FACTORS AFFECTING PRODUCTIVITY AND HARVEST RATES OF GREAT LAKES
MALLARDS
By
Howard V. Singer

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Fisheries and Wildlife—Master of Science
2014

ABSTRACT
FACTORS AFFECTING PRODUCTIVITY AND HARVEST RATES OF GREAT LAKES
MALLARDS
By
Howard V. Singer
Mallard (Anas platyrhynchos) populations in the Great Lake States; Michigan,
Minnesota, and Wisconsin (hereafter GLS), are monitored as part of a larger population known
as the Mid-continent mallard population. Regulations regarding harvest of Mid-continent
mallards in the mid-1990s became more liberal than at any time in the past 50-years. Waterfowl
managers of the GLS were concerned that liberalized regulations would lead to declining mallard
populations throughout the GLS. Resource managers in Michigan and Wisconsin adopted
regulations that were more restrictive than allowed under federal regulations for harvest of
female mallards. Research has suggested that declining mallard abundance throughout the GLS
could be linked to declining productivity and excessive harvest under liberalized regulations.
This research suggests that productivity rates of GLS mallards have remained relatively steady
from 1961-2011 despite environmental variation on the breeding grounds. I also found that kill
rates during the more liberal hunting seasons were lower than during years when regulations
were more restrictive, there were fewer mallards, and there were more duck hunters. Further
analyses suggest regulation changes at the state level will have less of an impact on harvest rates
than will regulation changes at the flyway level. My results suggest that the decline of GLS
mallard abundance is likely not linked to excessive harvest or declining productivity.

This work is dedicated to a dear friend, Al Macon, who unfortunately passed away while
this work was in progress. During my youth, Al used to pick me up from school and take me
duck hunting. It is probably worth mentioning that Al was not a duck hunter and had never
hunted ducks before taking me out; he was in his early 30s. We spent countless hours together in
the marsh during my childhood, and we both got our first duck together although in different
years. Al’s patience with me as a youth helped me become the person I am today and as cliché
as it sounds, I wouldn’t be here today if it weren’t for Al. Al was there when I bagged my first
duck and will be with me when I bag my last duck. Thanks for all the great memories Al. You
will always be missed, you were a great person.

iii

ACKNOWLEDGEMENTS

I would first like to thank my parents for being supportive of my hobbies as a kid even if
they did not always please everyone. My mom probably still has a stash of all the duck calls she
stole from me for blowing them too loudly in the house. Although my dad was not a hunter nor
did he own any guns, he always encouraged me to do what made me happy and he spent quite a
few days sitting in a duck marsh watching me hunt, mostly hoping that I did not shoot any ducks.
Thanks also to my brother Sam and sister Becky, they both have always encouraged me to do
what I love in life and have helped me throughout my life.
I would also like to thank the Michigan Department of Natural Resources for funding this
project as well as Michigan State University for working collaboratively with the MIDNR. I am
also thankful to all those who supplied me with data over the past two years and offered me ideas
and suggestions on completing my thesis.
I am forever indebted to Chris Nicolai, Mandy Van Dellen, and Jim Sedinger; all helped
me get to where I am at today. Chris had faith in me as a potential waterfowl biologist and, I am
pretty sure demanded that Mandy hire me to work as a research assistant for her. She hired me
and then introduced me to Jim, her advisor, who later advised me through the daunting task of
writing a manuscript that we eventually published. It was a great feeling to be Jim’s first
undergraduate student to publish a manuscript in a peer-reviewed journal. Thanks guys.
I would also like to say a big thanks to my committee members Drs. Scott Winterstein,
Chuck Nelson and in particular Dave Luukkonen for lengthy discussions about mallards among
other things over the past two and a half years. I gained a wealth of knowledge from all of our
discussions and travels that I couldn’t have got anywhere else. I appreciated the conversations
iv

that Chuck and I had regarding mallards and also Chuck trying to help me figure out the
direction that my career could go in after my tenure at MSU. And also thanks to Scott for
keeping an eye out for me and also spending countless hours making sure I was keeping up with
everything. This wouldn’t have been possible without the help from you all.
And lastly, if it weren’t for the ducks and waterfowl conservation, I would not be living
the life I am now; filled with days in the marsh listening to whistling wings as ducks pass by.

v

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................................ vii
LIST OF FIGURES ..................................................................................................................... viii
CHAPTER 1
THESIS INTRODUCTION .............................................................................................................1
Research objectives and thesis organization ........................................................................1
CHAPTER 2
CLIMATIC FACTORS INFLUENCING PRODUCTIVITY RATES OF MALLARDS
BANDED IN THE GREAT LAKE STATES .................................................................................2
Introduction ......................................................................................................................................2
Study area and methodology................................................................................................5
Results ................................................................................................................................10
Productivity rate estimates .....................................................................................10
PHDI model results ................................................................................................12
POND model results ..............................................................................................16
Discussion ..........................................................................................................................17
Management Implications ..................................................................................................22
LITERATURE CITED ......................................................................................................24
CHAPTER 3
FACTORS AFFECTING HARVEST RATES OF FEMALE MALLARDS BANDED IN THE
GREAT LAKE STATES ...............................................................................................................31
Introduction ........................................................................................................................31
Study area and methodology..............................................................................................35
Duck hunter analysis ..............................................................................................35
Kill rate analysis ....................................................................................................38
Results ................................................................................................................................42
Duck hunter analysis ..............................................................................................42
Kill rate analysis ....................................................................................................45
Discussion ..........................................................................................................................50
Duck hunter analysis ..............................................................................................50
Kill rate analysis ....................................................................................................52
Management Implications ..................................................................................................57
LITERATURE CITED ......................................................................................................60

vi

LIST OF TABLES

Table 1. Candidate models used to model productivity rates for both the (a) PHDI and (b)
POND data sets. In the POND data set I was interested in the effects of density
dependence on productivity and thus have included the covariate MALL in all models ....9
Table 2. Performance of models used to explain variation in productivity rates for the PHDI (34
random years) and POND (1968-2011, 1973-2011, and 1991-2011 for MN, WI, and MI,
respectively) data sets for each state (a) Michigan; (b) Minnesota; and (c)Wisconsin. The
number of model parameters (K) including an intercept, Akaike’s information criterion
adjusted for small sample size (AICc), AICc compared with the best model (ΔAICc), and
the likelihood of each model in a set (Akaike weight, Wi) are reported ............................12
Table 3. Number of years that a given state had various season lengths and daily hen mallard
limit combinations. Season lengths were categorized as follows: 1= <38 days; 2=≥ 38
days <50 days; 3=≥50 days. The limit was the total number of hen mallards allowed per
day by each state. Years with season limits of four were omitted from the analyses .......40
Table 4. A priori model set considered to predict kill rates for hen mallards in Michigan,
Minnesota, and Wisconsin from 1961-2011. All models included the covariate HUNT,
AGE, and an intercept. I used the standardized average of duck hunters from Michigan
and Wisconsin for those states respectively and used the standardized average of flyway
duck hunters for Minnesota ...............................................................................................40
Table 5. A priori model set considered to predict the influence on kill rates for hen mallards in
Minnesota from 1997-2011. Note that season length (variable DAYS) was not included
due to a constant season length of 60 days from 1997-2011 .............................................41
Table 6. A priori models used to determine the most predictive estimate of hunter numbers at
varying spatial scales to include in my kill rate analyses ..................................................44
Table 7. A priori models and model weights for predicting hen mallard kill rates in the GLS: (a)
Michigan; (b) Minnesota; and (c) Wisconsin. Model weights, wi, that sum to ~95% were
used to model average kill rate estimates for each state. All models contained an
intercept and the variable HUNT as the number of waterfowl hunters likely influences
kill rates. Also, the variable AGE was in all models as research has shown that juveniles
are more vulnerable to harvest than are adults...................................................................46
Table 8. Percent change in model-averaged predicted kill rates when regulations change from 1hen to 2-hens daily by state ................................................................................................50

vii

LIST OF FIGURES

Figure 1. Mean productivity rates for mallards in the 3 GLS (± 95% CI) 1961-2011 .................10
Figure 2. Annual productivity rates for mallards in the 3 GLS: (a) Michigan; (b) Minnesota; and
(c)Wisconsin with long-term average (solid horizontal line) and 95% CI (dashed
horizontal lines) .................................................................................................................11
Figure 3. Model-averaged parameter estimates ( ̂ ) and 95% CI from standardized predictor
variables across all models for the PHDI model set. For interpretation of the references
to color in this and all other figures, the reader is referred to the electronic version of this
thesis ..................................................................................................................................15
Figure 4. Model-averaged parameter estimates ( ̂ ) and 95% CI from standardized predictor
variables across all models for the POND model set .........................................................16
Figure 5. Acreage of land in the GLS that was enrolled in the CRP program from 1986-2013 ..20
Figure 6. Mallard population estimates for the GLS (red line) and the MC (blue line; not
including the GLS) from 1961-2011. Surveys began in 1968, 1973, and 1991 for
Minnesota, Wisconsin, and Michigan, respectively. Michigan supports the most breeding
mallards of the GLS which explains the large increase in 1991 when Michigan began
surveying mallard abundance ............................................................................................34
Figure 7. Federal duck stamp sales for the Great Lake States (Michigan= dashed line,
Minnesota=solid, thin line, Wisconsin=dotted line) and the entire Mississippi flyway (MS
flyway; thick, gray line) from 1961-2011 ..........................................................................42
Figure 8. Relationship between the estimate of Mississippi Flyway waterfowl hunters (y-axis)
and the Mid-continent mallard population (x-axis) from 1961-2011 (1961-1996 circles
and 1997-2011 squares). Variation explained by the relationship, r2, for 1961-1996 and
1997-2011 was 0.58 and 0.18, respectively .......................................................................43
Figure 9. Kill rates for female mallards banded in: (a) Michigan, (b) Minnesota, and (c)
Wisconsin, 1961-2011 (AHY=solid line and HY=dashed line) ........................................45
Figure 10. Model weighted predictions for: (a) Michigan, (b) Minnesota, and (c) Wisconsin
female mallard kill rates from top 2 ranked models with model weighted 95% CI ..........48
Figure 11. Mean kill rate for MN female mallards (AHY=adult and HY=juvenile) during years
when MN adopted a 1-hen daily limit (2005-2010) and a 2-hen daily limit (1997-2004
and 2011) and 95% CI .......................................................................................................49

viii

Figure 12. Relationship between the population estimate for MC mallards (not including GLS
mallards) and the Mississippi flyway estimate of duck hunters as derived from the duck
stamp sale from 1961-2011. Note how the trend between the two “breaks” around
1995………………………………………………………………………………………53

Figure 13. Annual mallard harvest per hunter as estimated from the Parts Collection Survey
harvest data for both male and female and juvenile and adult mallards and the estimate of
state waterfowl hunters from duck stamp sales for (a) Michigan , (b) Minnesota, and (c)
Wisconsin 1961-2011 ........................................................................................................54

ix

CHAPTER 1 THESIS INTRODUCTION
Research objectives and thesis organization
The purpose of this research is to understand how productivity and kill rates of mallards
breeding from 1961-2011 in the Great Lake States (Michigan, Minnesota, and Wisconsin;
hereafter GLS) are influenced by environmental factors, waterfowl hunter numbers, and
regulations. This research will address the needs of both state and federal waterfowl
management plans and may have important management implications for future waterfowl
harvest management strategies at multiple management scales.
In this research I will: 1) estimate annual mallard productivity rates, 2) model how
weather during the breeding season influences productivity rates and test for density dependence,
3) estimate annual female mallard kill rates, 4) model how hunter numbers at varying spatial
scales influence kill rate and finally 5) model how regulations at varying management scales and
hunter numbers impact kill rates for the GLS.
Chapter 2 will focus on estimating mallard productivity rates and modeling productivity
rates over the last fifty-one years (1961-2011) and also during years when states conducted
waterfowl and wetland surveys (Michigan 1992-present; Minnesota 1968-present; and
Wisconsin 1973-present). Chapter 3 will focus first on estimating kill rates and second,
modeling kill rates as a function of duck hunter numbers at varying spatial scales and finally
modeling kill rates as a function of regulations and waterfowl hunter numbers. There may be
some duplication of text in Chapters 2 and 3 as they are organized as separate manuscripts for
submission to scientific journals.

1

CHAPTER 2 CLIMATIC FACTORS INFLUENCING MALLARD PRODUCTIVITY
RATES IN THE GREAT LAKE STATES

Introduction
Changing climatic patterns, such as increased drought (Johnson et al. 2005), are likely to
have irreversible, negative impacts to wetland conditions, such as permanency and productivity
(Peterson et al. 1997; McCarty 2001; Hoegh-Guldberg and Bruno 2010). Wetlands throughout
North America have been lost due to a variety of anthropogenic factors (e.g. loss of habitat due
to roads, agricultural expansion, deforestation; Krapu 1974; Boyd 1985) and now more so,
environmental factors (Dahl 1990; Prince et al. 1992). With expected increases in global
temperatures (IPCC 2013), wetlands will likely continue to be lost. Johnson et al. (2005) found
that under different climate change scenarios, favorable habitat for breeding waterfowl will
likely shift to less productive areas of the northern prairies. Climate prediction models for the
upper mid-west suggest summer temperatures and precipitation will likely increase and decrease
respectively, thus exacerbating the rate of wetland loss. About 70% of wetlands in the Great
Lakes Region have been lost since historical times (Snell 1987).
An important function of wetlands is as wildlife habitat and breeding waterfowl depend
on wetlands for successful reproduction (Krapu and Reinecke 1992; Anteau 2012) as wetlands
provide food resources for breeding, concealment from predators, and function as brood rearing
habitat. It is well documented that the Prairie Pothole region (hereafter PPR) is an important
area for breeding waterfowl (Boyd 1981; Krapu et al. 1983; Reynolds 1987); unlike the PPR,
wetlands of the Great Lake States (Michigan, Minnesota, and Wisconsin; hereafter GLS) are
more permanent and sometimes considered less productive for dabbling ducks than those of the
PPR (Reynolds 1987; Kantrud et al. 1989; Euliss et al. 2004; Johnson et al. 2004; Simpson et al.

2

2005; Soulliere et al. 2007), although there is uncertainty about how productivity of dabbling
ducks breeding in the GLS varies with weather.
A substantial loss of wetlands could cause a reduction in breeding habitat availability and
quality and subsequently duck population declines (Larson 1994; Glick 2005). Annual variation
in duck populations have been highly correlated with wetland abundance indices during the
breeding season (Crissey 1969; Dzubin 1969; Leitch and Kaminski 1985; Johnson and Shaffer
1987; Kaminski and Gluesing 1987; Reynolds 1987; Batt et al. 1989; Bethke and Knudds 1995;
Krapu et al. 1997). Furthermore, Heitmeyer and Fredrickson (1981) concluded that mallard
production may also be influenced by wetland conditions on the wintering grounds, thus mallard
productivity and subsequently population size may respond to wetland levels throughout the
annual cycle. However, other factors may also contribute to variation in productivity rates.
Productivity rates of ducks are thought to be influenced by predation and environmental
and contamination factors which could potentially decrease duckling survival and subsequently
productivity rates (Talent et al. 1983; Ringleman 1992; Henny et al. 2000). Cold and wet
weather following hatching is known to decrease duckling survival, potentially influencing
productivity (Korschgen et al. 1996; Cox et al. 1998; Krapu et al. 2000; Pietz et al. 2003). Ducks
require high quality food resources and brooding cover for successful reproduction (Sedinger
1992; Cox et al. 1998) which can be provided by certain wetlands. Wetlands with open water
and flooded emergent vegetation are important for duckling survival (Weller and Spatcher 1965;
Bloom et al. 2012). Wetlands throughout the GLS have been invaded by a myriad of invasive
species (i.e. Purple loosestrife, (Lythrum salicaria); Phragmites (Phragmites australis); Eurasian
Watermilfoil (Myriophyllum spicatum) over the past 30 years (Welling and Becker 1993). The
ecological impact from aquatic invasive species is not well understood (Parker et al. 1999), but

3

invasive species may displace native species in wetlands, thus reducing breeding habitat
available to waterfowl (Thompson et al. 1987) and potentially displacing breeding ducks.
Female mallards often exhibit greater philopatry than do males (Clark et al.1997;
Greenwood and Harvey 1982; but see Doherty et al. 2002), but local hydrological conditions
may displace birds from natal breeding grounds (Hansen and Mcknight 1964; Crissey 1969;
Smith 1970; Jackson et al. 1985). Potential benefits of long-distance dispersal may outweigh
costs if higher quality nesting sites are acquired (Coulton et al. 2011). This may result in higher
annual emigration and immigration rates which in turn could affect local breeding population
size. Immigration and emigration are thought to be contributing factors to annual breeding
population size (Crissey 1969), which vary annually throughout the GLS. If density dependence
influences productivity rates, then changes in immigration and emigration rates could influence
productivity rates.
This research assesses the GLS mallard population which began to decline around 2001
after approximately 30-years of growth. This initial decline in abundance came at a time when
Lake Michigan-Huron water levels were at a 50-year low (GLERL 2013). Predictions regarding
mallard abundance suggest that with the continued loss of habitat, the GLS region could see a
19-39% decrease in duck abundance by the 2030s (Glick 2005). Given current climate change
prediction models and the current decline of mallard abundance, concern of the GLS mallard
population is lingering. Current hypotheses regarding the GLS mallard population decline are
focused around declining productivity (Van Horn et al. 2006) and non-breeding season survival
(i.e. hunting; Coluccy et al. 2008). However, Hoekman et al. (2002) found >75% of the variation
in population growth for mid-continent mallards can be attributed to nest success, adult hen
breeding season survival, and duckling survival which all may be influenced by predator-prey

4

interactions. However, our understanding of the relationship between climate and vulnerability
to predation during the breeding season and resulting effects on productivity is incomplete.
I estimated long-term (1961-2011) mallard productivity rates for birds breeding in the
GLS to understand how productivity rates have changed over time and have been affected by
wetland habitat availability (i.e., via variation in regional hydrology), weather, and mallard
abundance during the breeding season. Here, I define productivity rate as the ratio of young to
adult females in the pre-hunt late summer population within each GLS. I hypothesize that
indices of breeding habitat availability, temperature, and precipitation during the breeding season
would be positively related to productivity rates and that density dependence on the breeding
grounds could be influencing productivity rates. The quantity and quality of wetlands for
breeding mallards are thought to have declined in the GLS, so I predicted that productivity rates
had declined. Understanding the link, if one exists, between productivity rates and climatic
factors may help managers focus habitat management at particular times of the breeding season
or toward wetlands that provide habitat throughout the entire breeding season.
Study area and methodology
This study included mallards breeding in the GLS, which are managed as part of the midcontinental population of mallards (USFWS 2013). Roughly 80% of the GLS lies within the
Prairie Hardwood and Boreal Hardwood Transition ecoregions except for parts of western and
southern Minnesota and a small portion of southeastern Michigan and western and southern
Wisconsin, which are considered Prairie Pothole and Eastern-Tallgrass Prairie ecoregions
respectively (NABCI 2000). Weather patterns in this region are largely driven by strong lake
effects and wetlands are more permanent than those of the PPR. Mallards in this region
contributed 8-15% of the entire mid-continent mallard population (USFWS 2013).

5

To estimate population age ratios, I used samples of mallards harvested in the GLS and
collected as part of the United States Fish and Wildlife Service (USFWS) Parts Collection
Survey from 1961-2011. Estimates of harvest by age/sex cohorts was derived from data
collected during Mississippi Flyway “wing bees” and provided by USFWS Harvest Surveys
Section personnel; I restricted the analyses to birds taken during September 1-October 31 as
these months represent a period when harvest is largely comprised of local breeding birds and
large scale migrations have not begun into the Great Lakes region (Jessen 1970; Krementz et al
2013). Furthermore, T. Arnold and C. de Sobrino (unpub data) found that 75, 57, and 80 percent
of annual mallard harvest in Michigan, Minnesota, and Wisconsin, respectively is comprised of
local breeding mallards, which lends support to the assumption that harvested mallards in
September and October are primarily local breeding birds.
First, I estimated direct recovery rates for each state/age class (DRR) defined as, the
proportion of the banded sample that was shot and reported (Shot) during the hunting season
immediately following banding,

; where

is the number of banded

mallards in a year i in age class j of sex k. I then adjusted DRR estimates for reporting rates to
estimate harvest rate (HR) by the equation,

; where

is the annual reporting rate

as estimated from reward band studies (Nichols et al. 1991). I adopted reporting rates used by
Arnold (T. Arnold, unpub data; 38% 1961-1995 and 77% 1996-2011). Band reporting rates
increased in ~ 1996 when implementation of “toll-free” bands was adopted, allowing hunters to
report harvested birds via telephone where previously birds could only be reported via USPS
mail service (Royle and Garretson 2005). I then estimated cohort specific fall abundance as,
̂

̂
̂

, where ̂ is the harvest of cohort i estimated from the parts collection survey data and

6

̂ is the harvest rate of cohort i. Juvenile game birds are likely more vulnerable to harvest than
are adults so I adjusted harvest age ratios for differential harvest vulnerability of age cohorts
using ratios of direct recovery rates (March and Hunt 1978; Reynolds 1987; Zimmerman et al.
2010). Finally, to estimate productivity rates, I used

̂

, where

is the

estimate of juvenile female and male abundance, respectively divided by the estimate of adult
female abundance. Sample size of band recovery data in some year/states was small, so I used a
5-yr moving average of direct recovery rates to adjust harvest age ratios (Zimmerman et al.
2010).
I used the Mixed Models Procedure (PROC MIXED) in SAS (SAS Institute 2004) to
model response of productivity rate to temperature, precipitation, wetland hydrological
conditions, and mallard abundance. I excluded one and two years of data from WI and MI
respectively as these were extreme outliers (SAS Institute 2004). I used Palmer Hydrological
Drought Indices (hereafter PHDI; NCDC 2012) or wetland abundance estimates from spring
aerial waterfowl surveys for each state (hereafter POND) as measures of breeding habitat
availability. POND estimates were not available for the entire 1961-2011 time series because
state waterfowl surveys began later (MI 1992, MN 1968, and WI 1973). I tested for the effects
of precipitation and temperature on productivity rates by including average June and July
temperature (hereafter JUNET and JULYT respectively) and average June and July precipitation
(hereafter JUNEP AND JULYP respectively). These climatic variables represent a time period
post-hatch, when duckling survival is quite variable, thus potentially influencing productivity
rates. Density dependence in productivity is one of the current parameters used to evaluate
mallard population dynamics under the Adaptive Harvest Management protocol (USFWS 2013).
To test for the effects of density dependence, I included annual mallard abundance estimates
7

from each state (hereafter MALL). The POND variable provided a metric to assess available
breeding habitat which is thought to be a limiting factor to reproduction (USFWS 2013).
However, POND was not available across the 51-year study and both MN and WI do not survey
the entire state. Thus, I used a 3-month average (April-June) PHDI to index breeding habitat
conditions during peak settling and nesting (Bellrose 1976). The PHDI is a drought index that is
used to assess long-term moisture supply (NCDC 2012). Other researchers have found high
positive correlation between Lake Michigan-Huron water levels in year t-1 and current year
mallard abundance estimates (D. Luukkonen unpub data). To evaluate potential time lag effects,
I considered a one-year lag effect of PHDI and POND (PHDI_LAG and POND_LAG,
respectively). All predictor variables were treated as continuous and standardized with a mean
of zero and standard deviation of one.
The global model for the first set of candidate models (hereafter PHDI models) included
the variables: PHDI, PHDI_LAG, JUNET, JULYT, JUNEP, JULYP and an intercept term and
the global model for the second set of candidate models (hereafter POND models) included the
precipitation and temperature variables in addition to POND, POND_LAG, and MALL and an
intercept term.
The a priori model sets included only main-effects and consisted of 8 models for the
PHDI data set and 9 models for the POND data set (Table 1).

8

Table 1. Candidate models used to model productivity rates for both the (a) PHDI and (b)
POND data sets. In the POND data set I was interested in the effects of density
dependence on productivity and thus have included the covariate MALL in all models.
a. PHDI
PHDI+PHDI_LAG+JUNET+JULYT+JUNEP+JULYP
PHDI+PHDI_LAG+JUNET+JULYT
PHDI+PHDI_LAG+JUNEP+JULYP
JUNET+JULYT+JUNEP+JULYP
PHDI+PHDI_LAG
JUNET+JULYT
JUNEP+JULYP
INTERCEPT ONLY

b. POND
POND + POND_LAG + JUNET + JULYT + JUNEP + JULYP + MALL
POND + POND_LAG + JUNET + JULYT + MALL
POND + POND_LAG + JUNEP + JULYP + MALL
JUNET + JULYT + JUNEP + JULYP + MALL
POND + POND_LAG + MALL
JUNET + JULYT + MALL
JUNEP + JULYP +MALL
MALL
INTERCEPT ONLY
I compared models using Akaike’s information criterion corrected for small sample size
(AICc; Burnham and Anderson 2002) and reported model-averaged parameter estimates ( ̂ and
unconditional standard error (± SE). This approach allowed us to make inference from a set of

9

models instead of one model and is thought to be more useful for predictive modeling (Burnham
and Anderson 2002; Arnold 2010).
I used a Chi-square goodness-of-fit test, ̂

, where

is the chi-square test statistic

and df is the estimable parameters from the global model plus one for the estimation of ̂ (PHDI
global model and POND global model = 7 and 8 parameters, respectively) to check for
overdispersion in the data. I used a one-fold cross validation approach to assess model fit by
withholding 33% (n=17) of the data in the PHDI model set. I did not withhold data for the
POND data set as the number of observations was insufficient. I then assessed model fit by
fitting a linear regression model of the test and training data from the cross-validation test and
evaluate model fit using the r2 for each of the GLS.

Results
Productivity rate estimates
Mean mallard productivity rates were 2.9, 2.8, and 3.1 young/adult hen for MI, MN and
WI respectively (95% CI: ±0.35, ±0.27, and ±0.48 young/adult hen; Figure 1). Annual
productivity rates ranged from 1.3-7.7, 0.7-4.0, and 1.0-9.8 young/adult hen for MI, MN, and
WI, respectively (Figure 2).

1961-2011 mean
productivity rates
(±95% CI)

4
3.5
3
2.5
2
1.5
MI

MN

WI

Figure 1. Mean productivity rates for mallards in the 3 GLS (± 95% CI) 1961-2011.
10

11

2009

2006

2003

10

2000

1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

Productivity rate
10

1997

1994

1991

1988

1985

1982

1979

1976

1973

1970

1967

1964

1961

Productivity rate
1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

Productivity rate
10

a) Michigan

8

6

4

2

0

b) Minnesota

8

6

4

2

0

c) Wisconsin

8

6

4

2

0

Figure 2. Annual productivity rates for mallards in the 3 GLS: (a) Michigan; (b) Minnesota; and
(c)Wisconsin with long-term average (solid horizontal line) and 95% CI (dashed horizontal
lines).

PHDI model results
None of the productivity models accounted for more than ~55% of the model weight
based on AICc (Table 2).
Table 2. Performance of models used to explain variation in productivity rates for the PHDI (34
randomly selected years) and POND (1968-2011, 1973-2011, and 1991-2011 for MN,
WI, and MI respectively) data sets for each state (a) Michigan; (b) Minnesota; and
(c)Wisconsin. The number of model parameters (K) with all models including an
intercept, Akaike’s information criterion adjusted for small sample size (AICc), AICc
compared with the best model (ΔAICc), and the likelihood of each model in a set (Akaike
weight, Wi) are reported.

a) Michigan
PHDI model set
PHDI + PHDI_LAG + JUNET+JULYT
PHDI+PHDI_LAG
PHDI+PHDI_LAG+JUNET+JULYT+JUNEP+JULYP
JUNET+JULYT
INTERCEPT ONLY
PHDI+PHDI_LAG+JUNEP+JULYP
JUNET+JULYT+JUNEP+JULYP
JUNEP+JULYP

K
6
4
8
4
2
6
6
4

AICc
89.8
92.3
92.8
93.9
95.3
95.5
96
98.1

ΔAICc
0
2.5
3
4.1
5.5
5.7
6.2
8.3

Wi
0.55
0.16
0.12
0.07
0.04
0.03
0.02
0.01

POND model set
JUNET+JULYT+MALL
POND+POND_LAG+JUNET+JULYT+MALL
POND+POND_LAG+JUNET+JULYT+JUNEP+
JULYP+MALL
MALL
INTERCEPT ONLY
JUNET+MALL+JULYT+JUNEP+JULYP
POND+POND_LAG+MALL
JUNEP+JULYP+MALL
POND+POND_LAG+JUNEP+JULYP+MALL

K
4
7

AICc
50.5
51.1

ΔAICc
0
0.6

Wi
0.22
0.16

9
3
2
7
5
5
7

51.5
51.6
52.1
52.1
52.4
53.5
53.6

1
1.1
1.6
1.6
1.9
3
3.1

0.13
0.12
0.10
0.10
0.08
0.05
0.05

12

Table 2 (cont’d)

b) Minnesota
PHDI model set
PHDI+PHDI_LAG
PHDI+PHDI_LAG+JUNEP+JULYP
INTERCEPT ONLY
PHDI+PHDI_LAG+JUNET+JULYT
PHDI+PHDI_LAG+JUNET+JULYT+JUNEP+JULYP
JUNEP+JULYP
JUNET+JULYT
JUNET+JULYT+JUNEP+JULYP

K
4
6
2
6
8
4
4
6

AICc
90.2
90.4
92.8
93.4
93.9
95.1
95.5
97.9

ΔAICc
0
0.2
2.6
3.2
3.7
4.9
5.3
7.7

Wi
0.37
0.33
0.10
0.07
0.06
0.03
0.03
0.01

POND model set
INTERCEPT ONLY
POND+POND_LAG+MALL
MALL
JUNET+JULYT+MALL
POND+POND_LAG+JUNET+JULYT+MALL
POND+POND_LAG+JUNEP+JULYP+MALL
JUNEP+JULYP+MALL
JUNET+JULYT+JUNEP+JULYP+MALL
POND+POND_LAG+JUNET+JULYT+JUNEP+
JULYP+MALL

K
2
5
3
5
7
7
5
7

AICc
119
119.4
120.6
121.2
121.4
123.2
124.2
125.1

ΔAICc
0
0.4
1.6
2.2
2.4
4.2
5.2
6.1

Wi
0.31
0.26
0.14
0.10
0.09
0.04
0.02
0.01

9

125.3

6.3

0.01

13

Table 2 (cont’d)

c) Wisconsin
PHDI model set
PHDI+PHDI_LAG+JUNET+JULYT
PHDI+PHDI_LAG+JUNET+JULYT+JUNEP+JULYP
PHDI+PHDI_LAG
JUNET+JULYT
PHDI+PHDI_LAG+JUNEP+JULYP
INTERCEPT ONLY
JUNET+JULYT+JUNEP+JULYP
JUNEP+JULYP

K
6
8
4
4
6
2
6
4

AICc
130
130.9
131.3
131.8
132
132.3
132.5
133

ΔAICc
0
0.9
1.3
1.8
2
2.3
2.5
3

Wi
0.27
0.17
0.14
0.11
0.10
0.08
0.08
0.06

POND model set
INTERCEPT ONLY
JUNET+JULYT+MALL
MALL
POND+POND_LAG+MALL
POND+POND_LAG+JUNET+JULYT+MALL
JUNET+JULYT+JUNEP+JULYP+MALL
JUNEP+JULYP+MALL
POND+POND_LAG+JUNEP+JULYP+MALL
POND+POND_LAG+JUNET+JULYT+JUNEP+
JULYP+MALL

K
2
4
3
5
7
7
5
7

AICc
141.8
142.6
142.7
143.2
143.6
144
144.2
144.7

ΔAICc
0
0.8
0.9
1.4
1.8
2.2
2.4
2.9

Wi
0.23
0.16
0.15
0.12
0.09
0.08
0.07
0.05

9

144.8

3

0.05

Model averaged parameter estimates (Figure 3) showed that PHDI was positively related to
productivity rates and was the most influential predictor variable explaining variation in
productivity rates for all states, however 95% CI overlapped zero in WI. JULYT was positively
related to productivity rates in all 3 states with the greatest effect seen in MI and WI.
Productivity rates were weakly related to PHDI_LAG, JUNET, JUNEP, and JULYP with 95%
CI broadly overlapping zero. Chi-squared goodness-of-fit tests suggested that generally the data
were not overdispersed in MI or MN, but were overdispersed in WI ( ̂

1.03, 1.18, and 3.59 for

the global model from the PHDI data set for MI, MN, and WI, respectively and ̂

0.38, 1.47,

and 2.93 for the global model from the POND model set for MI, MN, and WI, respectively;
14

Burnham and Anderson 2002). However, the data were underdispersed in the POND data set for
Michigan ( ̂

0.38). Cross-validation results suggest that the top ranking model based on AICc

explained about 10% of the variation of productivity rates for mallards in Michigan.
Productivity rates in Minnesota and Wisconsin did not respond well to the top models (R2≤0.1).

MI

1.00
Model-averaged parameter estimates (±95% CI)

MN
WI

0.80
0.60
0.40
0.20
0.00
-0.20
-0.40
PHDI

PHDI_LAG

JUNET

JULYT

JUNEP

JULYP

-0.60
JUNEP

JULYP

Figure 3. Model-averaged parameter estimates ( ̂ ) and 95% CI from standardized predictor
variables across all models for the PHDI model set. For interpretation of the references
to color in this and all other figures, the reader is referred to the electronic version of this
thesis.

15

MI

Model-averaged parameter estimates (±95% CI)

1.00

MN

0.80

WI
0.60
0.40
0.20
0.00
-0.20
-0.40
-0.60
-0.80
POND

POND_LAG

JUNET

JULYT

JUNEP

JULYP

MALL

-1.00

Figure 4. Model-averaged parameter estimates ( ̂ ) and 95% CI from standardized predictor
variables across all models for the POND model set.

POND model results
No model accounted for more than ~31% of the model weight based on AICc (Table 2).
Model averaged parameter estimates overlapped zero suggesting large uncertainty evaluating the
effects of the variables on productivity rates (Figure 4).
Similar to the PHDI model set, temperature explained more of the variation in productivity rates
than did precipitation, but 95% CI overlapped zero for all variables. The JULYT variable
explained more variation than the other variables and was positive for all states, but CI
overlapped zero. There appears to be some evidence for a density dependence effect on

16

productivity rates as observed by a slightly negative estimate for the MALL covariate in
Michigan and Minnesota, but again 95% CI overlapped zero.

Discussion
My results suggest that annual variation in productivity rates is most influenced by
wetland hydrologic conditions, as indexed by the PHDI variable, more so than temperature,
precipitation, and mallard abundance during the breeding season. Hydrology is an important
component of wetlands that likely reflects habitat availability for mallards during settling,
pairing, and brood rearing. Mallard nest survival in the aspen parklands has been linked to the
amount of herbaceous vegetation on study areas and total precipitation for the 12 months prior to
nesting; also, nesting effort was positively related to wetland inundation in July and duckling
survival was positively related to the proportion of seasonal wetlands holding water in July
(Howerter et al. 2014). Others have found similar positive relationships between mallard
abundance and pond counts which might result from more birds settling to breed or greater
return of birds breeding or produced in previous years when landscapes are wet (Stoudt 1969;
Krapu et al. 1983; Kaminski and Prince 1984; Leitch and Kaminski 1985; Johnson and Grier
1988); however, the estimate of ponds in the POND model set explained less variation in
productivity than did the PHDI variable. Crissey (1969) found that mallard abundance in a given
year was correlated with July pond counts from the previous year and attributed this to increased
reproduction. Although I did not directly look at July pond counts, I found that July temperature
was positively related to productivity rates. My estimate of ponds did not cover the entire time
series and may not be a reliable indicator of breeding habitat availability as this data does not
classify wetlands as potential breeding habitat (i.e. sheet water in fields) and this variable may

17

too coarsely index changes in hydrology to be a useful predictor of productivity. Many wetlands
of the GLS are permanent or semi-permanent lakes, ponds, and rivers and since these typically
would not become dry, the count of ponds in the GLS may not capture variation in habitat
available for breeding mallards.
Cold and wet weather, especially for ducklings < 10 days old, is expected to negatively
affect survival (Mendenhall and Milne 1985; Krapu et al. 2006; Bloom et al. 2012). I did not
observe any significant effects of precipitation on productivity rates but I did observe effects of
temperature on productivity rates for Michigan and Wisconsin. However, productivity rates in
Minnesota were affected minimally by June or July temperature. Mallard duckling survival in
the aspen parklands was negatively related to the number of days in June and July when
minimum air temperature dropped below 10°C (Howerter et al. 2014). If a brood rearing season
was on the average wetter and colder and duckling survival was lower, productivity rates could
be affected. Extreme temperatures have been shown to influence thermoregulation in one-dayold mallard ducklings (Koskimies and Lahti 1964) and ducklings are more vulnerable to
temperature extremes as thermal regulation is incomplete during the first ten days post-hatch
(Orthmeyer and Ball 1990; Rotella and Ratti 1992). Although mallard ducklings are able to
thermoregulate within one day after hatch, they can require brooding for up to 3 weeks
(Untergasser and Hayward 1973). I studied effects of average monthly temperatures on
productivity rates rather than daily extremes. It is possible that monthly averages may be high
while daily extremes set record lows and vice versa, thus my methods would fail to capture these
extreme events. Furthermore, I estimated productivity at the state scale and there is often a great
deal of climatic variation within states. However, Jehl and Hussel (1966) did not observe any
negative effects on waterfowl reproduction (i.e. observed brood counts) during a summer with a

18

severe three day snow storm. My productivity rate estimates were at a coarse spatial scale (i.e.
individual states) and thus I could not capture climatic variation at certain areas where duck
production may have been higher or lower than a states average. It is likely that weather patterns
vary substantially across individual states.
Many of the model-averaged parameter estimates had 95% CI that overlapped zero, but
estimates from the PHDI model set which included the entire time series were likely more robust
than estimates from the POND model set. Furthermore, predation has been found to decrease
hen success, however was not found to negatively impact mallard productivity in the PPR
(Amundson et al. 2013). Mallards breed across a broad geographical area (Bellrose 1976), often
have multiple clutches during a single breeding season (Bellrose 1976), and initiate nests over a
broad temporal scale during a breeding season (Bellrose 1976), likely making them capable of
overcoming extreme weather events which could impact productivity of species that may be less
resilient.
Many factors that might affect mallard productivity changed over the course of the study,
but productivity rates stayed relatively stable. For example, the amount of Conservation Reserve
Program (CRP) acres on the landscape changed the amount of upland herbaceous cover in the
GLS during the mid-1980’s (Figure 5). Also, invasive wetland plant species such as purple
loosestrife and phragmites became established. While some studies suggest that the decline of
the GLS mallard breeding populations could be linked to a decline in productivity (Van Horn et
al. 2006), my results suggest that long-term productivity rates have remained relatively stable, or
increased in the case of Wisconsin, despite considerable annual variation. Although productivity
rates were not significantly different among the GLS, Michigan and Wisconsin were more
similar than were productivity rates in Minnesota. I suspect this is due to differences in

19

ecoregions amongst portions of the GLS. Portions of Minnesota that are ideal for breeding
waterfowl lie within the PPR and past research has demonstrated differences among effects of
climate in the ecoregions (Hokeman et al. 2005; Soulliere et al. 2007; Coluccy et al. 2008).
Furthermore, variation in vital rates effecting population growth rates are different for mallards
breeding in geographically diverse breeding regions (Hokeman et al. 2005; Coluccy et al. 2008;

Conservation Reserve Program
Enrolled Acres (1000s)

Howerter 2014).
3,500

Wisconsin

3,000

Michigan

Minnesota

2,500
2,000
1,500
1,000
500
0

Figure 5. Acreage of land in the GLS that was enrolled in the CRP program from 1986-2013.
Coluccy et al. (2008) found that breeding season parameters such as nest success and
duckling survival account for 63% of the variation in annual population growth rate. My results
suggest that July temperature and breeding habitat availability effect productivity rates and this is
likely due to increased nest success and duckling survival when breeding habitat is abundant and
July temperatures are warmer during the brood rearing stage. However, I did not find any strong
relationships with June temperature except a slight negative relationship for productivity rates in
MI. June is likely the month that the majority of mallard nests hatch (Bellrose 1976) and
predation rates on nests decline as the breeding season progresses (Greenwood et al. 1995).
20

Furthermore, natural disasters have been found to have minimal effects on nest loss in riparian
bird communities (Best and Stauffer 1980). Thus, I expect that June temperature will be linked
to productivity rates via decreased duckling survival. Perhaps during the month of June, nest
success is high and duckling survival is high thus there appears to be a minimal effect of June
temperature on productivity rates.
The importance of breeding habitat to successful waterfowl reproduction and abundance
is well documented. Furthermore, annual variation in breeding habitat availability is likely
greater in the PPR than the prairie and boreal hardwood transition regions (Kantrud et al. 1989,
Soulliere et al. 2007), suggesting that mallard breeding abundance likely varies more on an
annual basis in prairie type ecosystems than it does in the GLS. Variation in water levels across
the GLS has been linked to climatic change at multiple scales (Keough et al. 1999). My results
suggest that productivity rates are influenced by hydrological factors such that climatic changes
could potentially negatively affect productivity rates by reducing breeding habitat availability.
Recent climate change models predict areas of the upper Midwest are likely to experience dryer
and hotter summers (UCS 2009), which could lead to poorer quality habitat through decreased
water depth and consequently altered emergent vegetation patterns (Cowardin et al. 1988;
Simpson et al. 2007).
My productivity rate estimates were higher than biologically likely for some years (i.e.
productivity rates > 8 young/adult hen). Mallards typically lay a clutch of 9 eggs (Bellrose 1976)
and it would be unlikely for all eggs to survive until fledging. I could not determine why this is
so, but offer some speculation. Extremes in productivity rates were potentially do to: 1) an early
influx of migrating mallards, and 2) a larger than normal sample of harvested mallards from a
certain area (i.e. banding just prior to opening of hunting seasons on a managed waterfowl area).

21

Nonetheless, most of my estimates should be robust and show the importance of long-term
studies (Gibbs et al. 1999) to understand variation in reproductive rates as often times there is a
great deal of annual variation and trends correlating with climatic events.
Although productivity rates have remained relatively stable, future climatic changes
could cause productivity rates to decrease due to reduced breeding and brood rearing habitat
availability. My models suggest that productivity rates are influenced by weather patterns
throughout the brood rearing and nesting periods. With the predicted increase of summer
temperatures and decrease of summer precipitation (Hayhoe et al. 2009), evapotranspiration is
expected to increase resulting in drying of wetlands that once could have provided ideal breeding
and brood rearing habitat, leaving managers with the difficult decision of where to focus
management.
Management Implications
My main findings suggest that mallard productivity rates have remained relatively stable
despite varying climatic conditions and habitat alterations. The best predictor I found suggests
that management for suitable nesting and brood rearing habitat should be the focus for waterfowl
managers, but I caution that focus on one vital rate may influence another vital rate to either
decrease or increase (Hoekeman et al. 2014). Mallards seem to have wide tolerance for nesting
conditions and have been able to continue to reproduce effectively over the past 51-years despite
changes in habitat and weather. However, the estimate obtained by the annual spring surveys in
the GLS may not capture variation in breeding habitat availability as wetlands are more
permanent than those of the prairies (Soulliere et al 2007), thus further classification of wetlands
during spring surveys may help quantify breeding habitat availability.

22

Further studies could be conducted at smaller spatial and temporal scales so to quantify
the effects of hydrological variation at different regions of the GLS and ensure management is
being focused in the correct areas. However, I caution that long-term studies are needed in order
to capture variation in hydrological periods and weather cycles. The potential exists for extreme
variation in waterfowl productivity at certain locations, so focusing resources on areas of
generally high mallard abundance may be of importance under certain climate change scenarios.
Although intensive, managers in the GLS will want to consider the costs and benefits of creating
or enhancing wetlands to provide water level control as a hedge against drying conditions.

23

LITERATURE CITED

24

LITERATURE CITED

Amundson, C. L., M. R. Pieron, T. W. Arnold, and L. A. Beaudoin. 2013. The effects of
predator removal on mallard production and population change in northeastern North
Dakota. Journal of Wildlife Management 77:143-152.
Anteau, M. J. 2012. Do interactions of land use and climate affect productivity of waterbirds
and prairie-pothole wetlands? Wetlands 32:1-9.
Arnold, T. W. 2010. Uninformative parameters and model selection using Akaike’s information
criterion. Journal of Wildlife Management 74:1175-1178.
Batt, B. D. J., M. G. Anderson, C. D. Anderson, and F. D. Caswell. 1989. The use of prairie
potholes by North American ducks. Pp.204-227 in A. G. Van der Valk, ed. Northern
prairie wetlands. Iowa State University Press, Ames, Iowa.
Bellrose, F. C. 1976. Ducks, geese and swans of North America. Stackpole Books, Harrisburg,
PA.
Best, L. B. and D. F. Stauffer. 1980. Factors affecting nesting success in riparian bird
communities. Condor 82:149-158.
Bloom, P. M., R. G. Clark, D. W. Howerter, and L. M. Armstrong. 2012. Landscape-level
correlates of mallard duckling survival: Implications for conservation programs. Journal
of Wildlife Management 76:813-823.
Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: A
practical information-theoretic approach. Second edition. Springer, New York, New
York.
Coluccy, J. M., T. Yerkes, R. Simpson, J. W. Simpson, L. Armstrong, and J. Davis. 2008.
Population dynamics of breeding mallards in the Great Lake States. Journal of Wildlife
Management 72:1181-1187.
Coulton, D. W., R. G. Clark, D. W. Howerter, M. G. Anderson, and L. I. Wassenaar. 2011.
Costs and benefits of natal dispersal in yearling mallards Anas platyrhynchos. Journal of
Avian Biology 42:123-133.
Cowardin, L. M., D. H. Johnson, T. L. Shaffer, and D. W. Sparling. 1988. Applications of a
simulation model to decisions in mallard management. United States Fish and Wildlife
Service Technical Report 17.

25

Cox R. C. Jr, M. A. Hanson, C. C. Roy, N. H. Euliss Jr., D. H Johnson, and M. G. Butler. 1998.
Mallard duckling growth and survival in relation to aquatic invertebrates. Journal of
Wildlife Management. 62:124-133.
Crissey, W. F. 1969. Prairie potholes from a continental viewpoint. Saskatoon wetlands
seminar. Canadian Wildlife Service Report.
Dahl, T. E. 1990. Wetlands Losses in the United States 1780’s to 1980’s. U.S. Department of
the Interior, Fish and Wildlife Service, Washington.
Dahl, T. E. 2006. Status and trends of wetlands in the conterminous United States 1999-2004.
U.S. Department of the Interior, Fish and Wildlife Service, Washington D. C.
Dahl, T. E. 2011. Status and trends of wetlands in the conterminous United States 2004-2009.
U.S. Department of the Interior, Fish and Wildlife Service, Washington D. C.
Euliss, N. H. Jr., J. W. LaBaugh, L. H. Fredrickson, D. M. Mushet, M. K. Laubhan, G. A.
Swanson, T. C. Winter, D. O. Rosenberry, and R. D. Nelson. 2004. The wetland
continuum: a conceptual framework for interpreting biological studies. Wetlands 24:448458.
Gibbs, J. P., H. L. Snell, and C. E. Causton. 1999. Effective monitoring for adaptive wildlife
management: lessons from the Galapagos Islands. Journal of Wildlife Management
63:1055-1065.
GLERL (2013) Great lakes water level observations. Available online:
http://www.glerl.noaa.gov/data/now/wlevels/levels.html—accessed 5/2/2014.
Glick, P. 2005. The waterfowler’s guide to global warming. Washington, DC: National
Wildlife Federation.
Greenwood, R. J., A. B. Sargeant, D. H. Johnson, L. M. Cowardin, and T. L. Shaffer. 1995.
Factors associated with duck nest success in the prairie pothole region of Canada.
Wildlife Monographs 128:1-57.
Hansen, H. A. and D. E. Mcknight. 1964. Emigration of drought-displaced ducks to the arctic.
Transactions of the North American wildlife and natural resources conference 29:119127.
Hayhoe, K., J. VanDorn, V. Naik, and D. Wuebbles. 2009. Climate change in the Midwest. in
Confronting climate change in the U.S. Midwest. Union of concerned scientists.
Available online: http://ucsusa.org/assets/documents/global_warming/midwest-climateimpacts.pdf —accessed 1/9/2014.

26

Heitmeyer, M. E. and L. H. Fredrickson. 1981. Do wetland conditions in the Mississippi delta
hardwoods influence mallard recruitment? Transaction of North American Wildlife and
Natural Resources Conference 46:44-57.
Henny, C. J., L. J. Blus, D. J. Hoffman, L. Sileo, D. J. Audet, and M. R. Snyder. 2000. Field
evaluation of lead effects on Canada geese and Mallards in the Coeur d’Alene River
Basin, Idaho. Archives of Environmental Contamination and Toxicology. 39:97:112.
Hoegh-Guldberg, O. and J. F. Bruno. 2010. The impact of climate change on the world’s
marine ecosystems. Science 328:1523-1528.
Hoekman, S. T., L. S. Mills, D. W. Howerter, J. H. Devries, and J. J. Ball. 2002. Sensitivity
analyses of the life cycle of midcontinent mallards. Journal of Wildlife Management
66:883:900.
Howerter, D. W., M. G. Anderson, J. H. Devries, B. L. Joynt, L. M. Armstrong, R. B. Emery,
and T. W. Arnold. 2014. Variation in mallard vital rates in Canadian aspen parklands:
The prairie habitat joint venture assessment. Wildlife Monographs 188:1-37.
Jackson, D. H., H. H. Prince, and R. A. Bishop. 1985. Influence of drought on mallards
breeding in northern Iowa. Journal of Wildlife Management 49:442-448.
Jessen, R. L. 1970. Mallard population trends and hunting losses in Minnesota. Journal of
Wildlife Management 34:93-105.
Johnson, D. H. and J. W. Grier. 1988. Determinants of breeding distributions of ducks.
Wildlife Monographs 100:1-37.
Johnson, W. C., S. E. Boettcher, K. A. Poiani, and G. R. Guntenspergen. 2004. Influence of
weather regimes on the water levels of glaciated prairie wetlands. Wetlands 24:385-398.
Johnson, W. C., B. V. Millett, T. Gilmanov, R. A. Voldseth, G. R. Guntenspergen, and D. E.
Naugle. 2005. Vulnerability of northen prairie wetlands to climate change. Bioscience
55:863-872.
Kaminski, R. M. and H. H. Prince. 1984. Dabbling duck-habitat associations during spring in
Delta Marsh, Manitoba. Journal of Wildlife Management 48:37-50.
Kantrud, H. A., G. L. Krapu, and G. A. Swanson. 1989. Prairie basin wetlands of the Dakotas: a
community profile. U.S. Fish and Wildlife Service. Biol. Repo. 85 116pp.
Keough, J. R., T. A. Thompson, G. R. Guntenspergen, and D. A. Wilcox. 1999.
Hydrogeomorphic factors and ecosystem responses in coastal wetlands of the Great
Lakes. Wetlands 19:821-834.

27

Korschgen, C. E., K. P. Kenow, W. L. Green, D. H. Johnson, M. D. Samuel, and L. Sileo. 1996
survival of radio-marked Canvasback ducklings in northwestern Minnesota. Journal of
Wildlife Management 60:120-132.
Koskimies, J. and L. Lahti. 1964. Cold-hardiness of the newly hatched young in relation to the
ecology and distribution in ten species of European ducks. Auk 81:281-307.
Krapu, G. L., A. T. Klett, and D. G. Jorde. 1983. The effect of variable spring water conditions
on mallard reproduction. Auk 100:689-698.
Krapu, G. L. and K. J. Reinecke. 1992. Foraging ecology and nutrition. Pp 1-29 in B. D. J.
Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L.
Krapu eds. Ecology and management of breeding waterfowl. University of Minnesota
Press, Minneapolis, Minnesota.
Krapu, G. L., P. J. Pietz, D. A. Brandt, and R. R. Cox Jr. 2000. Factors limiting Mallard brood
survival in prairie pothole landscapes. Journal of Wildlife Management 64:553-561.
Krapu, G. L., P. J. Pietz, D. A. Brandt, and R. R. Cox, Jr. 2006. Mallard brood movements,
wetland use, and duckling survival during and following a prairie drought. Journal of
Wildlife Management 70:1436-1444.
Krementz, D. G., K. Asante, and L. W. Naylor. 2012. Autumn migration of Mississippi flyway
mallards as determined by satellite telemetry. Journal of Fish and Wildlife management
2:238-251.
Larson, D. L. 1994. Potential effects of anthropogenic greenhouse gases on avian habitats and
populations in the northern Great Plains. American Midland Naturalist 131:330-346.
Leitch, W. G. and R. M. Kaminski. 1985. Long-term wetland-waterfowl trends in
Saskatchewan grasslands. Journal of Wildlife Management 49:212-222.
Loekmoen, J. T., H. F. Duebbert, and D. E. Sharp. 1990. Homing and reproductive habits of
Mallards, Gadwall, and Blue-Wing teal. Wildlife Monographs 106:3-28.
March, J. R. and R. A. Hunt. 1978. Mallard population and harvest dynamics in Wisconsin.
Wisconsin Department of Natural Resources. Technical bulletin.
McCarty, J. P. 2001. Ecological consequences of recent climate change. Conservation Biology
15:320-331.
Mendenhall, V. M. and H. Milne. 1985. Factors affecting duckling survival of Eiders Somateria
mollissima in northeast Scotland. Ibis 127:148-158.
NCDC. 2012. Time bias corrected divisional temperature, precipitation, and drought index.
Available online: http://www.ncdc.noaa.gov/temp-and-precip/—accessed: 6/15/2013.
28

Orthmeyer, D. L. and J. J. Ball. 1990. Survival of mallard broods in Benton Lake National
Wildlife Refuge in northcentral Montana. Journal of Wildlife Management 54:62-66.
Parker, I. M., D. Simberloff, W. M. Londsdale, K. Goodell, M. Wonham, P. M. Kareiva, M. H.
Williamson, B. Von Holle, P. B. Moyle, J. E. Byers, and L. Goldwasser. 1999. Impact:
toward a framework for understanding the ecological effects of invaders. Biological
Invasions 1:3-19.
Prince, H. H., P. I. Padding, and R. W. Knapton. 1992. Waterfowl use of the Laurentian Great
Lakes. Journal of Great Lakes Research 18:673-699.
Rotella, J. J. and J. T. Ratti. 1992. Mallard brood survival and wetland habitat conditions in
south-western Manitoba. Journal of Wildlife Management 54:499-507.
Reynolds, R. E. 1987. Breeding duck population, production and habitat surveys, 1979-85.
North American Wildlife and Natural Resources Conference Transactions 52:186-205.
Ringleman, J. K. 2002. Identifying the factors that limit duck production. United States Fish
and Wildlife Leaflet. 13,7:1-8.
SAS Institute Inc. 2004. SAS 9.3. SAS Institute Inc, Cary, NC.
Scott, R. W. and F A. Huff. 1996. Impacts of the great lakes on regional climate conditions.
Journal of Great Lakes Research 22:845-863.
Sedinger, J. S. 1992. Ecology of prefledging waterfowl. Pp. 109-127 in B. D. J. Batt, A. D.
Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu eds.
Ecology and Management of Breeding Waterfowl. University of Minnesota Press,
Minneapolis, Minnesota.
Simpson, J. W., T. J. Yerkes, B. D. Smith and T. D. Nudds. 2005. Mallard duckling survival in
the Great Lakes Region. Condor 107:898-909.
Simpson, J. W., T. J. Yerkes, T. D. Nudds, and B. D. Smith. 2007. Effects of habitat on mallard
duckling survival in the Great Lakes region. Journal of Wildlife Management 71:18851891.
Smith, R. L. 1970. Response of Pintail breeding populations to drought. Journal of Wildlife
Management 34:943-946.
Snell, E. A. 1987. Wetland distribution and conversion in southern Ontario. Inland waters and
Lands Directorate, Environment Canada, Ottawa, ON, CanadaTable 6. Working paper
48, En 73-4/48E.

29

Soulliere, G. J., B. A. Potter, J. M. Coluccy, R. C. Gatti, C. L. Roy, D. R. Luukkonen, P. W.
Brown, and M. W. Eichholz. 2007. Upper Mississippi river and great lakes region joint
venture waterfowl habitat conservation strategy. U.S. Fish and Wildlife Service, Fort
Snelling, Minnesota, USA.
Sorensen, L. G., R. Goldberg, T. L. Root, and M. G. Anderson. 1998. Potential effects of global
warming on waterfowl populations breeding in the northern Great Plains. Climatic
change 40:343-369.
Stoudt, J. H. 1969. Relationships between waterfowl and water areas on the Redvers Waterfowl
Study Area. Page 123-131 in Saskatoon Wetland Seminar. Canada Wildlife Service
REP. Ser. No 6. Ottawa ON, Canada.
Talent, L. G., R. L. Jarvis, and G. L. Krapu. 1983. Survival of mallard broods in south-central
North Dakota. Condor 85:74-78.
Thompson, D. Q., R. L. Stuckey, and E. B. Thompson. 1987. Spread, impact, and control of
purple loosestrife (Lythum salicaria) in North American wetlands. U. S. Fish and
Wildlife Service. 55 pages. Jamestown, ND: Northern Prairie Wildlife Research Center
online. http://npwrc.usgs.gov/resource/plants/loostrf/index.htm (Version 04JUN99).
U. S. Fish and Wildlife Service. 2006. Migratory bird hunting activity and harvest during the
1999 and 2000 hunting seasons – Final report. US Department of the Interior,
Washington, DC.
U. S. Fish and Wildlife Service. 2013. Adaptive harvest management: 2013 Hunting season. U.
S. Department of the Interior, Washington, D. C. 62 pp. Available online at
http//www.fws.gov/migratorybirds/mgmt/AHM/AHM-intro.htm.
U. S. Fish and Wildlife Service. 2013. Trends in duck breeding populations, 1955-2013. U. S.
Department of the Interior, Washington, D. C. USA.
Van Horn, K., K. Benton, and R. Gatti. 2006. Waterfowl breeding population survey for
Wisconsin, 1973-2006. Wisconsin Department of Natural Resources, Wildlife
Management Administrative Report, Madison, WI.
Weller, M. W. and C. S. Spatcher. 1965. Role of habitat in the distribution and abundance of
marsh birds. Iowa Agricultural and Home Economics Experiment Station Special Report
43.
Welling, C. H. and R. L. Becker. 1993. Reduction of Purple loosestrife establishment in
Minnesota wetlands. Wildlife Society Bulletin 21:56-64.
Zimmerman, G. S., W. A. Link, M. J. Conroy, J. R. Sauer, K. D. Richkus, and G. S. Boomer.
2010. Estimating migratory game-bird productivity by integrating age ratio and banding
data. Wildlife Research 37:612-622.
30

CHAPTER 3 FACTORS AFFECTING HARVEST RATES OF FEMALE MALLARDS
BANDED IN THE GREAT LAKE STATES

Introduction
Wildlife management is focused around three guiding principles; managing populations,
habitat and people (Giles 1978) for wildlife-related benefits such as recreational hunting,
photography and wildlife viewing. Managers must focus attention on all three principles as each
potentially influences the others. Furthermore, populations are managed for continued
recreational opportunities for future generations of outdoor enthusiasts.
Waterfowl management is no different and mallards (Anas platyrhynchos) are a focal
species of waterfowl management (USFWS 2013). Mallard harvest is highest among duck
species taken in the Mississippi flyway (USFWS 2013) and duck hunting has substantial
economic, recreational, and aesthetic value (Martin et al. 1980; Virtiska 2013). Continued
management for sustaining populations of mallards and recreational opportunities is important.
Aerial surveys have been conducted since 1955 over the Prairie Pothole region (PPR) to
obtain spring breeding population and pond count estimates (USFWS 2013). This area is
thought to be one of the most important waterfowl breeding areas (Hochbaum 1983) and holds
the most breeding ducks in North America (Pospahala et al. 1974). Furthermore, large scale
banding operations were initiated in the 1950s and managers use data from recovered marked
individuals to obtain robust estimates of survival and harvest rates. Furthermore, aerial surveys
provide information that helps estimate population size for a myriad of species and this
information helps in setting regulations for waterfowl hunting seasons.
Considerable time and money is annually devoted to setting hunting season regulation
frameworks for waterfowl and this activity is a primary means of managing populations. Flyway

31

Councils, States, and the U.S. Fish and Wildlife Service (USFWS) set frameworks at broad
regional levels (i.e. flyways; Johnson et al. 1993; Conroy et al. 2002) and the States may adopt
regulations as liberal as the frameworks allow or may further restrict regulations. Waterfowl
abundance is thought to be influenced by habitat availability (Crissey 1969) and harvest
dynamics, thus monitoring and management of harvest and habitat should ensure continued duck
abundance, recreational opportunities and subsequent hunter participation. Interest in the role of
hunter harvest on waterfowl population dynamics has been a concern for decades (Nichols et al.
1995).
Waterfowl harvest management in the 1960s took a very simple approach. That is, when
spring breeding habitat conditions were good and duck abundance estimates were high, hunting
seasons were liberalized (i.e. long seasons and high daily limits; Babcock and Sparrowe 1989).
The opposite was true when duck numbers and pond abundance were down. This approach is
considered to be a “passive” approach to management (Sedinger and Herzog 2012). Over time,
as populations changed, seasonal regulations waxed and waned accordingly. Then, beginning in
the early 1990s, a new approach to waterfowl harvest management was introduced (Williams and
Johnson 1995).
Adaptive harvest management (AHM; USFWS 1995) was adopted as a means to reduce
the amount of uncertainty in; 1) population dynamics and 2) the impact of harvest when setting
waterfowl regulations (Walters 1986; USFWS 2014). Furthermore, it explores the relationships
between regulations, harvest, and waterfowl abundance (USFWS 2014) and is intended to reduce
uncertainties about how density dependence affects mallard population dynamics. Managers use
monitoring data to build models that predict population size under various population dynamic
scenarios and then compare these predictions to observed population levels. Furthermore,

32

harvest rates are predicted annually for Mid-continent (MC) adult male mallards for three
regulation packages and compared to those observed from band-recovery data. Mallards of the
Great Lake States (Michigan, Minnesota, and Wisconsin; GLS) are also considered in
management decisions for the MC mallard population.
Mallards of the GLS are managed under the AHM protocol as part of the MC mallard
population. The GLS population has been monitored as a whole since 1991, but monitoring
began in 1968 and 1973 for Minnesota and Wisconsin, respectively. During this time frame, the
mallard population increased despite harvest rates exceeding 50-year highs and record numbers
of waterfowl hunters (USFWS 2013; Figure 6). The population reached an all-time high in the
early 2000s before declining. Other measures of population trends such as the North American
Breeding Bird Survey also support the notion that mallards were on a long-term increase in the
GLS since 1968 (Sauer et al 2014.) Furthermore, the GLS mallard population has been prone to
very high local harvest; total GLS mallard harvest is often comprised of > 50% local mallards
(75, 57, and 80% for Michigan, Minnesota, and Wisconsin respectively; T. Arnold and C. de
Sobrino, unpub data).

33

12000
GLS

1200

MC

1000

10000
8000

800
6000
600
4000

400

2000

200

0
1955
1958
1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

0

MC mallard estimate (x1000s)

GLS mallard estimate (x1000)

1400

Figure 6. Mallard population estimates for the GLS (red line) and the MC (blue line; not
including the GLS) from 1961-2011. Surveys began in 1968, 1973, and 1991 for
Minnesota, Wisconsin, and Michigan, respectively. Michigan supports the most breeding
mallards of the GLS which explains the large increase in 1991 when Michigan began
surveying mallard abundance.
The apparent decline in GLS breeding mallard abundance has been attributed to nonbreeding season survival (i.e. high harvest pressure under AHM; Coluccy et al. 2008) and
declining productivity (Van Horn 2006. Furthermore, over the past 4 decades, hunter numbers
have declined in the GLS and throughout the Mississippi Flyway. Policy makers have been
concerned about recent trends for mallards and waterfowl hunters and Michigan and Wisconsin
have restricted mallard limits to include only 1 hen daily since the inception of AHM while the
federal framework for hen mallards has allowed 2 in the daily harvest; similarly Minnesota
restricted hen mallard daily limits, but for a subset of years during AHM. Little research has
been done to determine how this restriction beyond federal frameworks has influenced hen
mallard harvest rates.

34

The goal of this research was to shed some light on how harvest regulations and hunter
numbers have influenced mallard harvest rates in the GLS. In particular, have restrictive
regulations that have been implemented in the GLS since the inception of AHM been effective?
The main objectives of this chapter were twofold. First, I investigated effects of hunter
numbers at varying spatial scales on kill rates.
Second, I investigated impacts of regulations at varying scales (i.e. season length and
daily limit) and hunter numbers on kill rates. My main interest is to understand how decisions
regarding hen mallard daily limit at varying spatial scales have influenced kill rates.

Study area and methodology
This study included mallards breeding in the GLS, which are managed as part of the MC
population of mallards (USFWS 2013). The majority of the GLS lies within the Prairie
Hardwood and Boreal Hardwood Transition ecoregions except for parts of western and southern
Minnesota and southeastern Michigan which are considered Prairie Pothole and EasternTallgrass Prairie ecoregions, respectively (PPR; NABCI 2000). Weather patterns in this region
are largely driven by strong lake effects and wetlands are more permanent than those of the PPR.
Mallards in this region contributed 8-15% of the spring abundance of the entire MC mallard
population (USFWS 2013).
Duck hunter analysis
My goal was to select the estimate of waterfowl hunter numbers from different spatial
scales that was most predictive of kill rates under various management scenarios. I explored
three different scales of hunter numbers to determine the most appropriate scale.

35

I used samples of mallards banded in the GLS and harvested throughout North America
from 1961-2011 to estimate kill rates. Band and recovery data were obtained from the U.S.
Geological Survey; I restricted the analysis to birds banded during “pre-season” (JulySeptember) banding operations 1961-2011 in Michigan, Minnesota, and Wisconsin that were
marked with “federal numbered band only” or “captured by spotlighting” (Add_Info codes ‘00’
and ‘70’ respectively), and were “normal, wild-birds” (Status ‘3’). I omitted birds that were
banded with reward bands as reward bands are often reported at a higher rate than are standard
federal leg bands (Nichols et al. 1991). Recoveries were restricted to birds that were “shot” or
“found dead” (reporting codes “00” and “01”, respectively) from September 1-January 31. Birds
were classified into two age groups (1=hatch year and local and 2=after hatch year).
First, I estimated direct recovery rates for each state/age class (DRR) defined as, the
proportion of the banded sample that was shot and reported (Shot) during the hunting season
immediately following banding,

; where

is the number of banded

mallards in year i in age class j. I then adjusted DRR estimates for reporting rates to estimate
harvest rate (HR) by the equation,

; where

is the annual reporting rate as

estimated from reward band studies (Nichols et al. 1991). I adopted reporting rates used by
Arnold (T. Arnold, unpub data; 38% 1961-1995 and 77% 1996-2011). Band reporting rates
increased in ~ 1996 when implementation of “toll-free” bands was adopted, allowing hunters to
report harvested birds via telephone where previously birds could only be reported via USPS
mail service (Royle and Garretson 2005). Estimates of wounding loss are ~20% (Anderson and
Burnham 1976) so I adjusted HR to obtain kill rate (KR). The final equation to estimate KR is;
̂

.

36

I obtained estimates of the number of Mississippi flyway waterfowl hunters from the
USFWS Federal Duck Stamp office (USFWS 2012). This report documents estimates of
waterfowl hunters based on the sale of Federal Migratory Bird Hunter and Conservation Stamps
(hereafter duck stamp) for each state in the Mississippi flyway. . In this analysis, I was
interested in evaluating the impact of estimates of duck hunter numbers on kill rates at varying
spatial scales as numbers of waterfowl hunters, more specifically duck hunters, likely influence
kill rates of ducks at varying spatial scales (Conroy et al. 2005). Duck hunting regulations are
set according to the AHM protocol and thus are flyway wide, therefore kill rates could be
influenced by the number of duck hunters in the entire Mississippi flyway. I used the number of
duck stamps sold at the flyway scale (hereafter FHUNT) as one predictor of kill rates. T. Arnold
an de Sobrino (unpub data) suggest that mallard harvest in the GLS is largely (>50%) comprised
of local breeding birds, so the potential for kill rates to be more closely tied to local hunter
numbers exists. As an estimate of waterfowl hunters at the state scale, I used the number of duck
stamps sold in each state (hereafter SHUNT). Furthermore, if mallards from one state are
primarily harvested in an isolated region, I hypothesized that duck hunter numbers in a given
harvest region may have the greatest impact on kill rates. To obtain an estimate for this
hypothesis, I analyzed band recovery data for each of the states to find the top five harvest
states/provinces for each stock of mallards (mallards banded in a state). I used data from
mallards banded and recovered in the same year (direct recoveries) to select the states and
provinces. The province of Ontario was one of the top 5 kill states/provinces for MI banded
mallards. Since estimates of duck hunters in the Canadian provinces only date back to 1966, I
decided to not consider this variable in kill rate models (Environment Canada 2014). I then ran
separate analyses for all 3 states considering FHUNT and SHUNT as predictors of kill rates.

37

Duck hunter number estimates from duck stamp sales have some biases (K. Wilkins,
USFWS, personal communication). In recent years (2000-2011) the estimates are likely not as
robust because duck stamp sales from large corporations have not been tracked back to the state
of sale, but rather the corporations’ headquarters and also the United States Post Office reported
its’ duck stamp sales to regional offices which may not be contained in the state of sale (K.
Wilkins personal communication). Furthermore, speculation has been made that purchasing of
duck stamps by non-hunters has increased over time (K. Wilkins personal communication).
I fitted linear models in program SAS (SAS 9.3) to evaluate the effects of hunter numbers
at varying spatial scales on kill rates. I ranked models using Akaike’s information criterion
adjusted for small sample size (AICc; Burnham and Anderson 2002). The top ranking model for
each state would then allow us to select the appropriate scale to model hunter numbers in my kill
rate models. All estimates of duck hunter numbers were standardized with a mean of zero and
standard deviation of one. I used a Chi-square goodness-of-fit test, ̂

, where

is the chi-

square test statistic and df is the estimable parameters from the global model plus one for the
estimation of ̂ to check for overdispersion in the data.

Kill rate analysis
I used the Mixed Models Procedure (PROC MIXED) in SAS (SAS Institute 2004) to
model the response of kill rates to age class (hatch-year, and after hatch-year), season length
(hereafter DAYS), daily hen mallard limit at the state level (hereafter LIMIT), and hunter
numbers (hereafter SHUNT or FHUNT) at the flyway level for Minnesota and state level for

38

Michigan and Wisconsin. I attempted to include season opening dates and dates in which it ran,
but states have different zones that open on different dates. The DAYS and LIMIT variables
were categorical with DAYS divided into three categories (1- <38 days; 2- >= 38<50; 3->=50)
and LIMIT into two categories for all states based on the observed daily limit and season lengths
from state regulations (Table 3). States did not restrict season length beyond the federal
frameworks, so DAYS was relevant only at the flyway scale. I omitted two and three years of
data from Wisconsin and Minnesota respectively as they had seasons with a 4 hen mallard daily
limit and there were too few years with these limits to get meaningful estimates. However, the
LIMIT variable did vary temporally from the federal framework as states adopted more
restrictive daily limits in many cases. Prior to 1995, there was only one year (1968) when
federal regulations had multiple options regarding for the Mississippi flyway states regarding
season length, daily limits and species restrictions (Martin and Carney 1977).
Waterfowl managers in the GLS have suggested that restrictions of federal regulations at
the state level have little impact on kill rates (S. Cordts personal communication), but I felt
obligated to mention these fine details regarding variation in season length and daily limits.
Substantial variation in federal framework regulations and environmental conditions makes
understanding directly how changes in season length or daily limit influence kill rates convoluted
(William and Johnson 1995). Nonetheless, I modeled kill rates in response to coarse changes in
regulations that are most likely to impact kill rates. Data to determine state regulations used for
predictor variables was obtained from the Wisconsin Department of Natural Resources,
Minnesota Department of Natural Resources, and the Michigan Department of Natural
Resources.

39

Table 3. Number of years that a given state had various season lengths and daily hen mallard
limit combinations. Season lengths were categorized as follows: 1= <38 days; 2=≥ 38
days <50 days; 3=≥50 days. The limit was the total number of hen mallards allowed per
day by each state. Years with season limits of four were omitted from the analyses.
Limit
1

2
Length

Michigan
Minnesota
Wisconsin

1
8
8
8

2
6
7
6

3
17
9
17

1
2
2
2

2
5
5
5

3
13
17
11

I evaluated eleven a priori models and an intercept only model to test for the effects of
hunting regulations and hunter numbers on kill rates of GLS mallards (Table 4).
Table 4. A priori model set considered to predict kill rates for hen mallards in Michigan,
Minnesota, and Wisconsin from 1961-2011. All models included the covariate HUNT,
AGE, and an intercept. I used the standardized average of duck hunters from Michigan
and Wisconsin for those states respectively and used the standardized average of flyway
duck hunters for Minnesota.
KILL RATE MODELS
AGE + HUNT + LIMIT
AGE + HUNT + LIMIT + LIMIT*HUNT
HUNT + AGE*LIMIT
AGE + HUNT + DAYS + LIMIT
AGE + HUNT
AGE + HUNT + LIMIT + DAYS + LIMIT*HUNT*DAYS
AGE + HUNT + DAYS*LIMIT
AGE + HUNT + DAYS
HUNT + AGE*DAYS
AGE + HUNT + DAYS + DAYS*HUNT
HUNT + AGE*DAYS*LIMIT
INTERCEPT

All models included the estimate of duck hunter numbers that was most predictive of kill
rates from the hunter numbers analysis, the AGE variable in all models as there is strong
40

evidence for differential harvest vulnerability between age classes (Reynolds 1987), and finally
an intercept term.
I used the ESTIMATE function in SAS to predict kill rates under different harvest
regulation scenarios. I made kill rate predictions using a sub-set of models containing 95% of
the model weight and then reweighted these models for model-weighted prediction (Burnham
and Anderson 2002). To simulate recent levels of hunter participation, I used the 2007-2011
standardized average of FHUNT for Minnesota and SHUNT for Michigan and Wisconsin. I
followed Burnham and Anderson (2002) for model-weighted estimation and associated
prediction variances.
During AHM years (1997-2011), Minnesota was the only state to adopt the 2-hen daily
limit allowed under the federal framework while in other years (2005-2010) was more restrictive
than the federal framework by allowing only 1 hen daily limit. I ran a separate set of models for
AHM years in Minnesota to estimate the effects of state-specific regulation changes during
AHM (Table 5). I used the best ranking model based on AICc and then the ESTIMATE function
to predict kill rates during one hen and two hen daily limits for Minnesota banded female
mallards. This would help explain the affects that an individual state can have on kill rates when
further restricting harvest from the federal framework.
Table 5. A priori model set considered to predict the influence on kill rates for hen mallards in
Minnesota from 1997-2011. Note that season length (variable DAYS) was not included
due to a constant season length of 60 days from 1997-2011.
KILL RATE MODELS FOR AHM
AGE + FHUNT + LIMIT
AGE + FHUNT + LIMIT + LIMIT*FHUNT
FHUNT + AGE*LIMIT
AGE + FHUNT
INTERCEPT

41

Results
Duck hunter analysis
Duck hunter numbers initially increased, peaked during the 1970’s and then declined
throughout the GLS and the Mississippi flyway during the 51-year (1961-2011) period over

MI

MN

WI

MS flyway

1200

1000
800
600
400

200
0

MS flyway duck hunters (1000s)

200
180
160
140
120
100
80
60
40
20
0

1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

GLS duck hunters (1000s)

which I modeled effects of duck hunter numbers on mallard kill rates for the GLS (Figure 7).

Figure 7. Federal duck stamp sales for the Great Lake States (Michigan= dashed line,
Minnesota=solid, thin line, Wisconsin=dotted line) and the entire Mississippi flyway (MS
flyway; thick, gray line) from 1961-2011.
Furthermore, the relationship between waterfowl hunters and MC mallard population changed in
the mid-1990s after the AHM protocol was adopted (Figure 8).

42

MF waterfowl hunters (1000s)

1200
1000
800

600

1961-1996

400

1997-2011

200
0
0

5
10
MC mallard population (millions)

Figure 8. Relationship between the estimate of Mississippi Flyway waterfowl hunters (y-axis)
and the Mid-continent mallard population (x-axis) from 1961-2011 (1961-1996 circles
and 1997-2011 squares). Variation explained by the relationship, r2, for 1961-1996 and
1997-2011 was 0.58 and 0.18, respectively.
I found that the number of duck hunters at the flyway and state scale was an important predictor
of kill rates in Minnesota ( ̂ = 0.027, SE 0.006) and Wisconsin ( ̂ = 0.041, SE 0.006),
respectively when added to an age and intercept model (Table 6). Although the age only model
was the top ranking model in Michigan (Table 6), the second-ranked model contained the state
scale waterfowl hunter number variable ( ̂ = 0.013, SE 0.005). For evaluating models with
regulation predictors (e.g. season length and hen daily limit), I included the FHUNT variable in
all Minnesota models and the SHUNT variable in all Michigan and Wisconsin models.

43

Table 6. A priori models used to determine the most predictive estimate of hunter numbers at
varying spatial scales to include in my kill rate analyses for (a) Michigan , (b) Minnesota,
and (c) Wisconsin.
a) Michigan
AGE
AGE + SHUNT
AGE + FHUNT
INTERCEPT

K

AICc

ΔAICc

Wi

4
5
5
2

-287
-283.9
-283.4
-261.4

0
3.1
3.6
25.6

0.73
0.15
0.12
0.00

b) Minnesota
AGE + FHUNT
AGE + SHUNT
AGE
INTERCEPT
c) Wisconsin
AGE + SHUNT
AGE + FHUNT
AGE
INTERCEPT

44

K

AICc

ΔAICc

Wi

5
5
4
2

-264.0
-257.5
-255.7
-233.0

0
6.5
8.3
31

0.95
0.04
0.01
0.00

K

AICc

ΔAICc

Wi

5
5
4
2

-267.2
-254.6
-238.4
-201.2

0
12.6
28.8
66

0.99
0.00
0.00
0.00

Kill rate analysis
Kill rates were variable across the 51-year (1961-2011) time period for the GLS (Figure
9).

a) Michigan

Annual kill rate

0.5
0.4
0.3

AHY

0.2

HY

0.1
1961
1965
1969
1973
1977
1981
1985
1989
1993
1997
2001
2005
2009

0

b) Minnesota

Annual kill rate

0.5
0.4
0.3

AHY

0.2

HY

0.1
1961
1965
1969
1973
1977
1981
1985
1989
1993
1997
2001
2005
2009

0

c) Wisconsin

Annual kill rate

0.5
0.4
0.3

AHY

0.2

HY

0.1
1961
1965
1969
1973
1977
1981
1985
1989
1993
1997
2001
2005
2009

0

Figure 9. Kill rates for female mallards banded in Michigan (a), Minnesota (b), and Wisconsin
(c) 1961-2011 (AHY=solid line and HY=dashed line).
45

Kill rates were generally higher among hatch year females and kill rates were higher during the
1960s and 1970s, followed by a period of restrictive regulations in the mid-1980s when they
declined and then kill rates under AHM began to increase to historically intermediate levels.

Kill rates in all models were higher for hatch year birds than for after hatch year birds,
thus I included AGE in all models. Kill rates in Michigan and Wisconsin were influenced more
by hunter numbers and daily hen limit than by season length (Table 7). Kill rates generally
increased when daily limits increased from 1 hen to 2 hens. The top ranking model for
Minnesota supported an interaction term with DAYS and LIMIT while this interaction was not in
the top models for the other states (Table 7).
Table 7. A priori models and model weights for predicting hen mallard kill rates in the GLS: (a)
Michigan; (b) Minnesota; and (c) Wisconsin. Model weights, wi, that sum to ~95% were
used to model average kill rate estimates for each state. All models contained an
intercept and the variable HUNT as the number of waterfowl hunters likely influences
kill rates. Also, the variable AGE was in all models as research has shown that juveniles
are more vulnerable to harvest than are adults.

a) Michigan
AGE + HUNT + LIMIT
HUNT + AGE*LIMIT
AGE + HUNT + DAYS + LIMIT
AGE + HUNT + LIMIT + LIMIT*HUNT
AGE + HUNT + DAYS*LIMIT
HUNT + AGE*DAYS*LIMIT
AGE + HUNT + LIMIT + DAYS + LIMIT*HUNT*DAYS
AGE + HUNT
AGE + HUNT + DAYS
HUNT + AGE*DAYS
AGE + HUNT + DAYS + DAYS*HUNT
INTERCEPT
46

K
7
7
10
9
11
15
16
5
8
9
11
2

AICc ΔAICc
-310.4
0
-309.3
1.1
-307.2
3.2
-303.6
6.8
-302
8.4
-295.5 14.9
-286.7 23.7
-283.9 26.5
-275.2 35.2
-268.6 41.8
-266
44.4
-261.4
49

Wi
0.55
0.32
0.11
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00

Table 7 (cont’d)
b) Minnesota
AGE + HUNT + DAYS*LIMIT
AGE + HUNT + DAYS + LIMIT
AGE + HUNT + LIMIT
AGE + HUNT
HUNT + AGE*LIMIT
AGE + HUNT + LIMIT + LIMIT*HUNT
AGE + HUNT + DAYS
HUNT + AGE*DAYS*LIMIT
AGE + HUNT + LIMIT + DAYS + LIMIT*HUNT*DAYS
HUNT + AGE*DAYS
AGE + HUNT + DAYS + DAYS*HUNT
INTERCEPT

K
11
10
7
5
7
9
8
15
16
9
11
2

AICc ΔAICc
-279.2
0
-269.9
9.3
-267.2
12
-263.7 15.5
-262.8 16.4
-261.5 17.7
-258.1 21.1
-257.6 21.6
-247.9 31.3
-247.7 31.5
-246.3 32.9
-233
46.2

Wi
0.99
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

K
7
7
9
10
5
16
11
8
15
9
11
2

AICc ΔAICc
-283.4
0
-279.2
4.2
-278.3
5.1
-270.6 12.8
-267
16.4
-263.3 20.1
-262.5 20.9
-255.1 28.3
-245.7 37.7
-245.4
38
-243.9 39.5
-201.2 82.2

Wi
0.83
0.10
0.06
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

c) Wisconsin
AGE + HUNT + LIMIT
HUNT + AGE*LIMIT
AGE + HUNT + LIMIT + LIMIT*HUNT
AGE + HUNT + DAYS + LIMIT
AGE + HUNT
AGE + HUNT + LIMIT + DAYS + LIMIT*HUNT*DAYS
AGE + HUNT + DAYS*LIMIT
AGE + HUNT + DAYS
HUNT + AGE*DAYS*LIMIT
HUNT + AGE*DAYS
AGE + HUNT + DAYS + DAYS*HUNT
INTERCEPT

Predicted kill rates were generally higher during years when hen mallard daily limits were 2/day,
but confidence limits overlapped except in Wisconsin (Figure 10).

47

a) Michigan

0.4
0.35

Kill rate

0.3
0.25
0.2

1 hen

0.15

2 hen

0.1
0.05
0
HY

AHY

b) Minnesota

0.18
0.16
0.14
Kill rate

0.12
0.1

1 hen

0.08

2 hen

0.06
0.04
0.02
0
HY

AHY

c) Wisconsin

0.35
0.3

Kill rate

0.25
0.2

1 hen

0.15

2 hen

0.1
0.05
0
HY

AHY

Figure 10. Model weighted predictions for: (a) Michigan, (b) Minnesota, and (c) Wisconsin
female mallard kill rates from top 2 ranked models with model weighted 95% CI.

48

Confidence intervals for predicted kill rates for one and two hen daily limits in Minnesota
during AHM (1997-2011) overlapped and the relative change in kill rates going from one to two
hen daily limits as a state-imposed regulation restriction was much less than the relative change
during years when the frameworks allowed the entire flyway to change daily limits (Figure 11;
Table 8).
0.18
0.16
0.14

Kill rate

0.12
0.1

1 hen

0.08

2 hen

0.06
0.04
0.02
0
AHY

HY

Figure 11. Mean kill rate for MN female mallards (AHY=adult and HY=juvenile)
during years when MN adopted a 1-hen daily limit (2005-2010) and a 2-hen daily limit (19972004 and 2011) and 95% CI’s.

I also found that percent change in kill rates, (KRpre-AHM-KRAHM)/KRAHM*100, during
pre-AHM years (1961-1997) was greater than that during AHM years (1997-2011) for
Minnesota. I then summarized the percent change in kill rates for Michigan and Wisconsin to
compare those to Minnesota (Table 8). Percent increase for predicted kill rates under 2
compared to 1 hen daily in Michigan and Wisconsin was 43% and 36% for HY birds and 54%
and 68% for AHY birds, respectively (Table 8). Percent increases in kill rates going from 1 to 2

49

hens daily were greater for AHY birds compared to HY birds in all states and time periods
(Table 8).
Table 8. Percent change in model-averaged predicted kill rates when regulations change from 1hen to 2-hens daily by state.
Michigan
HY
AHY
1961-1996 (MN only)
1997-2011 (MN only)
1961-2011

43.18

Minnesota
HY
AHY
62.36
12.23

125.31
25.43

9.07

20.47

53.92

Wisconsin
HY
AHY

35.74

68.00

Discussion
Duck hunter analysis
Waterfowl hunter numbers have been temporally dynamic across the Mississippi flyway
and so it is important to consider effects of hunter numbers when evaluating harvest regulation
and kill rates. Waterfowl management policy makers have the capacity to restrict hunting
licenses (e.g. South Dakota Game Fish and Parks 2014), but regulations not explicitly intended
to restrict hunter numbers can also impact the number of hunters participating in migratory bird
hunting seasons (Luukkonen and Frawley 2010); however, regulations have been relatively
stable under AHM for 17 years and the recent declines in duck hunter numbers appear to be
largely disconnected from changes in the status of waterfowl populations or harvest regulations
(Vrtiska 2013). With fewer hunters, it is unlikely that liberal regulations would result in kill
rates comparable to historic levels with higher hunter numbers given the same regulation regime.
The average Mississippi flyway duck hunter hunts about 7.5 days out of the available 60 day
season (Raftovich et al. 2014) and relatively stable effort per hunter could contribute to the weak
relationship between historic season lengths and kill rates. Season lengths averaged ~39, 49, 41,
47, and 60 days during the 1960s-2000s, respectively.
50

My models showed that estimates of duck hunters at the state scale influence kill rates
more so in states with higher harvest derivations for local birds (Michigan and Wisconsin) and
estimates of total flyway duck hunters suggest the best fit for Minnesota kill rate models. This
was not surprising as local mallards make up the majority of the harvest in Michigan and
Wisconsin, thus local hunters may heavily influence kill rates.
Michigan and Wisconsin are unique in that harvest of locally-banded birds is higher than
in Minnesota and many surrounding states (T. Arnold and C. de Sobrino unpub data). The
contribution of local birds in the Wisconsin harvest increased over time (Munro and Kimball
1982; Trots 1987; R. Gatti unpub data). Zuwerink (2000) found that ~80% of Wisconsin’s
mallard harvest occurs from locally breeding birds. Speculation has been made for the GLS (T.
Arnold personal communication; D. Luukkonen personal communication) that harvest derivation
may be over-estimated as a large portion of the Hudson Bay lowlands are not included in large
scale banding operations and mallard densities there have been observed as high as 62
pairs/100km2 (Dennis 1974). Furthermore, the Hudson Bay Lowlands is the third largest marsh
in the world (Zoltai 1973), holding potentially vast numbers of breeding mallards that enter the
fall population unmarked. If a large portion of these birds migrate through or winter in Michigan
or Wisconsin and are harvested but not individually marked, then contribution of local mallards
to harvest derivations may be over-estimated.
Minnesota lies on the edge of the PPR and Prairie Boreal Hardwood transition zone and
hunters in this state likely harvest a large number of mallards migrating from more northerly
breeding areas (T. Arnold and C. de Sobrino, unpub data), thus subjecting local breeding
mallards to lower harvest vulnerability locally. As with all mallards that migrate farther south,
these birds may be equally prone to harvest in southern states during fall migration, but the

51

mixing of various mallard stocks in Minnesota during the fall, may explain why kill rates were
more responsive to flyway waterfowl hunter numbers instead of state waterfowl hunter numbers.
The top ranking model in the hunter number analysis for Michigan supported the AGE +
Intercept parameters and did not include an estimate of waterfowl hunter numbers. However,
models containing an estimate of flyway hunters and state hunters explained ~25% of the model
weight.
Kill rate analysis
I modeled the effects of regulations and duck hunter numbers on kill rates of female
mallards banded in the GLS and recovered throughout North America from 1961-2011. The
number of marked individuals in Michigan limited my ability to draw strong inference from the
Michigan data set. Brownie et al. (1985) found that generally a minimum of 300 banded
individuals would be required to analyze band recovery data. However, this number would vary
depending on species, recovery rates, etc. Sample size for AHY females in Michigan was often
<300. I ran an independent-samples t-test exploring direct recovery rates during years when
bandings were <300 and years >300. 95% CI overlapped for both age classes. Thus I feel
confident in my ability to draw inference from Michigan’s data set, however emphasis should be
made to obtain proper sample size for future research. Minnesota and Wisconsin had one year
each when AHY banding were <300.
Kill rates throughout the GLS have varied greatly over the past fifty years. Kill rate is the
proportion of the population harvested in a year and also takes into account the percentage of
birds that are killed but not recovered by hunters. Kill rates during the 1960s and into the 1970s
were at their peak (Figure 9). This came at a time when mallard populations in Minnesota and
Wisconsin (Michigan was not surveyed until the 1990s) were low, duck hunter numbers were at

52

50-year highs and harvest regulations were moderate with regards to daily hen limits. This
resulted in the highest kill rates in the past 50-years. This trend slowly declined until the mid1980s when duck populations in the MC declined and regulations became restrictive (i.e. 1 hendaily and 30-day seasons). This resulted in declining kill rates and declining Mississippi flyway
duck hunters. Furthermore, the relationship between waterfowl hunters and waterfowl

12000

1200

10000

1000

8000

800

6000

600

4000
2000

400
200
0

1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

0

MC mallard
population
Mississippi flyway
duck hunters

Mississippi flyway duck
hunters (1000s)

MC mallard population
(1000s)

populations became uncoupled (Figure 12).

Figure 12. Relationship between the population estimate for MC mallards (not including GLS
mallards) and the Mississippi flyway estimate of duck hunters as derived from the duck
stamp sale from 1961-2011. Note how the trend between the two “breaks” around 1995.
Higher kill rates are expected when hunter numbers are up and populations are low. Currently
we are at the opposite end of that spectrum where duck hunter numbers are down and duck
populations are high with lower kill rates despite liberal season lengths and limits. The result of
these recent conditions is relatively high annual mallard harvest per duck hunter in the GLS
compared to historic success (Figure 13).

53

Annual mallard
harvest/hunter

4
3.5
3
2.5
2
1.5
1
0.5
0

Annual mallard
harvest/hunter

4
3.5
3
2.5
2
1.5
1
0.5
0

a) Michigan

1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

Annual mallard
harvest/hunter

4
3.5
3
2.5
2
1.5
1
0.5
0

1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

b) Minnesota

1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009

c) Wisconsin

Figure 13. Annual mallard harvest per hunter as estimated from the Parts Collection Survey
harvest data for both male and female and juvenile and adult mallards and the estimate of
state waterfowl hunters from duck stamp sales for (a) Michigan , (b) Minnesota, and (c)
Wisconsin 1961-2011.
54

I found that daily limit has more influence on kill rate than season length in Michigan and
Wisconsin, but length was important in Minnesota. Variation of kill rates explained by season
length could partially be confounded due to classification of the DAYS variable, where the
majority of the third class was from AHM years when season length was longest. Nonetheless,
kill rate estimates were lower under longer seasons than they were under shorter seasons. I
examined models with interactions of hunter numbers and season lengths to understand if this
had an impact on kill rate but these models ranked ≤4.0ΔAICc from the top model. I conclude
that kill rates under 60-day seasons coupled with low hunter numbers are low and likely at a
sustainable level; this conclusion is supported by female mallard survival rates that have been
relatively stable for most Great Lakes states (T. Arnold et al. unpub) and productivity rates that
have not declined based on results of my previous chapter.
My results evaluating percent change in kill rates in response to change in daily limits
suggest that percent change was smaller during AHM years than pre-AHM for Minnesota,
suggesting that kill rates are more influenced by flyway wide regulation changes than they are
state regulation changes. Although harvest likely varies spatially, inference can be drawn from
Minnesota on the effect of more liberal daily hen limits on kill rates for mallards breeding in
Michigan and Wisconsin. Hunter survey data from Minnesota generally confirm that roughly
50% of the hunters sampled are happy with the regulations pertaining to hen mallard daily limits,
whether it be 2 or 1 (Fulton et al. 2002; Schroeder et al. 2012). Roughly 55% of survey
respondents favored going to a 2-hen daily limit or had no opinion in Michigan (Frawley 2004).
This can be viewed as trust in the DNR and more importantly use this as an opportunity to
promote waterfowl hunting by making regulations more favorable to waterfowlers. My data

55

suggest that the influence of states adopting a more restrictive limit than that of the federal
framework did not significantly impact kill rates. However, the impact from restricting, or in the
case of Michigan and Wisconsin, increasing the daily limit likely will vary spatially based upon
differences in harvest derivation of locally banded birds. I would hypothesize if Michigan or
Wisconsin were to increase the daily limit from one to two, this would result in a larger increase
in kill rate than in Minnesota, due to higher local harvest derivation, and perhaps local and
regional productivity rates, but likely not as high as my predictions from models that include
changes in frameworks that affect regulations in all states of the Mississippi Flyway. Other
studies, Arnold et al. (2013 unpub data) found that high harvest was negatively impacting
survival for Wisconsin AHY female mallards, but this occurred when kill rates were over 20%
during the 1960s and 1970s. Although I did not look at the relationship between harvest and
survival, it is evident that kill rates are much lower now than when they were impacting survival.
Models predicting the percent change in kill rates for the full time series suggest a
slightly larger change in Michigan and Wisconsin than in Minnesota. I speculate that this could
potentially be due to spatial and temporal variation in harvest of various mallard stocks of the
GLS.
Modeling of kill rates as a function of regulations is not as straightforward as it seems.
Complexity of the federal framework did not allow me to examine regulations at the flyway
scale for all years. For example, in 1970, the framework allowed for: 1) a 45-day season with a 4
duck daily limit, where 4 may be mallard, 2) a 55-day season with a 6 duck daily limit, where 2
may be mallard, and 3) a 50 day season with a point system, where mallards were either 90 or 20
points depending on the sex. Another caveat pertains to shooting hour restrictions in Minnesota

56

and Wisconsin. For example, Minnesota has had opening day shooting hours start later than the
rest of the year and also had four years when shooting hours ended at 4 P.M.
Mallard kill rates were at all-time highs during the 1960s when mallard populations were
near record lows and hunter numbers were near record highs. Despite the high kill rates during
this time, mallard populations continued to grow over the next forty years. I caution that the
Minnesota example be used solely as point of reference and the change in kill rates will likely be
different for Michigan and Wisconsin. I feel that if Michigan and Wisconsin adopted a 2-hen
mallard daily limit, kill rates would increase, but not likely as high as kill rates observed during
the 1960s-1970s.
I hope the results from this research will be used by waterfowl managers when
considering adopting regulations offered by the federal framework. If nothing else, I hope this
research will bring further discussion to the table for state waterfowl managers when setting
season regulations.

Management Implications
This research question stemmed from concern about the GLS mallards and if regulations
under AHM are sustainable. One expression of concern was the fact that the Michigan and
Wisconsin DNR were being more restrictive than the federal framework in regards to hen
mallard daily harvest limits. I attempted to evaluate the impact from various regulation changes
from the last 51-years and gained valuable information from Minnesota over the past 6-years
when they have permitted the maximum hen mallard harvest allowed by the Federal framework.
Survey data obtained from the Michigan Department of Natural Resources (Frawley
2013) suggests that ~30% of duck hunters are satisfied with the “number of ducks harvested”

57

and ~50% duck hunters would prefer a daily hen mallard limit greater than 1 or had no
preference for daily hen mallard limit (Frawley 2004). Minnesota has found that its duck hunters
are generally happy with hen mallard daily limits whether it is two or one daily (Fulton et al.
2002; Schroeder et al. 2012).
Past research such as Williams and Johnson (1995), and this research emphasize the
disparity between management actions and understanding of population dynamics. Williams and
Johnson (1995) noted the difficulty of understanding population dynamics when regulations
“chase” populations and habitat conditions. My results are similar in the lack of certainty
regarding the factors directly influencing kill rates. Regulations now are the most liberalized
they have been in the past 55-years and duck populations are at their highest. Potential
experimentation of altering regulations so they do not “chase” populations and habitat conditions
could help us improve our knowledge of the effects of regulations on population dynamics via
kill rates and survival rates (Sedinger and Herzog 2012).
My results suggest that female mallard kill rates in Michigan and Wisconsin would likely
increase slightly if these states decided to adopt a 2-hen daily limit; but I speculate that mallard
population abundance and future hunting success will not be affected by an increase in hen
mallard daily limits. Furthermore, there has been a negative relationship between hunter
numbers and mallard population estimates since the mid-1990s. Restricting hunting
opportunities further than the federal framework may be unnecessary to optimize cumulative
long-term harvest of mallard populations in the GLS.
As waterfowl managers, we are often the first to be aware of our direct management
actions as waterfowl are the most studied game species throughout the world (Nichols and
Johnson 1989). Given our knowledge of factors influencing mallard harvest rates, there appears

58

to be opportunities for additional sustainable mallard harvest in the GLS and current monitoring
programs (e.g. waterfowl abundance estimates, band return data, etc.) are in place to provide
feedback on outcomes of management experiments such as increasing the hen mallard daily
limit.

59

LITERATURE CITED

60

LITERATURE CITED

Anderson, D. R. and K. P. Burnham. 1976. Population ecology of the mallard. VI. The effect
of exploitation on survival. U. S. Fish and Wildlife Service Resource Publication 128.
66pp.
Babcock, M. B. and R. D. Sparrowe. 1989. Balancing expectations with reality in duck harvest
management. Transactions of the North American Wildlife and Natural Resources
Conference 54:594-599.
Brownie, C., D. R. Anderson, K. P. Burnham, and D. S. Robson. Statistical inference from band
recovery data: a handbook. U.S. Department of the Interior, Fish and Wildlife Service
Resource Publication 131.
Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: A
practical information-theoretic approach. Second edition. Springer, New York, New
York.
Coluccy, J. M., T. Yerkes, R. Simpson, J. W. Simpson, L. Armstrong, and J. Davis. 2008.
Population dynamics of breeding mallards in the Great Lake States. Journal of Wildlife
Management 72:1181-1187.
Conroy, M. J., M. W. Miller, and J. E. Hines. 2002. Identification and synthetic modeling of
factors affecting American black duck populations. Wildlife Monographs 150.
Conroy, M. J., C. J. Fonnesbeck, and N. L. Zimpfer. 2005. Modeling regional waterfowl
harvest rates using markov chain monte carlo. Journal of Wildlife Management 69:7790.
Crissey, W. F. 1970. Prairie potholes from a continental viewpoint. Pp. 161-171 in Saskatoon
wetlands seminar. Canadian Wildlife Service Series Report Series 6.
Dennis, D. G. 1974. Breeding pair surveys of waterfowl in southern Ontario. Pp. 45-52 in H.
Boyd, ed. Canadian Wildlife Service waterfowl studies in eastern Canada 1969-1973.
Canadian Wildlife Service Report Series 29:106.
Environment Canada. 2014. Permit sales. Available at http://www.ec.gc.ca/reommbs/default.asp?lang=En&n=C9046964-1.
Frawley, B. J. 2004. 2004 Waterfowl harvest survey. Michigan Department of Natural
Resources, Wildlife Division Report 3456.
Frawley, B. J. 2013. 2012 Waterfowl harvest survey. Michigan Department of Natural
Resources. Wildlife Division Report 3575.
Fulton, D. C., J. Vlaming, J. S. Lawrence, and E. W. Price. 2002. The 2000 waterfowl hunting
season in Minnesota: A study of hunters’ opinions and activities.

61

Giles, R. H. 1978. Wildlife management. W. H. Freeman Company. San Francisco, California.
416pp.
Hochbaum, G. S. 1983. Waterfowl of the prairie provinces and northwest territories:
Populations, habitat, and management. Proc. Western Hemisphere Waterfowl and
Waterbird Symposium. Int. Waterfowl Res. Bur., Supply and Services Canada, Catalog
CW66-63/1983E 1:34-36.
Johnson, F. A., B. K. William, J. D. Nichols, J. E. Hines, W. L. Kendall, G. W. Smith, and D.
Caithamer. 1993. Developing an adaptive management strategy for harvesting
waterfowl in North America. Transactions of the North American Wildlife and natural
Resources Conference 58:565-583.
Martin, E. M. and S. M. Carney. 1977. Population ecology of the mallard. IV. A review of
duck hunting regulations, activity, and success, with special reference to the mallard.
U.S. Fish and Wildlife Service Resource Publication 130.
Martin, F. W., R. S. Pospahala, and J. D. Nichols. 1980. Assessment and population
management of North American migratory birds. Environmental biomonitoring,
assessment, prediction, and management-certain case studies and related quantitative
issues. International Cooperative Publishing House, Fairland, Maryland, USA 187-239.
Munro, R. E. and C. F. Kimball. 1982. Population ecology of the mallard: VII. Distribution
and derivation of the harvest. U.S. Fish and Wildlife Service Resource Publication 147.
Nichols, J. D. and F. A. Johnson. 1989. Evaluation and experimentation with duck management
strategies. Transactions of the North American Wildlife and Natural Resources
Conference 54:566-593.
Nichols, J. D., R. J. Blohm, R. E. Reynolds, R. E. Trost, J. E. Hines, and J. P. Bladen. 1991.
Band reporting rates for mallards with reward band of different dollar values. Journal of
Wildlife Management 55:119-126.
Nichols, J. D., R. E. Reynolds, R. J. Blohm, R. E. Trost, J. E. Hines, and J. P. Bladen. 1995.
Geographic variation in band reporting rates for mallards based on reward banding.
Journal of Wildlife Management 59:697-708.
North American Bird Conservation Initiative. 2014. Bird conservation regions. Available at
http://www.nabci-us.org/map.html.
Pospahala, R. S., D. R. Anderson, and C. J. Hines. 1974. Population ecology of the mallard II.
Breeding habitat conditions, size of the breeding population, and production indices.
U.S. Fish and Wildlife Service Resource Publication 115:73.
Reynolds, R. J. 1987. Breeding duck populations, production and habitat surveys, 1979-85.
Transactions of the North American Wildlife and Natural Resources Conference 52:186205.
Royle, J. A. and P. R. Garretson. 2005. The effect of reward band value on mid-continent
mallard band reporting rates. Journal of Wildlife Management 69:800-804.
62

Sandilands, A. P. 2010. Birds of Ontario: Habitat requirements, limiting factors, and status.
UBC press, 2:60.
Schroeder, S. A., J. S. Lawrence, and S. D. Cordts. 2012. The 2010 waterfowl hunting season in
Minnesota: A study of hunters’ opinions and activities. University of Minnesota,
Minnesota Cooperative Fish and Wildlife Research Unit, Department of Fisheries,
Wildlife, and Conservation Biology.
SAS Institute Inc. 2004. SAS 9.3. SAS Institute Inc, Cary, NC.
Sauer, J. R., J. E. Hines, J. E. Fallon, K. L. Pardieck, D. J. Ziolkowski, Jr., and W. A. Link.
2014. The North American breeding bird survey, results and analysis 1966-2012.
Version 02.19.2014 USGS Patuxent Wildlife Research Center, Laurel, MD.
Sedinger, J. S. and M. P. Herzog. 2012. Harvest and dynamics of duck populations. Journal of
Wildlife Management 76:1108-1116.
Trost, R. E. 1987. A preliminary assessment of the recent distribution and derivation of mallard
harvest in the United States based on recoveries from breeding ground bandings, 197584. Unpublished report, Office of Migratory Bird Management, U.S. Fish and Wildlife
Service. 69pp.
U. S. Fish and Wildlife Service. 2012. Duck stamp sales by state 1934-2011. Available at
http://www.fws.gov/duckstamps/DuckStampSalesbyStateOctober2012.pdf.
U. S. Fish and Wildlife Service. 2013. Adaptive Harvest Management: 2013 Hunting Season.
U. S. Department of the Interior, Washington, D. C. USA.
U. S. Fish and Wildlife Service. 2013. Trends in duck breeding populations, 1955-2013. U. S.
Department of the Interior, Washington, D. C. USA.
Van Horn, K., K. Benton, and R. Gatti. 2006. Waterfowl breeding population survey for
Wisconsin, 1973-2006. Wisconsin Department of Natural Resources, Wildlife
Management Administrative Report, Madison, WI.
Vrtiska, M. P., J. H. Gammonley, L. W. Naylor, and A. H. Raedeke. 2013. Economic and
conservation ramifications from the decline of waterfowl hunters. Wildlife Society
Bulletin 37:380-388.
Walters, C. 1986. Adaptive management of renewable resources. Caldwell Press, Caldwell,
New Jersey, USA.
Williams, B. K. and F. A. Johnson. 1995. Adaptive management and the regulation of
waterfowl harvests. Wildlife Society Bulletin 23:430-436.
Zoltai, S. C. 1973. Vegetation, surficial deposits and permafrost relationships in the Hudson
Bay lowland. Proceedings of the Symposium on the Physical Environment of the
Hudson Bay lowland. Guelph Canada, University of Guelph. Pp 17-34.

63

Zuwerink, D. A. 2001. Changes in the derivation of the mallard harvest from the northern US
and Canada, 1966-1998. Master’s thesis, The Ohio State University, Columbus, Ohio
USA.

64