ohmi-

:
J

0 no: ’
, .

 

%
:i
l
'4
I
.

 

---. 3""
_ m.
_- a ' .-' -‘

7 ﬁud

1 V‘ ' “WW"
_. If n' ,' t . ' .3
15:33! ?g**'2§13§'
-9?!" ‘4; r '_ .

VI
1:", ..

ab,

. ~wﬁ
iii-1:;
23'

33:?“

'v‘." :y ‘ {‘1':
335;»: #59495;
V a £5 "

3 I

:2
”'3‘:
f x

a

a.
r

_ .
9.0.."

if»?

~. I

:(.;a

A -
Dv

- b

3:13;

$91;

' .‘
“( .-
. .'

v

’I
«I {‘"I‘IJ'

. "V.
.l'o’o

o

1 w:

|
.1“;
' ' 7'

.
\

. 3' ~:' 'V‘F
, ﬁ'21'1’€"‘=" '

.

, .

1* A
”I"
n"
1
a

..
1
,

3 ..

~..a_h .' .27 n.

As?"

. ~
~ .
'7 5'4"”!-
~
3',”
.oq‘n'n -

:—
d’w
3%}?
.

A "' V
‘ﬁgl'gitzﬁil:
- rmﬁm

"13. -'

‘wl .
“NJ '1
3 . .

 

IlllllllllllllllHllUHllllllllIIHIIIIIWIIHIIIHlIIHIHI

31293 017181060

This is to certify that the

thesis entitled

Chinook Salmon (Oncorhxnchus tshawytscha) Population
Dynamics in Lake Michigan, 1985 to 1996

 

presented by

Darren Matthew Benjamin

has been accepted towards fulﬁllment
of the requirements for

M.S. degree in Fish. & Wildl.

 

 

We. am

Major professor

Date April 28, 1998

 

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

LIBRARY

 

Unlverslty

Mlchlgan State

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINE return on or before date due.

 

 

 

 

MTEDUE J DATEDUE MTEDUE
F'B ~lnn
,- 15;) @f rtuul

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1198 alumna-961114

CHINOOK SALMON (Oncorhynchus tshawytscha) POPULATION
DYNAMICS IN LAKE MICHIGAN , 1985 TO 1996

By

Darren Matthew Benjamin

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

1998

ABSTRACT

CHINOOK SALMON (Oncorhynchus tshawytscha) POPULATION
DYNAMICS IN LAKE MICHIGAN, 1985 TO 1996

By

Darren Matthew Benjamin

Chinook salmon remains a popular sport ﬁsh in Lake Michigan despite the
collapse of the chinook ﬁshery beginning in 1987. This collapse is widely believed to
have been caused by a combination of nutritional stress and mortality from bacterial
kidney disease (BKD). This study involved a spatial analysis of trends in the chinook
ﬁshery, and a lake-wide stock assessment model for chinook sahnon. Fishery trend
analysis shows that declines in the ﬁshery were more severe in the western regions of the
lake than in the eastern regions. Evidence suggests that these spatial differences in
ﬁshery trends were due to changes in chinook spatial distribution rather than differences
in mortality.

I built a stock assessment model for chinook salmon in Lake Michigan using
1985-1996 recreational ﬁshery data in order to estimate population abundance, ﬁshing
mortality, and time-varying natural mortality. This statistical catch-at-age analysis model
is ﬁt to observed recreational ﬁshery and weir return data. Results indicate that the
natural mortality rate increased from 1986 to 1993, and declined from 1994 to 1996.
These results are inconsistent with estimates of BKD incidence from harvest weirs,
suggesting that causes of changing natural mortality are not entirely due to BKD. and

may be more complex than previously considered.

Copyright by
DARREN MATTHEW BENJAMIN
1998

 

ACKNOWLEDGMENTS

This work was made possible with funding provided by the Michigan Sea Grant
College Program. I would also like to recognize the cooperation by members from the
Lake Michigan Technical Committee of the Great Lakes Fishery Commission, who
readily provided the data to do this study. Additionally, Jerry Rakoczy and Roger
Lockwood of the Michigan Department of Natural Resources patiently worked with me
to understand the Great Lakes Creel Survey Program.

I would like to thank my committee members, Dr. James R. Bence, Dr. Daniel B.
Hayes, and Dr. Gary G. Mittelbach, for all of their helpful comments and suggestions
during the course of my program, in each of their classes, and for improving my thesis. I
especially thank Dr. Bence for his guidance, patience, and unselﬁsh commitment to his
students.

I also thank the faculty and graduate students in the Department of Fisheries and
Wildlife at Michigan State University for working to make the Department a challenging
and enjoyable place to learn. A special thanks goes to fellow graduate students Chris
Weeks, Mike Rutter, Shawn Sitar, and Ann Krause for their knowledge and insight, and
for joining me in our pursuit of research at alternate sites.

Finally, for her endless love and patience, I thank my wife, Kathy. With her

support, I know anything is possible.

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................ vii
LIST OF FIGURES ........................................................................................................ ix
CHAPTER ONE
INTRODUCTION .......................................................................................................... 1
Chinook Salmon Biology ............................................................................................ 3
Goals and Objectives .................................................................................................. 8
CHAPTER TWO
SPATIAL AND TEMPORAL CHANGES IN THE LAKE MICHIGAN
CHINOOK SALMON FISHERY, 1985-1996. ............................................................... 9
Introduction ................................................................................................................ 9
Methods ..................................................................................................................... 12
Stocking Data ........................................................................................................ 12
Monitoring of the Sport Fishery ............................................................................. 13
Creel survey data and estimates .......................................................................... 14
Charter report data ............................................................................................. 16
Harvest Ratio ..................................................................................................... 16
Lake Regions ......................................................................................................... 17
Results ....................................................................................................................... 19
Stocking History .................................................................................................... 19
Salmonine Fishery Lake-wide Trends .................................................................... 22
Chinook salmon ﬁshery lake-wide and regional trends ........................................... 26
Regional trends in salmonine effort for the sport ﬁshery ..................................... 26
Chinook salmon harvest ..................................................................................... 29
Chinook salmon targeted harvest rates as an index of abundance ........................ 32
Regional year class stocking, harvest, and harvest ratio (% return) ..................... 35
Discussion ................................................................................................................. 40
Conclusion ................................................................................................................. 49
CHAPTER THREE
LAKE-WIDE STOCK ASSESSMENT MODEL ........................................................... 51
Introduction ............................................................................................................... 5 1
Methods ..................................................................................................................... 53
Population Model ................................................................................................... 53

Abundance ......................................................................................................... 54

Natural Mortality ................................................................................................ 54
Fishing Mortality ................................................................................................ 56
Maturation (Spawning) Mortality ....................................................................... 57
Catch .................................................................................................................. 57
Effort ................................................................................................................. 58
Observed Data and Other Model Inputs .................................................................. 58
Recruitment ....................................................................................................... 58
Sport Fishery Information .................................................................................. 59
Weir Harvest Information ................................................................................... 60
Fitting the Model to Observed Data ........................................................................ 60
Results ....................................................................................................................... 62
Fishery Effort and Harvest ..................................................................................... 62
Age Compositions .................................................................................................. 66
Fishing Mortality ................................................................................................... 71
Maturation ............................................................................................................. 71
Natural Mortality ................................................................................................... 71
Total Mortality ....................................................................................................... 72
Population Abundance ........................................................................................... 73
Uncertainty of Parameter Estimates ........................................................................ 75
Discussion ................................................................................................................. 77
CHAPTER FOUR
CONCLUSIONS ........................................................................................................... 80
APPENDIX: ADDITIONAL TABLES ......................................................... 83
LIST OF REFERENCES ............................................................................................. 106

vi

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

LIST OF TABLES

Number of salmonine ﬁngerlings stocked in Lake Michigan, by species, 1963 to

1996. ............................................................................................................... 83
Number of salmonine yearlings stocked in Lake Michigan, by species, 1963 to
1996. ............................................................................................................... 84
Total number of salrnonines stocked in Lake Michigan, by species, from 1986-
1996. Includes fmgerlings, yearlings, and lake trout fry. ................................. 85
Number of chinook salmon ﬁngerlings stocked in Lake Michigan, by region,
from 1967 to 1996. .......................................................................................... 86

Lake Michigan total sport ﬁshery effort and targeted salmonine effort, total
salmonine harvest, and salmonine targeted harvest rate from 1986 to 1996.

Does not include stream ﬁshery. ...................................................................... 88
Salmonine harvest by the Lake Michigan sport ﬁshery, 1986 to 1996. Does not
include the stream fishery. ............................................................................... 89
Salmonine effort from the Lake Michigan sport ﬁshery, by region, from 1986 to
1996. Does not include stream ﬁshery. ........................................................... 90
Chinook salmon harvest by the Lake Michigan sport ﬁshery, 1986 to 1996.
Does not include stream ﬁshery. ...................................................................... 91

Chinook salmon annual targeted harvest rates for the Lake Michigan sport
ﬁshery, 1986 to 1996. Does not include stream ﬁshery. .................................. 92

Table 10. Estimated year-class harvest of chinook salmon for the Lake Michigan sport

fishery. Does not include stream ﬁshery. ........................................................ 93

Table 11. Estimated year-class harvest ratio (harvest per number stocked) of chinook

salmon for the Lake Michigan sport ﬁshery. Does not include stream ﬁshery. 93

Table 12. Comparison of model predicted vs. observed targeted effort and total chinook

salmon harvest. No lake-wide observed data are available prior to 1986. ........ 94

Table 13. Estimated annual ﬁshing mortality (P1:a ,y). ..................................................... 95
Table 14. Estimated annual instantaneous natural mortality rates. ................................. 96
Table 15. Estimated total annual mortality (A). ............................................................. 97

Table 16. Model estimated abundance-at-age. Age-0 abundance is equivalent to

recruitment. ..................................................................................................... 98

vii

Table 17. Standing stock biomass (pounds) as estimated from abundance-at-age from the
CAA model, and mean weight at annulus formation from CONNECT

(Rutherford 1997). .......................................................................................... 99
Table 18. Parameters and 95% conﬁdence intervals as estimated by the model. See

Methods for a discussion of confidence interval estimates. ............................ 100
Table 19. Estimated maturation and ﬁshery selectivity. Values were estimated by

logistic functions, with parameters estimated by the CAA model ................... 101
Table 20. Observed and predicted ﬁshery harvest age compositions ............................ 101
Table 21. Observed and predicted ﬁshery mature harvest age compositions. ............... 102
Table 22. Observed and predicted weir harvest age compositions. .............................. 103

Table 23. Parameter estimates from a sensitivity analysis on age-0 baseline natural
mortality. Age-0 natural mortality was increased by 25% from an intitial value
of 0.75 to 0.94. .............................................................................................. 104

Table 24. Parameter estimates from a sensitivity analysis on age-O baseline natural
mortality. Age-O natural mortality was decreased by 25% from an intitial value
of 0.75 to 0.56. .............................................................................................. 105

viii

Figure 1.
Figure 2.

Figure 3.

LIST OF FIGURES

Map of Lake Michigan divided into 7 regions: Green Bay, North, Northwest,
Northeast, Southwest, Southeast, and Illinois-Indiana. ................................... 18

Lake Michigan stocking levels for six species of salrnonines from 1963 to
1996. ............................................................................................................. 20

Total effort and salmonine effort (in millions of angler-hours), from the Lake
Michigan sport ﬁshery, 1986 to 1996. Other effort includes effort not directed
at salmonines, as well as effort reported by the charter ﬁshery. ...................... 22

Figure 4. Salmonine total harvest (in thousands of ﬁsh) and targeted harvest rate of

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

salrnonines from the Lake Michigan sport ﬁshery, 1986-1996. See Methods
for a description of targeted harvest rate ......................................................... 23

Lake-wide salmonine harvest (in millions of ﬁsh) by species for the Lake
Michigan sport ﬁshery, 1986-1996. (Chinook = chinook salmon, Coho = coho
salmon, Lake = lake trout, Rainbow = rainbow trout and steelhead, Brown =
brown trout) ................................................................................................... 25

Salmonine effort (in millions of angler-hours) from the Lake Michigan sport
ﬁshery, 1986 to 1996. Standard error bars are shown for regions within
Michigan’s waters only. See Figure 1 for a deﬁnition of lake regions. ........... 28

Chinook salmon harvest from the Lake Michigan sport ﬁshery, by lake region,
1986 to 1996. Standard error bars are shown for Michigan and Wisconsin
harvest only. See Figure 1 for a deﬁnition of the lake regions ........................ 31

Chinook salmon targeted harvest rates (targeted harvest per salmonine angler-
hour), by lake region, for the Lake Michigan sport ﬁshery, 1986-1996.
Standard error bars are shown only for Michigan (see Methods). See Figure l
for a deﬁnition of lake regions. ...................................................................... 34

Chinook salmon stocking and harvest, by year-class and region, for the Lake
Michigan sport ﬁshery. See Figure 1 for a deﬁnition of lake regions. ............ 38

Figure 10. Chinook salmon stocking and harvest ratio, by year-class and region, for the

Lake Michigan sport ﬁshery. See Figure 1 for a deﬁnition of lake regions. ...39

Figure 11. Relative alewife abundance from various regions of Lake Michigan, 1985-

1995. Data are from Great Lakes Science Center annual fall bottom trawl
surveys. Estimates are based on ﬁtting a general linear mixed model to these

data, including year and depth effects as well as port and year*port interactions
(Ann Krause, Michigan State University, unpublished results). ...................... 48

Figure 12. Observed and predicted values of sport ﬁshery effort in millions of angler-
hours (top) and chinook salmon harvest in thousands of ﬁsh (bottom) ............ 64

Figure 13. Loge-based residuals from model predictions of ﬁshery effort (top) and
chinook salmon harvest (bottom) for the Lake Michigan sport ﬁshery ............ 65

Figure 14. Standardized residuals for ﬁshery harvest age compositions of chinook
salmon in Lake Michigan. See text for calculation of standardized residuals. 68

Figure 15. Standardized residuals for weir harvest age compositions of chinook salmon
from Lake Michigan. See text for calculation of standardized residuals. ........ 69

Figure 16. Standardized residuals for ﬁshery mature harvest age compositions of
chinook salmon in Lake Michigan. See text for calculation of standardized
residuals ......................................................................................................... 70

Figure 17. Standing stock biomass (pounds) as estimated from abundance-at-age from
the CAA model, and mean weight at annulus formation from CONNECT
(Rutherford 1997). ......................................................................................... 75

Figure 18. Observed estimates of prevalence of Renibacterium salmoninarum, the
causative of agent of BKD, versus model-estimated time-varying instantaneous
natural mortality rate (TVM). Observed data was obtained from mature
chinook salmon sampled at Strawberry Creek, Sturgeon Bay, Wisconsin
(Marcquenski 1997). ...................................................................................... 79

CHAPTER ONE

INTRODUCTION

The chinook salmon (Oncorhynchus tshawytscha) has been an important top
predator in the Lake Michigan ﬁsh community for 30 years. In addition, their size and
fighting ability have made them a prized sport ﬁsh, and their presence in the lake has
helped to support a multi-million dollar ﬁshing industry. In the late 1980’s the future of
chinook salmon in the lake was in question when thousands of ﬁsh died from bacterial
kidney disease (BKD), and lakewide harvest declined.

Today, increases in chinook salmon salmon harvest may be an indication that the
population is beginning to show signs of recovery. Nevertheless, fundamental questions
remain regarding the future of ﬁshery management in Lake Michigan. Among these
questions is: “How can similar problems be avoided in the future?” This thesis
addresses the Lake Michigan chinook salmon ﬁshery from 1985 to 1996 in order to
understand the response of the ﬁshery to changes in the ﬁsh population. An analysis of
chinook salmon population dynamics using a stock assessment model provides an
additional understanding of trends in mortality and abundance.

The Lake Michigan ﬁsh community has changed dramatically in the past 60
years, and continues to change today (Eshenroder et al. 1995). Prior to the 1950’s, the

ﬁsh community consisted mostly of species endemic to the Great Lakes or its tributaries.

2
The inshore ﬁsh communities of Green Bay and other shallow embayments and large

river estuaries included lake sturgeon (A cipenserfulvescens), emerald shiner (Notropis
atherinoides), suckers (C atostomus sp.), yellow perch (Perca ﬂavescens), and walleye
(S tizostedion vitreum vitreum). The pelagic ﬁsh community included planktivorous lake
herring (Coregonus artedi) and six species of deepwater ciscoes (Coregonus sp.), two of
which had already suffered population declines prior to 1900 (Smith 1968). Lake trout
(Salvelinus namaycush) was the dominant piscivore of the pelagic community, and was
the major piscivore along with the burbot (Lora Iota) in the benthic community. Lake
trout and burbot preyed upon adult deepwater ciscoes and deepwater sculpin

(M yoxocephalus thompsoni) in the benthic community.

Commercial ﬁshing, habitat destruction, and invasions by exotic species disrupted
the ﬁsh community in the 1940’s and 1950’s (Smith 1972). Parasitism by the exotic sea
lamprey (Petromyzon marinas) contributed largely to the decline of lake trout and burbot.
The decline of these piscivores allowed the exotic alewife (A Iosa pseudoharengus) to
ﬂourish. The presence of alewife continued to disrupt the ﬁsh community, as alewife
preyed upon the larval ﬁsh of native species including: deepwater ciscoes, emerald
shiner, lake herring, yellow perch, deepwater sculpin, spoonhead sculpin (Cottus ricei),
burbot, and lake trout (Eek and Wells 1987; Crowder 1980; Eshenroder et al. 1995;
Krueger et al. 1995; Mason and Brandt 1996). By the 1960’s, the ﬁsh biomass in the lake
was dominated by alewife and another exotic species, rainbow smelt (Osmerus mordax).

The Great Lakes Fishery Commission began a sea lamprey control program in
1960, and by the late 1960’s successfully reduced the sea lamprey population in Lake

Michigan to 10-20% of its pre-1960 level (Holey et al. 1995). With sea lamprey numbers

3
reduced, and an overabundance of alewives, stocking programs for lake trout and Paciﬁc

salmon were initiated beginning in 1963. Five major salmonines have been stocked: lake
trout, chinook salmon, coho salmon, rainbow trout (steelhead), and brown trout. Brook
trout and splake (brook x lake hybrids) have been stocked in smaller quantities. Lake
trout were stocked with the goal of reestablishing self-sustaining populations. The other
salmonines were introduced as a biological control for alewives, and to provide a
recreational ﬁshery (Tody and Tanner 1966). The combination of predation by
salmonines and low recruitment effectively reduced the alewife population by the early
1980’s (Eek and Wells 1987; Jude and Tesar 1985). Increases in abundance of native
species including bloater, yellow perch, and deepwater sculpin, followed the decline of
alewives (Eshenroder et al. 1995).
Chinook Salmon Biology

Chinook salmon were ﬁrst successfully stocked in the Michigan waters of Lake
Michigan in 1967. Broodstock originated from Oregon and Washington, and after three
years of stocking, Michigan became self-sufﬁcient in the collection of eggs from Lake
Michigan chinook salmon. Michigan supplied eggs to Wisconsin from 1969 to 1971, and
to Illinois and Indiana beginning in 1970 (Keller et al. 1990). Illinois and Indiana
continue to stock chinook salmon from eggs collected at Michigan weirs.

Chinook salmon eggs are collected from mature ﬁsh returning to harvest weirs
from September through November. Fry are raised in hatcheries, and are stocked in the
following April and May as fmgerlings, just prior to reaching their critical smolt life

stage. Chinook salmon smolts undergo physiological changes that prepare them for life

4
in a large lake or ocean environment, and it is during the smolting process that chinook

salmon imprint on their natal streams.

Chinook salmon are semelparous, and generally complete their life cycle by
returning to the streams from which they were born (or stocked). Young-of-the-year
chinook salmon reside in the stream over the winter and early spring months. Natural
reproduction was observed as early as 1973, and is now signiﬁcant enough to warrant
management consideration (Rybicki 1973; Taube 1974; Elliott 1994). Improvements in
stream habitat, such as darn removal, run-of-the-river ﬂow regulation, and bank
stabilization, probably contributed signiﬁcantly to increases in natural reproduction in the
1980’s. Today the annual wild smolt production is estimated at 2.2 million ﬁsh
(Rutherford 1997), and the contribution of wild chinook salmon to the sport ﬁshery is
estimated to be 30% (Hesse 1994).

Natural reproduction is primarily limited to streams tributary to the Michigan
waters of Lake Michigan. Most of Wisconsin’s tributaries to Lake Michigan are blocked
by dams, lack suitable spawning substrate, and have large water level ﬂuctuations (Avery
1974). Illinois and Indiana streams also lack suitable spawning substrate.

Chinook salmon are relatively short-lived in Lake Michigan. Nearly all ﬁsh
mature by age-5, while some precocious males (jacks) mature at age-l. In the late 19803,
age-3 ﬁsh dominated the sport harvest and the weir harvest. Today, harvest is dominated
by 2-year olds. Chinook salmon growth rates may be density-dependent. Mean weight
and mean length increased for older ﬁsh after the collapse of the ﬁshery, suggesting that a

reduction in the population alleviated competition for food (Wesley 1996).

5
Within-lake estimates of mortality rates are not well documented for chinook in

Lake Michigan (Keller et al. 1990). Chinook salmon are faced with natural mortality
from different sources at different life stages. Probably the most critical life stage is from
age-0 to age-1, during which time chinook salmon are either stocked or spawned and
must make the journey to the lake. Survival of these young-of-the-year ﬁsh varies
spatially and temporally. Carl (1984) estimated a daily instantaneous mortality rate of
0.025 for wild juveniles in two Michigan tributary streams. Seelbaeh (1985) estimated a
planting to smolting mortality of 0-32% for hatchery chinook salmon in the Little
Manistee River. Variability in survival of young-of-the-year chinook salmon in Lake
Michigan is influenced by predation, water temperature, and date, location, and density of
stocked ﬁngerlings (Clark 1996).

Mortality from diseases and parasitism has been implicated as the major cause of
the decline in the chinook salmon population in the late 1980’s. Dead adult chinook
salmon began to wash up on the Lake Michigan shoreline beginning in the spring of
1986. Reports of dead ﬁsh on the beaches began in the southern end of the lake in the
spring and moved progressively northward during the season, as ﬁsh were carried by
strong South-to-North currents. The majority of sick ﬁsh examined had severe clinical
signs of bacterial kidney disease (BKD), which was considered the ﬁnal cause of death.
Because BKD occurs naturally in chinook salmon populations and does not necessarily
cause mortality, it is believed that some unknown stress weakened the ﬁsh and caused
BKD to become lethal. Many believe that this unknown stress was due to the decline in
alewife abundance, which may have caused nutritional stress and increased susceptibility

to disease (Marcquenski et al. 1997; Rybieki and Clapp 1996; Stewart and Ibarra 1991).

6
The number of visible deaths was estimated to be at least 10,000 ﬁsh in 1988, and at least

20,000 ﬁsh in 1989 (Nelson and Hnath 1990; Johnson and Hnath 1991). Nelson and
Hnath (1990) noted that the number of BKD-related deaths was small relative to the
ﬁshery harvest, and suggested that these deaths would not affect the ﬁshery. However,
an estimate does not exist of the number of BKD—related deaths that remained in the
bottom of the lake instead of washing up on the shore. The total number of BKD-related
deaths, including those that do not wash ashore, is likely to be much higher than
previously reported (Nelson and Hnath 1990).

Fishing mortality is not considered to be the major source of mortality in Lake
Michigan chinook salmon (Rutherford 1997). The sport ﬁshery grew rapidly through the
1970’s and early 1980’s, as the number of stocked salmonines increased and as ﬁshing
technology developed and anglers improved at targeting salmonines (Hansen et al. 1990).
Fishing effort for salmonines peaked in the mid- 1980’s before declining by 63% from
1986 to 1992. Salmonine effort has been relatively stable from 1992 to 1996. A more
detailed analysis of trends in the sport ﬁshery is presented in Chapter 2.

Mortality from commercial ﬁshery harvest of chinook salmon is low in
comparison to sport ﬁshery harvest. Commercial harvest is limited to tribal ﬁshery
harvest in the northern region of the lake and Grand Traverse Bay (Schorfhaar 1997;
Keller et al. 1990). From 1983 to 1987, annual tribal harvest of chinook salmon in the
Northern region of the lake ranged from 300 to 900 ﬁsh, or 4,000 to 10,000 pounds
(Keller et al. 1990). In contrast, estimated sport harvest of chinook salmon exceeded 8

million pounds annually over the same period (Francis 1996).

7
Mature chinook salmon that return to tributary streams are either harvested at the

weirs, harvested by the stream ﬁshery, or die naturally from spawning-related mortality.
Chinook salmon that return to streams with harvest weirs can be counted in order to
obtain an index of abundance of spawning ﬁsh. Data from Wisconsin weirs show a peak
in chinook salmon returns in 1987, followed by a sharp decline in 1989 (Eggold et al.
1997). This decline in chinook salmon returns from 1987 to 1989 corresponds with the
increase in natural mortality observed in the spring of 1988 and 1989, suggesting that
chinook salmon died before maturing and returning to the weirs.

Within-lake movements of chinook salmon have not been quantiﬁed. Patterns in
survey harvest rates suggest a seasonal northward movement from spring to summer
along the eastern shore (Jonas and Clapp 1998). Harvest rates also suggest an east to
west movement in the summer and back again in the fall (Elliott 1993). Winter
temperatures and forage abundance and distribution can inﬂuence these movement
patterns (Elliott 1993; Keller et al. 1990). Coded wire tagged (CWT) chinook salmon
recovered from the sport ﬁshery indicated movements both within-lake and between-
lakes. Chinook salmon planted in Michigan’s waters of Lake Michigan appeared in the
sport harvest in Wisconsin, Illinois, Indiana, and Michigan, as well as in Lake Huron.
However, CWT chinook salmon planted in Wisconsin’s waters were widely dispersed in
western Lake Michigan, but were not readily available to anglers in eastern Lake

Michigan (Lychwiek 1985; Keller et al. 1990).

8
Goals and Objectives

The goal of this project is to understand the population dynamics of chinook
salmon in Lake Michigan and how these dynamics affect the sport ﬁshery. It is intended
that this will serve as groundwork toward future modeling efforts that incorporate other
salmonine predators and their trophic interactions with forage ﬁshes in Lake Michigan
(e.g. Rutherford 1997 ; Koonce and Jones 1991). Additionally, information on
recruitment, harvest, and mortality, as derived from the stock assessment model, should
help improve the management of chinook salmon in Lake Michigan.

The objectives of this project were: (1) develop a lake-wide database of stocking
and harvest information for all salmonines in Lake Michigan, with the intention that other
salmonine species will be similarly analyzed in the future; (2) analyze trends in the
chinook salmon sport ﬁshery and make inferences about spatial patterns of abundance
and mortality; (3) build an age-structured stock assessment model speciﬁc to Lake
Michigan chinook salmon, drawing from existing catch-at-age models (e.g., Fournier and
Archibald 1982; Methot 1990); (4) evaluate temporal variations in recruitment and
mortality of chinook by ﬁtting the model to available data; (5) understand the limits of
parameter estimates as they are applied to the model; (6) evaluate the potential
importance of incorporating lakewide spatial variation of mortality into the model.

The following chapters are divided into three sections. Chapter 2 is an analysis of
trends in the salmonine sport ﬁshery, with a particular emphasis on chinook salmon.
Chapter 3 is a lake-wide stock assessment model I built for chinook salmon in order to
quantify mortality rates and abundance. The ﬁnal chapter is a discussion of a spatial

stock assessment model, implications for management, and directions for future research.

CHAPTER TWO

SPATIAL AND TEMPORAL CHANGES IN THE LAKE MICHIGAN
CHINOOK SALMON F lSHERY, 1985-1996.

Introduction

The present Lake Michigan ﬁsh community is complex and dynamic. The 1940s
and 1950s were periods of dramatic change, as native lake trout (Salvelinus namaycush)
and cisco (Coregonus sp.) populations either declined or became extinct due to invasions
by exotic species, commercial overﬁshing, and degraded spawning habitat (Wells and
McLain 1973). By the late 1950s, the ﬁsh community was of little economic or
recreational value. Successful management efforts to control exotic sea lamprey
(Petromyzon marinas), as well as the need to control overabundant alewives (Alosa
pseudoharengus), opened the door for the introduction of trout and Paciﬁc salmon in the
1960s. The introduction of salmonines served several purposes: to restore lake trout, to
control nuisance alewives, and to support a sport ﬁshery (Tody and Tanner 1966).

Lake Michigan’s modern salmonine stocking program began with the successful
introductions of rainbow trout (steelhead) (Oncorhynchus mykiss) in 1963. Lake trout
were re-introduced in 1965. Coho salmon (0. kisutch), brown trout (Salmo trutta), and
brook trout (S. fontinalis) were introduced in 1966, followed by chinook salmon (0.

tshawytscha) in 1967. Stocking of all salmonines increased from the 19608 to the 19803.

9

10
Stocking rates by some states increased more slowly or even declined in the mid-1980s

due to limits in hatchery production capacity and increased concerns about lake carrying
capacity (Keller et al. 1990; Kitchell and Crowder 1986; Stewart et al. 1981). Lake-wide
stocking of all salmonines has been relatively constant since the late 1980’s (Keller et al.
1990; Holey 1996).

The salmonine sport ﬁshery grew rapidly through the 1970s and 1980s; angler
effort increased by an order of magnitude, harvest rate doubled, and harvest increased 20-
fold in the Wisconsin waters of Lake Michigan (Hansen et al. 1990). Much of the ﬁshery
growth was driven by increases in annual stocking of salmonines. Of these salmonines,
chinook salmon was the most heavily stocked and was the most prized sportﬁsh because
of its size and ﬁghting ability. By the mid-1980s, Lake Michigan supported the most
spectacular sport ﬁshery in its history and contributed to an estimated $2 billion Great
Lakes ﬁshery (Keller et al. 1990).

As stocking levels continued to grow through the 1970s, biologists became
concerned that high levels of stocking would produce a predator-prey system in which
predator abundance would not be governed by prey dynamics, and leading to instability
(Stewart et al. 1981). Stewart et al. (1981) challenged Lake Michigan ﬁshery managers
to consider temporal ﬂuctuations in forage biomass and species composition when
determining stocking levels. Michigan created a plan to reduce forage consumption by
10% by reducing its overall stocking by 8.5% relative to the 1980-84 average, beginning
in 1985 and extending through 1990 (Keller et al. 1990). Wisconsin in turn planned to
reduce chinook salmon stocking rates by 10% in response to declines in the species’

condition and in alewife abundance (Hansen 1986; Keller et al. 1990).

1 1
In 1986 and 1987, dead chinook salmon were littering beaches along the

southeastern shoreline. By 1988, the number of visible dead chinook salmon was
estimated at 10,000 ﬁsh (Nelson and Hnath 1990; Johnson and Hnath 1991), and
increased to an estimated minimum of 20,000 in 1989. Clinical tests indicated that these
ﬁsh ultimately died from an infestation of Renibacterimn salmoninarum, a bacterium that
causes bacterial kidney disease (BKD) (Nelson and Hnath 1990). Because R.
salmoninarum is common even in healthy salmon, it is believed that some other
environmental stress weakened these ﬁsh to the point where BKD became lethal (Nelson
and Hnath 1990). It has been suggested that the additional stress is nutritional stress from
a reduced alewife population (Jones et al. 1993; Nelson and Hnath 1990; Rybieki and
Clapp 1996; Stewart and Ibarra 1991; Wesley 1996). Chinook salmon continue to die
from BKD today (Clark 1996), although the presence of dead chinook on the beaches has
declined (Marcquenski 1997). Increases in natural mortality of chinook salmon were
reﬂected in the sport ﬁshery, as harvest rates, harvest, and ﬁshery effort declined
beginning in 1987. By 1993, Lake Michigan chinook salmon harvest had severly
declined despite the maintenance of high stocking levels.

The purpose of this study is to describe more fully the spatial and temporal trends
in the Lake Michigan chinook salmon ﬁshery from 1986 to 1996, within the context of
the entire salmonine ﬁshery. A better understanding of the extent and location of harvest
declines, as well as a spatial understanding of ﬁshing mortality and chinook salmon

movements will aid in stocking decisions and in p0pulation modeling.

12

Methods
Stocking Data

Information on salmonine stocking was provided by Lake Michigan ﬁshery
management agencies. Illinois stocking data was provided by Rich Hess (Illinois
Department of Natural Resources). Indiana stocking data was compiled from Indiana
DNR stocking reports provided by J irn Francis (Indiana Department of Natural
Resources). Michigan stocking information from 1963 to 1978 was compiled from
summarized data provided by Bill McClay (Michigan Department of Natural Resources).
Stocking information from 1979 to 1996 was provided by Christine Larson through the
Fish Stocking Information System (Michigan Department of Natural Resources).
Wisconsin stocking information was compiled from Wisconsin DNR summary reports
(Hansen et al. 1991; Coshun 1991; Hansen 1988; Burzynski and Multhauf 1995;
Burzynski 1996).

Compiled lake-wide data was entered into a database and error-checked for
accuracy. The database was compared to an existing Lake Michigan stocking database
developed for the Great Lakes Fishery Commission (M. Holey, USFWS, personal
communication). The existing GLFC database was missing data for rainbow trout from
1963 to 1974, for brook trout from 1966 to 1975, and for brown trout from 1966 to 1974.
The GLFC database covered stocking of chinook salmon from 1967, and coho salmon
from 1966. For stocking years included in both databases, differences in stocking
numbers across databases were generally minor. For example, chinook salmon stocking

data differed in only 5 years between 1967 and 1988, and differences in those years were

13
less than 7%. Coho salmon stocking differed in 9 of the years between 1966 and 1988,

and most of those differences were less than 7% except for 1966 (60%) and 1985 (20%).
Discrepancies were most commonly due to double entry errors in the GLFC database,
while in other minor cases the GLFC database contained records of additional plants that
I could not account for. This second situation is not surprising since most of my stocking
numbers originated from summary reports and not raw data. Still, any errors in my
database would have originated within the summary reports themselves. The GLFC
database contained stocking records up through 1988 for all species except lake trout,
which contained records through 1992. My database contains records from 1963 to 1996
for all salmonines except lake trout. Results of this comparison and copies of the updated
database were presented to the Lake Michigan Technical Committee in 1996.
Monitoring of the Sport Fishery

Data and estimates on sport ﬁshery harvest, effort, and catch rates were carefully
reviewed. There are two primary sources for these data. The ﬁrst is from creel surveys
run by each of the states and the second source is from mandatory reports obtained from
charter operators. I begin by discussing data and estimates for the creel surveys run by
each of the state resource agencies. In Michigan, Illinois, and Wisconsin, these creel
surveys explicitly exclude the charter component (before 1990, Michigan’s charter
ﬁshery was covered as part of the creel survey). The charter trips are included as part of
the Indiana creel survey and these data are used to evaluate that component of the ﬁshery.
For the other states, information on the charter component of the ﬁshery comes from

mandatory charter reports.

14

C reel survey data and estimates

Annual creel surveys are conducted by each of the states surrounding Lake
Michigan in order to monitor the sport ﬁshery. Consistent estimates of total effort and
harvest are available from 1986 to the present. Wisconsin conducted a creel survey of the
salmonid ﬁshery in the Wisconsin waters of Lake Michigan from 1969 to 1985 (Hansen
et al. 1990), and began sampling the entire ﬁshery in 1986. Illinois began consistently
sampling its ﬁshery in 1986, though additional surveys were done in 1985, 1979, and
1969. Indiana has sampled its portion of Lake Michigan annually beginning in 1974,
though sampling methods have been consistent since 1986. Michigan began consistently
monitoring its Lake Michigan ﬁshery in 1985.

Austen et al. (1995) compare and contrast the creel survey methods from each of
the states. Creel surveys on Lake Michigan are generally conducted from April through
October, and ice ﬁsheries on Green Bay and Grand Traverse Bay are occasionally
sampled as well. Each survey approximates a two-stage sampling design, with sampled
days treated as the ﬁrst stage, and counts or interviews within days treated as the second
stage. Sampling is stratiﬁed by period (month or similar interval), day type (weekday or
weekend/holiday), area (port, site, or management area), and ﬁshing mode (boat, pier,
shore, stream, etc.). I grouped the boat ﬁshery to include data estimates from surveys of
launched boats, moored boats, and charter boats (see Charter report data). The shore
ﬁshery included surveys of shore, pier, and ice ﬁshery anglers. The stream ﬁshery was
not included in this study due to a lack of information for Michigan’s stream ﬁshery.

Fishing effort is estimated from interval counts at access sites or from

instantaneous counts of boats, pedestrian anglers, ice shanties, cars, or trailers. Average

15
daily counts are converted to a measure of ﬁshing effort (angler-hours), and ﬁshing effort

is estimated for a stratum by multiplying the average daily effort by the number of days
in the stratum. Harvest rates (harvest per angler-hour) are calculated from the angler
interviews within each stratum, and are multiplied by ﬁshing effort to estimate harvest for
a stratum. Summary harvest rates reported here are calculated by dividing the sum of the
annual harvest by the sum of the annual effort. Variances are available for Michigan and
Wisconsin’s surveys only. For Wisconsin, variances were provided by the WDNR for
total harvest by species and total ﬁshery effort. For Michigan, variances were calculated
for total harvest and targeted harvest by species, total effort, and salmonine effort.
Standard errors are reported here for Michigan and Wisconsin regions only.

Changes in harvest rates, or catch-per-unit-effort (CPUE), can be used to assess
trends in relative abundance. In spite of drawbacks in using CPUE as an index of
abundance (Malvestuto 1983), this is necessary in Lake Michigan since there is no lake-
wide ﬁshery-independent survey for salmonines. In Lake Michigan, the sport ﬁshery
effort is primarily directed towards salmonines and yellow perch. Targeted harvest rates
are used in this study to avoid bias due to changes in contribution of effort for yellow
perch or other species. Targeted effort is deﬁned as effort directed at the harvest of
salmonines. Targeted harvest is estimated from targeted effort, and targeted harvest rates
are calculated as the quotient of targeted harvest and targeted effort.

Summary information on the sport ﬁshery was provided by biologists ﬁ'om
Wisconsin, Illinois, and Indiana. Wisconsin data was provided by Brad Eggold
(Wisconsin Department of Natural Resources). Illinois creel data was obtained from

annual summary reports (e.g. Broﬂta and Marsden 1997), and additional data was

16
provided by Wayne Broﬂta (Illinois Natural History Survey). Illinois charter ﬁshery data

was provided by Rich Hess (Illinois Department of Natural Resources). Indiana data was
obtained from annual summary reports (Braun 1987; Palla 1997), and for recent years
was provided by Jim Francis (Indiana Department of Natural Resources).

Michigan’s creel survey estimates were recalculated from the raw data for this
study. Pre-existing methods utilized a mean-of-ratios catch rate estimator that is
inappropriate for Michigan access point angler surveys (Lockwood 1997). I recalculated
estimates for Michigan’s waters of Lake Michigan using a ratio-of-means catch rate
estimator and new variance estimators, as outlined in Lockwood et al. (in review).
Charter report data

Chatter ﬁshery information is generally obtained from harvest reports ﬁled by
licensed charter captains to their respective state. Wisconsin initiated a mandatory
reporting system in 1974, although because of early underreporting these reports were not
considered to be reliable until 1976 (Hansen et al. 1990). Illinois charter boat reporting
began in 1976. In Indiana, charter ﬁshery information is sampled in the creel. Charter
boat reporting for the Michigan waters began in 1990. Prior to 1990, the Michigan
charter ﬁshery information was sampled in the creel.

Harvest Ratio

Fundamental to management is knowing the percentage of stocked ﬁsh that are
harvested by the ﬁshery. This harvest ratio is deﬁned as the ratio of total number
harvested to the total number stocked (Hansen et al. 1990). To estimate a harvest ratio, I
calculated an annual harvest age composition for chinook salmon aged by the Michigan

creel survey. I applied these age compositions to the lake-wide harvest to estimate year-

17
class harvest for the 1985 to 1992 year classes. I assumed the Michigan age composition

applied to the lake-wide population, which seemed reasonable, based on similar length-
frequency data from the Wisconsin and Michigan creel surveys. In addition, coded wire
tag studies suggest that the chinook salmon population is highly mixed throughout the
lake (Bence et al. 1996). I included up to the 1992 year class because the 1993 year class
had not been completely harvested by the ﬁshery in 1996.
Law

I divided the lake into seven distinct regions for a spatial analysis of the chinook
salmon ﬁshery (Figure 1). These regions follow statistical district boundaries (Smith et
al. 1961), where aggregates of two or more statistical districts constitute a lake region.
The Green Bay region includes statistical districts WM-l and WM-2 from the Wisconsin
waters of Green Bay, and MM-l from the Michigan waters of Green Bay. The Northern
region includes Michigan statistical districts MM-2, MM-3, and MM-4 (Grand Traverse
Bay). The Northwestern region includes Wisconsin statistical districts WM-3 and WM-4
along the eastern shore of the Door Peninsula. The Northeastern region includes
Michigan statistical districts MM-S and MM-6. The Southwestern region includes
Wisconsin statistical districts WM-S and WM-6. The Southeastern region includes
Michigan statistical districts MM-7 and MM-8. The Illinois-Indiana region includes all

waters within Illinois and Indiana state boundaries.

 

NORTHWEST -------- - Arcana
m, NORTHEAST

     

 

- - -- ’ °° m
swine.”
8km
unto sum Pt.
........ “sued District Bournhty
oumwrssr1 — W” M"
W.
5m” 60
m3 3:3. """""""
............. - - Noland
means-I
ml:
ILLINOIS-
INDIANA
10 I 0 10 20 an “
Chicago E A
_m.£_ﬂ Statute mos

iLLmorsi mourn
1

Figure 1. Map of Lake Michigan divided into 7 regions: Green Bay, North,
Northwest, Northeast, Southwest, Southeast, and Illinois-Indiana.

19
Results

Stocking Histogy

Lake-wide stocking of salmonines in Lake Michigan has been presented
elsewhere (Keller et al. 1990; Holey 1996). I review the lake-wide stocking history of
salmonines in Lake Michigan here because past work has not always given information
on all life stages and all years stocked, and because there are inconsistencies in stocking
summaries derived from various sources. The procedure used to ensure that the stocking
summary presented here is as accurate as possible is described in the Methods. While the
trends presented are quite similar to other presentations, details do differ, and these
differences become more important when stocking is considered for sub-regions of the
lake.

Chinook salmon have historically dominated the stocking program They are
stocked almost entirely as spring ﬁngerlings. Stocking levels increased annually from
900,000 in 1967 to 6 million in 1980 (Figure 2). From 1980 to 1996, stocking ﬂuctuated
around 6 million fmgerlings with peak years in 1984 (7.7 million) and 1989 (7.9 million).

Signiﬁcant numbers of lake trout were stocked into Lake Michigan beginning
with 1.3 million yearlings in 1965, although 292,000 yearlings were stocked from 1959 to
1962 (Figure 2). Annual stocking of yearlings increased annually to 2.5 million in 1980.
Stocking of yearlings has ranged from 1.1 million to 2.8 million from 1981 to 1996.
Relatively few ﬁngerlings were stocked in comparison to yearlings in the 19608 and

1970s. Fingerling and fry stocking contributed from 10% to 63% of the total number

Brook Trout

I Yearlings E Fingedings

 

636669727578818487909396
Year

Chinook Salmon

 

636669727578818487909396

 

 

 

 

Year
Lake Trout
I Yeanings [:1 Fry E Fingerlings

6.0 }
A n
:3, 4.0 'ﬂ .
E,

3.0 -

2.0 -
a: 1.0 -

0.0 ‘

 

636669727578818487909396
Year

20

Stodted (millions)

3t

# Stodted (milliots)

# Stocked (millio

 

Brown Trout
2.5
2.0
1.5
1.0
0.5
0.0

636669727578818487909396
Year
Coho Sa|mon

 

 

 

 

Rainbow Trout (Steelhead)

I Yeaninos E Fingedings

 

636669727578818487909396
Year

Figure 2. Lake Michigan stocking levels for six species of salmonines from 1963 to

1996.

21
stocked in the 19803, and fry have not been stocked since 1987. Total stocking peaked at

5.4 million ﬁsh in 1989.

Coho salmon stocking began in 1966 and exceeded 3 million by 1969 (Figure 2).
From 1969 to 1996, coho stocking declined from 3.5 million to 2.5 million with a peak
year in 1979 (4.4 million). Coho have usually been stocked as yearlings, although the
proportion of ﬁngerlings stocked has increased from 1987 to 1996.

Annual rainbow trout (steelhead) stocking levels increased annually from 1963 to
1973, reaching a peak in 1973 at 3 million ﬁsh (Figure 2). Stocking ﬂuctuated between
1.2 million and 3.2 million ﬁsh from 1974 to 1984. From 1985 to 1996, stocking has
been relatively consistent, ranging from 1.5 to 2 million ﬁsh. The stocking ratio of
yearlings to ﬁngerlings was roughly 1:1 from 1970 to 1984. Since 1985, stocking has
been composed of roughly 75% yearlings.

Brown trout stocking began in 1966 and increased to 2 million ﬁsh by 1973
(Figure 2). Stocking ﬂuctuated between 500,000 and 1.5 million ﬁsh from 1974 to 1981.
From 1982 to the present, annual stocking levels have ﬂuctuated between 1.5 and 2
million ﬁsh. Brown trout are stocked both as ﬁngerlings and as yearlings.

Brook trout are stocked primarily in Wisconsin waters as both ﬁngerlings and as
yearlings, although they were occasionally stocked in Michigan waters until 1990, and in
Illinois until 1980. Fewer than 100,000 brook trout were stocked annually from 1966 to
1976 (Figure 2). In 1977, an additional 500,000 ﬁngerlings were stocked in Wisconsin
for a total of 623,000 - the most of any year. Stocking levels ﬂuctuated between 200,000
and 300,000 from 1978 to 1986, and between 100,000 and 500,000 from 1987 to 1989.

Stocking levels have been declining from 1990 to 1996.

22
Salmonine Fishery Lake-wide Trends
Sport ﬁshery effort in Lake Michigan declined from 1986 to 1996. Total effort

declined by 54%, from 14.1 million angler-hours in 1986 to 6.5 million angler-hours in
1996 (Figure 3). Salmonine effort comprised 61% of the total ﬁshery effort in 1986, but
fell to 41% in 1992 before returning to 49% in 1996. Salmonine effort declined by 63%
from 8.6 million angler-hours in 1986 to 3.2 million angler-hours in 1992. Salmonine

effort has been stable at 3.2 million angler-hours from 1992 to 1996.

 

II Salmonine Effort e Other Effort

 

16.0

14.0 -

 

12.0 -

 

 

10.0 -

.. 1,1 “l ‘l
l i: ‘llllzl‘l ll

 

 

 

 

8.0 —

 

 

 

6.0 -

l“ (Hi-‘31“ “
1ll ,mlj‘ml l' ”‘lll‘llll‘

11% ll Wm
llillll‘il l‘llllill‘illm

 

 

 

 

 

l‘llllll lll‘ iii

 

”ill lli‘lillll M:

4.0 -

Effort (millions of angler-hours)

2.0 ~

 

0.0 ~
86 87 88 89 90 91 92 93 94 95 96

Year

Figure 3. Total effort and salmonine effort (in millions of angler-hours), from the
Lake Michigan sport ﬁshery, 1986 to 1996. Other effort includes effort not directed
at salmonines, as well as effort reported by the charter ﬁshery.

23
Harvest of salmonines declined by 53% from 1.8 million salmonines in 1986 to

855,000 in 1990 (Figure 4). From 1990 to 1996, salmonine harvest has been relatively
consistent at 800,000, with a low harvest occurring in 1992 at 746,000 salmonines.
Targeted salmonine harvest rate (targeted salmonine harvest per salmonine angler-hour)
ﬂuctuated between 0.12 and 0.16 ﬁ'om 1986 to 1990 (Figure 4). Harvest rate has been

increasing ﬁom 1990 to a period high of 0.19 in 1996.

 

[m Harvest —Targeted Harvest Rate

 

 

0.25

g

d

'00

8
9.

1.500 .. J- 0-20

- -‘-'- A :-:-:-:- .................
.................................

. ' ‘ ' ' ' ‘ 3:5- .....................

. . . ._ ........

................................................

.j. ..........................................
...............................................................
' ' ' ' e'e'n I‘I ;‘_-_-:-“~ Iz.‘-_. ......................................
.'¢‘-'.'.' :‘(ﬁ-‘u‘ 3934's u‘fﬂv‘. ‘-:-'.:. ...........
.-. . .;.:.-.;.

........
..............
._. a. ..................
. . . . ' '3'? i. x.
................................
.............................
...........
.........
..............
.-.‘.-. ............
. . ~.-.-.‘.- . . . q . . . .
........................
.......
.....

.......
..........
.....
.............
.........
..................
................

...-
.....

Salmonine Harvest (thousands)
S
O
O

................
aux} .i- .......................
.... n... . n... ..t.. 3.... ~.\ ...........
.........................................
......................................
......................................
...............................................

.........
......
..............

...............................

3‘. . o.‘ . e n o q... . . - - . .3. .........................

xi f; Dix", //
/ xix ; 57x 3 xx ,ﬂ’
935650
r .
i f I
S (r v ff. ’5.
we:
3 2: x
‘9’. x it
1 4
I ﬂ
0 o
8 d
0
Targeted Harvest Rate
(harvest per angler-hour)

 

\v ‘ *- -:-:‘:~:-. 1 5.3;. ...............
..........................................

 

 

Figure 4. Salmonine total harvest (in thousands of ﬁsh) and targeted harvest rate of
salmonines from the Lake Michigan sport ﬁshery, 1986-1996. See Methods for a
description of targeted harvest rate.

24
Lake-wide harvest of coho salmon peaked in 1989 at 407,000 and declined to

155,000 in 1991. Harvest levels from 1992 to 1996 ranged ﬁ'om 181,000 to 295,000 with
an average of 237,000. Contribution of coho salmon harvest to the total salmonine
harvest increased from 18% in 1986 to 35% in 1993.

Lake trout are generally not preferred by anglers, but are relied upon when ﬁshing
for other salmonines is poor (Lange et al. 1995). Lake trout harvest comprised 13% to
27% of the salmonine harvest between 1986 and 1996. Peak harvest was in 1989 at
347,000 while 1996 was the lowest harvest year at 115,000.

Rainbow trout harvest was limited to a stream ﬁshery in the mid- to late-19803.

In 1986, the lake harvest of rainbow trout was 68,000, less than 5% of the salmonine
harvest. An offshore ﬁshery developed for rainbow trout as anglers learned to target
rainbow trout along surface temperature breaks, and harvest increased to a peak of
172,000 in 1993, comprising 20% of the salmonine harvest. Harvest declined to 142,000
in 1996.

Brown trout harvest has accounted for 6% to 13% of salmonine harvest from
1986 to 1996. Harvest declined from 171,000 in 1986 to 73,000 in 1988. Harvest has
ﬂuctuated between 63,000 and 110,000 from 1987 to 1996. Most brown trout are
harvested in Wisconsin and Michigan.

Brook trout harvest has historically accounted for less than 1% of lake-wide
salmonine harvest. Most brook trout harvest is concentrated in Wisconsin, although a
small fraction is harvested in Michigan. Harvest has ranged from 500 to 6,000 from 1986

to 1996.

25

 

[- Chinook a Coho [:1 Lake at Rainbow [:1 Brown]

 

N
o

 

 

_L
m
1

1.6~

1.41'

 

 

_L
N
1

llllllllllilliillii

on .1. 5
V i

 

 

  

 

1.0+

 

 

 

.6 .c
0) CD
1 L

1
ll! 3‘ If
the :1
‘._—.—'T
-.—

Harvested (millions of fish)

.6 .o
O N
l l

lllllllllilllllllli I :z: I

9
.h
l

I
l

 

 

 

cc
co
co
co
on
(D
(D
O
(D
d
0
N
(D
0)
(D
.A
CD
01

Figure 5. Lake-wide salmonine harvest (in millions of fish) by species for the Lake
Michigan sport ﬁshery, 1986-1996. (Chinook = chinook salmon, Coho = coho
salmon, Lake = lake trout, Rainbow = rainbow trout and steelhead, Brown = brown
trout)

26
Chinook salmon ﬁshery lake-wide and regional trends

In the subsections that follow I present detailed information on temporal trends in
harvest, effort, and catch rate for chinook salmon for each region of Lake Michigan.
These detailed results show the following general patterns. First, there were substantial
differences in how the overall collapse of the chinook salmon ﬁshery unfolded. Most
notably, the decline was greater on the eastern (Michigan) shore in comparison with the
western (Wisconsin) shore, with the greatest decline in the southeast. This generalization
applied to harvest (Figure 7), harvest rate (Figure 8), and harvest ratio (percent of a
stocked year class harvested (Figure 10), but less so to the amount of ﬁshing effort on
salmonines (Figure 6). These results suggest some spatial changes over time in either the
distribution or survival of chinook salmon. I return to this topic in the Discussion. In
addition, comparisons of stocking, harvest, and harvest ratio provided no evidence that
regional in-lake harvest was closely tied to regional stocking numbers.

Regional trends in salmonine eﬁ'ort for the sport ﬁshery

Trends in salmonine effort for each region of Lake Michigan generally followed a
lake-wide trend of declining effort in the late 1980’s, followed by a consistently low level
of effort in the early to mid- 1990’s. The major differences between regions were the
years in which the declines actually began and ended, and the overall extent of the
declines (Figure 6).

Salmonine effort in Green Bay increased from 1986 to 1988 before declining by
69% from 384,000 angler-hours in 1988 to 119,000 angler-hours in 1996. Effort was
relatively consistent from 1992 to 1996. In comparison to the lake-wide salmonine

effort, effort in Green Bay contributed 3-7% of the lake-wide total from 1986 to 1996.

27
Salmonine effort in the Northern region was at a period-low level in 1986 in

contrast to a period-high level lake-wide. Effort peaked in 1987 at 245,000 (i 20,000)
angler-hours and ﬂuctuated between 157,000 (1 16,000) and 194,000 (i 17,000) angler-
hours from 1988 to 1993. Effort has been declining from 194,000 (i 17,000) in 1991 to
119,000 (i 6,900) in 1996. Similar to Green Bay, effort in the northern region
contributed between 3% and 6% of the lake-wide effort from 1987 to 1996.

Salmonine effort in the Northwest region declined by 55% from 984,000 angler-
hours in 1986 to 439,000 angler-hours in 1990, similar to the lake-wide rate of decline of
53% over the same years. Effort from 1990 to 1996 has been relatively stable at 400,000
to 450,000 angler-hours. Effort in the northwest comprised 9-14% of the lake-wide total
from 1986 to 1996.

The high rate of decline of salmonine effort in the Northeast region was second
only to the Southeast, declining by 69% from 1.6 million (1 172,000) angler-hours in
1986 to 510,000 (i 33,000) angler-hours in 1992. Period-low salmonine effort occurred
in 1995 at 424,000 (i 34,000) angler-hours. Effort in the northeast comprised 13-23% of
the lake-wide total from 1986 to 1996.

Salmonine effort in the Southwest did not begin to decline until 1988, and
declined by 62% from 1.4 million angler-hours in 1987 to 542,000 angler-hours in 1990
(Figure 6). Effort remained low from 1990 to 1993, and increased slightly to 620,000
angler-hours from 1994 to 1996. Effort in the southwest comprised 13-20% of the lake-

wide total from 1986 to 1996.

28

   

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Green Bay North
g 05 t 0.3
° 0.3
m 0.4 E 02
o 0.3 .
.s 3 0.2
5 0‘2 5 0.1
5 0.1 g 0.1
0.0 0.0
8687888990919293949596 8687888990919293949596
Year Year
Northwest Northeast
12
g 1.0 g
0.8
.3 06 .2
04 5
02 5 is; .2;
00 0.0 I I ID: I :2.- r ._-. I '24:
8687888990919293949596 8687888990919293949596
Year Year
Southwest Southeast
t 1.5 § 4.0
8
LU m 3.0
1.0 « m
.3 .5 2.0
J
2 0.5 »: ;:; if. :_~., :3“ 7:. “3.: E 1'0
a 0.0 Fifi I :1. I 2.: I I]; r if I It; I :3}:- r ’1' r": a 0.0
8687888990919293949596 8687888990919293949596
Year Year
Illinois-Indiana Lake-wide
E 2.0 1: 10.0
m 1.5 - 5,. g 8.0
.2 1.0 1 g 6'0
8 323% 213’ , 8 4.0
0.5 « iééa cf§§—-§r E
:2“; E3: as; .. 8 2”
0.0 H I i' l H I l l I". T'“ T I I 0.0
8687888990919293949596 8687888990919293949596

Year Year

Figure 6. Salmonine effort (in millions of angler-hours) from the Lake Michigan
sport fishery, 1986 to 1996. Standard error bars are shown for regions within
Michigan’s waters only. See Figure 1 for a definition of lake regions.

29
The greatest declines in Lake Michigan salmonine effort occurred in the southeast

region. Salmonine effort declined by 77% from a peak of 2.75 million (i 241,000)
angler-hours in 1986 to a period low level of 621,000 ( i 38,000) angler-hours in 1992.
Effort has remained below 715,000 angler-hours from 1992 to 1996 (Figure 6). Effort in
the southeast once comprised 32% of the lake-wide total in 1986, but declined to 19% by
1993. Salmonine effort in the Illinois-Indiana region is large relative to its lake area, due
primarily to the high human population density along almost it’s entire shoreline. In
1992, 1994, and 1996, this region reported more salmonine effort than any other region in
the lake (Figure 6). Salmonine effort declined by 55% from 1.6 million angler-hours in
1986 to 706,000 angler-hours in 1990. Effort has been relatively stable from 1990 to
1996 at 682,000 to 782,000 angler-hours. The relative contribution of effort in the
Illinois-Indiana region to the lake-wide total increased from 15-18% from 1986 to 1991,
to 21-24% from 1992 to 1996.
Chinook salmon harvest

The decline in salmonine harvest from 1986 to 1988 was driven by declines in
chinook salmon harvest (Figure 5). In 1986, chinook salmon harvest comprised more
than 50% of the total salmonine harvest. By 1993, chinook salmon comprised only 16%
of the salmonine harvest. Lake-wide chinook salmon harvest declined by 86% from
950,000 in 1986 to 132,000 in 1993 (Figure 7). Harvest increased from 226,000 in 1994
to 304,000 in 1996, but remains less than one-third of the peak harvest in 1986. Trends
in chinook salmon harvest differ across regions and do not follow a general lake-wide
trend. In general, harvest declines were greater in the eastern regions of the lake than in

the western regions.

3O
Chinook salmon harvest in Green Bay increased from 27,000 in 1986 to 42,000 in

1989, while the lake-wide harvest declined over the same period. Harvest declined by
46% from 1989 to 1990, and period-low harvest of 6,000 occurred in 1993 for an overall
decline of 86% from 1989 to 1993. Harvest in Green Bay was 3% of the lake-wide total
in 1986, increased to 12% in 1989, and has ﬂuctuated between 3 and 10% from 1990 to
1996.

Only a small fraction of the lake-wide salmonine ﬁshery is contained in the
Northern region of the lake, probably because the region is less densely populated, fewer
salmonines are stocked there, and because tribal ﬁsheries and lake trout refuges limit
sport ﬁshing effort. Relatively few chinook salmon are harvested in the northern waters
of Lake Michigan. From 1986 to 1996, harvest in the northern waters contributed 1-6%
of the lake-wide harvest. Still, declines in harvest generally followed the lake-wide trend,
with a peak in harvest of 23,000 (i 3,700) in 1987 and a low harvest of 2,400 (i 300) in
1994 for an overall decline of 90% (Figure 7).

Chinook salmon harvest in the northwest region contributed 11% to 32% of the
lake-wide harvest from 1986-1996 (Figure 7). Peak harvests were 102,000 in 1986 and
113,000 in 1987. Harvest from 1988 to 1996 was lower than previous years, with
additional peak years in 1989 and 1996. Period-low harvest occurred at 42,000 in 1992
for an overall decline of 63% from 1987 to 1992. Harvest has been increasing annually

from 1993 to 1996.

Goon Boy

 

oamuomumdu)
o 3 3 8 8 8

587888990919293949596

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

You
Northwest
guo
120
i
2100 .3 ‘
c. ”-19”. '
r sow-e ‘
‘o-n-l'h—t; - .— ' ’_‘.
201;; a w i I, »'~,_—, 1 ;,_3
I o ’ '7' VAVHTIYH “11"?
o:
8687888990919293949596
You
Somhweot
140
1204
100~
é eo-
00‘ f;
E ‘0‘ :' T .1
20+ ?: l—(p—n
I o r’ 1,,rf'" IHI‘. U 1’1 T7
o
8687888990919293949596
You
Illinois-Indiana
80
50
£40
30
20
10
0
o

8687888990919293949596
You

31

 

 

North
30
25
220
$15
am
55
i 0
8687888990919293949596
You
Northont
gtoo
300
.3.
9200
£100
13 0
eeetaeeeeoeiezeaeeesee
You
Burrito-t

 

#Honrootodhhouoncb)
9 § “2’ § §§

8687888990919293949596
You

Ldto-w'ido

d

o§§§§§

 

tr Harvested (thaw)

8687888990919293949596
You

Figure 7. Chinook salmon harvest from the Lake Michigan sport ﬁshery, by lake
region, 1986 to 1996. Standard error bars are shown for Michigan and Wisconsin
harvest only. See Figure 1 for a deﬁnition of the lake regions.

32
Second only to the southeast, chinook salmon harvest in the northeast region

declined more than any other region. Harvest in the northeast region peaked in 1986 at
304,000 (i 46,000) and declined by 95% from 1986 to 15,000 (i 1,100) in 1994 (Figure
7). Harvest increased to 70,000 in 1996, the highest level of harvest since 1988.

Chinook salmon harvest in the southwest region increased from 115,000 in 1986
to 128,000 in 1987 before declining by 53% in 1988 (Figure 7). Harvest continued to
decline to a period low of 24,000 in 1993 — an overall decline of 81% between 1987 and
1993. Harvest increased to 75,000 in 1995 and 1996.

Chinook salmon harvest in the southeast region declined by 63% from a peak of
348,000 (i 41,000) in 1986 to 129,000 (i 19,000) in 1987. Harvest continued to decline
to a low of 14,000 (i 1,400) in 1992 — an overall decline of 96% from 1986 to 1992.
Harvest increased from 16,000 (i 1,600) in 1993 to 40,000 (i 3,100) in 1996 (Figure 7).
Along with declines in salmonine effort, chinook salmon harvest declined more in the
southeast than any other region of the lake.

Chinook salmon harvest from the Illinois and Indiana waters followed a decline
similar to the lake-wide trend from 1986 to 1994 (Figure 7). Peak harvest occurred in
1986 at 49,000 and declined by 85% to 7,000 in 1994. Harvest has increased ﬁom 1994
to 1996. From 1986 to 1991, contribution of harvest from the Illinois and Indiana waters
to the lake-wide harvest increased from 5% to 10% before declining to 5% again in 1996.
Chinook salmon targeted harvest rates as an index of abundance

Lake-wide targeted harvest rates (targeted chinook salmon harvest per salmonine
angler-hour) of chinook salmon suggest that relative abundance declined from 1986 to

1993, and increased from 1994 to 1996 (Figure 8). Harvest rate declined concurrently

33
with declines in harvest, from 0.087 in 1986 to 0.027 in 1993, and increased to 0.064 by

1996. There were regional differences in harvest rate trends, namely, declines occurred
in the north and eastern regions of the lake, while declines in the western regions were
not as severe and in some cases harvest rate actually increased. Because much of the
ﬁshery is concentrated in the eastern regions, these regions had the most inﬂuence on the
lake-wide harvest rate trend. Regional differences in harvest rate trends suggest a change
in the spatial distribution of chinook salmon rather than simply a decline in lake-wide
abundance.

Targeted harvest rates in Green Bay ranged from 0.030 to 0.083 ﬁ'om 1986 to
1996, but do not show a declining trend, as peak rates occurred in 1989, 1991, and 1995
(Figure 8). Harvest rates in the northern region declined by 82% from 0.086 (i 0.017) in
1987 to 0.016 (i 0.002) in 1994. Harvest rates in the northwest showed the largest
decline from 1987 to 1988 but have ﬂuctuated between 0.046 and 0.075 from 1989 to
1995. Harvest rates increased from 1992 to 1996, with a period-high harvest rate of 0.11
in 1996. Harvest rates in the northeast similarly declined from 0.13 (i 0.024) in 1986 to
0.024 (1 0.003) in 1994. Harvest rates increased in 1995 and 1996, surpassing the 1987
level. In the southwest region, harvest rates peaked at 0.061 in 1987 before declining to a
low of 0.020 in 1994 - a 70% decline. However, by 1995, harvest rates returned to 1986-
1987 levels. Harvest rate in the southeast declined by 82% from 0.12 (1' 0.018) in 1986
to 0.021 (:t 0.003) in 1992. By 1996, harvest rate had returned to the 1987 level of 0.063
(i 0.006). In the Illinois-Indiana region, harvest rates declined from 0.031 in 1986 to

0.010 in 1994 before increasing to 0.024 in 1996.

34

   

   

   

   

Green Bay North
0.16 0.16
2 S
g 0.12 g 0.12
g 0.08 g 0.08
5 0.04 E 0.04
I I
0.00 0.00
8687888990919293949596 8687888990919293949596
Year Year
Northwest Northeast
2 0.16 2 0.16
g 0.12 g 0.12
{g 0.08 ‘g’ 0.08
g 0.04 5; 0.04
0.00 0.00
8687888990919293949596 8687888990919293949596
Year Year
Southwest Southeast
0.16 0.16
9 2
g 0.12 g 0.12
g 0.08 g 0.08
S 0.04 S 0.04
I I
0.00 0.00
8687888990919293949596 8687888990919293949596
Year Year
Illinois/Indiana Lake-wide
0.16 0.16
2 2
g 0.12 g 0.12
g 0.08 g 0.08
I 0.04 2% 0.04
°-°° 0.00
8687888990919293949596 8687888990919293949596
Year Year

Figure 8. Chinook salmon targeted harvest rates (targeted harvest per salmonine
angler-hour), by lake region, for the Lake Michigan sport fishery, 1986-1996.
Standard error bars are shown only for Michigan (see Methods). See Figure 1 for a
deﬁnition of lake regions.

35
Regional year class stocking, harvest, and harvest ratio (% return)

From 1985 to 1988, lake-wide stocking levels ﬂuctuated by 10% from 5.4 million
to 5.9 million. Harvest of those year classes, however, declined by 55% from 464,000 for
the 1985 year class to 209,000 for the 1988 year class (Figure 9). Stocking increased to
an all-time high of 7.85 million in 1989 while the harvest of that year class was 185,000
and the harvest ratio (percent of stocked ﬁsh harvested by the ﬁshery) fell below 3%.
Harvest ratio remained below 3% for the 1990 to 1992 year classes (Figure 10). All
regions of Lake Michigan experienced declining year class harvest from the 1985 year
class to the 1992 year class. Additionally, changes in regional year class harvest do not
appear to have been affected by local (within region) changes in stocking. If year class
harvest was affected by stocking levels, it was masked by the inﬂuence of changes in
stocking outside the local region, which further suggests that chinook salmon spatial
distribution was changing, and that this change had an effect on the ﬁshery.

Year-class harvest in Green Bay declined from the 1985 to the 1988 year classes
concurrent with declines in stocking (Figure 9). Stocking was highest in 1989 before
declining again through 1992. Increased stocking levels in 1989 and 1990 did not
improve year class harvest. Harvest ratio for the 1985 to 1988 year classes was relatively
constant at 5.1-5.7%, and dropped below 3% for the 1989 to 1992 year classes (Figure
10). The low harvest ratios for these four year classes were comparable to the lake-wide
values.

Harvest in the northern region was highest for the 1985 year class, and declined
for the 1986 and 1987 year classes (Figure 9). Year class harvest was relatively

consistent for the 1987 to 1991 year classes before declining again for the 1992 year

36
class. Stocking increased ﬁ‘om 1985 to 1989 before declining slightly from 1990 to 1992.

Harvest ratio showed a similar trend to year class harvest. Harvest ratios in the North
peaked at 2.75% for the 1985 year class and declined to 0.5% for the 1992 year class
(Figure 10).

Year class harvest in the northwest region declined by 31% from 72,000 for the
1985 year class to 50,000 for the 1987 year class, despite consistent stocking levels of 1.1
million (Figure 9). Harvest has been relatively consistent for the 1987 to 1992 year
classes at 42,000 to 53,000. Stocking levels were reduced in 1988 to 728,000 but peaked
in 1989 at 1.2 million ﬁngerlings before declining again from 1990 to 1992. Harvest
ratio ranged from 4.96 to 6.42 from the 1985 year class to the 1988 year class, but
declined to 3.63 for the 1989 year class. The harvest ratio increased for the 1990 to 1992
year classes, reaching a peak of 10.6 for the 1992 year class (Figure 10).

Approximately 800,000 chinook salmon were stocked annually in the northeast
from 1985 to 1987, while harvest of those three year classes declined by 65% from
130,000 to 45,000 (Figure 9). Stocking increased each year from 1988 to 1990, while
year class harvest remained consistently below 50,000. Harvest ratio for the 1985 year
class exceeded 15%, and declined for each subsequent year-class to a low of 1.8% for the
1992 year class (Figure 10).

Year class harvest in the southwest declined by 61% from 78,000 for the 1985
year class to 31,000 for the 1991 year class (Figure 9). Stocking declined by 59% from
1.1 million in 1985 to 455,000 in 1988, concurrent with the decline in year-class harvest.

Increases in annual stocking of 1.1 million in 1989 and 1990 did not cause an increase in

37
year-class harvest. Harvest ratio has ﬂuctuated from 3.2% for the 1990 year class to

8.6% for the 1988 year class (Figure 10).

Harvest in the southeast declined by 85% from 115,000 for the 1985 year class to
17,000 for the 1991 year class (Figure 9). Harvest ratio followed the same trend as year
class harvest, declining by 89% from 9.3% to 1.0% from the 1985 year class to the 1991
year class (Figure 10). Stocking from 1985 to 1988 was relatively constant at 1.3 million.
Stocking increased to 1.8 million in 1989 and declined to 1.5 million in 1992. Changes
in stocking did not increase year class harvest.

Harvest in the Illinois-Indiana region declined by 42% from 26,000 to 15,000
from the 1985 year class to the 1987 year class, despite increases in stocking by 54%
from 1985 to 1987 (Figure 9). Harvest ratio similarly declined by 63% from 4.7% to
1.8% over the same period (Figure 10). Harvest ratio has been consistently low at 1 to

2% from the 1987 to the 1992 year class.

38

 

 

   

 

 

 

   

 

 

 

 

   

 

 

   

Green Bay North
0.8 - 1.0
8306 4 3g gﬁu
8 a 8g x506
g E 0-4 £523 5 .8 E
m E as: I 3 a) g 0-4
0.0 ' ‘ 0.0
85 86 87 88 89 90 91 92 85 86 87 88 89 90 91 92
Year Class Year Class
Northwest Northeast
1.5
A B " A
gg §g Egrm
m? ﬁg m§0&
% v ‘ V
‘ V
0.0-
85 86 87 88 89 90 91 92 85 86 87 88 89 90 91 92
Year Class Year Class
Southwest Southeast
1.5 * 2.0 150
3 g 10 § .8 E g 1.5 ‘00 g
32 5g 3210 i
a: ' 0.5 m ‘ 50
0.0 ‘ 0.0 0
8586878889909192 8586878889909192
Year Class Year Class
Illinois-Indiana Lake-wide
1.5 10.0 500
8 A 3 3‘ ‘3 i? 8.0 ‘ 400 ‘3 A
£510 8g 35 m ems
mios §8 m: m 2mi§
*v *5 even 1W‘V
0.0 0.0 0
85 86 87 88 89 90 91 92 85 86 87 88 89 90 91 92
Year Class YearClaee

 

 

Figure 9. Chinook salmon stocking and harvest, by year-class and region, for the
Lake Michigan sport ﬁshery. See Figure 1 for a deﬁnition of lake regions.

39

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

   

 

 

 

 

 

 

 

Green Bay North
6.0 A ’3‘ 1.0
4 5.0 t, g 0.8 4..
" "0 ‘39 E 0.6 .
,. '1’ 3.0 E g ‘
.1 2° 3 0 °"
01.0; e 02*
i J. “ 0.0 ‘ 0.04
91 92
Northwest Northeast
A j.‘ A .
5 1.2- g g .
E 10* 0 EE- .
v 0.8 1 E v _
0.61 i g .
0.4< o .
5 0.2] i 5 .
‘ 0.0 ‘ .
é - 5 2
IE: - E ’ I?
I!
i; - r .
6 . 80'
O §
87 88 88
Year Class
lllinole-lndena
A 5.0 A
g 4.0 g g t
E , 3°85: E. E
2.0 g g
' O
6 0.2 ‘-° i 6'5
.

0.0

   

 

 

Figure 10. Chinook salmon stocking and harvest ratio, by year-class and region, for
the Lake Michigan sport ﬁshery. See Figure l for a deﬁnition of lake regions.

40
Discussion

The Lake Michigan salmonine ﬁshery has changed dramatically from 1986 to
1996. Lake-wide effort declined from 1989 to 1992 and has been consistently low from
1992 to 1996. Salmonine harvest declined from 1986 to 1990, and remained relatively
stable from 1990 to 1996. An increase in the targeted salmonine harvest rate from 1990
to 1996 is an indication that the salmonine ﬁshery was not completely dependent upon
the success of the chinook salmon harvest. Anglers shifted their efforts towards other
salmonines and maintained high harvest rates. The harvest rate of 0.14 in 1996 was
higher than the peak harvest years of 1986 and 1987. Hansen et a1. (1990) reported a
salmonine harvest rate exceeding 0.15 from 1982 to 1985 for the Wisconsin waters,
suggesting that the lake-wide ﬁshery may have peaked prior to 1986.

The question remains as to why the chinook salmon ﬁshery collapsed in the early
19908. Keller et al. (1990) suggest that the collapse was driven by changes in the
geographical distribution of chinook salmon, poor year class survival, and increased
mortality due to disease. This study provides information on the extent and location of
the declines in the ﬁshery as well as some additional insight into the causes of the ﬁshery
collapse. I believe that declines in the Lake Michigan chinook salmon ﬁshery were the
result of changes in ﬁshing effort, natural mortality, and the spatial distribution of the
salmon.

Declines in salmonine effort from 1986 to 1996 were a lake-wide phenomenon
with relatively little difference in the rate of decline across lake regions. Salmonine
harvest likely declined as a result of declining effort, although rates of harvest decline

were not consistent across species. Chinook salmon harvest declined far more than

41
harvest of any other salmonine, indicating that chinook salmon harvest was driven by

more than simply changes in effort.

While following trends in salmonine effort eliminates bias associated with effort
for yellow perch or other species, changes in salmonine effort may not accurately track
changes in effort targeted at chinook salmon. Anglers contend that they use different
ﬁshing methods to target lake trout, rainbow trout, and salmon by ﬁshing different
depths, ﬁshing with different lures or colors, or by ﬁshing along temperature breaks
(Bence and Smith in press; personal observation). Anglers increasingly targeted chinook
salmon in the early 1980’s, but shifted their effort towards other salmonines when
chinook salmon ﬁshing was poor (Bence and Smith in press).

Further analysis suggests that salmonine effort shifted away from chinook salmon
and towards other species during the late 1980’s. In 1986, 10% of angling parties
interviewed in Michigan’s boat ﬁshery indicated that they were speciﬁcally targeting
chinook salmon (Jerry Rakoczy, Michigan DNR, unpublished data). By 1993, only 1%
of anglers were targeting chinook salmon. Similarly, the percentage of boat anglers that
were speciﬁcally targeting salmon was 26% in 1987, and declined to 8% by 1992. In
contrast, the percentage of boat anglers targeting trout in general increased from 1% in
1986 to 8% in 1994. Boat anglers may have also become less speciﬁc as the ﬁshery
changed in the 1980s and 19903. The percentage of boat anglers targeting salmon and
trout increased from 24% in 1986 to 34% in 1991. Finally, the percentage of anglers that
indicated they were not targeting anything at all increased from 3% in 1986 to 12% in

1992.

42
Because of the popularity of chinook salmon in Lake Michigan, and because they

are the most important salmonine in terms of numbers stocked and harvested, declines in
harvest rates for chinook salmon probably contributed to the initial cause of the decline in
salmonine effort from 1986 to 1988 (Bence and Smith in press). Successful anglers were
able to redirect their effort towards other salmonines, while unsuccessful anglers reduced
their ﬁshing effort or left the ﬁshery altogether. The result has been an increasing
salmonine harvest rate from 1988 to 1996 (Figure 4).

Increasing public knowledge of contaminants in Great Lakes ﬁsh and ﬁsh
consumption advisory publications may have played a role in declines in effort. In 1989,
the National Wildlife Federation (NWF) published a controversial Lake Michigan ﬁsh
consumption report that had an immediate impact on the ﬁshery and caused a cascade of
media coverage (Associated Press 1989; Campbell 1989; NWF 1989). Reports of dead
chinook salmon on Lake Michigan beaches from 1987 to 1989 could have also served as
a message to the angling public that the ﬁsh in Lake Michigan were not healthy to eat and
therefore not worth the effort and money required to catch them. Consumption issues are
unlikely to be the cause of the continued low levels of ﬁshery effort. A 1996 survey of
Great Lakes anglers revealed that concerns about ﬁsh contamination was the least likely
reason for low ﬁshery effort. A lack of free time was cited as the most likely reason,
followed by low catch rates (Michigan Sea Grant 1998).

Another explanation for declines in ﬁshing effort is that the pattern on Lake
Michigan reﬂects a trend that goes beyond what is happening on either Lake Michigan or
the Great Lakes in general. It could reﬂect part of a national trend for the public to spend

less time in activities such as ﬁshing and hunting (Bence and Smith in press).

43
Poor year class survival has been implicated as one of the causes of the poor

chinook salmon ﬁshery in the late 1980’s (Keller et al. 1990). Poor returns to the sport
ﬁshery and to the weirs are evidence of poor year class survival beginning with the 1984
year class, although the causes are unknown. Most likely, though, the poor survival was
a result of in-lake processes and was not caused by changes in the condition of the
hatchery product (Keller et al. 1990). Since no marked changes in growth rates were
observed for chinook salmon prior to 1985 and the onset of BKD (Wesley 1996), it is
likely that poor survival prior to 1985 was due to early life mortality. Higher mortality
rates probably affected the older age classes after 1985 because most chinook that
washed up on beaches in the late 1980’s were age 2 or older (Nelson and Hnath 1990;
Johnson and Hnath 1991). Further, growth rates of older chinook salmon signiﬁcantly
increased after the BKD outbreak than before the outbreak, suggesting that density-
related stress immediately prior to the BKD outbreak may have slowed growth rates and
triggered increased mortality (Wesley 1996). Finally, the age structure of the harvest in
Michigan’s waters shifted towards younger age classes in the late 1980’s (see Chapter 3).
Quantifying these increased mortality rates has been difﬁcult. Because BKD was
implicated as the ultimate cause of death for chinook salmon on beaches in the late
1980’s, managers have monitored the incidence of BKD in an attempt to monitor natural
mortality rates. Incidence of BKD is monitored in chinook salmon returning to the weirs
and in ﬁshery-independent surveys. Fish are examined for clinical signs of disease, and
blood samples are tested specifically for BKD (Clark 1996). While this monitoring is
intended to provide an index of in-lake BKD mortality, the statistic “percent positive with

BKD” is difﬁcult to interpret because it could mean one of two things. First, a decrease

44
in BKD incidence could reﬂect in-lake decreases in BKD mortality, which assumes that

the sampled ﬁsh are representative of the population. Second, a decrease in BKD
incidence could instead reﬂect in-lake increases in BKD mortality, which assumes that a
greater proportion of infected ﬁsh die than survive to be tested (Clark 1996). Because of
this dichotomy, “percent infection rates” should not be used as the only index of BKD
mortality rates (Clark 1996).

Tests for the presence of Renibacterium salmoninarum, the causative agent of
BKD, at the Strawberry Creek weir in Sturgeon Bay, Wisconsin, have shown a decline in
the percentage of positive chinook salmon from a peak of 67% in 1988 to a low of 2% in
1994 (Marcquenski 1996). Incidence of clinical signs of BKD in chinook salmon
returning to Michigan weirs was about 85% in the late 1980s and declined to less than
10% by 1992 (Clark 1996). Clinical signs of BKD returning to the Manistee weir in
1992, however, were greater than 20% and declined to less than 10% in 1995. Visual
signs of BKD in chinook salmon collected from a ﬁshery-independent survey from 1990
to 1996 showed a peak level of about 37% and declined to less than 5% in 1996 (Clapp
1997). Laboratory tests for BKD of survey-caught ﬁsh in 1996, however, showed greater
than 10% incidence. Visual estimates of BKD incidence from surveys were consistently
higher than visual estimates from Michigan weirs (Clapp 1997) and could be an
indication that fewer BKD-infected ﬁsh survive to maturity.

Keller et al. (1990) noted that catch of chinook salmon in 1987 occurred in the
northern regions of the lake one month earlier in the season than normal. They suggested
that chinook salmon were more evenly distributed throughout the lake than normal due to

milder winter temperatures, and that this change in distribution contributed to the poor

45
1987 chinook salmon ﬁshery. An even distribution of chinook salmon throughout the

lake should be reﬂected by similar trends in regional catch rates. Poor survival would
decrease abundance lake-wide, and similar declines in regional catch rates would reﬂect
this. However, catch rates did not decline similarly across all regions, which suggests
that chinook salmon were not evenly dispersed but were in fact spatially congregated.

Temperature and food seem to be the two driving factors that inﬂuence chinook
salmon distribution (Keller et al. 1990; Elliott 1993). Chinook salmon prey primarily
upon alewife, bloater, and smelt, but there is debate about whether chinook salmon prefer
alewife (Jude et al. 1987), or whether they are opportunistic (Elliott 1993; Rybicki and
Clapp 1996). Forage abundance in Lake Michigan varies seasonally and spatially
(Brandt et al. 1991). In particular, alewife and rainbow smelt have been more abundant
and constitute a larger proportion of the forage abundance in the northern and western
waters of the lake. Bloaters are abundant throughout the lake but are dominant in the
eastern waters. Regional diets of sport-caught chinook salmon reﬂect regional forage
abundance (Hagar 1984; Toneys 1992; Elliott 1993; Peeters 1993; Rybieki and Clapp
1996).

Alewife spatial distribution in Lake Michigan shifted between 1985 to 1995 (Ann
Krause, Michigan State University, unpublished results). Alewives were abundant across
western and eastern regions of the lake in the mid-1980’s, as indicated in trawl surveys
conducted by the Great Lakes Science Center. Abundance then declined in the eastern
regions of the lake in the early 1990’s as abundance in the western regions increased
(Figure 11). Trends in alewife distribution and abundance appear to match trends in

chinook salmon harvest and targeted harvest rates (Figure 7 and Figure 8), and suggest

46
that the spatial distribution of chinook salmon changed as the spatial distribution of

alewife changed. This is further supported by preliminary survey data which show a
correlation between high chinook catch rates and a high proportion of alewives in their
stomachs (Dave Clapp, Michigan DNR, personal communication). Earlier studies
showed seasonal and spatial differences in chinook diets that corresponded with forage
abundance and species composition (Elliott 1993), suggesting that chinook salmon
demonstrated a seasonal migration in the spring away from eastern waters and back again
in the fall. If chinook salmon prefer alewife as their primary prey, then changes in prey
distribution would cause changes in predator distribution and would be reﬂected in the
ﬁshery. Chinook salmon that successfully migrated in order to continue to prey on
alewives survived, while those did not follow alewives were forced to prey on other
species - namely bloater and rainbow smelt. Chinook salmon that preyed primarily upon
species other than alewife may have been more susceptible to nutritional stress and
subsequent mortality.

While localized increases in mortality may have been possible, especially in the
southeast region of the lake, it is not accurate to think of the lake as consisting of several
distinct populations suffering different mortality rates. Chinook salmon that tend to stay
in a given area may suffer different mortality rates than ﬁsh in other areas, but the ﬁsh in
each area is a mix of ﬁsh that originated from different stocking and spawning locations,
and the mix is itself likely to be dynamic as forage abundance changes spatially over
time. This highly migratory nature of chinook salmon suggests that changes in spatial
distribution are likely to have caused most of the regional differences in how the ﬁshery

changed.

47
Attempts to increase local yields in Lake Michigan by increasing local stocking

are likely to lead to frustration. Regional increases in stocking levels did not improve
regional year class harvest. This was particularly true for the Northeast and Southeast
regions, where year class harvest continued to decline for the 1985 to 1993 year classes
despite increases in stocking levels from 1985 to 1992 (Figure 9). During the study
period, lake-wide increases in numbers stocked for a year class also did not lead to lake-
wide increases in harvest. If anything, lake-wide increases in stocking led to declines in
harvest, CPUE, and other measures of ﬁshery success. This is probably due to density
dependent processes, which although not proven, is consistent with the data. For
example, BKD infection rates are positively related to stocking levels (Clark 1996). Of
special importance, the harvest ratio observed for the 1989 through 1992 year classes
(about 2.5%) represented a substantial decline over that seen for the 1985 (about 8%).
Harvest ratios probably had already declined for the 1985 year class in comparison with
earlier cohorts. This year class was impacted by BKD mortality, and the harvest ratios
for the 1985 year class calculated for Wisconsin’s waters were already substantially

below those reported for Wisconsin for the 1969-1982 year classes (Hansen et al. 1990).

48

 

 

   

 

 

 

 

 

   

 

 

 

   

Northwest Region Northern Region
l—Smrseon Ben
70 70
8 5° g 60
*1 5° 1350
§ 40 g 40
e 30 o 30
_ _>
g 20 g 20
1! 1° ‘3 10
0 0
85 86 87 88 89 90 91 92 93 94 95 85 86 87 88 89 90 91 92 93 94 95
Year Year
Southwest Region Northeast Region
|— Port Washington] [ Frankfort ------ Ludingion]
70
O
8 6° 2
C so ‘
e 40 g
in a
52° 5
8'. 1° 8
0
858687888990 9192 93 94 95 858687888990 9192 93 94 95
Year Year
Illinois-Indiana Region Southeast Region
L 3......” ------ 3...... Hanna
70
i“ 2
2 .53 E.
a 30 2
e o
B 20 S
1 2
:2 1° 63
0
8586878889909192939495 8586878889909192939495
Year Year

Figure 11. Relative alewife abundance from various regions of Lake Michigan,
1985-1995. Data are from Great Lakes Science Center annual fall bottom trawl
surveys. Estimates are based on ﬁtting a general linear mixed model to these data,
including year and depth effects as well as port and year‘port interactions (Ann
Krause, Michigan State University, unpublished results).

49
Conclusion

The collapse of a ﬁshery is often caused by overﬁshing, but this was not the case
with the Lake Michigan chinook salmon ﬁshery from 1987 to 1992. The chinook salmon
population is driven by annual stocking, and returns to the ﬁshery declined despite the
maintenance of high stocking levels. Chinook salmon suffered high mortality rates due
in part to bacterial kidney disease, while the underlying cause of the disease is probably
related to nutritional stress due to a decline in the abundance of alewives. Additional
stress may be temperature-related, as most visual accounts of mortality occur in early
spring, when water temperatures are coldest.

The decline of the ﬁshery differed across regions. A complete collapse of the
ﬁshery was seen in the eastern regions of the lake, although declines in effort and harvest
were observed in all regions. The greatest declines occurred in the Northeast and
Southeast regions that traditionally had the highest levels of stocking, effort, harvest, and
harvest rates. With a decline in the ﬁshery came a change in the distribution of the
harvest. For example, 21% of the lake-wide chinook salmon stocking in 1985 occurred
in the Southeast region, and accounted for 25% of the lake-wide harvest of the 1985 year
class. By 1992, stocking in that region increased to 27% of the lake-wide total, while
year class harvest fell to 13% of the lake-wide total. In contrast, stocking in the
Northwest region of the 1985 year class was 19% of the lakewide total, and year class
harvest was 16%. By 1992, stocking in the Northwest region decreased to 9% of the
lake-wide total, and year class harvest increased to 34% of the lake-wide total. The
relative contributions of the Green Bay, North, and Illinois-Indiana regions to the lake-

wide harvest remained relatively constant for the 1985 to 1992 year classes.

50
Trends in the sport ﬁshery data suggest that a change in the spatial distribution of

chinook salmon was the driving force behind regional differences in the decline of the
ﬁshery. Increases in lake-wide mortality probably contributed to these declines, but
spatial differences in mortality are unlikely to be the primary cause of these differences.
Tagging studies and similar harvest size distributions show that chinook salmon do not
form distinct subpopulations, but rather mix widely. Most likely chinook salmon
migrated in response to local stresses, and concentrated in the western regions of the lake

when alewife abundance in the eastern regions declined.

CHAPTER THREE

LAKE-WIDE STOCK ASSESSMENT MODEL

Introduction

The Paciﬁc salmon stocking program in Lake Michigan began in the late 1960s in
response to the extirpation of native lake trout (Salvelinus namaycush) stocks and high
abundance of exotic alewives (A Iosa pseudoharengus) (Eshenroder et al. 1995). The
immediate popularity of these introduced salmonine species to the sport ﬁshery, as well
as a growing ﬁshery-related industry, prompted state agencies to maintain an intensive
stocking program. Numbers of salmonines planted in Lake Michigan peaked in 1984 at
17 million ﬁsh (Holey 1995; Chapter 2). However, beginning in 1987 the thriving
salmonine ﬁshery of the mid-1980s underwent a substantial decline in angling success
and harvest. This was largely due to a dramatic decline in the chinook salmon population
despite the maintenance of consistent stocking levels (Bence and Smith in press; Chapter
2). While explanations for the decline have included the prevalence of a bacterial kidney
disease, the root of the problem may be that current levels of stocking are too high,
leading to a scarcity of forage ﬁsh, nutritional stress, and BKD-induced mortality
(Stewart et al. 1981; Stewart and Ibarra 1991; Koonce and Jones 1994). Little is known,
in the quantiﬁable sense, about the impact of BKD-related mortality on the population

(Clark 1996).

51

52
The Michigan DNR Division of Fisheries has recently expressed the need for

using more rigorous modeling analyses to develop lake-wide stocking plans for
salmonines (Clark 1996). Similarly, the Lake Michigan Committee of the Great Lakes
Fishery Commission has called for more mathematical modeling of existing data to help
establish detailed management plans, and species-speciﬁc harvest levels for salmonines
in Lake Michigan (Eshenroder et al. 1995). Ultimately, decisions on what species to
stock, how many to stock, and where to stock them could be crucial toward reviving the
chinook salmon ﬁshery in Lake Michigan.

I have approached this problem using a population model based on catch-at-age-
analysis (CAA) (Megrey 1989). This study is unique to salmonine dynamics of Lake
Michigan because it incorporates an integrated modeling and data analysis approach not
previously used for Lake Michigan salmonines (Sitar 1996).

The goal of this study was to quantify the contributions of three sources of
mortality, (1) ﬁshing mortality, (2) spawning mortality, and (3) time-varying natural
mortality, to the total annual mortality of chinook salmon from 1985-1996. Furthermore,
model parameters and estimates of age-speciﬁc abundance can be used as improved
inputs in tropho-dynamic models for Lake Michigan (e. g. SIMPLE, Koonce and Jones
1994).

The Michigan DNR (MDNR) is currently addressing this issue with a multi-
species model, CONNECT, which is designed to predict ideal stocking levels required to
meet ﬁsh community objectives for salmonines in Lake Michigan (Rutherford 1997).
While novel in its approach, the current CONNECT model for chinook salmon was not

rigorously ﬁt to all available data using an optimization routine, as would be done in a

53
CAA model. Other recent modeling work on chinook salmon in the Great Lakes

involves a CAA model built for chinook in Lake Huron (Bence and Meehan
unpublished). The Lake Michigan chinook model presented here builds on the

information used by CONNECT and the structure of the Lake Huron CAA model.

Methods

The stock assessment model is an age-structured, deterministic model that
estimates abundance for multiple cohorts. Initial cohort abundance is assumed to be
known, and the model accounts for changes in abundance due to various sources of
mortality.

Population Model

The basic idea of the population model is that population abundance at the start of
a given year is equal to the abundance at the start of the previous year, multiplied by the
proportion of that population that survives the year. Survival is a function of a

continuous-time instantaneous mortality rate, such that:
N N '2’
y+l = ye (1)

Lake Michigan chinook salmon population dynamics do not follow equation 1 because
not all sources of mortality function in a continuous fashion over a yearly period.
Preliminary analysis of sport harvest data indicates that most of the in-lake ﬁshing
mortality occurs in July and August. Similarly, analysis of weir return data suggests that
most of the spawning-related mortality occurs in September and October. Chinook

salmon population dynamics in Lake Michigan can be more accurately modeled using an

54
approach that combines both continuous-time and discrete-time sources of mortality

(Kope 1987; Bence and Meehan unpublished).
Abundance

Chinook salmon abundance for a cohort at the start of a calendar year was
assumed to be a function of abundance of the cohort at the start of the previous year,

natural mortality, ﬁshing mortality, and maturation mortality, such that:
N-.. = N..e‘”°~' <1- p“ >(1- P...) (2)

where M... is an instantaneous natural mortality rate (for age-a and year-y), Pp... is the
annual proportion of the population removed by the ﬁshery, and PM". is the annual
proportion of the population that matures and returns to the streams to spawn and die.
Natural Mortality

Natural mortality (M...) is an instantaneous annual rate and is assumed to operate
independently of ﬁshing mortality (Hilborn and Walters 1992). Most CAA models
assume a constant natural mortality rate that applies to all ages (Megrey 1989), however,
we know that the natural mortality rate for chinook salmon in Lake Michigan increased in
the late 1980’s in response to an outbreak of bacterial kidney disease (Johnson and Hnath
1991). In order to quantify the changes in natural mortality during the study period, I
modeled mortality as a sum of a constant component, and a time-varying (TVM)
component, such that the total natural mortality rate (MM) is:

Ma.y =Ma +M7vM,_, (3)

Age-speciﬁc natural mortality (M,) was assumed to be a known constant, estimated prior
to the onset of BKD-related mortality (Table 14). I derived an estimate of the age-0

mortality rate based upon previous modeling work on Lake Michigan salmonines during

55
pre-BKD mortality years (Stewart et a1. 1981). Mortality rate estimates for ages 1-3 were

based upon estimates from west-coast populations (Rutherford 1997). I assumed that
mortality rates for ages 4-5 were equal to age-3 mortality, because any increase in
mortality rates for older ﬁsh would be accounted for by TVM and maturation mortality.
Age- and year-speciﬁc TVM was estimated by the model. I assumed that TVM
affected age groups in a logistic fashion, with ages 05 being increasingly affected. This
assumption reﬂected observations of BKD mortality (Nelson and Hnath 1990), and
approximated assumptions of BKD mortality rates from CONNECT (Rutherford 1997)

The logistic mortality function was:

7’.
1+ Lima-ﬂ)

(4)

MWMa.y =

where a is age, 7, is a year-speciﬁc TVM intensity parameter, and 0t and B are parameters
that determine the shape of the logistic function. The model estimated ln(yy), ln(or), and
ln(B) as formal parameters. This logistic model forced a relationship between age-
speciﬁc rates within years, while the year-speciﬁc TVM intensity parameters were
unrelated between years. BKD mortality did not appear to be a signiﬁcant source of
mortality in Lake Michigan chinook until about 1987 (Clark 1996). Since that time,
BKD-infected chinook have been observed in the population at varying levels of
incidence (Clark 1996). I allowed the model to estimate TVM from 1985 to 1996. Initial
parameter values were chosen that matched age- and year-speciﬁc TVM with age- and
year-speciﬁc estimates of BKD mortality rates reported in the CONNECT model

(Rutherford 1997).

56
Fishing Mortality

Fishing mortality is assumed to be an instantaneous event, occurring at the end of
July. I felt that this ﬁshery more closely represents a seasonal, or pulse, ﬁshery as
opposed to a continuous ﬁshery. An identical approach was used for Lake Huron
chinook (Bence and Meehan unpublished). The proportion of the population removed by

the ﬁshery (Pp...) was estimated as:
PF“ = 1- e “" (5)

where Fmy is an instantaneous mortality rate per unit time that occurs over an
inﬁnitesirnally short time unit. Fa,y operates under a separability assumption (Megrey
1989) and is a function of age-speciﬁc ﬁshery selectivity (S,) and year-speciﬁc ﬁshing
intensity (f,), such that:

F, =Saf. (6)

Selectivity to the sport ﬁshery was assumed constant over time. Selectivity was also
assumed to operate in a logistic fashion where ages 0-5 were increasingly selected for by
the ﬁshery, such that:

1
So 1—‘— ‘7’

where or and 6 determined the shape of the logistic curve, and ln(a) and ln(B) were
formal parameters estimated by the model. A maximum selectivity value of 1 indicated
that an age group was fully selected to the ﬁshery.

Year-speciﬁc ﬁshing intensities (fy) were estimated by the model, with ln(f,)

estimated as formal parameters from 1985 through 1995. For 1967 to 1984, ﬁshing

inst
tat:
fror

fun:

whe
salrr
ﬁsh:

Catt

durir
mom
Comr
an a;
that
the a;
clital;

rental

57
intensity was assumed to increase linearly from zero to the 1985 level estimated by the

model (Jones et al. 1993).
Maturation (Spawning) Mortality

Similar to ﬁshing mortality, maturation mortality (MAT) was assumed to be an
instantaneous event, occurring immediately after ﬁshing mortality and before additional
natural mortality. Age-specific maturation was assumed to increase in a logistic fashion
from age 0 to age 4, while MAT for age-5 was assumed to equal 1. The maturation
function for ages 0 through 4 was:

1
Pm. = W (8)

where ln(or) and ln(ﬁ) were estimated as formal parameters. I assumed all chinook
salmon reached maturity by age-5 because few age-6 ﬁsh are observed in the ﬁshery and
ﬁshery-independent surveys.
Catch

Sport ﬁshery catch, or harvest, was assumed to occur as an instantaneous event,
during which time the population was subject to no other sources of mortality, and after 7
months of natural mortality had taken place (Bence and Meehan unpublished). A
common approach in catch-at-age analysis is to assume that ﬁshing mortality operates in
an approximately continuous fashion, thereby warranting the use of the standard Baranov
equation to estimate catch (Hilborn and Walters 1992). For intensely seasonal ﬁsheries,
the approach taken here can provide a better approximation of catch than the Baranov
equation (Mertz and Myers 1996). Catch is the proportion of the population abundance

remaining after seven months that dies from ﬁshing, and is estimated by:

C = N ..,e-M°°’7“2PF0 (9)

To better estimate age-speciﬁc maturation, I estimated the age composition of mature
chinook salmon harvested by the ﬁshery. Mature chinook harvested by the ﬁshery is the
proportion of the harvest that has reached maturity, and is estimated by:

CHAT“ = Ca.y PHATO (10)

Eﬁort
Sport ﬁshery effort is related to year-speciﬁc ﬁshing intensity divided by an

assumed constant catchability coefﬁcient (q), such that:

f.

E=-—'-
4

.V

(11)

where ln(q) is a formal parameter estimated by the model.
Observed Da_t_a__ and Other Model Inguts
Recruitment

Chinook salmon are stocked in the spring as age-0 ﬁngerlings. Lake-wide
stocking data from 1967 to 1996 was collected ﬁom the Departments of Natural
Resources ﬁ'om Wisconsin, Illinois, Indiana, and Michigan (Chapter 2). In the modern
era, no chinook were stocked in Lake Michigan prior to 1967. Recruitment to age-0 of
naturally reproduced chinook has been steadily increasing over time (Clark 1996). Input
data for natural recruitment were taken from Rutherford (1997), and are based on
estimates of natural reproduction from Carl (1980; 1982; and 1984), Seelbach (1985 and
1986), Zafft (1992), and Hesse (1994). Age-0 recruitment is the sum of stocked

ﬁngerlings and estimated wild smolts. Recruitment is an input into the model and is not

59
used as observed data to the ﬁt the model. The model assumes that recruitment to age-0

occurs at the beginning of the year.
Sport Fishery Information

Harvest and effort information was compiled from data collected by creel survey
programs run by each of the four states surrounding the lake (Chapter 2). I attempted to
use effort that was directly targeting chinook salmon, although I had to compromise due
to differences in creel survey programs. For Wisconsin data, effort was estimated from
interviews in which anglers speciﬁcally indicated they were targeting chinook salmon.
For Illinois and Indiana, I used effort targeted at all salmonine species. I estimated
Michigan effort from raw creel data using interviews in which anglers indicated they
were targeting chinook salmon, coho salmon, salmon in general, or salmon and trout in
general. Targeted effort used for this model differs from targeted effort reported in
Chapter 2, which was deﬁned as effort that was targeting all salmonines. The model uses
harvest and effort data to estimate changes in population abundance. I used chinook
salmon effort in the model to avoid any bias due to possible changes in effort not directed
at chinook salmon. Post hoc comparison, however, of chinook harvest rates calculated
from chinook salmon effort versus salmonine effort showed the same trends. Chinook
salmon total harvest estimated from total angling effort (including non-targeted effort)
was used for each state.

Michigan is the only one of the four states that collects substantial harvest age
composition data. Wisconsin collects length composition data. Preliminary analysis of
length compositions between Michigan and Wisconsin indicate similar harvest length

compositions. Therefore, I assumed that age composition of the lake-wide harvest could

60
be reasonably estimated by the age composition of the Michigan harvest. Chinook

salmon migrate widely within Lake Michigan and even between Lake Michigan and Lake
Huron (Clark 1996), suggesting that stocks are reasonably mixed, and therefore that the
age composition of the population may be fairly homogenous.

Michigan also collects data on maturity of chinook salmon sampled from the
ﬁshery. From these data, the age composition of mature ﬁsh harvested by the ﬁshery was
estimated. These age compositions of mature chinook were based on sampling data from
July 15 to August 31. This time period was chosen so that maturation was advanced
enough that identification of maturity would not be difﬁcult, though not so advanced that
aging error due to scale erosion would be a problem. Model estimates of mature age
composition were ﬁt to empirical estimates in order to provide additional information on
maturation schedules of chinook.

Weir Harvest Information

Harvest data from Michigan and Wisconsin weirs are available from 1985 to the
present. Prior to 1985, only Michigan collected weir information. I generated an
estimate of the lake-wide weir harvest age composition for 1985 to 1996, weighted by the
number of chinook sampled from each of the weirs around the lake. Questions about the
validity of sampled age compositions, coupled with the inconsistencies of reported age
compositions between years, prevented the use of weir harvest information in the model
prior to 1985.

Fitting the Model to Observed Data
Model estimates of effort, total harvest and mature harvest, harvest and mature

harvest age compositions, and weir harvest age compositions were ﬁt to observed data

61
from 1985 to 1996. Model parameters were iteratively and independently adjusted in

order to provide the best ﬁt. Fit was measured with a log-likelihood function, and best ﬁt
was reached when the log-likelihood function was maximized (Methot 1990). A quasi-
Newton search algorithm was used to ﬁnd the maximum likelihood, with forward
differencing used to estimate the partial derivatives of the objective function. Parameters
were estimated using quadratic extrapolation. The log-likelihood equation was:
L=L,+L,+L3+L,+L5 (12)
where L, was the log-likelihood of the model ﬁt to observed ﬁshery effort data, L; was
the log-likelihood of the model ﬁt to observed ﬁshery harvest, 14 was the log-likelihood
of the model ﬁt to observed ﬁshery harvest age composition, L. was the log-likelihood of
the model ﬁt to observed ﬁshery mature harvest age composition, and L5 was the log-
likelihood of the model ﬁt to observed weir harvest age composition. No external
weighting was applied to any of the likelihood functions (see Methot 1990). Errors were
assumed to be log-normally distributed for L1 and L; such that the log-likelihood

functions were deﬁned as:

111M?” ) - ln( 115”" )

2
0'

.V

(13)

 

LL, = 4.5):
y

where 113’” and 15'“ are the observed and predicted effort and harvest. The standard

deviation (0) was set at 0.06 for effort and 0.08 for harvest, and was estimated as:

 

a = Jln[(CV)2 +1] (14)
Law and Kelton (1982). I estimated an average coefﬁcient of variation (CV) from

observed annual effort and harvest estimates from Michigan’s waters of Lake Michigan

(See Chapter 2).

62
Errors were assumed to be multinomially distributed for the age composition log-

likelihood functions, such that they were deﬁned (ignoring constants) as:
L..345 =£ny 2P3? ln(Ppryed) (15)

”red are the observed and

where try is the effective sample size in year y, and p...Obs and p.y
predicted proportions at age a in year y. The effective sample size in the likelihood
functions for the harvest, mature harvest, and weir harvest age compositions was set to
100, 50, and 50 respectively. These values represent subjective judgements about the
accuracy of the observed data. (For a discussion of this issue see Fournier and Archibald
(1982)). I have more conﬁdence in the observed harvest age composition than the mature
harvest and the weir harvest age compositions because of aging error caused by (1) scale

erosion or (2) problems associated with the use of an age-length key to estimate weir

harvest age compositions.

Results

Fishery Effort and Harvest

The model ﬁt the observed ﬁshery effort and harvest data reasonably well, but
had more difﬁculty fitting 1985-1988 data versus 1989-1996 data (Figure 12). Chinook
salmon effort and harvest increased from 1985 to 1986 before declining through 1994.
Effort was relatively constant from 1994 to 1996, while harvest increased. The model
was generally able to follow these declines. There was an obvious tradeoff as the model
attempted to ﬁt ﬁshery effort and avoid overestirnating harvest for 1985-88. The result

was an underestimation of the decline in effort and an overestimation of the decline in

 

63
harvest. A pattern in effort and harvest residuals shows that effort and harvest errors are

correlated, which is to be expected because both are related to the ﬁshing intensity

parameter, fy (Figure 13).

 

[morn -o-Pred.

 

     

xx”...

    

xxx" .\

 

 

 

 

[mom -o-Pred.|

 

 

 

 

1,200 -
1

1,000 *-

meoo. 5 .85I

89

87

 

ions of

III [III

Figure 12. Observed and predicted values of sport ﬁshery effort
angler-hours (top) and chinook salmon harvest in thousands of ﬁsh (bottom).

0.35 -

0.30 -

0.25 <

0.20 —

0.10 4

0.05 .

Eﬂort Residuals

0.00 -

-0.05 -

-0.10 4

 

-0.15

0.20 .

0.15 -

0.10 -

0.05 .

0.00 .

v0.05 .

-0.10 -

Harvest Residuals

-0.15 -

-020 .

-0.25 ‘

 

-0.30 .

Figure 13. Loge-based residuals from model predictions of ﬁshery effort (top) and
chinook salmon harvest (bottom) for the Lake Michigan sport ﬁshery.

old:
wei
fror
in 11
com

5011‘

{15111

(Tat

age l

the:

Com;

Cigar

them

66
Age Comitions

 

The onset of additional natural mortality in the late 1980’s resulted in a decline of
older age classes from the population, and this decline was reﬂected in the ﬁshery and
weir age composition data. Fishery harvest consisted primarily of 2 and 3-year old ﬁsh
from 1985 to 1989, with more 3 and 4-year old ﬁsh harvested than 1 and 2-year old ﬁsh
in most years (Table 20). From 1990 to 1995, the harvest age composition shifted to
consist primarily of 1 and 2-year old ﬁsh, with 1-year old ﬁsh dominating the harvest in
some years, and with more 1-year old ﬁsh harvested than 3-year old ﬁsh in all years.

Similar trends were seen in the age composition of mature ﬁsh harvested by the
ﬁshery and in the weirs. Age 3-4 ﬁsh dominated the mature harvest from 1985 to 1989
(Table 21). In 1990, the age composition shifted to mostly age 2-3 ﬁsh along with an
equal proportion of age-1 and age-4 ﬁsh. Weir harvest age compositions were comprised
of age 2-4 ﬁsh from 1985 to 1990, but shifted to age 1-3 ﬁsh from 1991 to 1996 (Table
22).

I examined patterns in standardized residuals to evaluate model ﬁt to observed

age composition data. Standardized residuals were estimated as:

prod

pf: - p...

SR = d ed (16)
MSG-125T. Nu.

 

 

where p” is the proportion at age a in year y from the observed and predicted age

composition data, and lief is the effective sample size.
The model had difﬁculty ﬁtting the ﬁshery age composition data, as evidenced by
clear patterns in the standardized residuals (Figure 14). The proportion of age-0 ﬁsh in

the harvest was consistently overestimated, while the proportion of age-1 ﬁsh was

67
consistently underestimated. The model also consistently overestimated the proportion of

age-2 ﬁsh. Residuals are more randomly distributed and the magnitude of the residuals
decreases for ages 3-5. These patterns are probably a result of the difficulty the model
has when estimating the true selectivity with a logistic function. The logistic function
cannot follow the slope of the true selectivity function, so it compromises by
overestimating age-0, underestimating age-1, and overestimating age-2. One solution
would be to allow the model to estimate age-speciﬁc selectivity without the constraint of
a forcing function. An immediate solution would be to set age-0 selectivity to zero, since
very few age-0 ﬁsh are observed in the ﬁshery harvest.

Similar patterns in the standardized residuals are also evident in the weir harvest
age composition data. Most notable is the model’s tendency to overestimate age-0 and
underestimate age-1 ﬁsh, and was likely due to the inability of the logistic function to
follow the true maturation rates. Residual patterns for age 2-4 ﬁsh show a deﬁnite
transition between 1990 and 1991, and reﬂect the model’s inability to follow the abrupt
change in the observed age composition data (Figure 15).

Abrupt changes in age composition data are also reﬂected in the standardized
residuals for the ﬁshery mature harvest age compositions. Most notable is the increase in
the residuals for age-1 ﬁsh from 1989 to 1990, as the model cannot follow the rapid
increase in age-1 harvest from 1990 to 1993 (Figure 16). Residuals appear to be
randomly distributed for ages 2-4. The model consistently overestimates the proportion

of age-0 and age-5 ﬁsh, although differences from observed data are small.

‘Il‘ll1u‘1c 1llu‘l. III:

Fi

68
LIAge O DAge ﬂ

3.
2“ l l |
1.

0- I

85 86 87 88 89 90 91 92 93 94 95 96

 

 

 

 

 

Standardized Residuals

 

 

LIAge 2 DAge 3]

 

Standardized Residuals

 

 

85 86 87 88 89 90 91 92 93 94 95 96

 

IIAge 4 ElAge 5]

 

Standardized Residuals

 

85 86 87 88 89 90 91 92 93 94 95 96

Figure 14. Standardized residuals for ﬁshery harvest age compositions of chinook
salmon in Lake Michigan. See text for calculation of standardized residuals.

0.03305”; Bitumen-.- =~Mw

W. a: {New GI BLIP-WU: C.“

Fig“,

531ml

69

[£060 DAge 1]

 

3.0 -
2.5 « F r

2.0 - j- l— l-

1.5 r

1.0 ‘
(15" II
Oll- II

-0£5.
-11).
4.5d
-20 1

 

 

 

 

 

 

 

 

 

 

SI I I'edli 'I I

 

85 86 87 88 89 90 91 92 93 94 95 96

 

ILAge 2 DAgeﬁ

 

 

 

85 86 87 88 89 90 91 92 93 94 95 96

 

IIAoe 4 DAgﬁl

 

 

 

85 86 87 88 89 90 91 92 93 94 95 96

Figure 15. Standardized residuals for weir harvest age compositions of chian
salmon from Lake Michigan. See text for calculation of standardized residuals.

70

 

IAge 0 CIAge1
5 .
5 .
4 q
3
3 .
.1
B 2
a
g 1 .
o -
.1 .
.2 .1
85 86 87 88 89 90 91 92 93 94 95 96
IAge 2 DAge 3
4 .
3 -

Standardlzed Rosldlals
o n

r
N
I

r
(J
r

 

85 86 87 88 89 90 91 92 93 94 95 96

IAge 4 DAge 5

S‘lnndlmllzod Health-h
m

 

85 88 87 88 89 90 91 92 93 94 95 96

Figure 16. Standardized residuals for ﬁshery mature harvest age compositions of
chinook salmon in Lake Michigan. See text for calculation of standardized
residuals.

71

Fishing Mortality

Fishing mortality is estimated by the model for ages 0-5 from 1967 to 1996.
Fishing mortality has had relatively little impact on age 0 and age 1 chinook salmon, with
mortality never exceeding 3% in any year for either age class (Table 13). Thus although
the model did not accurately estimate harvest of the younger ages (0, 1), these errors had
only minor inﬂuence on the predicted dynamics. Age-2 ﬁshing mortality reached a peak
of 13% in 1986, and declined to 5% by 1992. Ages 3-5 chinook suffer peak ﬁshing
mortality levels in 1986, from 30% and 41%, and declined to 13-18% by 1992.
Maturation

Model-estimated proportions of chinook salmon that matured for ages 0-4 were
0.00, 0.02, 0.13, 0.51, and 0.87. Age-5 maturity was set to 1 (Table 19). Note that these
proportions do not indicate the proportion that return to the streams, since some mature
ﬁsh are harvested by the ﬁshery. Total harvest of mature ﬁsh from weirs and from the
ﬁshery were not ﬁt to observed data because observed weir harvest data does not account
for all chinook that run up all streams tributary to Lake Michigan. Therefore, only the
age compositions of the ﬁshery mature harvest and weir harvest are used to compare with
observed data (Table 21; Table 22).
Natural Mortality

I established baseline age-speciﬁc natural mortality rates that I assumed to operate
over the entire study period. These rates were set at 0.75 and 0.30 for ages 0 and 1,
respectively, and 0.10 for ages 2 to 5 (See Methods). Mortality rates were held constant
for older ages because I assumed that any additional mortality for older ﬁsh would be

accounted for by TVM and maturation.

72
BKD-related deaths of chinook salmon were not observed in Lake Michigan until

1986, but I allowed the model to estimate TVM beginning in 1985. The model estimated
a TVM rate of 0.00 for ages 0 and 1, and estimated that the same TVM rate applied for
ages 2 to 5. TVM increased for ages 2 to 5 from 0.00 in 1985 to a peak of 1.70 in 1993
before declining to 0.29 in 1996 (Table 18).

There is some concern that the logistic model applied to TVM may have been too
restrictive. In particular, the fact that the model estimates equal values for ages 2 to 5
may suggest that TVM for older ages, if given the freedom, might actually decline. I did
not test this by ﬁtting a different curve to the model, but I did test the baseline CAA
model against a model that estimated TVM separately for each age and year. I compared
the baseline CAA model against the new model using a likelihood ratio test (Seber and
Wild 1989). Allowing the model to independently estimate TVM for each age and year
signiﬁcantly improved model ﬁt (P<0.005). TVM estimates for ages 0 and 1 continued
to be relatively small, but estimates for ages 2 to 5 were markedly different across ages,
with no consistent trends across ages or years. If there is time-varying mortality among
ages, I would expect it to have some systematic pattern so that in a given year close ages
would respond the same way. Instead, mortality rates varied without pattern. This may
not reﬂect time-varying mortality but rather an over-parameterization of this alternative
model, which uses these new parameters to explain other process errors such as
differential catchability or aging errors.

Total Mortality
The model does not allow survival past age 5, therefore total annual mortality of

age 5 is 100%. Age 0 and age 1 chinook are not exposed to TVM mortality and very

73
little ﬁshing mortality, that total annual mortality of these ages has remained relatively

steady from 1967 to 1995, at 53% and 28% for age 0 and age-1, respectively (Table 15).
Estimated TVM appears to have the greatest effect on ages 2 and 3. Pre-TVM mortality
averaged 25% and 61% for ages 2 and 3, respectively. Total annual mortality increased
substantially for these age groups during TVM years, averaging 66% and 83% for ages 2
and 3, respectively. Annual mortality before 1985 for age 4 chinook salmon averaged
91%, and increased to 96% after 1985. Highest total annual mortality for all ages was
observed from 1991 to 1994, which corresponds with the lowest harvest years, though
not the highest ﬁshing mortality years.

A comparison of total number of deaths in each year from 1985 to 1996 indicates
that relative contributions of different sources of mortality shifted over time after 1985.
For age-3 ﬁsh, for example, natural mortality accounted for 12% of the total deaths, but
increased to 54% after 1985. Fishing mortality accounted for 20% prior to 1985 and
declined to 15% after 1985. Spawning mortality comprised 68% of the total annual
mortality for age-3 ﬁsh prior to 1985, but declined to 31% after 1985 as natural mortality
increased. It is clear that the increase in natural mortality caused a decline in maturation
deaths and harvest, as most BKD infected ﬁsh were not surviving to reach the weirs or to
be caught by anglers.
Population Abundance

Assuming that ﬁshing intensity (fy) increases linearly from 1967 to 1985, I used
the model to estimate abundance from 1967 to 1996 (Table 16). Chinook salmon were
ﬁrst stocked into Lake Michigan in 1967; therefore, age-5 ﬁsh do not appear in the

population until 1972. Total recruitment reached a peak in 1989 at 10 million chinook

74
salmon. Recruitment ﬂuctuated between 7.5 million and 10 million from 1986 to 1996,

and was driven by stocking and steady increases in estimated natural reproduction. The
model estimated that the population size was less than 1 million in 1967, and surpassed
10 million by 1978. Population size ﬂuctuated between 13.5 million and 17.9 million
from 1980 to 1996, with a peak abundance in 1990.

Standing stock biomass was estimated from the model’s estimate of abundance-
at-age and the estimated mean weight at annulus formation ﬁ'om the CONNECT model
(Rutherford 1997). Stock biomass increased from 1967 to a peak level of 50 million
pounds in 1986 before high mortality rates on older chinook salmon caused the biomass
to decline beginning in 1987 (Figure 17; Table 17). Stock biomass declined by nearly
50% from 1986 to 26 million pounds by 1993. Natural mortality rates have been
declining in 1995 and 1996, and stock biomass has increased to 39 million pounds in

1996.

75

 

60

 

 

 

 

.., |||||

.. Illlllllllllllll

. Illllllllllllllllllll
.- l||||||||||||||||||l|l|

676971737577798183858789919395
Year

Biomass (millions 01 lbs.)

 

 

 

Figure 17. Standing stock biomass (pounds) as estimated from abundance-at-age
from the CAA model, and mean weight at annulus formation from CONNECT

(Rutherford 1997).

Uncertainty of Parameter Estimates

I estimated 95 % conﬁdence intervals for each parameter by inverting the
likelihood ratio test (Seber and Wild 1989) (Table 18). Uncertainty for the time-varying
mortality parameter (TVM).) was variable across years, with the largest uncertainty
associated with the 1985 and 1996 parameters. High variability for the 1996 parameter
was probably due to the lack of information on mortality in 1997. Estimating uncertainty
about the parameters determining the shape of the mortality function proved difﬁcult.
Nevertheless, the model showed a strong tendency to set age 0-1 mortality to zero and
make age 2-5 mortality equivalent, although this could be accomplished with various

combinations of a and [3. The lower 95% conﬁdence limit on the logistic function

76

parameter (1 resulted in zero TVM for age 0 and a small level of time-varying natural
mortality for age 1 chinook salmon, with essentially equal (and higher) TVM for ages 2-
5. The upper 95% conﬁdence limit went to inﬁnity as mortality of ages 0-1 went to zero
and mortality of ages 2-5 was constant (i.e. a step function). Conﬁdence limits on the
logistic function parameter )3 could not be estimated because the model would increase or
until the denominator in the TVM function (Equation 4) approached zero, causing the

model to crash. I found this to be a limitation with Microsoft Excel©, because

preliminary model runs in GaussTM did not crash.

77
Discussion

Severe declines in the Lake Michigan chinook salmon ﬁshery in the late 1980’s
prompted ﬁshery managers to evaluate chinook salmon management efforts in an attempt
to revive the ﬁshery (Clark 1996). Declines in the chinook salmon ﬁshery were likely
due to density-dependent mortality, as evidenced by an outbreak of BKD and declines in
ﬁshery catch rates, and were probably caused by nutritional stress due to declines in the
alewife population.

This model is a ﬁrst attempt to quantify relationships between different sources of
mortality on chinook salmon during a period of critical and substantial changes in
chinook population dynamics. In order to use this model to make predictions or
projections for the future, I would need to add to the model additional assumptions about
how these population dynamics will operate in the future. One assumption would be that
these rates would remain constant at their current (1996) values, but my analysis of the
past suggests that this is probably not correct (Chapter 2). Accurate forecasts would need
to account for how vital rates change in response to chinook salmon abundance and other
factors, and this would require an improved mechanistic understanding. However,
mortality and abundance estimates made by the model can be used to improve existing
Lake Michigan multi-species models (e.g. Rutherford 1997; Stewart et a1. 1981; Jones et
al. 1993).

I conducted a sensitivity analysis of the age-0 baseline natural mortality rate
(0.75). Age-O was increased and decreased by 25% and the model was re-ﬁt to the data
for both trials. Model output from both trials was compared to the original results to see

if adjusting the natural mortality rate would result in the same qualitative conclusions of

78
chinook salmon population dynamics. Both trials in fact yielded the same qualitative

results as the original model (Table 18; Table 23; Table 24). However, a 25% reduction
in age-0 natural mortality appeared to signiﬁcantly improve model ﬁt, and raised
concerns about the appropriate estimate of natural mortality. Because empirical evidence
(declines in observed weir returns per ﬁsh stocked; Chapter 2) points to an increase in
age-0 mortality in the mid-1980s, simply decreasing the input of natural mortality in
order to improve model ﬁt does not seem justiﬁed.

The results of this model suggest that chinook salmon suffered very high
mortality in the late 1980s and early 1990s, with most of the mortality due to increasing
natural mortality. Ages 2 through 5 were subject to equivalent TVM rates within years,
and all four age groups suffered severe declines in abundance over a year. Consequently,
the mode in the age composition of the harvest shifted from age 3 to age 2.

Initial values for time-varying mortality were obtained from existing estimates of
BKD mortality rates (Rutherford 1997). However, because the model does not ﬁt any
observed data on BKD mortality, inferences about BKD mortality based on model-
estirnated time-varying mortality should be made carefully. In particular, the model
estimated an annual increase in TVM from 1985 to 1993, followed by a decline from
1993 to 1996 (Figure 18). In contrast, observed data show greater levels of BKD
incidence in the late 1980’s, followed by declines in the early 1990’s (Marcquenski 1997;
Clark 1996). This discrepancy between the model estimates of mortality and observed
incidence rates of BKD could mean one of two things. First, observed incidence rates of

BKD are not an index of BKD mortality rates (Clark 1996). Second, causes of mortality

79
rates may be more complex than originally thought, and cannot be estimated by simply

observing one of the symptoms (i.e., BKD incidence).

High in-lake natural mortality rates on age 2-5 chinook are a real problem in Lake
Michigan, and the reduction of natural mortality, similar to reduction of sea lamprey
mortality, should be an immediate management goal If mortality is density dependent
(Clark 1996), then reducing the population density of chinook salmon is a viable method
for reducing mortality and improving the ﬁshery. Because chinook salmon recruitment is
governed by stocking in Lake Michigan, management decisions regarding stocking will

inﬂuence population abundance.

 

[— -n. Salmoninarum -—TVM RE]

 

80 -.I. . - -- w—n 1.80
70 . .. 1.60
-1.40
’3 5° ’
.5 ..F.
(L 50 -
5 T 1.00 i
E 40
1...:
C
g 3° ’ J- 0.60
‘3 20 ' «r 0.40
10 -- - 0.20
o . 0.00

 

 

 

 

Figure 18. Observed estimates of prevalence of Renibacterium salmoninarum, the
causative of agent of BKD, versus model-estimated time-varying instantaneous
natural mortality rate (TVM). Observed data was obtained from mature chinook
salmon sampled at Strawberry Creek, Sturgeon Bay, Wisconsin (Marcquenski
1997).

CHAPTER FOUR

CONCLUSIONS

The chinook salmon population in Lake Michigan underwent dramatic changes
between 1986 and 1996. These changes were most directly felt by the sport ﬁshery, as
harvest and harvest rates for chinook salmon began declining in 1987, triggering a
decline in sport ﬁshery effort, which led to a cycle of further declines in harvest. Greatest
declines in the ﬁshery were seen in the Michigan waters of the lake along the eastern
shoreline, where chinook salmon harvest declined by 95%. Complete collapse of the
entire salmonine sport ﬁshery, however, was avoided. The ﬁshery that was once
dominated by chinook salmon harvest was able to diversify and maintain high harvest
rates by targeting other salmonine species. Part of the reason for spatial differences in
trends in the chinook salmon ﬁshery was due to changes in the spatial distribution of
chinook salmon, as evidenced by spatial differences in harvest rate trends. It is likely that
chinook salmon concentrated in the western regions of the lake in response to spatial
changes in the distribution of alewives, their primary forage.

The sport ﬁshery was not the only place where changes in the chinook salmon
population were felt. Dead chinook salmon washed up on the southern Lake Michigan
shoreline in the late 1980’s, suggesting an increase in lake-wide natural mortality. While

most of these ﬁsh ultimately died from BKD, it is likely that another environmental or

80

8 1
nutritional stress affected their resistance to disease. Regardless of the cause, modeling

results show that this increase in lake-wide natural mortality was a signiﬁcant source of
mortality in older (age 2-5) chinook salmon, accounting for upwards of 70% of the deaths
in some years. The increased mortality of older chinook salmon could not be explained
by overﬁshing, which is often to blame for ﬁshery collapses. Because of increased
natural mortality of older ﬁsh, ﬁshery harvest declined, and ﬁshery and weir harvest age
compositions shifted towards younger ages, as fewer older ﬁsh survived to be harvested.
Finally, estimated standing stock biomass declined by about 50% from peak levels in the
mid-1980’s, to the early-1990’s. The population appears to be recovering in recent years
as harvest and harvest rates are increasing, and age compositions are slowly shifting
towards older ﬁsh. Preliminary analysis of 1997 ﬁshery data show further increases.
While the initial objective of this study was to build a spatial model of chinook
salmon population dynamics in Lake Michigan, I soon realized that there was much more
work to be done compiling the necessary data than was previously thought. A spatial
analysis of trends in the chinook salmon ﬁshery grew out of the need to re-estirnate the
time series of harvest and effort from the Michigan sport ﬁshery, and showed that there
were distinct spatial differences in ﬁshery trends that may be explained by chinook
movements. Coded wire tagging studies support this possibility, and both studies could
be combined to further understand chinook salmon movements. More important, theories
about movements could lay the groundwork towards a spatial model A spatial model of
chinook salmon population dynamics could estimate spatial differences in mortality rates

and abundance, and would be useful in determining different stocking scenarios.

82
What is the future of chinook salmon in Lake Michigan? It seems that as the

ﬁshery grew in the 1970’s and 1980’s, chinook salmon became an indispensable species
in the ﬁsh community. Because chinook salmon were originally stocked in part to
control alewives, it is ironic to think that as the chinook ﬁshery collapsed, the commercial
ﬁshery for alewives was reduced. It seems that the success of the salmonine ﬁshery in
Lake Michigan has dictated ﬁshery management goals. Fish community objectives, as
outlined by the Great Lakes Fishery Commission (Eshenroder et a1. 1995), have called
for a diverse salmonine ﬁsh community capable of sustaining an annual yield of 6 to 15
million pounds. Included is a short-term goal of annual yields of chinook salmon of
about 6.8 million pounds. Also included is the goal of an increased reliance upon
naturally reproduced salmonines. With what appears to be a recovery in the chinook
salmon population in recent years, their popularity in the ﬁshery, and their ability to
naturally reproduce, chinook salmon will continue to be an integral part of the Lake
Michigan ﬁsh community.

Management of chinook salmon in the next 10 years will continue to be
challenged by a number of issues, and several unanswered questions remain. For
example, chinook salmon health continues to be monitored by measuring BKD incidence
rates in surveys and in the weirs, although it is not clear what the relationship is between
incidence rates and mortality rates. Reducing stocking rates could alleviate mortality
rates (Clark 1996; Keller et al. 1990), but exactly how many ﬁsh should be stocked and in
what species combinations is an area for further research. The answer is complicated by

uncertainty surrounding estimates of forage abundance.

APPENDIX:

ADDITIONAL TABLES

Table 1. Number of salmonine ﬁngerlings stocked in Lake Michigan, by species,

 

 

 

1963 to 1996.

Year Brook trout Brown trout Chinook salmon Coho salmon Lake trout Rainbow trout m
1963 0 0 0 0 0 0 0
1964 0 0 0 0 0 0 0
1965 0 0 0 0 0 0 0
1%6 20.0111 16.3111 0 0 0 81.299 117.599
1967 0 1 2.540 802 .390 0 569.600 74.695 1 .459.225
1%8 0 1 72.400 686.692 0 0 0 859.092
1969 0 57.2111 717 .585 0 0 22.2111 796.985
1970 0 94.540 1 903.492 0 0 362.088 2.360.120
1971 0 531.804 2.215.198 0 208.1110 702.579 3,657,581

197 2 9.980 722.740 2.032.128 0 405.400 465.832 3.636.080
197 3 0 1.313.842 3045.767 313 .700 301010 1.532.270 6,505,579
1974 4.1110 469.300 3.578.053 0 260.250 1.261.815 5.573.418
1975 0 82.647 4275.782 156.200 149.1110 894.151 5.557.690
1976 61.29!) 387 .922 3 302.057 352.728 0 392.669 4.496.“
1977 524.772 362.200 2.818.561 0 47.5111 143.661 3.896.694
197 8 30.0111 854.217 5.365.263 0 65.0111 1 284,753 7,599,263

1979 0 663 .947 5.184.271 511.506 120.271 1.667.085 8,147,080
1980 2.560 753 .074 6.105.924 244.486 268.700 1.620.094 8.994.838
1981 89.070 578.440 4.747.799 101 .953 560.500 847.700 6.925.462
1982 193.477 1.516.793 6.146.427 245.581 707.347 1.410.712 10.220337

1983 210.035 1.578.114 6.291.913 127.555 31.480 1.709.163 9,948,260
1984 85,481 1 .149. 1 78 7.709.792 439.704 445.920 1.755.442 11.585517

1985 130.739 1.127.110 5.955.523 139.018 1.158.423 631.128 9.141.941

1986 25.460 719.318 5.692.678 246.352 822.600 629,729 8,136,137

1987 53 .277 81 l .485 5.800.757 299.429 24.984 378.371 7.368.303

1988 135.050 783.652 5.416.870 939.153 623.600 371.960 8.270.285

1989 6010 753 .140 7 859.479 608.324 3 371.122 536.978 13.135.043

1990 208.700 936.747 7 .128.723 1.206.152 0 418,722 9.899.044
1991 203010 639.296 6.237.562 815.515 0 654.428 8.549311

1992 109.700 765.382 5.795.465 1.225.339 673 .621 385.399 8954316
1993 142.300 869915 5.529.950 130.105 0 417.558 7.089.818
1994 1 19.400 1 244.853 5.892.950 710.082 1.357.821 874,559 10.199.665
1995 271 .932 1.014.458 6.590.976 1 030.639 0 287 .990 9.195.995
1996 105.330 816.765 6.193.377 1 021.630 143 .629 345 .336 8.626.157

 

84

Table 2. Number of salmonine yearlings stocked in Lake Michigan, by species, 1963
to 1996.

 

 

 

Species
Year Brook trout Brown trout Coho salmon Lake trout Rainbow trout Total
1963 0 0 0 0 9.200 9.200
1964 0 O 0 0 15.000 15.000
1965 0 O 0 1.273.878 24.830 1.298.708
1966 29.240 21.700 659.356 1.766.190 194.290 2.670.776
1967 32.809 35.935 1.732.298 1.854.820 40.230 3,696,092
1968 49.481 79.190 1.183.872 1.875.900 389.349 3.577.792
1969 33.518 84.377 3.237.856 1.999.805 409.454 5,765,010
1970 49.500 129.820 3.535.930 1.960.000 294.189 5,969,439
1971 93.048 177.31 1 2.743.046 2.135.545 665.849 5.814.799
1972 94.782 203.469 2.619.908 2.520.120 850.220 6,288,499
1973 50. 150 598.953 2.265.257 2.209. 150 1.546.452 6.669.962
1974 30.250 363.358 3.230.972 2.137.100 905.888 6,667,568
1975 61.300 425.345 2.368.691 2.428.424 734.928 6.018.688
1976 25.820 653.188 2.843.671 2.547.800 1,473,445 7,543,924
1977 98.480 793.525 3,088,218 2.370.100 1,058,108 7,408,431
1978 218.225 655.202 2.658.941 2.474.400 651,767 6,658,535
1979 192.970 548.202 3.832.337 2.376.601 865.394 7.815.504
1980 205.000 554.564 2.698.884 2.522.600 1.040.] 19 7.02 1 . 167
1981 1 19.397 591.242 2.349.478 2,081,530 1.094.020 6,235,667
1982 51.226 642.821 1.934.960 2.038.790 1.1 16.517 5.784.314
1983 87.403 670.682 2.2368 17 2.209.590 1.016.864 6.221.356
1984 147.561 653.768 2.514.343 1.1 19,140 1.360.818 5.795.630
1985 185.226 670.437 2.519.665 2.623.399 1,193,695 7,192,422
1986 17 1.436 7 14.7 35 2.045.045 2.474.406 1.671.942 7.077.564
1987 79.000 529.684 2.005. 142 1.973.350 1.447.628 6.034.804
1988 361.936 761.627 2.243.742 1.922.628 1.058.959 6.348.892
1989 144.100 750.835 1.725.601 2.005.600 1.308.187 5,934,323
1990 191.448 841.024 1.173.901 1.317.115 1.181.337 4,704,825
1991 123.100 743.983 1.655.396 2.779.482 1.320.495 6.622.456
1992 162.720 849.225 1.516.87 1 2.761.244 1,437,414 6,727,474
1993 151.794 888.817 1.578.646 2.697.835 1.422.809 6.739.901
1994 149.185 927.527 761.291 2.545.512 1.376.435 5,759,950
1995 56.025 861.602 1.367.189 2.264.428 1,762,601 6.31 1.845
1996 69.464 969.98 1 2,075,803 1.97 1.448 1,499,149 6,585,845

 

Table 3. Total number of salmonines stocked in Lake Michigan, by species, from
1986-1996. Includes ﬁngerlings, yearlings. and lake trout fry.

 

 

 

Year Brook trout Brown trout Chinook salmon Coho salmon Lake trout Rainbow trout “'13:?
1953 0 0 0 0 0 9.200 9.200
1%4 0 0 0 0 0 150(1) 15.000
1965 0 0 0 0 1 273.878 24.830 1 298.708
1956 49.240 38.0“) 0 659.356 1.766.190 275.589 2.788.375
1967 32.809 48.475 802.390 1.732.298 2.424.420 1 14.925 5.155.317
1968 49.481 251.590 686.692 1.183.872 1.875910 389.349 4.436.884
1969 33.518 141.577 717.585 3.237.856 1.999.805 431.654 6.561.995
1970 49.51!) 224.360 1 .‘X13 .492 3 .53 5.930 1.9601110 656.277 8.329.559
1971 93 .048 709.115 2.215.198 2.743.046 2.343.545 1.368.428 9,472,380
197 2 104.762 926.209 2.032.128 2.619918 2.925.520 1.316.052 9.924.579
1973 50.150 1.912.795 3.045.767 2.578.957 2.509.150 3.078.722 13,175,541
1974 34.250 832.658 3.578.053 3.230.972 2.397.350 2.167.703 12.240.986
1975 61.3111 507.992 4.275.782 2.524.891 2.577.424 1.628.989 11,576,378
1976 87.110 1.041.110 3.302.057 3.196.399 2.547.800 1.866.114 12,0405”
1977 623.252 1.155.725 2.818.561 3.088.218 2.417.600 1.201.769 11,305,125
1978 248 .225 1.509.449 5.365.263 2.658.941 2.539.400 1.936.520 14.257 .798
1979 192,970 1.212.149 5.184.271 4.343.843 2.496.872 2.532.479 15.962.584
1980 207.560 1.307.638 6.105.924 2.943.370 2.791.300 2.660.213 16.016.005
1981 208.467 1.169.682 4.747.799 2.451.431 3.142.030 1.941.720 13,661.13
1982 244.703 2.159.614 6.146.427 2.180.541 3.176.137 2.527.229 16,434,651
1983 297.438 2.248.796 6.291.913 2.364.372 2.541.070 2.726.027 16.469.616
1984 233.042 1.802.946 7.709.792 2.954.047 2,195,130 3.116.260 18,011,147
1985 315.965 1.797.547 5.955.523 2.658.683 5.081.822 1.824.823 17,634,363
1986 196.896 1.434.053 5.692.678 2.291.397 4.197016 2.301.671 16,113,701
1987 132.277 1.341.169 5.800.757 2.304.571 3.298.334 1.825.999 14.703.107
1988 496.986 1.545.279 5.416.870 3.182.895 2.546.228 1.430.919 14,619,177
1989 150.100 1 .503 .975 7.859.479 2.333.925 5.376.722 1 .845.165 19,069,366
1990 “11.148 1.777.771 7.128.723 2.380.053 1.317.115 1.600.059 14.603.869
1991 326.100 1.383.279 6.237.562 2.470.91 1 2.779.482 1.974.923 15.172.257
1992 272.420 1.614.607 5.795.465 2.742.210 3.434.865 1,822,813 15,682,331
1993 294.094 1.758.722 5.529.950 1,708,751 2.697.835 1.840.367 13.829.719
1994 268.585 2.172.380 5.892.950 1.471.373 3913.333 2,250,994 15,959,615
1995 327.957 1.876.1h0 6.590.976 2.397.828 2.264.428 2.050.591 15.507 .840
1996 174.794 1.786.746 6.193.377 3.097.433 2.115.077 1.844.485 15.211.912

 

86

.33 3 33 E8.— ..552 .3 ...auEﬂE 8.3 E coo—8% muﬁtowE— 55.8 3.855 ..c ..anaz .9 935,—.

87

1

 

 

55m.m5~.e 555.85 58.58 58.584 555.85 8m£mﬁ~ 58.85 8.3m 5555
555.536 35.555 555.35 5mm.mm5.~ mmmﬁme m3 .584 8.55 «55.nom m55—
5m5.a55.m 555.85 55325 58.5554 555.55m 25.5.3.5 $5.555 555.5% v55—
5m5.5nm.m 35.58 555. 555 555.com; 35.5mm 85.35; 58.35 555.5% 5.55—
m5¢.m55.m N2 .85 N55. 35 35.58; 5mw.m5v Soﬁa; 5.3.85 55 .5mm N555
«5053.5 Qinvoé 23.555 8.584 m 3.555 85.584 5m5.m55 58.8w ~55~
35.8 _ .5 35.85 555.3 _.~ 55 @354 85.555 85.28; 58.555 58.3 m 555—
5515525 85.55 a; 58.3 5." 555.5654 25.8 5; a55.NN_._ 555.35 $5.58 555_
555.53% 3 5.5m5 m3 .mmv 58.53; 53 .85 3985 53.85 58.5mm 553
5m5.555.m 58.565 86.85 58.58; 85.584 555.85 555.6% 58.5w; 5555
555.5556 35.5mm 55085 3: .m am; 85.85; we“ .95 555.85 58.nmm 5555
mmm.mm5.m 35.5% 85.59 .5 85.584 85.354 85.58 N _5. Ev 5m5.m5m 855
555.5555 «8.35 85.52.“ 85.85; 856mm; 55— .2mm.— 55~ .5mm 58.58 #55—
m 55. ~5N6 85. 55 58. 5 E 85.5 5 m4 858.5 5.3.555.— m5v.m5m 5556mm 855
53.53 .5 531.655 4 85.584 555.864 55m .555 M5m._5~.~ 5.55.85 555.33V ~53
555.35.“. 555655 58 .mN5 3 m.m~ 5; 85 .5mm ~m~.555 85.5w 555.5 5 5555
155.85 .5 ~mm.m55 58.555 555. _ 95.5 58.555 35.534 5m" .5m _ 555.655 5555
55%...» ﬁn 555.25 25.85 8.8a; 58.85 555.85.“ 85.5w 58.m5m 5555
m5~.m5m.m 58.3.5 53. ~55 N5_.m5m._ 58.85 N588.— 58.8_ 586mm 5555
Gm.” 5 5N 555.85~ 85.9w 35.655 5863.. www.mdm m55.m~ 53.55m 5555
5m5.m5m.m 555.55— 8. EN 555 .85 58.555 58.55...“ 555.8 55.8w 5555
N55.m5~.v 53.5mm m5m.8m v8.58; 555.86 3N.:5 $5.an 58.5mm m55—
mm5.w5m.m _55.m 55 8.8m 58.555 8.5mm Quinn 55m. SN 555.53 V53
85.5.5.5” _~m.m5~ 58.55— 55m.:w _.~ 8.3m 53.85 85.85 5m5.~3 m555
5N _ .Nm5.~ 559mm 8&5 _ 36.85 58.555 55N.55m 5 wumiﬁ N55~
w5— .SNN 8.555 85.55 Nike; 58.63 515.58 5 35.9: 53
55v.m55.~ 58.5— 5 5 58.85 8.5: 555.59. 38.8 58.8_ 5555
www.555 5 5 mam. Rm 85.8 8.8m 5 5 553
8.555 5 5 555.35 5 N 55. Rm 5 5 83
55m.8 5 5 58.53 5 5mm. ~5m 5 5 505—
th. «55.54855 583585 $85.55 «83582 385.82 5.82 5am :85 5%

 

Siam

 

88

 

 

 

 

In...

2525 54.555 255.5525 5554555 5552
552 .5 555.5N5 53452.5 5555555 5552
452 .5 555.555 555.5555 5555555 5552
552 .5 555.555 555.5555 53.355 5554
552 .5 555.555 252.5555 25.855 N552
5525 555.555 55542.». 5554555 2552
NE .5 552.555 55543.5 5555555 5552
552 .5 552.354 55553.5 53.55.52 5552
52 2.5 555.5554 53.5555 55.53.52 5552
53 .5 555.5554 35.5555 55555552 5552
m5— .5 52 5.5554 53.5555 555452 .3 5552
93m 3952 59953.5 3.92:: 539—. comm 59.9th 5.65m 30.5 39>

woﬂsgm

 

$995.5 859:3 95:.9:_ 8: 8.5 .5552 3 5552 Sc...— 93: 69.59: 5395...:
95:55.3 5:: .5822— 9:_::E_8 .33 5.8.59 95:259. 59.9595 5:5 29.59 E955 :25 .33 5529.2 99.3 .5 9.55.5

89

 

 

Eﬁwnw 3mg: "3.3— %deN mawﬁcm wamﬁe 3m 83
2&er wnmdm _ Swag 2N. — ma 3m.m- mmcéh cmvg mag
antmcw $5.02 0362 maﬁa ~86»: @369 8nd $3
wmmde com. _ S 03:52 cSﬁaN mg; mg 39.3 new; mac—
Ehdg E 552 35.02 nwwﬁvm SHED §.Ne v36 83
ammdww 3&3; Emdvm mmeéﬂ mmmﬁmn 3an cam; 33
o2 .mmw Nada weNJmm 3 Ndmm ands wads 3%.». 83
@269..— Nmm. d m“ mwaé‘nm 3 ~63. 53.— mm comma 8.3 $3
mmndcoé mad: @9ch v5.3a mmmﬁmm Gowns «96 $3
mmmdmv; ”atom. emcﬁmm Nomwﬁ m c2 .89 wﬁvdo mg." 53
3 ”Kg.— owaﬁe vi. _ ﬂ m Nmeémm m _ ad; amad: mcwm £3
38. :55 33qu Bob 33 8.53 230 88:8 £885 :55 E85 59: 485 30>

 

888m

52.3 58...... 2: 2.22: .2. 8.5 .83 3 £3 55.7.: 5% 595...: 3.3 2: 3 .85.. 0585.8 a 2.5

 

9O

 

 

 

 

 

Emma; 2133 $53 mango 83mm omnawm 83: $3: $2
$32.", 33% ”~58 83 K M863. 89m? $33 :32 mg
838m 53.85 23% 388 «$.94 83% 833 83.2 32
$335 33% :5? $33 23.3. 3.3% 822 on?! 83
5.8% $33 :3.on Buds was? 52% 83M: mag 8.:
832+ $255 $32. 3%me 2”.va 8&9 $32 23.8 5:
833m 833 23.9% 2.28; E3? 812: 3m.§ 33mm 82
$3.33 23% £33. :33: £33. 3.8a $32 «858 $2
$333 $38; 8334 232d 233 83.8 5.32 «8.93 ”we
mmméamﬁ 856mm.“ Zmﬁsl Stun: Manama 353m; Sﬁmvn 93mm $3
@858.” 8;»? RES; 5.32 can? $384 :33 83% one
38. “Sasasaa .woaaam .553m 52,582 .8382 582 55 82c .8»
Sang
stoic

58.5. 25.2.. 8: 8.5 .33 3 33 E9... £33.. .3 S35: 95% :«uEﬂ—Z 8.3 2: :8: 9.8% ”5:253 .h 035,—.

91

 

 

 

 

mawﬁom 3m6~ mg. mm 50.3 awed» 32K Nave». mm.» 33
www.mam ”3.2 I: .2. wwmém S mdm v3.3 93A” main mae
~86? E; H mamdm m—Ndn $5.3 $0.3 Eta a2; 33
wNa. g 2 am g .m acém maﬁg 893. 3: Km 256 new mag
Sodm d m d QQ emmﬁm $8.3 93.? mmm. ~ m Smd $3.2 Nag
ammNmN mdeN v3.3 813 39$ 39% mag. Smdm 53
amde “3.2 now. g v wmadv 3%? 3a.? wmvd v8.2 6&3
Ra; mm 95.3 maodm Rad» www.mw 3 w.mm mead Rm. 3. $3
mNmem 35.8 H mcdm v3.3 89: ﬁve.» 39$ Smdm £3
3— due www.mm Smﬁn— wu— .anﬁ magma: 2 Mama www.mm mmoém $3
maﬁa 3%? @362 $3.53” 3 m. :: 35.2% and mowén £3
38. «amvﬂémoﬂa .moaagm “manna—5m 503582 .m 335.62 6.52 ham 80.5 30>
8. um

 

stoic 559.3 25.2: 8: 8.5 .eaau 8 33 5.2.2.. 95% gum—.32 8—3 a... 3 “meta: 5.5—3 €2.30 .w cash.

92

 

 

wood vmod Sod need 2 g .o :36 emod omod 82
536 m 56 awed Rod mad mmcd muod owed mag
weed Sod cued omod ~86 #86 Sad v86 33
89: ~86 >86 mgd cmcd sued Sad 28.: 83
95.6 n 5.: god Ecd god ~36 35.: Sad N22
$56 086 v3.6 god mad «8.: med $96 32
$56 v86 ﬁ 86 636 536 Rod 98.: 036 Gag
amod Sod $56 935 $36 weed $56 «8.: $3
25.: cued v86 #86 omcd mood $86 086 $3
mead awed wood mood $8.6 $86 98.6 Sod 3%
god Rad Sod S a .o 2b.: 53 .o de need 33
023.834 0.333.065:— amoaﬂam 335:8 3035.52 3005.52 5.52 hi 80.5 30>

 

Baum

50%: 500.5
25.2: .2. 8.5 .33 3 £2 $.28 5% 50.5.2 3.3 2: 5 8.2 22:2 2305 3:53 55.8 032.26 a cash

 

 

 

 

0.0 2 0.0 00 0.2 3 0.0 0.0 000.
0.0 3 0.0 0.. 0.0 0.0 0.0 0.0 30.
0.0 0.. 0.0 3 0.0 ..m 0.0 0.0 002
0.0 E 2 3 0.0 ...m 0.0 0.0 000.
0.0 0.. 0.0 0.0 0.0 3. 3 0.0 000.
3 2 3. 0... 0.0 0.0 0._ 0.0 000.
0.0 0.0 0.0 0.0 0.0 «.0 0.. 0.0 32
0.0 0.0. 0.0 0.0 0.0 3 0.0 0.0 003
85-93 80.63835 09.580 .2380 083552 08552 582 :0 820 £6 a...»
8.000

 

200:0: E0030 0.5.0:. .0: 0000. 5.2.00.

93

22.0 =0u._.0..>. 00.04 0... 00.. 5.5.00. 0.00:...0 ..0 600.080 000.5... 00.. .0030... 0.00.. 08:0: 30.0.0000 00.05.03. .2 030...

 

 

 

 

0.0.00. .010 000.00 000.0. 000.00 30.00 000.0 000.0. 000.
£0.03 0.0.0. 30.00 000.0. 80.0.. 000.00 000.0 000.0. .00.
.840. 05.0. 08.00 000.00 08.0.. 000.00 80.0 000.0. 000.
000.3: 000.0. 80.00 000.00 000.... 000.00 02.0 00...: 000.
000.000 03.0. 08.00 000.; 000.0... 08.? 0.0.0 000.0. 000.
000.000 000.3 000.; 000.00 20.00 .0000 .006 000.00 000.
000.000 000.0. 000.00 000.00 30.00 000.00 02.0. 80.00 000.
.0300 08.00 000.00 03.0. . 000.0 000.00. 000.0. 03.00 000.
.30... 08.03-0605. 083580 000580 083582 300582 582 00m 08.0 $06 08>
830
020:0:

E0050 0.5.0:. .0: 800. .0005: 0.2.0 =0u.:0..>. 3.0a. 0... .8. 005.00 0.0055 ..0 000.000.. 9.0.0.0000 00.05.03. .3 0.0.0...

94

Table 12. Comparison of model predicted vs. observed targeted effort and total
chinook salmon harvest. No lake-wide observed data are available prior to 1986.

 

 

 

 

Effort Harvest
Year Observed Predicted Observed Predicted
1967 299,325 96
1968 598,649 835
1969 897,974 4,961
1970 1,197,299 16,804
1971 1,496,623 29,321
1972 1,795,948 47,133
1973 2,095,273 84,892
1974 2,394,597 123,256
1975 2,693,922 159,299
1976 2,993,247 220,913
1977 3,292,572 293,890
1978 3,591,896 348,764
1979 3,891,221 364,416
1980 4,190,546 411,812
1981 4,489,870 535,138
1982 4,789,195 638,076
1983 5,088,520 714,634
1984 5,387,844 767,788
1985 7,874,006 5,687,169 810,849 859,907
1986 9,006,345 8,128,072 948,915 1,07 6,21 1
1987 8,186,326 7,494,890 680,126 777,502
1988 6,820,940 5,813,923 357,325 466,441
1989 5,388,977 5,534,408 351,937 329,065
1990 3,661,795 3,710,544 220,399 215,425
1991 3,919,424 4,174,903 252,589 225,712
1992 2,929,243 3,070,331 158,097 145 .432
1993 3,030,855 3,134,258 131,928 124,445
1994 2,868,956 2,966,548 136,921 129,493
1995 2,795,866 3,058,135 225,564 192,744
1996 2,881,261 3,010,768 303,893 282,749

 

 

 

 

Table 1

_YEar_
1967
1968
1969
1970
197 l
1972
1973
1974
197.
197
197

Table 13. Estimated annual ﬁshing mortality (Paw).

 

 

 

Age
Year 0 1 2 3 4 5
1967 0.000
1968 0.000 0.002
1969 0.001 0.003 0.015
1970 0.001 0.004 0.020 0.052
1971 0.001 0.005 0.025 0.064 0.088
1972 0.001 0.006 0.030 0.077 0.104 0.1 11
1973 0.001 0.007 0.034 0.089 0.121 0.128
1974 0.001 0.008 0.039 0.101 0.137 0.145
1975 0.002 0.009 0.044 0.113 0.153 0.162
1976 0.002 0.011 0.049 0.125 0.168 0.178
1977 0.002 0.012 0.054 0.136 0.183 0.194
1978 0.002 0.013 0.058 0.148 0.198 0.210
1979 0.002 0.014 0.063 0.159 0.213 0.225
1980 0.003 0.015 0.068 0.170 0.227 0.240
1981 0.003 0.016 0.072 0.181 0.241 0.255
1982 0.003 0.017 0.077 0.192 0.255 0.270
1983 0.003 0.018 0.082 0.202 0.268 0.284
1984 0.003 0.019 0.086 0.213 0.282 0.298
1985 0.004 0.020 0.091 0.223 0.295 0.311
1986 0.005 0.028 0.127 0.303 0.393 0.413
1987 0.005 0.026 0.118 0.283 0.369 0.388
1988 0.004 0.020 0.093 0.228 0.300 0.317
1989 0.003 0.019 0.088 0.218 0.288 0.305
1990 0.002 0.013 0.060 0.152 0.204 0.216
1991 0.003 0.015 0.067 0.169 0.226 0.240
1992 0.002 0.011 0.050 0.128 0.172 0.182
1993 0.002 0.011 0.051 0.130 0.175 0.186
1994 0.002 0.010 0.048 0.123 0.167 0.177
1995 0.002 0.011 0.050 0.127 0.171 0.182
1996 0.002 0.011 0.049 0.125 0.169 0.179

 

 

Table

m
196'
1981
198
198
198
195
19‘.
19‘.
19
19
19

96

Table 14. Estimated annual instantaneous natural mortality rates.

 

 

 

Age
Year 0 1 2 3 4 5
1967-85 0.750 0.300 0.100 0.100 0.100 0.100
1986 0.750 0.300 0.397 0.397 0.397 0.397
1987 0.750 0.300 0.503 0.503 0.503 0.503
1988 0.750 0.300 0.763 0.763 0.763 0.763
1989 0.750 0.300 1.182 1.182 1.182 1.182
1990 0.750 0.300 1.031 1.031 1.031 1.031
1991 0.750 0.300 1.523 1.523 1.523 1.523
1992 0.750 0.300 1.616 1.616 1.616 1.616
1993 0.750 0.300 1.801 1.801 1.801 1.801
1994 0.750 0.300 1.285 1.285 1.285 1.285
1995 0.750 0.300 0.658 0.658 0.658 0.658
1996 0.750 0.300 0.394 0.394 0.394 0.394

 

 

Tabb 15

Year
1967
1968
1969
1910
1971
1972
1973
1974
197!
197
197
19'
19'
19

Table 15. Estimated total annual mortality (A).

 

 

 

Age
Year 0 1 2 3 4 5
1967 0.529
1968 0.529 0.277
1969 0.529 0.278 0.227
1970 0.530 0.279 0.230 0.578
1971 0.530 0.279 0.234 0.583 0.897
1972 0.530 0.280 0.238 0.589 0.899 1.000
1973 0.530 0.281 0.242 0.594 0.900 1.000
1974 0.530 0.282 0.246 0.600 0.902 1.000
1975 0.530 0.282 0.249 0.605 0.904 1.000
1976 0.530 0.283 0.253 0.610 0.906 1.000
1977 0.530 0.284 0.257 0.615 0.907 1.000
1978 0.530 0.285 0.261 0.620 0.909 1.000
1979 0.530 0.285 0.264 0.625 0.911 1.000
1980 0.530 0.286 0.268 0.630 0.912 1.000
1981 0.531 0.287 0.272 0.635 0.914 1.000
1982 0.531 0.288 0.275 0.640 0.916 1.000
1983 0.531 0.288 0.279 0.645 0.917 1.000
1984 0.531 0.289 0.283 0.650 0.919 1.000
1985 0.531 0.290 0.286 0.654 0.920 1.000
1986 0.532 0.296 0.491 0.769 0.949 1.000
1987 0.531 0.294 0.537 0.787 0.952 1.000
1988 0.531 0.290 0.633 0.823 0.959 1.000
1989 0.531 0.290 0.757 0.882 0.973 1.000
1990 0.530 0.285 0.709 0.851 0.964 1.000
1991 0.530 0.286 0.824 0.911 0.979 1.000
1992 0.530 0.283 0.836 0.915 0.979 1.000
1993 0.530 0.283 0.864 0.929 0.983 1.000
1994 0.530 0.283 0.772 0.881 0.971 1.000
1995 0.530 0.283 0.573 0.777 0.946 1.000

 

 

Tab‘

NCI‘

4 I
‘I‘ 1 ‘Il CIII

Table 16. Model estimated abundance-at-age. Age-0 abundance is equivalent to

 

 

 

recruitment.
C

Year 0 1 2 3 4 5 Total
1967 802.390 0 0 0 0 0 802.390
1968 686.692 377.697 0 0 0 0 1.064.389
1969 717.585 323.177 273.070 0 0 0 1.313.831
1970 1.913.492 337.653 233.405 211.215 0 0 2.695.765
1971 2.265.198 900.210 243.602 179.632 89.179 0 3.677.822
1972 2.102.128 1.065.474 648.774 186.543 74.843 9.213 4.086.974
1973 3.245.767 988.588 767.065 494.330 76.695 7.591 5.580.037
1974 3.978.053 1.526.136 710.959 581.541 200.553 7.637 7.004.879
1975 4.875.782 1.870.105 1.096.384 536.312 232.818 19.606 8.631.006
1976 4.002.057 2.291.707 1.342.070 822.924 211.874 22.346 8.692.978
1977 3.618.561 1.880.692 1.642.890 1.002.297 320.807 19.965 8.485.212
1978 6.165.263 1.700.160 1.346.812 1.220.827 385.571 29,679 10,848,313
1979 5.984.271 2.896.177 1.216.240 995.811 463.432 35.021 11.590.952
1980 7.305.924 2.810.634 2.069.638 894.774 373.021 41,326 13,495,317
1981 6.247.799 3.430.740 2.006.382 1.515.002 330.745 32.658 13,563,325
1982 7.646.427 2.933.319 2.446.454 1.461.358 552.608 28.429 15.068.594
1983 7.791.913 3.589.304 2.089.529 1.772.982 525.998 46.634 15,816,360
1984 9.229.792 3.656.919 2.554.109 1.506.746 629.732 43.580 17.620.879
1985 7.475.523 4.330.946 2.599.468 1.832.548 528.099 51.224 16.817.808
1986 7.692.678 3.507.131 3.075.331 1.855.469 633.700 42.167 16,806,476
1987 7.800.757 3.603.561 2.468.933 1.565.936 427.794 32.366 15.899346
1988 7.616.870 3.655.622 2.542.507 1.143.267 334.139 20,440 15,312,845
1989 10.059.479 3.573.163 2.594.625 933.360 202.614 13.644 17.376.885
1990 9.876.937 4.719.836 2.538.605 629.375 110.149 5.536 17,880,438
1991 8.643.262 4.639.423 3.374.990 738.005 93.629 3.913 17.493.222
1992 8.030.679 4.058.771 3.312.037 595.667 65.802 1.978 16.064.934
1993 7.729.950 3.773.688 2.908.863 542.438 50.815 1.355 15,007,109
1994 8.092.950 3.632.229 2.703.935 395.358 38.334 866 14,863,671
1995 8.790.976 3.803.194 2.604.125 617.832 47.190 1.106 15.864.424
1996 8.393.377 4.130.990 2.725.821 1.112.095 137.478 2.535 16.502.296

 

 

Table
from
(Rut!

99

Table 17. Standing stock biomass (pounds) as estimated from abundance-at—age
from the CAA model, and mean weight at annulus formation from CONNECT
(Rutherford 1997).

 

 

 

Age
Year 1 2 3 4 5 Total
1967 0 0 O 0 0 0
1968 453.237 0 0 0 0 453,237
1969 387,812 1,556,497 0 0 0 1,944,310
1970 405,184 1,330,407 3,062,616 0 0 4,798,206
1971 1,080,252 1,388,530 2,604,671 1,426,872 0 6,500,325
197 2 1,278,569 3,698,010 2,704,880 1, 197 .483 176,351 9,055,293
1973 1,186,306 4,372,272 7. 167 .789 1,227 , 124 145,304 14,098,795
1974 1,831,363 4,052,469 8,432,350 3,208,851 146,188 17,671,220
1975 2,244,126 6,249,386 7,776,523 3,725,091 375,309 20,370,434
1976 2,750,049 7 .649,801 1 1,932,392 3,389,984 427,752 26,149,977
1977 2,256,831 9,364,473 14,533,312 5,132,909 382,180 31,669,705
1978 2,040,192 7 676,829 17,701,993 6. 169, 140 568,132 34,156,287
1979 3,475,413 6,446,070 10,854,342 6,580,737 630,376 27,986,937
1980 3,372,761 10,969,081 9,753,040 5,296,894 743,869 30,135,646
1981 4,1 16,888 10,633,823 16,513,517 4,696,580 587 .839 36,548,647
1982 3,519,982 12,966,204 15,928,801 7,847 .028 511,723 40,773,738
1983 4,307,165 1 1,074,504 19,325,505 7 .469. 176 839,409 43,015,758
1984 4,388,303 13,536,777 16,423,534 8,942,200 784,434 44,075,248
1985 5,197,135 13,777,181 19,974,778 7 ,499,012 922,027 47,370,133
1986 4,208,557 16,299,253 20,224,616 8,998,544 759,010 50,489,981
1987 4,324,273 13,085,343 17,068,700 6,074,676 582,585 41,135,577
1988 4,386,746 13,475,286 12,461,609 4,744,777 367 .921 35,436,339
1989 4,287,796 14,789,362 13,533,716 3,241,821 261,182 36,113,877
1990 5,663,803 14,470,050 9,125,936 1,762,384 105,964 31,128,137
1991 5,567,307 19,237,444 10,701,077 1,498,058 74,902 37,078,789
1992 4,870,525 18,878.61 1 8,637,176 1,052,834 37 .860 33,477,005
1993 4,528,426 16,580,517 7,865,353 813,045 25,940 29,813,281
1994 4,358,675 15,412,428 5,732,688 613,344 16,577 26,133,711
1995 4,563,833 14,843,515 8,958,561 755,037 21,181 29,142,127
1996 4,957,188 15,537,178 16,125,380 2,199,645 48,533 38,867,924

 

 

 

Tabl
Metl

1.;
WA

100

Table 18. Parameters and 95% conﬁdence intervals as estimated by the model. See
Methods for a discussion of conﬁdence interval estimates.

 

 

Parameter Symbol Lower 95% Estimate Upper 95%
TVM 71935 0.000 0.000 0.052
71936 0.086 0.297 0.507
71937 0.192 0.403 0.615
Y1988 0.426 0.663 0.904
11.989 0.833 1.082 1.336
71990 0.675 0.931 1.193
7,991 1.163 1.423 1.689
Y1992 1.255 1.516 1.782
71993 1.440 1.701 1.969
71994 0.941 1.185 1.433
71995 0.277 0.558 0.838
71996 0.000 0.294 0.747
a 7.859 45.014 00
5 * 1.281 *
Fishing Intensity f1985 0.288 0.378 0.511
f1986 0.418 0.541 0.710
f1937 0.381 0.499 0.661
f1988 0.295 0.387 0.514
f1989 0.281 0.368 0.490
f1990 0.188 0.247 0.329
fml 0.211 0.278 0.371
f1992 0.155 0.204 0.273
f1993 0.158 0.209 0.279
171994 0.150 0.197 0.264
f1995 0.154 0.203 0.273
f1.”6 0.151 0.200 0.269
Selectivity a 1.646 1.789 1.942
[3 2.385 2.609 2.863
Maturation a 1.805 1.913 2.024
6 2.898 2.984 3.072
C atchabiliry C oeﬁ‘icient q 5.142E-08 6.653E-08 8.747E—08

 

 

 

Table 1
logistic

—

Matur

Selec:

121111

g
191
191

101

Table 19. Estimated maturation and ﬁshery selectivity. Values were estimated by
logistic functions, with parameters estimated by the CAA model.

 

Age-0 Age-1 Age-2 Age-3 Age-4 Age-5

Maturation (MATa ) 0.00 0.02 0.13 0.51 0.87 1.00

Selectivity (S a) 0.01 0.05 0.25 0.67 0.92 0.99

Table 20. Observed and predicted ﬁshery harvest age compositions.

Observed Fishery Harvest Age Composition

 

 

 

 

 

Year A e-0 A e-l A e-2 A e-3 A e-4 A e-5
1985 0.00 0.13 0.33 0.43 0.10 0.00
1986 0.00 0.07 0.22 0.54 0.17 0.00
1987 0.00 0.14 0.27 0.35 0.23 0.01
1988 0.00 0.18 0.24 0.35 0.22 0.01
1989 0.00 0.22 0.32 0.29 0.17 0.01
1990 0.00 0.30 0.37 0.25 0.07 0.00
1991 0.00 0.36 0.33 0.24 0.06 0.00
1992 0.01 0.45 0.28 0.20 0.07 0.00
1993 0.00 0.33 0.33 0.32 0.02 0.00
1994 0.00 0.39 0.36 0.24 0.02 0.00
1995 0.00 0.29 0.43 0.26 0.02 0.00
1996 0.00 0.21 0.30 0.35 0.13 0.00
Predicted Fishery Harvest Age Composition
Year A e-0 e-l A e-2 A e-3 A e-4 A e-S
1985 0.02 0.08 0.26 0.45 0.17 0.02
1986 0.02 0.08 0.29 0.41 0.18 0.01
1987 0.03 0.10 0.28 0.43 0.15 0.01
1988 0.04 0.13 0.32 0.36 0.14 0.01
1989 0.07 0.18 0.35 0.31 0.09 0.01
1990 0.07 0.24 0.39 0.24 0.06 0.00
1991 0.06 0.25 0.42 0.23 0.04 0.00
1992 0.07 0.25 0.44 0.20 0.03 0.00
1993 0.08 0.28 0.42 0.20 0.03 0.00
1994 0.07 0.25 0.48 0.18 0.02 0.00
1995 0.06 0.18 0.46 0.28 0.03 0.00
1996 0.04 0.13 0.38 0.39 0.07 0.00

 

102

Table 21. Observed and predicted ﬁshery mature harvest age compositions.

Observed Mature Harvest Age Composition

 

 

 

 

 

Year Age-0 Age- 1 Age-2 Age-3 Age-4 Age-5
1985 0.00 0.00 0.15 0.73 0.12 0.00
1986 0.00 0.00 0.00 0.41 0.59 0.00
1987 0.00 0.00 0.00 0.29 0.71 0.00
1988 0.00 0.00 0.02 0.40 0.57 0.01
1989 0.00 0.00 0.06 0.60 0.31 0.03
1990 0.00 0.13 0.29 0.42 0.16 0.00
1991 0.00 0.11 0.28 0.50 0.11 0.00
1992 0.00 0.08 0.19 0.49 0.25 0.00
1993 0.00 0.16 0.19 0.62 0.03 0.00
1994 0.00 0.04 0.23 0.66 0.07 0.00
1995 0.00 0.02 0.24 0.68 0.06 0.00
1996 0.00 0.02 0.24 0.68 0.06 0.00
Predicted Mature Harvest Age Composition
Year A e—0 A e-l A e-2 A e-3 A e-4 e-5
1985 0.00 0.00 0.08 0.53 0.35 0.04
1986 0.00 0.00 0.09 0.50 0.38 0.03
1987 0.00 0.01 0.09 0.54 0.33 0.03
1988 0.00 0.01 0.12 0.51 0.34 0.02
1989 0.00 0.01 0.16 0.54 0.27 0.02
1990 0.00 0.02 0.22 0.53 0.21 0.01
1991 0.00 0.03 0.26 0.55 0.16 0.01
1992 0.00 0.03 0.30 0.53 0.14 0.00
1993 0.00 0.03 0.30 0.54 0.12 0.00
1994 0.00 0.03 0.35 0.50 0.11 0.00
1995 0.00 0.02 0.26 0.61 0.11 0.00

1996 0.00 0.01 0.16 0.64 0.18 0.00

 

 

103

Table 22. Observed and predicted weir harvest age compositions.

Observed Weir Harvest Age Composition

 

 

 

 

 

Year A e-0 A e-l A e-2 e-3 A e-4 e-S
1985 0.00 0.06 0.17 0.60 0.16 0.01
1986 0.00 0.02 0.07 0.62 0.29 0.00
1987 0.00 0.12 0.12 0.40 0.33 0.03
1988 0.00 0.18 0.16 0.53 0.12 0.00
1989 0.00 0.23 0.13 0.39 0.25 0.00
1990 0.00 0.29 0.22 0.34 0.14 0.00
1991 0.00 0.35 0.43 0.21 0.01 0.00
1992 0.00 0.28 0.43 0.27 0.02 0.00
1993 0.00 0.28 0.46 0.25 0.02 0.00
1994 0.00 0.21 0.68 0.11 0.00 0.00
1995 0.00 0.20 0.45 0.35 0.01 0.00
1996 0.00 0.20 0.45 0.35 0.01 0.00
Predicted Weir Harvest—Age Composition
Year A e-0 A e-l A e-2 A e-3 A e-4 e-5
1985 0.01 0.06 0.21 0.48 0.22 0.02
1986 0.01 0.05 0.24 0.45 0.23 0.02
1987 0.02 0.07 0.24 0.47 0.19 0.02
1988 0.02 0.09 0.28 0.41 0.19 0.01
1989 0.04 0.13 0.32 0.37 0.13 0.01
1990 0.04 0.18 0.37 0.31 0.09 0.01
1991 0.04 0.20 0.40 0.30 0.06 0.00
1992 0.05 0.20 0.43 0.27 0.05 0.00
1993 0.05 0.22 0.41 0.27 0.04 0.00
1994 0.05 0.19 0.47 0.24 0.04 0.00
1995 0.04 0.13 0.43 0.36 0.04 0.00
1996 0.02 0.09 0.32 0.47 0.09 0.00

 

 

 

104

Table 23. Parameter estimates from a sensitivity analysis on age-0 baseline natural
mortality. Age-0 natural mortality was increased by 25% from an intitial value of
0.75 to 0.94.

 

 

 

 

Parameter Value Parameter Value
Fishing Intensity TVM
f1985 0.462 71935 0.002
f1986 0.597 71986 0.003
f1987 0.548 71937 0.457
f1988 0.434 71938 0.590
f1989 0409 71989 1-021
flggo 0.274 7,990 0.871
fl”, 0.311 7199, 1.384
f1992 0.228 71992 1.472
f1993 0.233 71993 1.651
f1...)4 0.220 71994 1.145
f1995 0.226 71995 0.495
8996 0.223 71996 0.239
a 31.325
Selectivity 6 1.183
a 1.784
5 2.576 C atchability C oefﬁcient
q 7.406E-08
Maturation
a 1 909

[3 2.978

105

Table 24. Parameter estimates from a sensitivity analysis on age-0 baseline natural
mortality. Age-0 natural mortality was decreased by 25% from an intitial value of

0.75 to 0.56.

 

 

 

 

Parameter Value Parameter Value
Fishing Intensity TVM
f1985 0309 71985 0000
f1986 0.456 7,936 0.338
f1987 0.420 71987 0.418
f1988 0.325 7,988 0.678
11989 0.310 7,939 1.087
f1990 0.207 71990 0.932
f1991 0.233 7199, 1.420
11992 0.172 71992 1.510
f1993 0.175 71993 1.695
f1.”4 0.166 71994 1.177
f1995 0.171 71995 0.555
8996 0.168 71996 0.291
01 41.752
Selectivity [3 1.468
a 1.755
[3 2.642 Catchability Coeﬂicient
q 5.592E-08
Maturation
a 1.910
6 2.947

 

 

LIST OF REFERENCES

106

LIST OF REFERENCES

Associated Press. 1989. Skippers reel under warning, cancer risk report staggers lake’s
sport ﬁshing industry. Detroit News. July 27, 1989.

Austen. D., W. Brofka. J.E. Marsden. J. Francis. J. Palla, J.R. Bence. R. Lockwood, J.
Rakoczy, K. Smith, and B.T. Eggold. 1995. Lake Michigan creel survey
methods. Prepared for the Lake Michigan Technical Committee, July 20. 1995.

Avery, EL. 1974. Reproduction and recruitment of anadromous salrnonids in Wisconsin
tributaries of Lake Michigan. Wisconsin Department of Natural Resources.
Federal Aid in Fish Restoration, F-83-R. Study 108. Final Report, Madison.

Bence, J.R., J.A. Clevenger Jr.. P. Gelderblom. and D. Wesander. 1996. Marking of
chinook salmon. Pages 40-58 in Fisheries Research and Management Report,
Michigan Waters of Lake Michigan. 1995. Great Lakes Fishery Commission,
Lake Michigan Annual Committee Meeting, Duluth, Minnesota. March 20-21.
1996.

Bence. J .R. and RM. Meehan. unpublished. Simple model development for Lake
Huron. Project Progress Report. Michigan State University. July 10, 1996.

Bence, JR. and K.D. Smith. In press. An overview of sport ﬁsheries of the Great
Lakes.

Burzynski, T. 1996. Wisconsin’s Lake Michigan salrnonid stocking program.
Wisconsin Department of Natural Resources. April, 1996. 38p.

Burzynski, T. and R. Multhauf. 1995. Wisconsin’s Lake Michigan salmonid stocking
program. Wisconsin Department of Natural Resources. April, 1995. 37p.

Brandt. S.B., D.M. Mason, E.V. Patrick, R.L. Argyle. L. Wells, P.A. Unger. and DJ.
Stewart. 1991. Acoustic measures of the abundance and size of pelagic
planktivores in Lake Michigan. Canadian Journal of Fisheries and Aquatic
Research 48: 894-908.

Braun, K. 1987. A creel survey of the Indiana waters of Lake Michigan. 1986. Fisheries
Section, Indiana Department of Natural Resources, Division of Fish and Wildlife.

 

107

Brofka, W.A, and J .E. Marsden. 1997. A survey of sport fishing in the Illinois portion
of Lake Michigan. Annual report to the Illinois Department of Natural Resources.
F-52-R11. Aquatic Ecology Technical Report 97-7.

Campbell, B. 1989. Report of toxins cuts into ﬁsh sales. dealers face trouble over
disputed study of Lake Michigan trout. Detroit Free Press. July 6, 1997.

Carl, L.M. 1980. Aspects of the population ecology of chinook salmon in Lake
Michigan tributaries. Ph.D. Dissertation, The University of Michigan, Ann
Arbor.

Carl, L.M. 1982. Natural reproduction of coho salmon and chinook salmon in some
Michigan streams. North American Journal of Fisheries Management. 4:375-
380.

Carl, L.M. 1984. Chinook salmon density. growth, mortality, and movement in two
Lake Michigan tributaries. Canadian Journal of Zoology. 62:65-71.

Clapp, DR 1997. Chinook salmon population dynamics in Michigan’s waters of the
Great Lakes. Pages 21-24 in Fisheries Research and Management Report.
Michigan Waters of Lake Michigan, 1996. Michigan Department of Natural
Resources, Fisheries Division.

Clark. R.D. Jr. 1996. Status of chinook salmon in the Upper Great Lakes, 1995.
Michigan Department of Natural Resources, Fisheries Division. Administrative
Report. June 12, 1996.

Coshun, M. 1991. Wisconsin’s Lake Michigan salmonid stocking program. Wisconsin
Department of Natural Resources. August, 1991.

Eggold, B.T.. S. Hogler. P. Peeters. 1997. Wisconsin’s 1996 weir harvest. Pages 5-8 in
Lake Michigan Management Reports, Wisconsin Department of Natural
Resources. Presented to: Great Lakes Fishery Commission, Lake Michigan
Committee, 1997 Annual Meeting, March 20—21, 1997.

Eck, G.W. and L. Wells. 1987. Recent changes in Lake Michigan’s ﬁsh community and
their probably causes, with emphasis on the role of the alewife (A losa
pseudoharengus). Canadian Journal of Fisheries and Aquatic Sciences 44(Suppl.
2): 53-60.

Elliott, RF. 1993. Feeding habits of chinook salmon in eastern Lake Michigan.
Michigan State University, Department of Fisheries and Wildlife. MS. Thesis.
108 pp.

 

 

108

Elliott, RF. 1994. Early life history of chinook salmon in Lake Michigan. Pages 286-
308 in Anonymous, editor. Michigan sport ﬁsh restoration program annual
reports for projects F-35-R-19 and F-53-R-10, April 1, 1993 to March 31, 1994.
Michigan Department of Natural Resources, Fisheries Division, Lansing. 322 p.

Eshenroder, R.L., M.E. Holey, T.K. Gorenﬂo, and RD. Clark, Jr. 1995. Fish-
community objectives for Lake Michigan. Great Lakes Fishery Commission,
Special Publication 95-3. 56 p.

Fournier, D., and GP. Archibald. 1982. A general theory for analyzing catch at age data.
Canadian Journal of Fisheries and Aquatic Sciences 39: 1195-1207.

Hagar, J.M. 1984. Diets of Lake Michigan salmonids: an assessment of the dynamics of
predator-prey interaction. MS. Thesis, University of Wisconsin, Madison, 97 pp.

Hansen, M.J. 1986. Size and condition of trout and salmon from the Wisconsin waters
of Lake Michigan, 1969-84. Wisconsin Department of Natural Resources,
Bureau of Fisheries Management, Fish Management Report 126, Madison.

Hansen, M.J. 1988. Wisconsin’s Great Lakes planting program. Wisconsin Department
of Natural Resources, 101 South Webster Street, Box 7921, Madison, Wisconsin
53707

Hansen, M.J., P.T. Schultz, and BA. Lasee. 1990. Changes in Wisconsin’s Lake
Michigan salmonid sport ﬁshery, 1969-1985. North American Journal of
Fisheries Management 10:442-457.

Hansen, M.J., P.T. Schultz, and B.A. Lasee. 1991. Wisconsin’s Lake Michigan
salmonid sport ﬁshery 1969-85. Wisconsin Department of Natural Resources.
Fisheries Management Report Number 145.

Hesse, 1A 1994. Contribution of hatchery and natural chinook salmon to the eastern
Lake Michigan ﬁshery, 1992-1993. Michigan State University. MS. Thesis.

Hilborn, R. and C]. Walters. 1992. Quantitative Fisheries Stock Assessment: Choice,
Dynamics, and uncertainty. Routledge, Chapman & Hall, Inc., New York, NY.

Holey, ME. 1995. Summary of trout and salmon stocking in Lake Michigan 1976-1994.
Pages 45-54 in: Moore, J. (ed.). Great Lakes Fishery Commission, Lake
Michigan Committee 1995 Annual Meeting, Minutes. GLFC, Ann Arbor,
Michigan.

Holey, M.E., R.W. Rybieki, G.W. Eck, E.H. Brown Jr., J.E. Marsden, D.S. Lavis, M.L.
Toneys, T.N. Trudeau, R.M. HorralL 1995. Progress toward lake trout
restoration in Lake Michigan. Journal of Great Lakes Research 21(1): 128-151.

 

 

109

Holey, ME. 1996. Summary of trout and salmon stocking in Lake Michigan 1976-1995.
Pages 43-52 in Lake Michigan Committee 1996 Annual Meeting. March 20-21,
1996, Duluth, Minnesota.

Johnson, DC, and J .G. Hnath. 1991. Lake Michigan chinook salmon mortality — 1988.
Michigan Department of Natural Resources, Fisheries Technical Report 91-4,
Ann Arbor.

Jonas, J.L., and BF. Clapp. 1998. Assessment of chinook salmon and steelhead
populations in eastern Lake Michigan. Pages 12-20 in Fisheries Research and
Management Report, Michigan Waters of Lake Michigan, 1997. Michigan
Department of Natural Resources, Fisheries Division.

Jude, J .J ., and F]. Tesar. 1985. Recent changes in the inshore forage ﬁsh of Lake
Michigan. Canadian Journal of Fisheries and Aquatic Sciences 42: 1154-1157.

Keller, M., K.D. Smith, and R.W. Rybieki (eds). 1990. Review of salmon and trout
management in Lake Michigan. Fisheries Special Report No. 14. Michigan
Department of Natural Resources. Charlevoix Fisheries Station. Box 205.
Charlevoix, Michigan 49720. April 1990. 254 p.

Kitchell, J .F. and LB. Crowder. 1986. Predator-prey interactions in Lake Michigan:
model predictions and recent dynamics. Environmental Biology of Fishes 16(1-
3):205-2l 1.

Koonce, IF. and M.L. Jones. 1994. Sustainability of the intensively managed ﬁsheries
of Lake Michigan and Lake Ontario. Final Report of the SIMPLE Task Group.
Great Lakes Fishery Commission. Ann Arbor, MI.

Kope, R.G. 1987. Separable virtual population analysis of paciﬁc salmon with
application to marked chinook salmon, Oncorhynchus tshawytscha, from
California’s Central Valley. Canadian Journal of Fisheries and Aquatic Sciences.
44: 1213-1220.

Krueger, C.C., D.L. Perkins, E.L. Mills, and J .E. Marsden. 1995. Predation by alewives
on lake trout fry in Lake Ontario: role of an exotic species in preventing
restoration of a native species. J oumal of Great Lakes Research 21 (Supplement
1): 458-469.

Law, AM. and W.D. Kelton. 1982. Simulation modeling and analysis. McGraw-Hill,
New York, NY.

Lockwood, RN. 1997. Evaluation of catch rate estimators from Michigan access point
angler surveys. North American Journal of Fisheries Management 17:611-620.

 

110

Lockwood, R.N., D.M. Benjamin, and J .R. Bence. In review. Estimating angling effort
and catch from Michigan roving and access site angler survey data. Michigan
Department of Natural Resources, Fisheries Division, Fisheries Research Report.

Lychwick, T]. 1985. Status of the chinook salmon coded wire recoveries-1984. Report
to the Great Lakes Fishery Commission, March 1985. Wisconsin Department of
Natural Resources, Sturgeon Bay.

Malvestuto, SP. 1983. Sampling the recreational ﬁshery. Pages 397-419 in LA.
Nielsen and D.L. Johnson, editors. Fisheries Techniques. American Fisheries
Society, Bethesda, Maryland.

Marcquenski, S. 1996. Lake Michigan ﬁsh health update. State of Wisconsin
Department of Natural Resources. Lake Michigan Management Reports.
Prepared for the Great Lakes Fishery Commission, Lake Michigan Committee,
Annual Meeting, March 20-21, 1996.

Marcquenski, S., J. Hnath, R. Horner, B. Eggold, G. Boronow, and P. Peeters. 1997.
Effect of spring epizootics on the Lake Michigan chinook sport ﬁshery. 59th
Midwest Fish and Wildlife Conference, Milwaukee, Wisconsin, December 7-10,
1997.

Mason, D.M. and SB. Brandt. 1996. Effect of alewife predation on survival of larval
yellow perch in an embayment of Lake Ontario. Canadian Journal of Fisheries
and Aquatic Sciences 53: 1609-1617.

Megrey, BA 1989. Review and comparison of age-structured stock assessment models
from theoretical and applied points of view. Pages 8-48 in: Edwards, BF. and
BA. Megrey(eds.). Mathematical Analysis of Fish Stock Dynamics. American
Fish. Soc., Bethesda, Maryland.

Mertz, G. and RA. Myers. 1996. An extended cohort analysis: incorporating the effect
of seasonal catches. Canadian Journal of Fisheries and Aquatic Sciences.
53: 159- 163

Methot, RD. 1990. Synthesis model: an adaptable framework for analysis of diverse
stock assessment data. Pages 257-277 in: L. Low (ed.) Proceedings of the
symposium on applications of stock assessment techniques to Gadids.
International North Paciﬁc Fisheries Commission. Bull. 50.

Michigan Sea Grant. 1998. Anglers share views on Great Lakes sport ﬁshery.
Upwellings 20(1): 4-5.

National Wildlife Federation. 1989. Lake Michigan sport ﬁsh, should you eat your
catch? National Wildlife Federation, Lake Michigan Sport Fish Consumption
Advisory Project. Washington, DC. 16pp.

111

Nelson, DD. and J .G. Hnath. 1990. Lake Michigan chinook salmon mortality - 1989.
Michigan Department of Natural Resources, Fisheries Technical Report 90-4,
Ann Arbor.

Palla, J. ‘1997. Indiana’s Lake Michigan creel survey results 1996. Indiana Division of
Fish & Wildlife. March 17, 1997.

Peeters, P. 1993. Salmonid diet study status report. Wisconsin Department of Natural
Resources, Sturgeon Bay, Wisconsin.

Rutherford, B. 1997. Evaluation of natural reproduction, stocking rates, and ﬁshing
regulations for steelhead Oncorhynchus mykiss, chinook salmon 0. tschawytscha,
and coho salmon in Lake Michigan. Michigan sport ﬁsh restoration program
annual reports for projects F—35-R-22 and F-53-R-13, April 1, 1996 to March 31,
1997. Michigan Department of Natural Resources, Fisheries Division, Lansing.

Rybieki, R.W., and DR Clapp. 1996. Diet of chinook salmon in eastern Lake Michigan,
1991-93. State of Michigan Department of Natural Resources, Fisheries Division.
Research Report No. 2027. 22p.

Seber, G.A.F. and OJ. Wild. 1989. Nonlinear Regression. John Wiley and Sons, New
York, NY.

Seelbach, P.W. 1985. Smolt migration of wild and hatchery-raised coho and chinook
salmon in a tributary of northern Lake Michigan. Michigan Department of
Natural Resources. Fisheries Research Report No. 1935.

Seelbaeh, P.W. 1986. Population biology of steelhead in the Little Manistee River,
Michigan. Ph.D. Dissertation, The University of Michigan, Ann Arbor.

Sitar, S. 1996. Estimation of lake trout (Salvelr'nus namaycush) abundance and mortality
due to sea lampreys (Petromyzon marinas) and ﬁshing in the main basin of Lake
Huron, 1984-1993. Michigan State University. MS. thesis.

Smith, S.H., H.J. Buettner, and R. Hile. 1961. Fishery statistical districts of the Great
Lakes. Great Lakes Fishery Commission. Technical Report No. 2. September,
1961.

Smith, SH. 1968. Species succession and ﬁshery exploitation in the Great Lakes.
Journal of the Fisheries Research Board of Canada 25(4): 667-693.

Stewart, DJ. and M. Ibarra. 1991. Predation and production by salmonine ﬁshes in
Lake Michigan, 1978-88. Canadian Journal of Fisheries and Aquatic Science
48:909-922.

 

 

112

Stewart, D.J., J .F. Kitchell, and LB. Crowder. 1981. Forage ﬁshes and their salmonid
predators in Lake Michigan. Transactions of the American Fisheries Society
1102751-763.

Tody, W.I-I. and HA. Tanner. 1966. Coho salmon for the Great Lakes. Fish Manage.
Rep. No. 1. Fish Div. Mich. Dept. Nat. Res., Lansing, MI. 38 p.

Toneys, M. 1992. Salmonid diet study status report. Wisconsin Department of Natural
Resources, Sturgeon Bay, Wisconsin.

Wells, L., and AL. McLain. 1973. Lake Michigan: man’s effects on native ﬁsh stocks
and other biota. Great Lakes Fishery Commission. Technical Report No. 20. 56

p.

Wesley, J .K. 1996. Age and grth of chinook salmon in Lake Michigan: veriﬁcation,
current analysis, and past trends. State of Michigan Department of Natural
Resources, Fisheries Division. Research Report No. 2029. 93p.

Zafft, DJ. 1992. Migration of wild chinook and coho salmon smolts from the Pere
Marquette River, Michigan. MS. Thesis, Michigan State University, East
Lansing.

"111111111"11111111111“