan. The knowledge gained through
this research will aid managers by identifying components of the lake ecosystem most
inﬂuenced by LSD. My hope is that this study will stimulate further research, as well as
provide a starting point for aquatic ecosystem managers seeking to quantify and
potentially mitigate the effects of loss or alteration of critical lake habitat through

proactive management strategies in north temperate lakes.

CHAPTER 1: LOCAL EFFECTS OF RESIDENTIAL LAKESHORE DEVELOPMENT
ON LITTORAL HABITAT

Introduction

The loss of natural habitat in lakes has become an area of concern for ﬁsheries
and wildlife managers. For north temperate lakes, the ever-increasing residential
development of lake riparian zones is a major factor driving habitat loss. Fish and plant
species in the littoral zone of north temperate lakes can be adversely affected by
lakeshore modiﬁcation and development (Jennings et al. 1999; Schindler et al. 2000;
Hatzenbeler er al. 2004; Scheuerell and Schindler 2004), in part because residential
development of lakes often does not end where the land meets the water. Rather,
developers and homeowners typically modify their shoreline and the related littoral zone
of the lake for recreational and aesthetic purposes. Although small amounts of lakeshore
development (LSD) may beneﬁt aquatic organisms by diversifying littoral habitat,
extensive and uniform development along shorelines may be detrimental to aquatic biota
(Jennings et al. 1999). Because individual lake-dwellers modify their nearshore areas to
varying degrees, and such modiﬁcations contribute to and are confounded with
cumulative LSD, it is still uncertain what Spatial scale (i.e., lakewide development versus
development type for individual stretches of shoreline) is most important in assessing
habitat loss in lakes.

Littoral ﬁsh Species may be particularly vulnerable to effects of habitat loss given
their use of near-shore habitat for nesting, foraging, and refuge. A reduction of complex
littoral habitat may affect ﬁsh assemblages in complex ways because speciﬁc habitat

requirements or preferences vary among littoral ﬁsh Species, and within ﬁsh species,

habitat requirements vary with life stage. For example, juvenile, non-nesting adult, and
nesting adult largemouth bass (Micropterus salmoides) exhibit signiﬁcantly different
patterns of habitat use (Annett et al. 1996). Bluegill (Lepomis macrochirus) habitat use is
also dependent upon size (Mittelbach 1984). However, the degree to which LSD
compromises littoral habitat, and associated ﬁsh assemblages and life stages, is poorly
understood.

Lakeshore development may affect available habitat structures in the littoral zone
including coarse woody material (CWM) abundance, macrophyte cover, and substrate
particle size. Coarse woody material is an important physical structure contributing to
habitat diversity in both riverine and lake ecosystems. In lotic systems, CWM provides
habitat for invertebrates (Phillips and Kilambi 1994) and numerous ﬁsh species (Cunjak
and Power 1987; Neumann and Wildman 2002). The loss or removal of CWM from
rivers can be detrimental to aquatic organisms by reducing suitable in-stream habitat
(Gumell et al. 1995). Impacts of CWM in lakes, however, are poorly understood.
Coarse woody material located in the littoral zone of lakes may provide important refuges
for prey ﬁsh from predators (Savino and Stein 1989a) and may also promote substrate
habitat for macroinvertebrates which serve as prey for some ﬁsh species (Bowen et al.
1995). For example, age-0 ﬁsh may beneﬁt from the interstitial spaces and invertebrate
prey found among or associated with CWM. Further, the ﬁndings of Schindler et a1.
(2000) suggest a positive relationship between growth rates of bluegill and largemouth
bass and density of CWM. Christensen et al. (1996) determined that LSD in northern
Wisconsin and Michigan lakes had strong negative effects on abundance of CWM not

only at the whole-lake scale, but also at the local scale. Speciﬁcally, forested (or

undeveloped) sites had greater amounts of CWM than developed sites, and lakes with
more highly developed shorelines had lower CWM (Christensen et al. 1996).

The impacts of LSD on macrophyte communities may be more complex than
those on CWM. Residential development of lakes may indirectly increase aquatic plant
abundance through increased inputs of nutrients important for plant growth. Conversely,
LSD and associated activities may directly reduce macrophyte abundance through
chemical treatments, mechanical removals, or increased wave action due to boating
activity. Speciﬁc types of macrophytes may be particularly vulnerable to LSD. For
example, increased LSD has been found to reduce abundance of emergent and ﬂoating-
leaf vegetation (Radomski and Goeman 2001; Hatzenbeler et al. 2004). Further,
increased recreational use of lakes resulting from LSD has been associated with declines
in emergent plants (Ostendorp et al. 1995). Lakeshore development can potentially affect
macrophytes on both individual site and whole-lake levels.

Substrate composition determines the quality of spawning habitat and cover for
many ﬁsh species and inﬂuences benthic macroinvertebrate and periphyton composition
and production (McMahon et al. 1996). In lakes, some littoral ﬁsh species, such as those
of the family Centrarchidae, exhibit strong preferences for speciﬁc substrata for nest
construction (Balon 1975). For example, Hunt et al. (2002) found spawning male
largemouth bass frequently construct nests in sites dominated by sand or gravel substrata.
Similarly, smallmouth bass (M. dolomieu) nest success and nest density in four Au Sable
River reservoirs, Michigan, were highest in areas with gravel substrate (Wills et al.
2004). Residential development of lake Shorelines may alter the substrate composition in

areas critical for bass nest construction. Suitable Spawning habitat may be reduced by

shoreline modiﬁcations which remove rocks and gravel (e.g., for aesthetic purposes), add
ﬁne sand or Silt (e. g., through building of beaches), or disrupt natural sedimentation (e. g.,
through construction of retaining walls). For example, Jennings et al. (2003) found
littoral sediments at developed sites contained more ﬁne particles compared to
undeveloped sites within the same lake.

The effects of human development on lakes have typically been examined at local
Spatial scales with a focus on LSD effects on particular habitat characteristics, with the
exception of a few studies examining lakewide habitat response to LSD. Radomski and
Goeman (2001) used aerial photography to quantify the effects of LSD on ﬂoating-leaf
and emergent vegetation. Christensen et a1. (1996) studied coarse woody debris in
relation to amount of LSD and type of development. However, studies generally have
failed to investigate the effects of speciﬁc lakeshore modiﬁcations on substrate
composition of the littoral zone within a lake (but see Jennings et a1. 1996); and few
studies have examined the direct effects of riparian development and lakeshore
modiﬁcations on littoral habitat or identiﬁed the indirect effects of such LSD on littoral
ﬁshes. In this study, I quantiﬁed littoral habitat features along lake shorelines composed
of different modiﬁcation types in southeast Michigan lakes that are limnologically
similar, but differ according to overall level of LSD. Human development patterns and
habitat characteristics may be inﬂuenced by natural features as well. Therefore, my main
objectives were to answer two questions: Is development randomly distributed in relation
to natural features? And, how does development at the local and lakewide scale affect
lake littoral habitat? Based on previous research examining the effects of LSD on littoral

habitat, I generated a number of expectations. First, CWM in the littoral zone of southeast

Michigan lakes will be lower along developed shorelines and decrease with increasing
LSD, similar to lakes in northern Wisconsin (although overall abundance of CWM may
be lower, due to differences in land cover and vegetation between regions). Second,
macrophyte cover will be lower along developed shorelines and decrease as overall LSD
increases. Third, substrate particle size will be smaller in littoral areas adjacent to

residential development and decrease in Size as overall LSD increases.

Methods

Study lakes

I selected 15 lakes within a single ecoregion and within a single major river
watershed, in southeastern Michigan, USA (Figure 1), in order to control as much as
possible for differences in climate, geology, lake morphology and aquatic ﬂora and
fauna. The Huron River watershed contains lakes with a wide range of LSD, thus
providing a mix of highly developed, moderately developed, and undeveloped lakes.
I used visual observations by boat to quantify the number of riparian dwellings within 50
m of each lake, and then divided number of dwellings by lake perimeter (km) to calculate
LSD (dwellings-km"). Lakes were chosen to have similar composition of land use types
in their watershed (< 70% of either agricultural, forested, or urban land use within a 500
m buffer). In doing so, I aimed to select lakes that were biologically and physically
similar and that differed primarily in the extent of their LSD (Table 1). To standardize
lake size, and assure that all lakes were of sufﬁcient depth for summer stratiﬁcation, I

selected a subset of six study lakes that ranged 29 to 100 hectares and had a mean depth

>3m. The six lakes represent a gradient of LSD based on dwellings-km", ranging from
7.76 to 22.31 dwellings-km“.
Site Selection

I used a differential GPS unit (Trimble GeoExplorer®) to record length of
different features. At each site, I recorded natural feature attributes, and modiﬁcation type
of Shoreline segments along the perimeter of the six study lakes. I classiﬁed three types of
shoreline modiﬁcations: undeveloped (including unaltered or natural shoreline),
developed maintained (including beach, lawn, or groomed shoreline), and developed
retaining wall (including seawall and rip-rap). I recorded slope of the riparian area within
30 m of shore as low grade (<30°), or high grade (>30°), and determined wind exposure
based on shoreline exposure to the region's prevailing southwesterly wind (shoreline
facing SE, E, NE, N, and NW = high wind, all other shoreline = low wind).

I then entered these lakeshore attribute data into a Geographic Information
System (GIS). Although lakeshore frontage depends on local zoning ordinances, typical
residential lot size for the study lakes is roughly 40 m of shoreline. Therefore, using GIS,
I separated continuous stretches of each shoreline modiﬁcation type into 40m increments.
Due to relatively ﬂat topography in the study area, I only considered sites with low
terrestrial slope in sample site selection because high slope sites were so rare. I randomly
selected study sites within each lake, with three replicates per modiﬁcation*wind
exposure combination for a total of 18 sites per lake. This design allowed me to examine

littoral habitat variation both within and among lakes.

Field methods

I sampled study lakes and subsequent sites within each lake in random order. In
order to determine the effects of Shoreline modiﬁcation type at different wind exposures
and across a gradient of LSD, I quantiﬁed natural habitat features at each Site. I
quantiﬁed CWM using transect-intercept methods similar to Christensen et al. (1996). I
conducted sampling transects along the 0.5 m depth contour once at each sampling Site
during May and June 2003. I used a caliper to identify CWM >5 cm in diameter, and
counted the number of intersections per meter of transect to provide a relative measure of
CWM abundance.

I visually measured substrate composition at each sampling site using a modiﬁed
Wentworth scale (adapted ﬁ‘om Cummins 1962) and then classiﬁed substrate as one of
four groups based on particle size: silt (<0.0625mm), sand (0.0625-2mm), gravel (2-
32mm), or cobble (>32mm). During May and June 2003, I recorded substrate using a 1 m
quadrat placed every 5 meters along the 0.5 m and l m depth contours at each site, and
calculated mean substrate size along each contour at each site.

Macrophyte composition and cover at each Site was recorded during August 2003,
by an observer from a boat along the 0.5 m and 1 m depth contours. I surveyed emergent,
ﬂoating, and submersed vegetation every 5 m along each contour (0.5 m and 1 m) and
assigned a qualitative value for cover (0, 1, 2, or 3; where 0 = <5% cover, 1 = 5-33%, 2 =
34-66% cover, and 3 = 67-100% cover). For each depth contour for each site I then

calculated the mean for each vegetation type.

Statistical analyses

I analyzed all data using SAS version 8.0 (SAS Institute Inc. 2000). When
appropriate, I transformed data to meet the necessary distributional assumptions. In order
to determine if residential LSD was randomly distributed in relation to natural features, I
conducted a Chi-square analysis for each modiﬁcation type comparing expected (percent
of shoreline in each natural feature of SIOpe-wind combinations) versus observed (percent
distribution of modiﬁcation type in each natural feature category; SAS Institute 2000)
patterns of LSD relative to natural lake features. For example, if a particular development
type is distributed randomly with respect to natural shoreline features, then the percent of
lake Shoreline in each natural feature category (expected values) will equal the percent of
shoreline with the development type in each natural feature category (observed values).
To examine littoral habitat response to Shoreline development at local and lakewide
scales, I used a mixed-effect analysis of covariance (ANCOVA, SAS Institute 2000). For
each natural habitat feature (CWM abundance, substrate composition, and macrophyte
cover), I used the ANCOVA model to determine if habitat varied predictably among sites
as a function of site modiﬁcation type, overall LSD, or their interaction. For all mixed-
effect models, I treated shoreline modiﬁcation type and wind exposure as ﬁxed effects.
Lakeshore development was treated as the covariate. To be conservative, I used two-

tailed statistical tests, with rejection criterion set at a=0.05 for all analyses.

Results

Is lakeshore development distributed evenly relative to natural shoreline features?
The proportions of Shoreline composed of different combinations of wind

exposure and terrestrial slope varied among the six study lakes (Figure 2). With the

10

exception of one lake (North Lake: 62% high slope) the study lake Shorelines were
dominated by low terrestrial slope (range 38 to 91%). Predominance of low slope was to
be expected given the relatively ﬂat topography in the study area. As LSD increased,
there was a general increase in the proportion of retaining wall shoreline along with a
decrease in the proportion of undeveloped Shoreline. The distribution of each shoreline
modiﬁcation type relative to Slope and wind features, however, was not consistent across
the six study lakes and varied with increasing LSD (Table 2, Figure 3). Nearly all of the
Chi-square comparisons yielded signiﬁcant differences, indicating that the proportion of
lake shoreline in each wind-slope combination differed from the proportion of a given
modiﬁcation type in each wind-slope combination. However, patterns differed among
lakes. For example, in East Crooked Lake, retaining wall shoreline was
disproportionately prevalent along high wind-high slope, whereas in Halfrnoon Lake,
retaining wall shoreline was disproportionately prevalent along low wind-low Slope. In
only two cases did the Chi-square indicate even distribution of shoreline modiﬁcation
relative to natural features (East Crooked and Patterson Lake maintained shorelines;
Table 2).
What are the effects of local and lakewide lakeshore development on littoral habitat?
Coarse woody material

Coarse woody material abundance along 0.5 m depth study site transects ranged
from 0 to 0.53 intersections-m'land varied predictably among shoreline modiﬁcation
types (Table 3). In all study lakes, mean coarse woody material abundance was higher at
undeveloped sites (mean = 2.91 intersections-m") compared to maintained (mean = 0.43)

and retaining wall sites (mean = 0.20; ANCOVA, F = 6.48 p = 0.0024; Figure 4a).

11

Although not signiﬁcant, there were negative trends for LSD and also for the interaction
between modiﬁcation type and LSD (ANCOVA, F = 2.61 p = 0.079; Table 3), indicating
that differences in CWM abundance among modiﬁcation types are smaller in high LSD
lakes (Figure 4a), due primarily to a reduction in CWM at undeveloped Sites with
increasing LSD.
Macrophyte cover

Macrophyte cover of study sites was typically dominated by submersed
vegetation. Submersed cover ranged from 0 to ~70% among study sites. Submersed
macrophyte cover along the 0.5 m depth contour varied among modiﬁcation types
(ANCOVA, F = 6.88 p = 0.0017) and along the gradient of LSD (ANCOVA, F = 34.74 p
< 0.0001; Table 3, Figure 4b). Mean submersed cover did not differ between
undeveloped and maintained sites, but both modiﬁcation types had greater submersed
cover than retaining wall sites as determined by pairwise comparisons (p < 0.001 , and p =
0.027 respectively). This ﬁnding did not meet my expectations. Submersed macrophyte
cover along the 1 m depth contour resulted in ﬁndings similar to macrophyte cover along
the 0.5 m depth contour (Table 3). Similar to the 0.5 m contour, mean submersed cover at
the 1 m contour did not signiﬁcantly differ between undeveloped and maintained sites,
but both modiﬁcation types had greater submersed cover than retaining wall sites (p =
0.005, and p = 0.019 respectively). Also, submersed cover at l m study sites increased
with LSD (ANCOVA, F = 26.87 p < 0.0001) for all modiﬁcation types (ANCOVA, F =
4.47 p = 0.0142).

Floating macrophyte cover along the 0.5 m depth contour varied among

modiﬁcation types (ANCOVA, F = 7.33 p = 0.001 1) and along the gradient of LSD

12

(ANCOVA, F = 12.44 p = 0.0007; Table 3, Figure 5a). Mean ﬂoating cover did not
Si gniﬁcantly differ between undeveloped and maintained sites, but both modiﬁcation
types had greater ﬂoating cover than retaining wall Sites as determined by pairwise
comparisons (p = 0.0005, and p = 0.0022 respectively). Contrary to my expectations,
ﬂoating cover increased with LSD. Additionally, I noted a trend that as LSD increases,
ﬂoating vegetation at retaining wall sites appears to respond differently than at
undeveloped and maintained sites, but this interaction was not signiﬁcant (ANCOVA, F
= 1.21 p = 0.3033). Floating macrophyte cover along the 1 m depth contour varied among
modiﬁcation types and along the gradient of LSD. Similar to submersed cover, effects of
modiﬁcation type and LSD on ﬂoating cover were comparable along both the 0.5 and l
m depth contours (Table 3). As with the 0.5 m depth contour, mean ﬂoating cover at the 1
m depth contour was not signiﬁcantly different between undeveloped and maintained
sites, but both modiﬁcation types had higher ﬂoating cover than retaining wall sites (p =
0.0018, and p = 0.01 respectively). Lakeshore development as a covariate was also
signiﬁcant in explaining additional variation of ﬂoating macrophyte cover among study
sites at the 1 m depth contour (ANCOVA, F = 9.9 p = 0.0023), indicating that as LSD
increased, ﬂoating macrophyte cover generally increased. Although ﬂoating macrophyte
cover was low at retaining wall sites regardless of LSD (Figure 5a), the interaction
between modiﬁcation type and LSD was not Si gniﬁcant in explaining additional variation
in ﬂoating macrophyte cover among study sites (ANCOVA, F = 1.15 p = 0.3203).
Emergent macrophyte cover along the 0.5 m depth contour varied among
modiﬁcation types, but not along the gradient of LSD (Table 3, Figure 5b). Mean

emergent cover did not signiﬁcantly differ between maintained and retaining wall Sites,

13

but undeveloped sites had higher emergent macrophyte cover than both (p < 0.0001, and
p < 0.0001 respectively). LSD as a covariate was not signiﬁcant in explaining additional
variation in emergent macrophyte cover at study sites (ANCOVA, F = 1.96 p = 0.165).
Similar to the 0.5 m contour, emergent macrophyte cover along the 1 m depth contour
varied among modiﬁcation types, but not along the gradient of LSD. Mean emergent
cover at the 1 m depth contour was not signiﬁcantly different between maintained and
retaining wall sites, but both of these types had lower emergent cover than undeveloped
sites (p < 0.001, and p = 0.0029 respectively). As with the 0.5 m depth contour, LSD as a
covariate was not signiﬁcant in predicting emergent macrophyte cover along the 1 m
depth contour of study sites.
Substrate composition

Substrate size along the 0.5 m depth contour varied as a function of modiﬁcation
type, but not LSD (Table 3, Figure 6). Mean substrate size was not signiﬁcantly different
between undeveloped and maintained sites, but both of these types had smaller substrate
size than retaining wall sites (p = 0.0024, and p = 0.024 respectively), contrary to
expectations. LSD as a covariate was not signiﬁcant in explaining additional variation in
substrate Size of study sites (ANCOVA, F = 1.16 p = 0.285). Substrate size along the 1 m
depth contour resulted in ﬁndings similar to substrate size along the 0.5 m depth contour
(Table 3). Mean substrate size did not signiﬁcantly differ between undeveloped and
maintained sites, but both of these types had smaller substrate size than retaining wall
sites (p = 0.0045, and p = 0.09 respectively). As with substrate size at the 0.5 m depth
contour, LSD as a covariate did not explain a signiﬁcant amount of variation among Sites

at the 1 m depth contour (ANCOVA, F = 0.26 p = 0.615).

14

Discussion

Although LSD was not distributed evenly relative to natural features (wind
exposure and terrestrial slope) of the shoreline, consistent patterns in its distribution
relative to natural features were not evident. I expected retaining wall structures to be
more prevalent along high wind exposed shorelines, but this was not the case as only
three out of 10 observations across lakes had a higher proportion of retaining wall along
high wind shoreline than expected. In only two instances was the observed modiﬁcation
type distributed evenly with regards to amount of available Shoreline. The inconsistencies
among the Si gniﬁcant Chi-square tests indicate that natural features of the Shoreline are
not driving lakeshore modiﬁcation in any consistent pattern. All six study lakes had at
least some shoreline adjacent to public land that is not available for riparian development.
Thus, the pattern of public versus private shoreline might constrain my ﬁndings. For
example, a high amount of public land (not developed) along high wind exposure
shorelines could inﬂuence the lakewide patterns that were documented. Although natural
features of the shoreline did not explain variation in location of riparian development,
wind exposure and terrestrial slope were still included in the habitat sampling study
design to control for potential abiotic differences between sites that may inﬂuence littoral
habitat variables.

My results indicate that shoreline modiﬁcation type at the site-level signiﬁcantly
affected all littoral habitat variables quantiﬁed (CWM, substrate composition, and
submersed, ﬂoating and emergent macrophyte cover). Overall lakewide LSD was
signiﬁcant in explaining variation in submersed and ﬂoating macrophyte cover only,
while other littoral habitat variables were not signiﬁcantly correlated with LSD. In all

cases, littoral habitat was signiﬁcantly correlated with modiﬁcation type along both the

15

0.5 and 1 m depth contours, indicating that effects of residential lakeshore modiﬁcations
on littoral habitat extend well beyond the land-water interface (at least to 1m depth for
plants and substrate, and at least to 0.5 m depth for CWM). In general, interaction terms
were not signiﬁcant, indicating that site-level effects of development did not depend on
the lakewide level of development.

The importance of CWM as a habitat feature utilized by ﬁsh and aquatic
invertebrates in lotic systems is well documented. Research in riverine systems has
examined the negative impacts of CWM removal on aquatic organisms (Gurnell et al.
1995), but analysis of CWM as a habitat feature in lakes is lacking. Reduction of CWM
in north temperate lakes due to LSD has been documented in 16 Wisconsin and northern
Michigan lakes (Christensen et al. 1996). In a similar analysis of 34 northern Wisconsin
lakes, Jennings et al. (2003) also found less CWM with increasing LSD, as well as less
CWM at undeveloped sites compared to developed sites within lakes. My results from
CWM analyses in six southeastern Michigan lakes are consistent with ﬁndings from these
other studies; however, the relative abundance of CWM in my lakes is substantially
lower. Christensen et al. (1996) found CWM density at forested Sites to be 0.38 logs-m",
while developed sites had 0.057 logs-m'l. In my study, undeveloped sites contained
0.0773 logS°m", and developed sites (both maintained and retaining wall) had only
0.0078 logs-m". Because efforts were made to control for natural differences among
modiﬁcation types, my results combined with others indicate that riparians remove CWM
from the nearshore areas adjacent to their dwellings, but that differences in land cover or
geographic location also inﬂuence lakewide levels of CWM. On the lakewide scale,

CWM relative abundance decreased with increasing LSD, particularly at undeveloped

16

sites. This reduction in CWM at undeveloped sites may be due to removal by riparian
lakeshore owners in close proximity to such sites, but this scenario is unlikely. A more
plausible explanation for decreased abundance of CWM with increasing LSD involves
the generation and transport of CWM within a lake. If CWM is generated along
undeveloped shoreline and then is transported within a lake due to wind or wave action to
a different shoreline, the chance that CWM will end up settling along a developed
Shoreline (where it is likely to be removed by riparians) increases with LSD. This
explanation, however, requires two assumptions: that developed shorelines do not
generate substantial amounts of CWM; and that lakeshore owners remove CWM from
their nearshore areas rather expeditiously. Higher amounts of CWM at undeveloped sites
compared to other modiﬁcation types and decreased relative abundance of CWM with
increasing LSD met my expectations concerning woody material in the littoral zones of
north temperate lakes.

Human development of lake shorelines may include the removal or reduction of
aquatic vegetation in order to establish swimming areas or meet lakeshore owners’
aesthetic goals. With a few notable exceptions (Radomski and Goeman 2001; Jennings et
al. 2003, Hatzenbeler et al. 2004), the response of littoral vegetation to LSD of north
temperate lakes has not been examined. In a study of 44 Minnesota lakes, Radomski and
Goeman (2001) documented a decrease in the amount of ﬂoating-leaf and emergent
vegetation present at developed sites compared to undeveloped lakeshore. In a more
recent study, Jennings et al. (2003) found that not only were ﬂoating-leaf and emergent
vegetation reduced at developed Sites compared to undeveloped sites, but also in relation

to lakewide LSD. While Jennings et al. (2003) did not ﬁnd submersed vegetation to be

17

affected by site level differences, I found submersed, ﬂoating-leaf, and emergent
macrophyte cover were greater at undeveloped sites compared to sites with retaining
walls. These results might be explained due to my characterization of three separate
modiﬁcation types that may represent a gradient of disturbance (with retaining wall
shoreline having a higher disturbance to the littoral habitat than maintained shoreline). I
also found emergent macrophyte cover to be greater at undeveloped sites compared to
both developed modiﬁcation types. Based on these results, I surmise two possible
explanations for the observed patterns in macrophyte cover. First, submersed and ﬂoating
macrophyte assemblages as a whole are perhaps more resistant than emergent vegetation
in response to disturbance caused by maintained shorelines, but all cover types are still
reduced by retaining wall shorelines. Second, the modiﬁcation types I examined may
indicate a gradient of disturbance in which undeveloped, maintained, and retaining wall
shorelines represent low, intermediate, and high disturbance respectively. In contrast to
the ﬁndings of Jennings et al. (2003), I found a positive correlation between macrophyte
cover (submersed, ﬂoating-leaf, and emergent) and LSD, which did not meet my
expectations. Although I did not assess speciﬁc mechanisms, one possible explanation for
this observation would be increased non-point nutrient inputs due to increases in
cumulative LSD that could lead to higher incidence of macrophytes in the littoral zone.
This is a plausible explanation considering that the total phosphorus values for the Six
lakes in this study increased, although not signiﬁcantly (p = 0.149), with increasing LSD
(see Chapter 2).

Substrate Size and composition have often been identiﬁed as critical habitat

features for many stream-dwelling organisms. Alteration in the Size and structure of

18

substrata due to anthropogenic stressors has rarely been studied in lakes (but see Jennings
et al. 2003). My results indicate larger substrate size along retaining wall Sites, compared
with both undeveloped and maintained Sites, but no relationship was detected between
substrate and LSD. The larger substrate along retaining wall Sites may be due to reduced
sediment transport from land to water caused by the shoreline structure. I expected
substrate at undeveloped sites to be larger than at developed Sites reﬂecting greater
siltation at developed sites, but this was not the case in my study. Currently, it is unclear
how changes in substrate caused by LSD will affect nesting ﬁsh. The ﬁsh communities of
the Six study lakes are dominated by black crappie Pomoxis nigromaculatis, bluegill,
largemouth bass, pumpkinseed Lepomis gibbosus, rock bass A mblopites rupestris, and
smallmouth bass (J ubar, unpublished data), which construct nests in nearshore littoral
areas (Balon 1975) where the effects of LSD are most apparent. Understanding the direct
effects of modiﬁcation type and LSD on substrate composition of north temperate lakes

should lend insight to the mechanisms inﬂuencing ﬁsh nesting and nest distribution.

19

CHAPTER 2: LAKEWIDE EFFECTS OF RESIDENTIAL LAKESHORE
DEVELOPMENT ON WATER CHEMISTRY, MACROPHYTE COVER, AND
LITTORAL FISH GROWTH

Introduction

The alteration of lake shorelines by humans has increasingly become an area of
concern for aquatic ecosystem managers throughout North America. The impacts of
activities associated with lakeshore development (LSD) can affect littoral habitat
features, such as abundance of coarse woody material (Christensen et al. 1996), aquatic
macrophyte cover (Bryan and Scamecchia 1992, Radomski and Goeman 2001), substrate
particle size and composition (Jennings et al. 1996, Jennings et al. 2003) and also species
composition and spatial distribution of ﬁshes (Jennings et al. 1999, Scheuerell and
Schindler 2004). Studies contrasting developed and undeveloped lakeshore sites have
focused on the local effects of LSD on ﬁshes and aquatic habitat, and such studies and
their applicability to management have recently been emphasized in the primary
literature. Fewer studies have examined the cumulative impacts of LSD, which may
affect lake habitat and associated biota on a broader scale. Such cumulative impacts of
LSD may indirectly affect water chemistry, whole-lake macrophyte cover, and ﬁsh
growth, although these effects remain poorly understood. The alteration or loss of littoral
habitat in nearshore areas due to incremental LSD may combine with changes in
vegetative cover and water chemistry at the whole—lake scale to negatively affect littoral
ﬁsh species. However, much remains unknown regarding ﬁSh response to LSD as few
studies have addressed this issue (but see Jennings et al. 1999, Schindler et a1. 2000,

Scheuerell and Schindler 2004).

20

The effects of various land uses on nutrient concentrations in streams and rivers
have been well documented (Omemik 1976, Roth et al. 1996), as have more localized
effects of riparian inﬂuences on lotic systems (Roth et al. 1996, Nakamura et al. 2000).
Similar to riverine ecosystems, lake interactions with riparian and land use changes are
potentially complex. However, lake research has emphasized the broader watershed
scale. At the watershed scale, research has examined the effects of human land use on
lake productivity and ﬁsh habitat (Evans et al. 1996, Siver et al. 1996, Gunn and Sein
2000). For example, increased phosphorus levels due to increased fertilizer use or other
non-point nutrient inputs can lead to decreased dissolved oxygen and increased mean
temperature of lakes (Evans et al. 1996) causing eutrophication. Such a shift in trophic
status may reduce suitable habitat for some ﬁsh species. Relatively little is known
regarding the cumulative effects of residential LSD on water chemistry and plankton in
lakes, although in a study of New England lakes, Stemberger and Lazorchak (1994)
determined that percentage of disturbed lakeshore may be important in explaining
variability in zooplankton assemblages among lakes. Overall, LSD is likely to result in
higher nutrient input into lakes, due to increased fertilizer use by riparian landowners and
increased sediment transport across the land-water interface resulting from lakeshore
degradation and reduced riparian vegetation.

Lakeshore development can potentially affect macrophyte communities on both
individual site and whole-lake levels, in potentially complex manners. Lakeshore
development may indirectly increase aquatic plant abundance through increased inputs of
nutrients important for plant growth. For example, in a study of six southeast Michigan

lakes, I noted an increase in submersed vegetation at developed sites with increasing

2]

et al. 1992) because bass foraging success is higher at moderate as opposed to high plant
abundance. Sparse amounts of vegetation, however, may constrain bass diet composition
and growth, because relatively few macroinvertebrates and prey ﬁshes may persist
(Crowder and Cooper 1979, Anderson 1984). In addition to aquatic macrophytes, other
littoral habitat features, such as coarse woody material, also provide important refuge and
foraging areas for many aquatic organisms (Savino and Stein 1989b, Bowen et a1. 1995).
In a study of 14 northern Wisconsin and Michigan lakes, Christensen et al. (1996)
showed that CWM was negatively correlated with LSD, with a similar pattern noted in
southeast Michigan lakes (see Chapter 1). In a more recent study on the 14 northern
lakes, Schindler et al. (2000) recorded a decrease in bluegill growth rates as LSD
increased, and saw a Similar though weaker trend in bass. To date, however, no study has
examined the response of macrophyte cover to cumulative increases in LSD and the
subsequent effects on growth of littoral ﬁshes.

In order to quantify the effects of residential LSD on water chemistry, whole-lake
macrophyte cover, and growth of littoral ﬁshes, I used a comparative approach of lakes
along a gradient of residential development. I measured water chemistry, macrophyte
cover, and growth rates of bluegill and bass in 15 southeast Michigan lakes that have
similar morphometry and watershed land use, but differ according to overall amount of
LSD. Based on results of previous studies, I expected total phosphorus to increase and
Secchi disk depth to decrease in response to increasing LSD. I expected whole-lake
macrophyte cover to decrease with increasing LSD, particularly for emergent and
ﬂoating-leaf cover types which are more likely to be negatively affected by activities

associated with LSD. Finally, I also expected bluegill growth rates to decrease

23

et al. 1992) because bass foraging success is higher at moderate as opposed to high plant
abundance. Sparse amounts of vegetation, however, may constrain bass diet composition
and growth, because relatively few macroinvertebrates and prey ﬁshes may persist
(Crowder and Cooper 1979, Anderson 1984). In addition to aquatic macrophytes, other
littoral habitat features, such as coarse woody material, also provide important refuge and
foraging areas for many aquatic organisms (Savino and Stein 1989b, Bowen et al. 1995).
In a study of 14 northern Wisconsin and Michigan lakes, Christensen et al. (1996)
showed that CWM was negatively correlated with LSD, with a similar pattern noted in
southeast Michigan lakes (see Chapter 1). In a more recent study on the 14 northern
lakes, Schindler et al. (2000) recorded a decrease in bluegill grth rates as LSD
increased, and saw a Similar though weaker trend in bass. To date, however, no study has
examined the response of macrophyte cover to cumulative increases in LSD and the
subsequent effects on growth of littoral ﬁshes.

In order to quantify the effects of residential LSD on water chemistry, whole-lake
macrophyte cover, and grth of littoral ﬁshes, I used a comparative approach of lakes
along a gradient of residential development. I measured water chemistry, macrophyte
cover, and growth rates of bluegill and bass in 15 southeast Michigan lakes that have
similar morphometry and watershed land use, but differ according to overall amount of
LSD. Based on results of previous studies, I expected total phosphorus to increase and
Secchi disk depth to decrease in response to increasing LSD. I expected whole-lake
macrophyte cover to decrease with increasing LSD, particularly for emergent and
ﬂoating-leaf cover types which are more likely to be negatively affected by activities

associated with LSD. Finally, I also expected bluegill growth rates to decrease

23

systematically as LSD increases, with a similar response for bass because LSD may
reduce the ﬁsh production capacity of lakes by reducing the amount of habitat available

for forage, refuge, and spawning.

Methods

Study lakes

I selected 15 lakes within a single ecoregion and within a single major river
watershed, in southeastern Michigan, USA (Figure 7), in order to control as much as
possible for differences among lakes in climate, geology, morphometry and aquatic ﬂora
and fauna. Study lakes were selected to exhibit similar surface areas (29 to 106 hectares)
and sufﬁcient depth for summer stratiﬁcation (mean depth >3m, with the exception of
two lakes). Additionally, lakes were chosen to have similar composition of land use
types in their watershed (< 70% of either agricultural, forested, Or urban land use within a
500 m buffer). In doing so, I aimed to select lakes that were biologically and physically
similar so that lakes differed primarily in the extent of their LSD (Table 4). The Huron
River watershed contains lakes with a wide range of LSD, thus providing a mix of highly
developed, moderately developed, and undeveloped lakes. Visual observations by boat
were used to quantify the number of riparian dwellings within 50 m of each lake. Number
of dwellings for each lake was then divided by lake perimeter (km) to calculate LSD
(dwellings-km").
Water chemistry

To evaluate the effects of LSD on lake productivity, I sampled water chemistry

and clarity of the 15 study lakes during August and September of 2003 and 2004. Secchi

24

disk depth, a measurement of water clarity, was surveyed from the shaded side of the
boat. For each lake, the epilimnion depth was estimated from a temperature proﬁle at the
deepest area of the lake. I measured total phosphorus and total alkalinity using
epilimnetic water samples collected using a tube sampler. Total alkalinity (mg-L'l
CaCO3) was measured immediately with a titration test kit (LaMotte). Total phosphorus
samples were frozen for later analysis. In the lab, total phosphorus was measured using a
persulfate digestion (Menzel and Corwin 1965) followed by standard colorimetry
(Murphy and Riley 1962).
Macrophyte sampling

Whole-lake macrophyte sampling was conducted in the 15 study lakes during
August and September of 2003. I sampled lakes in mid-late summer because the lakes are
likely to be well stratiﬁed during that time, and macrophytes typically are at or near
maximum growth. To assess macrophyte cover at the whole-lake scale, I used a
modiﬁcation of the point-intercept method (Madsen 1999, Spence Cheruvelil 2004). I
used a Geographic Information System (GIS) to overlay a grid of sample points on each
lake. The sample points, each with a corresponding latitude and longitude, were located
40 or 50 m apart, depending on lake surface area (40 m for lakes with a surface area <65
ha, or 50 m for lakes with a surface area >65 ha). A handheld global positioning system
(GPS) unit was used to locate each sample point in the ﬁeld. At each sample point, water
depth was recorded, and macrophyte cover was assessed either by visual inspection (sites
where macrophytes were visible below the water surface) or by 2-sided rake (deeper
sites, where macrophytes were less visible). Areal cover estimates at each site were

assigned for three macrophyte categories: submersed macrophytes, ﬂoating leaf

25

macrophytes and emergent macrophytes. Cover estimates were based on qualitative
density of each macrophyte category, and were given a cover score ranging from 0 to 3 (0
= 0-5%, 1 = 5-33%, 2 = 33-67%, 3 = 67-100% cover) with a cover score of 2 or 3
indicating dense cover. The presence or absence of the invasive Eurasian waterrnilfoil
(Myriophyllum spicatum) was also noted at each site. From the whole-lake surveys, I
calculated eight macrophyte metrics: percent cover (lake), percent cover (littoral), percent
dense cover (lake), percent dense cover (littoral), percent submersed cover (littoral),
percent ﬂoating leaf cover (littoral), percent emergent cover (littoral), and percent
Eurasian waterrnilfoil cover (littoral) (Table 5).
Littoral ﬁsh sampling

To quantify the effects of LSD on littoral ﬁsh growth, I collected scales from
bluegill and bass in summer 2003 and 2004. Fish were sampled using nighttime
electroﬁshing (7 amps pulsed D.C., 120hz) transects conducted along the 1 m depth
contour at haphazardly selected sites within each lake. Transects averaged ten minutes
and were conducted along both developed and undeveloped stretches of lakeshore. In
2003, sampling on all lakes began at dusk and continued until at least 50 bluegill
(>80mm total length) and 50 bass (>100mm total length) were captured; however, 8 of
the sampling surveys produced fewer than 50 bass. To supplement sample sizes,
additional nighttime electroﬁshing was conducted during summer 2004 until more than
50 bass were collected per lake. Approximately ten scales were taken from posterior to
the pectoral ﬁn and just above the lateral line from at least 50 bluegill and 50 bass
ranging the entire Size distribution of ﬁsh captured. Lengths of all bluegill and bass

captured were recorded to the nearest millimeter. In order to collect smaller bluegill and

26

bass, supplemental sampling including beach, purse, and bag seining was conducted in all
lakes during summer 2004.
Fish growth analysis

In the lab, ﬁsh scales were mounted on slides for use in age determination and
grth analysis. I measured scale incremental grth distances using an Optimas 6.5
image analysis system and a Nikon Eclipse E600 compound microscope and camera at
20x magniﬁcation.

For the bluegill and bass collected, I back-calculated lengths at previous ages
using three different methods: Fraser-Lee method (Carlander 1982), body proportional
hypothesis (BPH), and scale proportional hypothesis (SPH; Francis 1990). For the Fraser-
Lee'back-calculations, a common intercept (“a” below, representing ﬁsh total length at
the onset of scale formation) was estimated separately for bluegill and bass sampled from
all 15 lakes, and then back-calculated lengths estimated using the formula:

LC 9- a I I.

XSt

 

Lt=a+

SC
where a is the intercept of the relationship LC = a + b(SC) obtained by linear regression,
Sc and S, are scale size at capture and for the year of back-calculation, respectively, and
LC and L, are total length at capture and for the year of back-calculation, respectively. The
BPH method assumes a constant proportional deviation in ﬁsh length from mean ﬁsh
length expected for a ﬁsh throughout life. To back-calculate length for a given scale size,

I used the formula:

Ll = w. XLC
, a+bSC

27

where a, SC, 5,, LC, and L, are as deﬁned above for the F raser-Lee method. The SPH
method assumes a constant proportional deviation in scale size from mean scale Size for a
ﬁsh throughout life. To back-calculate length based on the assumptions of the SPH

method, I used the formula:

., =-l—:~l+ch+—3-lxl%;-l

where c and d are the intercept and slope of the relationship Sc = c + d(LC) obtained by
linear regression. For each bluegill and bass, I used all three methods to estimate back-
calculated total length (mm) at each scale annulus and determined growth increments
(mm/yr) from the difference in estimated total lengths between consecutive annuli.

In order to test for Lee’s phenomenon, wherein younger ﬁsh from a sample would
appear to be exhibiting greater growth than ﬁsh of the same age from an earlier year-
class, I used conventional qualitative methods (DeVries and Frie 1996). Lee’s
phenomenon appears evident in my bluegill data (Table 6). Therefore, to minimize
possible error that would be associated with back—calculated lengths, I restricted bluegill
growth analysis to back-calculated lengths from ﬁsh during the 2001 and 2002 growth
years. A qualitative analysis of bass back-calculated total lengths did not clearly indicate
evidence of Lee’s phenomenon (Table 7). However, in order to minimize possible errors
associated with back-calculating lengths of older ﬁsh (and possible errors associated with
recent grth years) I restricted bass growth analysis to back-calculated lengths of ﬁsh

for the 1998 to 2001 grth years.

28

Statistical analyses

All data were analyzed using SAS version 8.0 (SAS Institute Inc. 2000). When
appropriate, data were transformed to meet the necessary distributional assumptions. In
order to determine if residential LSD affected whole-lake macrophyte cover, each
macrophyte metric was regressed against dwellings-km'l (PROC REG, SAS Institute
2000). I conducted similar analyses to determine if water chemistry (total phosphorus,
total alkalinity) and water clarity (Secchi disk depth) were correlated with LSD. To
determine if littoral ﬁsh grth varied predictably as a function of LSD, I compared
mean back-calculated total length estimates for bluegill and bass for the 15 study lakes
(PROC MIXED, SAS Institute 2000). To account for the effects individual lakes may
have on ﬁsh within that lake, I used a model which nested ﬁsh within lakes. For the
mixed-effect models, I treated back-calculated total length and growth year as ﬁxed
effects. Because multiple growth increments were calculated from a single individual,
and thus these observations were not statistically independent, I included individual ﬁsh
as a random effect in the model. Lakeshore development was treated as the covariate. To
determine if littoral ﬁsh growth varied predictably as a function of macrophyte cover or
water chemistry, separate analyses were conducted with macrophyte cover metrics or
water chemistry as covariates (PROC MIXED, SAS Institute 2000). Rejection criterion

was set at a = 0.05 for all analyses.

Results

Water chemistry
Total phosphorus in the 15 study lakes was regressed against LSD and year,

resulting in a non-Signiﬁcant LSD*year interaction (p > 0.5). Therefore, I calculated the

29

mean total phosphorus between the two samples years and used the values for subsequent
analyses. Mean total phosphorus in the 15 study lakes in the summers of 2003 and 2004
ranged from 10.7 — 21.3 rig-l" (Table 4), with Secchi disk depth varying from 0.7 — 3.25
m in 2003 and 1.6 — 5.15 m in 2004. Total phosphorus levels and water clarity indicate
the trophic state of the study lakes to be predominantly mesotrophic, with some lakes
leaning towards oligo-mesotrophic and others meso-eutrophic in condition (Wetzel
2001). Although Secchi disk depth tended to decrease with increasing total phosphorus
during 2003 and 2004, I detected no signiﬁcant correlation for either year (p-values >
0.05). Mean total phosphorus did not vary predictably with LSD (r2 = 0.05, p > 0.39), nor
did Secchi disk depth in 2003 (p > 0.77) or 2004 (p > 0.36).
Whole-lake macrophyte cover

Total plant cover in the 15 study lakes varied from 13 — 94% (Table 5). Total
dense cover of macrophytes ranged from 1 — 89%, and littoral plant cover in the 15 lakes
ranged from 19 — 96%. Because emergent and ﬂoating macrophytes only occurred in
relatively shallow regions of the study lakes, I only calculated littoral cover for these
plant types, which ranged from 2 - 40% and 3 — 80% for emergent and ﬂoating plants,
respectively (Table 5). Because multiple macrophyte growth form categories can occur at
the same location (e. g. emergent and ﬂoating cover present at the same site), the
macrophyte categories are not mutually exclusive; therefore percent cover across growth
forms for an individual lake does not sum to 100. Contrary to my expectations, only
littoral ﬂoating macrophyte cover was negatively correlated with LSD (r2 = 0.27, p =
0.045; Figure 8a). Emergent macrophyte cover exhibited a similar, though not signiﬁcant

response to LSD. All other macrophyte metrics did not vary predictably with LSD (p-

30

values > 0.05; Figure 8b-d). Additionally, mean total phosphorus did not explain a
signiﬁcant amount of variation in whole-lake plant cover (p > 0.51).
Littoral fish growth

Back-calculated total lengths of bluegill and bass derived from the three separate
analyses (Fraser-Lee, body-proportional hypothesis, and scale-proportional hypothesis)
were quite similar (for all comparisons r2 > 0.97, with slope not signiﬁcantly different
from 1). Because I sampled ﬁsh using multiple gear types and scales collected represent a
wide range of body sizes (35 — 260 mm for bluegill, 84 — 505 mm for largemouth bass),
the BPH method of back-calculating total length is appropriate for my study (Francis
1990). Growth increments used in statistical analyses are thus calculated from the BPH
method.

Although annual growth increment generally declined with increasing ﬁsh size, as
would be expected, ﬁsh size explained relatively little variation in growth increments,
which ranged broadly among bluegill (Figure 9) and bass (Figure 10) for any given ﬁsh
size. The analysis resulted in a signiﬁcant three-way interaction of bluegill back-
calculated total length“ grth year*LSD (p < 0.0001), indicating that the effect of LSD
on growth increment varies with ﬁsh size and growth year. The interaction of bluegill
back-calculated total length*LSD was also signiﬁcant (p < 0.036), indicating that the
effect of LSD on growth increment varies with ﬁsh Size. Lakeshore development as a
main effect was signiﬁcant (p < 0.004) in predicting bluegill growth among lakes, and
contrary to my expectations, showed a trend to increase (positive parameter estimate)
with increasing LSD. For bass growth among the 15 lakes, the three-way interaction of

back-calculated total length* growth year* LSD and the interactions of back-calculated

31

total length“ growth year and growth year* LSD were not signiﬁcant, therefore I removed
those interactions from the mixed-model in SAS. Following appropriate statistical
adjustments, bass growth among the study lakes resulted in a signiﬁcant interaction
between back-calculated total length*LSD (p < 0.006; positive parameter estimate)
indicating that effects of LSD on bass growth vary depending on ﬁsh size. When
examined as a main effect, bass growth was marginally signiﬁcant and negatively
correlated with LSD (negative value for parameter estimate, p = 0.094) indicating
decreased bass growth with increasing LSD, which met my expectations. In order to
facilitate interpretation of the bluegill and bass growth data, I constructed plots (Figures
11 and 12) comparing annual grth increment and back-calculated total length for both
ﬁsh species, while accounting for three levels of LSD (low development <13
dwellings-km", medium development 13-20 dwellings-km’i, and high development >20
dwellings-km"). Whole-lake macrophyte cover, dense cover, and total phosphorus did

not explain variation in grth among lakes for either bluegill or bass (p-values > 0.05).

Discussion

I show that cumulative amount of LSD does not signiﬁcantly affect lake water
chemistry and water clarity variables (total phosphorus and Secchi disk depth). I expected
to note an increase in total phosphorus and a decrease in Secchi disk depth with
increasing LSD, similar to results of studies examining riparian and land use changes on
the watershed-scale (Evans et al. 1996, Siver et al. 1996). The lack of variation in water
chemistry among lakes of differing LSD may be due to mechanisms extending beyond

the scale of my study. For example, watershed-scale effects on water chemistry have

32

been well documented in lakes, but I only controlled for extreme proportions of
individual land use within 500 m of each study lake. A more quantitative analysis of
individual lake watershed features (such as land use, surface water ﬂow, etc.) may shed
insight into mechanisms controlling water chemistry within lakes. I expected water
clarity to be correlated with total phosphorus (Wetzel 2001), but although I found a
negative trend between Secchi disk depth and total phosphorus, the relationship was not
signiﬁcant. The lack of a signiﬁcant correlation between total phosphorus and Secchi
depth may be due to the relatively small number of lakes sampled in this study and the
relatively narrow range of total phosphorus values involved. Also, I sampled water
chemistry and water clarity only once during the summers of 2003 and 2004, and samples
obtained for one survey date may not be representative of the seasonal averages for each
lake. Finally, differences among lakes with regards to prevalence of septic versus sewer
waste disposal might also prevent a straightforward relationship between LSD and
nutrient levels.

Wholeélake plant cover in the 15 study lakes did not vary signiﬁcantly with LSD
for most macrophyte forms surveyed, with only littoral ﬂoating macrophyte cover
negatively correlated with LSD. My ﬁndings are consistent with other studies which
failed to detect a response in submersed macrophyte cover (Jennings et al. 2003), but
found signiﬁcant declines in ﬂoating and emergent cover (Radomski and Goeman 2001;
Jennings et al. 2003) in relation to increasing LSD. The general lack of response of
macrophyte cover types to LSD may be due to contrasting effects to lakes resulting from
activities associated with LSD. For example, while riparian alteration of the lakeshore

may contribute to increased nutrient inputs contributing to macrophyte growth, activities

33

associated with LSD (such as chemical or mechanical harvest of plants, or increased
lakeshore degradation due to recreational activities) may act to reduce macrophyte cover
or otherwise inhibit aquatic plant growth. Regardless of the possibility of increased
nutrients positively affecting macrophyte growth, one consequence of LSD remains
relatively certain, that being the reduction of ﬂoating and emergent vegetation (see
Chapter 1). This may be due to the fact that many emergent and ﬂoating macrophyte
assemblages are composed of plant species intolerant to environmental degradation
(Hatzenbeler et al. 2004). Additionally, the Spatial distribution of emergent and ﬂoating
macrophytes (generally associated with nearshore areas) brings these plants in closer
proximity to residential lakeshore alterations, whereas some submersed plants may be
able to inhabit regions of the lake bottom that are relatively isolated from the immediate
effects of LSD and associated human activities.

The indirect effects of LSD on growth of littoral ﬁshes are potentially
complicated, because productivity of ﬁsh populations is often inﬂuenced by such factors
as lake surface area, water chemistry, ﬁsh population density, predator-prey interactions,
and angler harvest (Shuter et al. 1998; Tomcko and Pierce 2001). By selecting lakes of
similar size, morphometry, and presumed angling pressure (all study lakes were public
access), but which differed primarily in amount of LSD, I attempted to control for
inherent variation among lakes that might inﬂuence ﬁsh growth. More than 50 bluegill
and 50 bass were collected from each lake, and these species comprised the vast majority
of ﬁshes caught while electroﬁshing, indicating their predominance in all study lakes.
The results of my study indicate that bluegill growth is dependent upon LSD, but the

response varied with back-calculated total length of the ﬁsh and growth year. When

34

considering only LSD, bluegill growth was positively correlated with LSD, which is
contrary to the ﬁndings of Schindler et al. (2000) that indicated a negative relationship
between LSD and bluegill growth. Bass growth in my study lakes declined with
increasing LSD, but the slope of the relationship varied with growth year. This is similar
to the ﬁndings of Schindler et al. (2000), who found a marginally signiﬁcant negative
relationship between growth of the largest size class of bass and LSD. The study of
Schindler et al. (2000), however, only examined bass growth in nine lakes, with ﬁve of
the lakes represented by relatively small sample sizes (n < 30). The ﬁndings of my study
and Schindler et al. (2000) are qualitatively similar, but because we included more lakes
and larger sample sizes per lake, we likely had greater statistical power to detect an effect
of LSD. Also, there may be habitat differences between my study lakes and the lakes
examined by Schindler et al. (2000), with respect to macrophyte cover, substrate
composition, and particularly coarse woody material (see Chapter I), which may underlie
why we detected a stronger response.

The ﬁnding that bluegill grth increases while bass growth decreases along the
LSD gradient was an unexpected result in my study. One reason for the opposite effects
of LSD on growth may be due to the strong predator-prey interactions between bluegill
and bass (Savino and Stein 1982, 1989a). In a study of seven Wisconsin lakes, Olson
(1996) found bluegill growth has strong effects on the structure of bass populations, due
to the inﬂuence of growth rate on bluegill population size-distributions. For example, in
lakes with Slow bluegill growth, bass growth was strongly Size-dependent (small bass had
low growth, while larger bass had higher growth). In lakes with high bluegill growth

rates, bass growth rates were more uniform across all bass Sizes. Such strong predator-

35

prey interactions may explain why I found bluegill and bass growth responded differently
to LSD.

Determining the cumulative effects of LSD on lakes is difﬁcult because time lags
may exist between disturbance to the ecosystem and the response, and because the type
of development differs among lakes and over time. Numerous instances regarding river
and stream system response to riparian disturbance over time exist in the primary
literature. For example, Nakamura et al. (2000) believe the network structure of stream
and riparian systems may lend resilience in response to major disturbances by providing
widely distributed refuges. In contrast, lakes with relatively closed systems may not
exhibit similar resilience to disturbance. For example, Christensen et al. (1996) conclude
that 200 years would be required for natural coarse woody material in northern WI and
MI lakes to return to pre-settlement levels. The relatively low resilience of lakes
compared to lotic systems may be more apparent in such habitat characteristics as coarse
woody material and substrate composition. Aquatic vegetation, however, may exhibit
relatively high resilience to disturbance due to the relatively short life histories of
macrophytes. Degradation of water quality as a result of long-term changes to watersheds
due to human activity has been documented (e.g., Siver et al. 1996), but little is known

regarding the long-term effects of LSD on aquatic ecosystems.

Conclusion

I quantiﬁed the lakewide response of water chemistry, macrophyte cover, and ﬁsh
growth to a gradient of LSD. Ideally, a study of this nature would seek to identify aquatic

ecosystem responses both before and after shoreline alteration. Due to restraints imposed

36

by time and resources, such experimental studies are not feasible at present. The 15 lakes
comprising the gradient of LSD used in this study represent a wide range of disturbance
levels, from virtually undeveloped lakes, to nearly complete lakeshore development. If
conditions in low to moderately developed lakes on this gradient are considered to be
representative of conditions that may have once been found in more highly developed
lakes, then studies of this nature provide a glimpse of what many lakes were like prior to
extensive human development. Furthermore, ecosystem conditions in highly developed
lakes can be viewed as the result of development should low development lakes
experience increased levels of human lakeshore modiﬁcation. I feel this approach is
important for aquatic ecosystem managers seeking to determine past (and predict future)

responses of lake habitat and biota to residential LSD.

37

APPENDIX

Tables and Figures

38

ASKS

M ED :cm 6888.: n com 383:3th u $3 8:: some wEEsotsm 5&3 8 com a 55:5 8: :52 Beacon“
doom :5 moon Mo 50883 meta: 3:26:00 @323 03: .«o AEV nae: OE: Eooom SSE 2: mm Ecoem.V
.38 :5 Sam :0 E2555 mats: coo—S BEES o3: me 3096 HRS: 56:33 :88 2.: mm v35.

doom :5 moom :0 $0883 mag: :83 SEER. 03: mo 3R3 3.838;: :39 :88 2: 2 my

.oE—ouonm 8:: 2: mo bud—smut: “88% 8865 85:;
$5885 .286 ﬁﬁtavfﬁ \ $685: 83 n mam «38.8.: 2: Go: :82528 :92: Evade—go: 3:82? 2: a ”Em _

 

 

WE 5.va :6 Wm ow: mgm RNN m: on; mm.m 8% we :owcozmm
mum m.mm Ymm _.m 03 v.3 ﬁlm 3: no; mm.m 3% 5 5.82
w; m6: NS Wm com v.2 Sat :2 EN :6 mmmn no 209533
0.3 me m.mm m.m n2 ”3 mg: no em; 3M wmmo oh: :oxoocu mam,
o; odm m.mm 9N 0: Wm: mmd em mm; mo.“V ommm om 95m
2 ”S 0% o.m I; «.2 02. em N; San Sow mm :85
man: when me< :6on v5< A; mwhmﬁuwuv mug—.03.: ham Amhwmwﬁv LWMWMHW. Amowwwwoﬁ 8:5
..\e .x. ..\e v n N Gmx— _ :32 3:5 «gram

 

.moxﬂ beam 0 2: mo co>oo :5: :5 EmEoso 833 Ade Evade—26: 80:80:: $3323.85 .8753 A «Bah.

39

8588.7. :8, 8532 :88:— 023 5:5 0880: 88828 on 8: :800 858-5...

 

 

 

 

88.: v 2.2” 8.3 3.: 3.: 3.2 Nﬁmh No.3 o 820:5
50:6 v 52 3.8" 5.0— o 2.3 c 5.? 3.? 582
88.: v mud: 2.- 2.: 2w: 2.3 NONM 3.2 3.3. 805:5
2.2.0.088:
58.: v 8.5 :13. Ram— o 36.: Nmﬁm 3.3 o 880:0 :85
58.: v Eda 2.3 2.2 5.2. 3.: wodm 3.2 mi: :55
wooed 34mm 212 2.: o mms: wméw o 0 5:5
82 .o 3.2” 3.2m 8.: c 25m mmdm No.2 3.2 820:5
Seed v 52 3mm 3...: 2..- 2.2 QM 5.2. 34mm 582
58.: v mbév 3.5m 2 .: mm.:~ :dm 2.: 3.2 3.2 805.8: 855552
8: Nd 8.5 we. _ m ~52 3w:— wwdv ween: 0N2 3.x :0880 :85 . .
p.25: 3.2 :v.mm w :2 o 3. 3 5.»: mm .2 :96— :55
58.: v 3 .mm 2.4M 26 3.3 mmﬁ: w~.mm o o 535
Sood v No.2” ”ham 3.: o «Yum 3.2 8.2 mm.- 820:5
88.: v 5.2 wvﬁ 3.2 8.2 2.2 05.3. 5.2. 85m 5.82
58.: v mud: 5.5 2.: o 2.9m 9.2 3.2 o . 805.8: =9:
58.: v 2:." has 8.2 mbén 5.9. 8.; 3.2 :33” 880:0 :85 :55305
58.: v 34mm mdm 2.2 m6: 3.; o 5.2 o :55
*<\Z 2.3 o :d o mmﬁ: o o o 52.5
2::
2:85 8:895 :Qcomno 88095 82080 8:895 83080 8:895 :02080 8:5 8:85:02

.x.

a\o

o\o

o\o

.x.

.x.

o\o

o\o

 

 

>8— .

087.
:53 30.—

.8.

85m

65.: >84

28.0

E. as: :05

28..."

an: as: 8:

 

805888: 20:88— w58205 0. 25:88: 88:8
2: 80:3 820.5: 3:80:83,“ 203 5: 2858800 80.: 28> 885 2: 8:885 :85 .5938 288-50 a 50: :38: 08 828-5 888008:
882:8 0: :88 88888: :otomno .3 :2885 on :80? 888.: 888 03588 0: 3:22 89¢ 8:85:05 8 8:855: 8>m .25
8:85:05 :08 .8 a: 0:8 5 :2: 888.: 888: .8 885880 88258 a 5553 05—088 8 585m 2: 882:5 :03280 3% 80085 .88?
30:85 :8 :53 30555 8.88.: 888 .8 8:88:80 :08 8: 00:: a 5 05—08% .8 888m 2: 882:5 :28on :5 8885 80:: 28%

one. 2: .8 :08 5 2:28—28:28 :8 805858 .83 M55895 2;: 8:85:05 :08 8.: 052085. 83030 :8 :2088 8 888855 .N 28,—.

40

58:80 58: 5 md 0:: 58:. 8:03:80 .28 83 5588.. 5:085 8003 00800...

 

 

 

 

d. .9 a: .9 E .9 S. .9
8... :8... 38... 8... 82.... :8... 828.88
m2 m2 22.2.8
2. .9 E. .9 c. .9 E. .9
m2 8:. 5...... 5...... m 8. .o :.........v 8...... v .38 05.8235
Z 805055
.9 .9 E. .9 2. .9 E. .9
m2 5...... 5...... 5...... 02 2.2.... :8... «2...... .88 088.85
8.8.”.
2. .9 E. .9 c. .9 E. .9
m2 8...... v 5.2.... S.......v 02 2.2.... v 5...... ..........v .28 ”Siege
850556
.9 2. .9 a... .9
5.... So... 5...... 8.......v .522.
*<\Z *<\Z *(QZ §<\Z 3:003 08on
:8 .232 88.. .252
22.885: 08 5.2.2.58: :5 .2 0 .8..§..=32 :8 5.85.52 ..a .2 0 2.55»
50:80 :80: 5 2 58:80 :80: 5 m...
:8: 5 08

8.8.75 8855me 5058—08: 08580:... :8 0%: 8:85:05 803:0: 8:885 05 0: 50.8: 5558:8582
A50<mw5=035 5mm: 8058208: 08:80:: :8 2:...» 55.8.. :8 805.3585 608.0858: 0%: 8:85:05
05.085 0:505 885 0:: 5 m:00.:.:0 :82 .8980 :80: 5 2 :8 5 md m8... 8:05:80 $000.8: 55858

8.: <>OUZ< 50: 838: 8085885 5: 50:00: 8 8080: 8.50008 :8 8.8.7.2 0.8.8.: 0.5-553 .m 28,—.

41

33:. n £2 28 62.0.28 M 8m 29332»? n $3 8.2 :08 wﬁvgotsm than E can a :22? um: 95 888mm
.33 can Sow mo 32:83 wage “58:28 @333 95 go E: 596 V36 Euoom ESE 06 ﬂ Euoom.

.38 ES moom .8 32:83 waist :83 mug—9:3 95 .8 300.3 AREV 3828:,“ :38 05 a v35

.38 ES 80m .3 $0883, wﬁhﬁ :83 8383 95 mo 3E3 mEonamona :39 SEE 2: ﬂ ABA

65—80% 83 05 mo 5:2:qu 3.3on 2365
mos—S, @5885 .233 32%;ka \ 538:3 83 M “5m £258 05 89m town—=38 88a Engage—gov 3:085 05 a ”Em.

 

 

ONV QMN NdN ﬁn wk; ON— bhwm NV— ow; GNM wO—m cc ~3st
m. : vdc YN Nd VON ~.m— mmNm 32 no; :6 mhwm a0 05—03mm
W: ﬁwv o6 M.N ow— MAN _M.NN m: on; wmﬁ ccow v0 cemuutmm
wﬁm N.©— NH WM _0_ 9.9 w—Nm @2 ma; vow mm: 2: NA.Funffwbm
_.N.N WNM vﬁm _.m on: 0.»; NY—N _N_ Nb.— Mm.m mmcm —o 5.52
06— QVN ~.w eé oa_ N.N_ —©.w_ va~ vmé ow; hNX: 0v 98::
w._ 02 N.mo TN OON «A: :k; ~m~ N—N 55.0 9mm. no scanty—«I
~60 We N.m— M...” _m— m.©~ veg; we NNA mm; wwwm on 536—.
©.MN QwN m.mm m.m 52 m4.— mmgl mo vwé bod wmmc CA: tuxOBU “mam
0.x. WVN ﬂow ON 3; NA: hwgmﬁ m2 Fwd mod mmo: E. @8380 $03
~.NN Naom v.3 NA 2N _.o~ OWNﬁ ~o mvN omﬁ vowv mm 5:92.03
0; CNN N.mm o.N o: m.N_ de VN NM; moé ONmN ON 95m
_._ w.~_ Woo o.m :_ «.2 05K. VN NA 3mm _oom mm Eam
o. _ m o.w— 0.: _.m NOH ON— NYN 2 no; ~w.m Obmm Nw 520m
WON Nd— N.Nm v...” N: 0.: o~.— w E..— mo.m 02v 0v noxoouU
an“: AM.“— mnwed‘ ..Euoow «Mi 1:. mama—m—otwiv uni—3.33 .mnm Amhwmwﬂv ..WMHHWW Amouwwwoé 3:3
Gmd :32 8:5 35.5m

 

.muxﬂ x95 2 2.: .«o :38 ES— 28 3355 83>» .Aomd Hague—gov 80583 5556820 3293; .v 935,—.

42

 

 

m n n 2 2 M3 2 mm :2 8&3: ES
_ _ m N 3 on a : 3 Sam 5&5: 2:83.
mm x mm mm mm : 2 cm :22 5&5: 5:33..
E 3 2 8 3. 8 2 8 2.8 8&5: $3325
mm o m ow 2 ow a a 3.: 382:3 ﬁzz
n S a 8 Q E 3. mm :5: 8&3: 25::
h 3 mm 8 t 8 m a ﬁt 8&5: 525::
z o v a 8 a aw a 3.: 58283 .58..
S a w S ow S E av 2.: 8335 8:85 :5
mm ON a 8 a 3 M: 8 52 5&5: 3.86 .83
o M: 8 x mm : 2 cm 2.2 8355 52683
~ m 2 t N 2 _ 2 N3 380323 ES
2 n R S t 8 a S c: E8533 5.5.
2 2 M: a 3 8 E M3 N2 395233 :58
o ow mm 8 E a $ 2 a: 325283 8x85
3: E; 33 Q; 3.; 32 A5 4.x;
L0>¢U 59,90 LQ>OU LQ>OU LQ>OU LG>OU hu>cU ho>°U
23H MES—worm «GOMAQEH vengeansm own—0A— .:af— awn—on— :37— :3—
:285 .23.: .235 .22.: .33.: .235 3.; 3.; 35:02— 3.:30 3:3

 

.moom HonEoEom Ea ﬁsm=< macaw 860:8 803 8% :85 .bowoﬁo QSLQBQNE

b8 Sm m Ho N no 888 630 33382: 2:3 8% mo 088.880 6083 05 8:865 .358 SEQ .3568
33037.50 35:. £033 8 596 05 9 80% Sch no.8 2: mm 926% £ 28w 38:5 .33 2285 of E :5on
203 £53 £033 a $an BEES mo owﬂcoouom 05 mm “960 :83 38:: 380:? .83 83:0 2: 53 8383
$53 («0 082.880 :8ch 05 83065 630 Ema 30H .Eomoa 0.8? $53 :03? 8 3E3 338% ho owmﬁoeom
05 8 8&2 560 damEomE 585:8 Emoxm1u8m 2 2: mo 85:808an own—£883 03.38an .m «Bah

43

 

ad: ﬁmmﬁ Ewm— ad: 2% . Ewe . NS“ 32

 

 

$2 32 3: ma . m8 . 8 82
. 0.9: 32 58 . in . a: $2
. 32 32 . was . is 32
. $2 . wow . mow ooom
. . n: . a: 58
mam . a? 3m 88
. . 9? . 88
mo 8 8 8 8 3 mo 3 8
5 Q3. e on? m om< v 0% m ow< N 9% ﬂ w? :20
.30?

 

.8320 so» SS. 23 SS. soc
Ewes—n 8m mﬁwﬁx wouﬁaoﬁoiomn wow: Eco H 4523 gonna 55805.3 @004 omsmoom .mmﬂo EPA @6383

53, 8393 mo 3% a £515 .8 8598 we .89» wEudEoE 53, 320 So» a 55:5 853% own 332th a “a
36 33:58 2: 5:3 @8865 on 2:95 55805.3 Mood .330 30» Ba Avooméoomv “yuan—«30-8% 3 8%:
33m :m Soc Ewes—n Sm $605093 33988933 05 3 BEESoﬁ m3 mﬁwao— wowﬂzoﬁoiomm .e 22:.

44

 

Vwmv 2w? mdnm 0min mNVm md—m imam @ch m.mm~ YNNN 0.3: ”D: 03 mag ﬁmo 9mm 32

 

 

2.x? :3 was 3.2 9H 3.2 28 3% 32 32 «we 3: E: 3.2 n; 82
SM; . to: 3mm 3% 8m 3.: $8 $2 3.: 9o: 22 3o 68 82
. 3mm . Eon 0.2m 33 was 3”: $2 $2 in: 2m 3w M32
. wax . 3% «EN 28 3mm v.3; 32 ow 0.3 32
. 22 . «EN 32 :2 3: Q8 32 88
. £2 . $2 2: 23 M32 52
no: . woo 4.8 88
. . mg . 82
3 8 3 8 3 8 3 8 3 8 3 8 3 8 3 8
m a“? b 03 o a? m ou< v of. m ~me N on? _ ~u< $20
hao>

 

.8320 EPA SON 8 mom“ Soc was 6:95me— mo mama": cos—soﬁoiomn vow: Eco H 4523 9:3an cocoEocosm MBA
838m .320 89A @6883 53, 0:5me .«0 So» a £55, 8 8398 mo H8» $6855 55 $20 .89» a £532 moi—own own
332:8 a “a 36 388:8 2: :23 38065 3 2:03 cocoEococm Moog .320 .80» can cuooméoomv oSEmo¢oL8> an 3%:
32¢, 2m Soc 33 5208098— uou €052: £853»: 38938933 3 Bﬁgﬂov m3 was“: cos—so_mo-xomm H «Bah

45

Figure 1. Map illustrating the location of the six study lakes. The six study lakes are
located within the Huron River Watershed, southeast Michigan.

    
      

East Crooked ~ ~>'

 
 

Patterson

' Livingston Co.
Brulr‘ Halfmoon

Washtenaw Co

   
   

_)z

46

oxaq

coﬂmtmn. 5.52 .5955: 69.090 «mmm cc=m 53m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WV ”/1” _t
wwwm Wv:.

...... ,L ..,, .. $8
“aw. J
0
9
u
: a m
2.; c9; 687. :96 w a , $9. 5
H .. a
me; 26. .807. 2923 _ O
I,
as; :9. .82» 32 n. a m

_

1’
DE; 26. .306 26.- ” a. $00 m.l
:1 it .. :l ll ‘ -l : .lL s
u.
0
J
a m.-
.xom w

. $09

.tgoEonmuoE 83 E 838$“? 2 26 BSQEmG 3:96

Ho: mm 2:888 9:3 ”802 ATEonwE—EB—o _m.- n acmuozmmv qu 52E: 9 Atﬁo—awﬁzoac Emu u Eamv Qmu H832 80¢
@8585 8m 8wa £83 33m xmm 05 .«o 8:282? 05 mac—m 853$ 353: Mo 2832588 So.“ 2: mo mowﬁcoooom .N 255

47

ox~4

:88ch £82 :89.me 8x090 83

u ; ill i i Lilli!

EEm

‘
lrl l till til illt

-..°\°o

 
 
 

 

 

 

 
 

 

 

 

.‘ $8 d
_ 3
J
O
a
w.
«ll. ll .-ill .llJ _ 2
88.902: .T $8 6
v w 9
8595220 H O
w . I...
:95 359$. n A, ..O...
19
ll ill .i Ill .l e
,. £000 a“.
u.
0
J
W
m.
-. $8 a
. so?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ﬁgs 08 50ng 9 62v m8 832
80¢ @8885 En 8&3 .882 3:5 Em 2: mo 8:289? 05 mac.“ 8gb cosmommcoﬁ Echobmv 8E“ 05 mo 832888 .m charm

48

Figure 4. Relationship between lakeshore development (#dwellings/km) and coarse
woody material abundance (mean intersections per meter, a.), and submersed macrophyte
cover (cover score, b.), :tl SE, along the 0.5 m depth contour for the three shoreline
modiﬁcation types: undeveloped, maintained, and retaining wall. Note difference in y-
axes between coarse woody material abundance and submersed cover.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0'25 ' <> Undeveloped
a. qr —
g 02 . EL liaintained
E
3 -- A Retaining wall
g 0.15 - <>
.2
‘8'
g 0.1 -
2
.E
E 0.05 -
U
0 .

2.5 '-
2
o
o
2 2 -
o
>
o
o
E 1.5-
.c
a
2
o
a 1 -
U
o
8
“E’ 0.5-
.n
a
a)

o I I T j
5 1O 15 20 25

#dwellingslkm

49

Figure 5. Relationship between lakeshore development (#dwellings/km) and ﬂoating
macrophyte cover (cover score, a.), and emergent macrophyte cover (cover score, b.),
:tISE, along the 0.5 m depth contour for the three shoreline modiﬁcation types:
undeveloped, maintained, and retaining wall. Note difference in y-axes between floating
macrophyte and emergent macrophyte cover.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.5 -
2 3' <> Undeveloped <
8 >
g 2 ‘ D Maintained :
> - - - -
o
o A Retaining wall
m 1.5 d IIIIIIIIIIIIII
3. C]
.c
a.
2
3 1 -
E
a)
IE
7,: 0.5 -
2
u.

0 j
E
O
o
m
8
>
O
o
o
‘5.
.1:
a
2
o
w
E
E
o
E’
o
E
Ill
25

 

#dwellingslkm

50

Figure 6. Relationship between lakeshore development (#dwellings/km) and mean
substrate particle size, :tl SE, along the 0.5 m depth contour for the three shoreline
modiﬁcation types: undeveloped, maintained, and retaining wall.

 

 

 

 

 

 

 

O Undeveloped
[:1 Maintained
Sandy/_ — - - '-
Gravel A Retaining wall

0

.L“.

3 Sands

ﬂ

9!

H

in

g .

a» :22;

C

m

d)

E

Silt f . . <> 1
5 10 15 20 25
#Dwellingslkm
51

I—

Figure 7. Map illustrating the location of the ﬁfteen study lakes. The ﬁfteen study lakes
are located within the Huron River Watershed, southeast Michigan.

 

52

535335

 

 

 

 

 

 

 

 

 

on an 3 o
p p n so
0 Y 53“
1 $0?
0
. $00
ea . $8
«3. "5....
$6 ”in. r :8.
n .8... i
P b h *o
O
T (can
0
7 $3
0
. $8
m0.o ".5.“— . *3
$4... ”in o
2.3... "1 $32

 

 

.Aemtaoaaa 5:5 8% 35889: can: 38p 98 82? EucmtSm .83? 338 8a @8585 8.5 AB 850 382—53 .895 new .on 850
38880 ~89»: A5 850 urns—mobs: 83-20:? x<v 850 wage: 120:: do.“ qu 9 0809.2 uanobnE 838—055 .m «Suﬁ

 

d JaAoa peuamqns [moan %

.IaAoo eiﬂqdonnw %

 

 

 

 

 

 

 

 

 

 

 

 

5555535
c
.\.o
. $8
0 e . s3
1 $3
on “5.“.
23 "5.. £8
«2... ”1 .
u {000'
. $0
1 {.95.
' $3
0
y $90
«3. "3.".
m3... ".2... . . .s.
R... ”a.
r *2:

 

.IaAoo JuaBJauia “noun %

0'

Juan Sugieou [moan %

<

53

ASE. .: 833.8.qu

 

 

oom omN CON 00—. GOP on O
P h n P n h o
u or
o 00 o
o , cm
0
.8 w
m
o -9. m
w
I om m.
w
100 m.
o w.
o 000 my on (w‘
o o o o
O 1 ow
o
o a om
u 00?

 

ARE yucca—EU .580 waﬁoﬁn “a uni 8865 $2 .am so 353 Sun $83 355 2 of E voom
98 moom E 360:8 Ewes—n .«0 A88 .ﬁmﬁ: .88 3~m~so_ao-xoa£ 3?. sum 9 2522 3:5 E0882: 53on .352 d 953%

54

E5: 4... Bum-:uﬁoiuum

 

omv 00v 0mm com 9.“ CON on F 00 F cm 0
o o o o o
O O. O 00 0 ON
.. ., . o s M» .. . .. .
o o 3.» «a , .3 o. o «M088 no son... a
o o o o I. Q ,. I . . 0 ’00000 x 00 0, av
o I 0’. O 34‘ "a: o‘éﬁo» 0 vw. 0 no. 0 V
. 9 31.3.. . o. 3:... A} 4.. T8 m
o A. 3&9 RV i.1 u m.-
o 905 Q o 9‘0““ we 5 o B
o. I o no ,8 ..
O 00 O O 0‘. O
o a do 0 n W
O O O O” ‘0.“ .0 Q . N u-
o o 0 78* I.
. . .. x. m .u.
o o
A g m
o. o o ﬂow. w
0
o
o 09 M ..Uo
o )
Yo: w
0 * w

,, 08

ARE BEE—.30 .88@ 9:53: “a 86 2365 to. .3 so 358 «:5 £83 >95. 2 2t E voom can moom
E 880:8 33 5308092 mo 3:: .595. :38 ﬁnesse—8&8“: 36 GE 9 9538 955 808805 53on .352 .3 953%

55

AEEV ._._. umﬁﬁoﬁoiomm
8m 9.: 2: cm

b n b n

O
O
F

I
I

I
In
P

I
O
N

I
In
N

I
O
p)

 

  

mes. 93 :9:

'llll'l ’

3.2 8.. 52.3.2

I
ID
(0

 

I
O
1'

 

I (IA/mm) Iuaweuau! LﬂMOJE)

8%.. o3 26..

 

 

 

.25 353.985 05 an 38323“ 98 A_-Ev_.mw:_=o>6 ONAV v.83 “cacao—gov $3 28 6:: 353-023 05 3 858898
8.» ATExawﬁzokc omé 3 m8?— EoEQBgou 8:62: .2: 23m 05 ‘3 39:32.6“ 8m bah—.mwﬁzozé QVV m8?— EoEgBogw

33 .ﬁwco— =38 wouﬂsoﬁoioan mo 5on a ma Goon was Sow $89» 53va Ewes—n you EoEEoE 53on R552 .2 PSME

56

AEEV ._._. uﬁmioﬁoiumm

 

 

 

 

 

on» can . cmu 8m :3 2: on c

P T - p p n - ON 9
/ / m
/ w
. 3 u.
m.
T co m
w
/ m
33. am... =9... w:
.IIIIII VSFw
3.2 8.. 52.85. .M
1

m3“. 8.. 26.. r cu?

 

 

 

dc: nonmavéonm 05 3 360832 03 ATEu—awcéoav oNAv m8?— EoEQEgoc :wE can 6:: 3:8?023 05 3 360362

0.8 A -ExawE—EEV 8-2V new: EoEQBoBc 8:62: 6:: 23m 05 ‘3 328052 0.8 Maximum—326 mﬁvv mow—E “sumac—gov 33
b: 8.“ E0882: 53on 3:52 .2 «Suﬁ

.3 cﬂ :32 vegaoﬁoioaa mo 530:3 a 3 Zoom 98 wood E8» £26th mann— :3ch

57

LITERATURE CITED

58

LITERATURE CITED

Anderson, 0. 1984. Optimal foraging by largemouth bass in structured environments.
Ecology. 65(3): 851-861.

Annett, CA, J. Hunt, and ED. Dibble. 1996. The compleat bass: habitat requirements
for all stages of the life cycle of largemouth bass. Pages 306-314 In L.E. Miranda
and DR. DeVries, editors. Multidimensional approaches to reservoir ﬁsheries
management. American Fisheries Society, Symposium 16, Bethesda, Maryland.

Balon, E.K. 1975. Reproductive guilds of ﬁshes: a proposal and deﬁnition. Journal of the
Fisheries Research Board of Canada 32: 821-864.

Bain, M.B., and SE. Boltz. 1992. Effect of aquatic plant control on the microdistribution
and population characteristics of largemouth bass. Transactions of the American
Fisheries Society. 121 : 94-103.

Bettoli, P.W., M.J. Maceina, R.L. Noble, and R.K. Betsill. 1992. Piscivory in largemouth
bass as a function of aquatic vegetation abundance. North American Journal of
Fisheries Management. 12: 509-516.

Bowen, S.H., E.V. Lutz, and M0. Ahlgren. 1995. Dietary protein and energy as
determinants of food quality: Trophic strategies compared. Ecology 76: 899-
907.

Bryan, M.D., and D.L. Scamecchia. 1992. Species richness, composition, and abundance,
of ﬁsh larvae and juveniles inhabiting natural and developed shorelines of a
glacial Iowa lake. Environmental Biology of Fishes 35: 329-341.

Carlander, K.D. 1977. Handbook of freshwater ﬁshery biology. Volume 2. The Iowa
State University Press, Ames, Iowa.

Carlander, K.D. 1982. Standard intercepts for calculating lengths from scale
measurements for some centrarchid and percid ﬁshes. Transactions of the
American Fisheries Society 111:332-336.

Christensen, D.L., B.J. Herwig, D.E. Schindler, and SR. Carpenter. 1996. Impacts of
lakeshore residential development on coarse woody debris in north temperate
lakes. Ecological Applications. 6: 1143-1149.

Crowder, LB, and WE. Cooper. 1979. Structural complexity and ﬁsh-prey interactions
in ponds: a point of view. Pages 2-10 in D.L. Johnson and RA. Stein, editors.
Response of ﬁsh to habitat structure in standing water. North Central Division,
American Fisheries Society, Special Publication 6, Bethesda, Maryland, USA.

59

Cummins, K.W. 1962. An evaluation of some techniques for the collection and analysis
of benthic samples with special emphasis on lotic waters. American Midland
Naturalist 67: 477-504.

Cunjak, R.A., and G. Power. 1987. Cover use by stream-resident trout in winter: a ﬁeld
experiment. North American Journal of Fisheries Management 7: 539-544.

DeVries, DR, and RV. Frie. 1996. Determination of age and growth. Pages 483-512 In
B.R. Murphy and D.W. Willis, editors. American Fisheries Society,
Bethesda, Maryland, USA.

Evans, D.O., K.H. Nicholls, Y.C. Allen, and M.J. McMurtry. 1996. Historical land use,
phosphorus loading, and loss of ﬁsh habitat in Lake Simcoe, Canada. Canadian
Journal of Fisheries and Aquatic Sciences. 53: 194-218.

Francis, R.I.C.C. 1990. Back-calculation of ﬁsh length: a critical review. Journal of Fish
Biology. 36: 883-902.

Gotceitas, V., and P. Colgan. 1987. Selection between densities of artiﬁcial vegetation by
young bluegills avoiding predation. Transactions of the American Fisheries
Society. 116: 40-49.

Gotceitas, V., and P. Colgan. 1989. Predator foraging success and habitat complexity:
quantitative test of the threshold hypothesis. Oecologia 80: 158-166.

Gunn, J .M., and R. Sein. 2000. Effects of forestry roads on reproductive habitat and
exploitation of lake trout (Salvelinus namaycush) in three experimental lakes.
Canadian Journal of Fisheries and Aquatic Sciences 57: 97-104.

Gurnell, A.M., S.V. Gregory, and GE. Petts. 1995. The role of coarse woody debris in
forest aquatic habitats: implications of management. Aquatic Conservation:
Marine and Freshwater Ecosystems 5: 143-166.

Hatzenbeler, G.R., J .M. Kampa, M.J. Jennings, and BE. Emmons. 2004. A comparison
of ﬁsh and aquatic plant assemblages to assess ecological health of small
Wisconsin lakes. Lake and Reservoir Management 20: 211-218.

Hunt, J ., N. Bacheler, D. Wilson, and E. Videan. 2002. Enhancing largemouth bass
spawning: behavioral and habitat considerations. American Fisheries Society
Symposium. 31: 277-290.

Jennings, M.J., M.A. Bozek, G.R. Hatzenbeler, E.E. Emmons, and MD. Staggs. 1999.
Cumulative effects of incremental shoreline habitat modiﬁcation on ﬁsh
assemblages in north temperate lakes. North American Journal of Fisheries
Management 19: 18-27.

60

Jennings, M.J., E.E. Emmons, G.R. Hatzenbeler, C. Edwards, and M.A. Bozek. 2003. Is
littoral habitat affected by residential development and land use in watersheds of
Wisconsin lakes? Lake and Reservoir Management 19: 272-279.

Jennings, M., K. Johnson, and M. Staggs. 1996. Shoreline protection study: a report to
the Wisconsin state legislature. Wisconsin Department of Natural Resources,
Publication PUBL-RS-921-96, Madison.

Madsen, J .D. 1999. Point intercept and line intercept methods for aquatic plant
management. APCRP Tech. Rep. TN APCRP-Ml-OZ. US. Army Engineer
Research and Development Center, Vicksburg, MS. 16 pp.

McMahon, T.E., A.V. Zale, and DJ. Orth. 1996. Aquatic habitat measurements. Pages
83-120 In B.R. Murphy and D.W. Willis, editors. Fisheries Techniques. American
Fisheries Society, Bethesda, Maryland. .

Menzel, D.W., and N. Corwin. 1965. The measurement of total phosphorus in seawater
based on the liberation of organically bound fractions by persulfate oxidation.
Limnology and Oceanography. 10: 280-282.

Mittelbach, G.G. 1981. Foraging efﬁciency and body size: a study of optimal diet and
habitat use by bluegills. Ecology 62: 1370-1386.

Mittelbach, G.G. 1984. Predation and resource partition in two sunﬁshes (Centrarchidae)
Ecology 65: 499-513.

Murphy, J ., and LP. Riley. 1962. A modiﬁed single solution method for the
determination of phosphate in natural waters. Analytica Chimica Acta. 27: 31-36.

Nakamura, F., F]. Swanson, and SM. Wondzell. 2000. Disturbance regimes of stream
and riparian systems - a disturbance-cascade perspective. Hydrological Processes
14: 2849-2860.

Neumann, R.M., and TL. Wildman. 2002. Relationships between trout habitat use and
woody debris in two southern England streams. Ecology of Freshwater Fish
11: 240-250.

Olson, M.H. 1996. Predator-prey interactions in size-structured ﬁsh communities:
implications of prey growth. Oecologia 108: 757-763.

Omemik, J. 1976. The inﬂuence of land use on stream nutrient levels. Corvallis, Oregon,
US. Environmental Protection Agency Document: l-65.

Ostendorp, W., C. Iscli, M. Krauss, P. Krumscheid-Plankert, J .L. Moret, M. Rollier, and
F. Schanz. 1995. Lake shore deterioration, reed management and bank restoration
in some Central European lakes. Ecological Engineering 5: 51-75.

61

Phillips, EC, and RV. Kilambi. 1994. Use of coarse woody debris by Diptera in Ozark
streams, Arkansas. Journal of the North American Benthological Society 13:
151-159.

Pothoven, S.A., B. Vondracek, and D.L. Pereira. 1999. Effects of vegetation removal on
bluegill and largemouth bass in two Minnesota lakes. North American Journal of
Fisheries Management 19: 748-757.

Radomski, P., and T.J. Goeman. 2001. Consequences of human lakeshore development
on emergent and ﬂoating-leaf vegetation abundance. North American Journal of
Fisheries Management. 21: 41-46.

Roth, N.E., J .D. Allan, and D.L. Erickson. 1996. Landscape inﬂuences on stream biotic
integrity assessed at multiple spatial scales. Landscape Ecology 11: 141-156.

SAS Institute Inc. 1999. SAS user’s guide: statistics, version 8.0. SAS Institute, Cary,
North Carolina.

Savino, J .F., and RA Stein. 1982. Predator-prey interaction between largemouth bass
and bluegills as inﬂuenced by simulated, submersed vegetation. Transactions of
the American Fisheries Society. 111: 255-266.

Savino, J .R., and RA. Stein. 1989a. Behavior of ﬁsh predators and their prey: habitat
choice between open water and dense vegetation. Environmental Biology of
Fishes. 24: 287-293.

Savino, J .F., and RA. Stein. 1989b. Behavioral interactions between ﬁsh predators and
their prey: effects of plant density. Animal Behaviour 37: 311-321.

Scheuerell, M.D., and DE. Schindler. 2004. Changes in the spatial distribution of ﬁshes
in lakes along a residential development gradient. Ecosystems. 7: 98-106.

Schindler, D.E., S.I. Geib, and MR. Williams. 2000. Patterns of ﬁsh grth along a
residential development gradient in north temperate lakes. Ecosystems. 3: 229-
237.

Shuter, B. J., M. L. Jones, R. M. Korver, and N. P. Lester. 1998. A general, life history
based model for regional management of ﬁsh stocks: the inland lake trout

(Salvelinus namaycush) ﬁsheries of Ontario. Canadian Journal of Fisheries and
Aquatic Sciences 55: 2161-2177.

Siver, P.A., R.W. Canavan IV, C.K. Field, F.J. Marsicano, and A.M. Lott. 1996.

Historical changes in Connecticut lakes over a 55-year period. Journal of
Environmental Quality 25: 334-345.

62

Spence Cheruvelil, K. 2004. Examining lakes at multiple spatial scales: predicting ﬁsh
growth, macrophyte cover and lake physio-chemical variables. Ph.D. Dissertation,
Michigan State University.

Stemberger, RS, and J .M. Lazorchak. 1994. Zooplankton assemblage responses to
disturbance gradients. Canadian Journal of Fisheries and Aquatic Sciences
51 : 2435-2447.

Tomcko, C. M., and R. B. Pierce. 2001. The relationship of bluegill growth, lake
morphometry, and water quality in Minnesota. Transactions of the American
Fisheries Society 130: 317-321.

Wetzel, R.G. 2001. Limnology: Lake and river ecosystems (3rd edition) Academic Press,
San Diego, CA.

Wiley, M.J., R.W. Gorden, S.W. Waite, and T. Powless. 1984. The relationship between
aquatic macrophytes and sport ﬁsh production in Illinois ponds: a simple model.
North American Journal of Fisheries Management 4: 111-119.

Wills, T.C., M.T. Bremigan, and DB. Hayes. 2004. Variable effects of habitat

enhancement structures across species and habitats in Michigan reservoirs.
Transactions of the American Fisheries Society 133: 399-411.

63

   

ulililllilllﬂljljlﬂﬂl111113111