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

 

 

 

 

DATE DUE DATE DUE DATE DUE
r a m
‘Bhgié 0:22 APR 0 8 2009

ADI-“(1214 613.93

 

 

 

 

 

 

 

 

 

 

 

 

moo 6.1090031008965114

 

SURVIVAL, FALL MOVEMENTS, AND HABITAT USE OF
HUNTED AND NON-HUNTED RUFFED GROUSE
IN NORTHERN MICHIGAN
By

Margaret E. Clark

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Fisheries and Wildlife

2000

ABSTRACT
SURVIVAL, FALL MOVEMENTS, AND HABITAT USE OF
HUNTED AND NON-HUNTED RUFFED GROUSE
IN NORTHERN MICHIGAN
By

Margaret E. Clark

The ruffed grouse (Bonasa umbellus) is pursed by approximately 130,000 hunters
per year in Michigan. Hunters and wildlife biologists have many questions about the
effects of hunting on grouse and how to better manage forests for this popular bird. This
project takes a comprehensive ﬁve-year database (1993 — 1998) of fall grouse locations
in sites open and closed to hunting and combines it with a habitat quality analysis to
describe the characteristics of grouse behavior and survival in northern Michigan.

Ruffed grouse were radio-collared during the fall of each year in the Maltby Hills
and Pigeon River study sites in Northern Michigan. Survival data were collected from
August through May. In the ﬁrst study season (1993 - 1994), ruffed grouse in the 2 sites
closed to hunting had higher survival than did those in the 2 sites open to hunting. In all
other years, survival in the sites open to hunting was comparable or higher than survival
in the sites closed to hunting. In all sites, avian predation was the leading cause of
mortality. In the open sites, hunting was generally the second leading cause of mortality,
comprising 12 — 21% of known annual mortalities in the Maltby Hills site and I4 - 35%
of known mortalities in the Pigeon River site.

Collared birds were located several times per week to measure fall activity range

size and dispersal distances. Juvenile grouse tended to move longer distances and have

larger activity ranges than did adult grouse. More females dispersed and they moved
longer distances than did males. Activity range sizes were comparable to other studies in
the Northern United States, but dispersal distances in the Pigeon River were shorter than
recorded elsewhere.

Habitat quality was assessed in each site using the H31 model for ruffed grouse in
Michigan. Medium aged aspen (1 l - 29 yr.) and lowland conifers ranked highest in
cover quality, while upland hardwoods, oak, and old pine (> 30 yr.) ranked lowest. The
model’s spatial variable, distance to aspen, dominated the model when applied on a
landscape scale.

Using a geographic information system, habitat used (within activity ranges) was
compared with habitat available, and habitat quality within activity ranges was compared
to that within random circles. Ruffed grouse in the Maltby Hills study sites preferred
young pine (O — 29 yr.) and young aspen (O — 10 yr.), and grouse in the Pigeon River sites
preferred medium aged aspen (11 — 29 yr.) and jack pine. Birds in all sites avoided
upland hardwoods. In all sites, non-dispersing grouse selected areas that had HSI scores
higher than random. In 3 of the 4 sites, birds that chose high quality sites (according to
the model) lived significantly longer than did birds that chose low quality sites.

The Cox Proportional Hazards model was used to evaluate the potential risks
associated with various parameters. Juveniles had a higher risk of mortality than adults.
The amount of young aspen within activity ranges positively affected survival in 3 of the
4 study sites, whereas in some cases, use of lowland areas and mature aspen negatively
affected survival. This information can and should be used to improve ruffed grouse

habitat modeling and management to ensure the future of this popular game bird.

ACKNOWLEDGEMENTS

I would like to begin by thanking the funding sources that made the Michigan
Ruffed Grouse Project possible. Funding for this project was provided by the Federal
Aid in Wildlife Restoration Act, Pittman-Robertson Project W-127-R, and administered
through the Michigan Department of Natural Resources. Additional support was
provided by Michigan State University and the Ruffed Grouse Society.

I extend my sincere appreciation to my graduate committee. To Dr. Henry
Carnpa, Dr. Donald Beaver, and Dr. David Lusch, thank you for your advice, and for
your patience on this long-term project. I would like to specially thank my major
advisor, Dr. Scott Winterstein. I was truly lucky to have been “passed on” to Scott in
1992. Thank you, Scott for 8 years of guidance, support, and most importantly,
friendship.

I extend special thanks to Allison Gorrnley and Mike Larson, fellow graduate
students who worked on the grouse project in the Maltby Hills site. The entire Maltby
Hills data set presented in this report was coordinated and collected by Allison and Mike,
and I thank them for their years of hard work.

So many interns and ﬁeld workers came and went throughout this project, and I
thank them all for building grouse traps in the rain. Thanks, Amy, for working through
the winters, and Margi, for working through every season (even without pay). Thanks
also to everyone at the PRC SF headquarters, especially Joe Jarecki, and to the many
Department of Natural Resources personnel who contributed their time and efforts above

and beyond the call of duty.

iv

I’d also like to thank my friends (both old and new) in the Fisheries and Wildlife
Department. Thanks Christine, Delia, Katherine, and Kelly, for being wonderful friends
for 8 years! Thanks also to Sarah, Kendra, Annelise, Gabi, Michelle, Steve, Daniel,
Meredith, Marc, and Chris. Your friendship was invaluable, and I was very lucky to have
met you all.

Lastly, I’d like to thank my family for their love and support across the miles.
Thank you for the pictures, the phone calls, and the plane tickets, and thank you for your

faith in my ability to achieve this degree.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... viii
LIST OF FIGURES ......................................................................................................... xiii
INTRODUCTION .............................................................................................................. 1
OBJECTIVES ..................................................................................................................... 4
STUDY AREA ................................................................................................................... 5
CHAPTER 1: Survival and causes of mortality ................................................................. 8
Introduction ..................................................................................................................... 8
Methods ........................................................................................................................ II
Results ........................................................................................................................... l4
Trapping ................................................................................................................ 14
Sample population used for survival analysis ....................................................... 18
Survival - Maltby Hills ......................................................................................... 23
Survival — Pigeon River ........................................................................................ 35
Causes of mortality ............................................................................................... 47
Discussion ..................................................................................................................... 52
CHAPTER 2: Fall activity range and movements ............................................................ 57
Introduction ................................................................................................................... 57
Methods ........................................................................................................................ 60
Results ........................................................................................................................... 63
Type I non-dispersers ............................................................................................ 63

Type II non-dispersers .......................................................................................... 69
Dispersers .............................................................................................................. 75
Discussion ..................................................................................................................... 81
CHAPTER 3: Habitat quality and composition ............................................................... 85
Introduction ................................................................................................................... 85
Methods ........................................................................................................................ 88
Results ........................................................................................................................... 91

Vi

Habitat composition .............................................................................................. 91

Habitat quality — ﬁeld variables ............................................................................ 95
Habitat quality — random circles ........................................................................... 97
Discussion ................................................................................................................... 101
CHAPTER 4: Habitat use and preference ...................................................................... 104
Introduction ................................................................................................................. 104
Methods ...................................................................................................................... 107
Results ......................................................................................................................... 109
Forest type preference ......................................................................................... 109
Habitat quality preference — non-dispersers ....................................................... 113
Habitat quality preference — dispersers ............................................................... 119
Discussion ................................................................................................................... 124
CHAPTER 5: Relating habitat factors to survival ......................................................... 128
Introduction ................................................................................................................. 128
Methods ...................................................................................................................... 131
Results ......................................................................................................................... 133
Activity range composition ................................................................................. 133

HSI value and days survived ............................................................................... 137

Cox model results ............................................................................................... 143
Discussion ................................................................................................................... 147
MANAGEMENT IMPLICATIONS .............................................................................. 152
APPENDIX A ................................................................................................................. 161
APPENDIX B ................................................................................................................. 162
APPENDIX C ................................................................................................................. 165
LITERATURE CITED ................................................................................................... 170

vii

LIST OF TABLES

Table 1. Summary of trapping effort and success in the Maltby Hills open site, 1993 —
1997. ........................................................................................................................ 16

Table 2. Summary of trapping effort and success in the Maltby Hills closed site, 1993 —
1997. ........................................................................................................................ 16

Table 3. Summary of trapping effort and success in the Pigeon River open site, 1993 -
1997. ........................................................................................................................ 17

Table 4. Summary of trapping effort and success in the Pigeon River closed site, 1993 -—
1997. ........................................................................................................................ 17

Table 5. Summary of grouse excluded from analyses because mortality or loss of signal
before the 5-day mark by site. ................................................................................. 20

Table 6. Sex and age classes of collared ruffed grouse used in the survival analysis in the
Maltby Hills open site, 1993 —- 1997. ...................................................................... 21

Table 7. Sex and age classes of collared ruffed grouse used in the survival analysis in the
Maltby Hills closed site, 1993 — 1997. .................................................................... 21

Table 8. Sex and age classes of collared ruffed grouse used in the survival analysis in the
Pigeon River open site, 1993 — 1997 ....................................................................... 22

Table 9. Sex and age classes of collared ruffed grouse used in the survival analysis in the
Pigeon River closed site, 1993 — 1997. ................................................................... 22

Table 10. Overall survival of ruffed grouse in the Maltby Hills study sites for each year
(August 1 through May 15 of the following year). ................................................. 34

Table 11. Overall survival for ruffed grouse in the Maltby Hills open study site for each
year (August 1 through May 15 of the following year) by sex and age class. ........ 34

Table 12. Overall survival for ruffed grouse in the Maltby Hills closed study site for each
year (August 1 through May 15 of the following year) by sex and age class. ........ 34

Table 13. Overall survival of ruffed grouse in the Pigeon River study sites for each year
(August 1 through May 15 of the following year). ................................................. 46

Table 14. Overall survival for ruffed grouse in the Pigeon River open study site for each
year (August 1 through May 15 of the following year) by sex and age class. ........ 46

Table 15. Overall survival for ruffed grouse in the Pigeon River closed study site for each
year (August 1 through May 15 of the following year) by sex and age class. ........ 46

viii

Table 16. Fate of all birds used in survival analysis and percentage of known mortalities
for each year in the Maltby Hills open study site, as of May 15 of the following
year. ......................................................................................................................... 49

Table 17. Fate of all birds used in survival analysis and percentage of known mortalities
for each year in the Maltby Hills closed study site, as of May 15 of the following
year. ......................................................................................................................... 49

Table 18. Fate of all birds used in survival analysis and percentage of known mortalities
for each year in the Pigeon River open study site, as of May 15 of the following
year. ......................................................................................................................... 50

Table 19. Fate of all birds used in survival analysis and percentage of known mortalities
for each year in the Pigeon River closed study site, as of May 15 of the following
year. ......................................................................................................................... 50

Table 20. Sex and age classes of collared ruffed grouse shot in the Maltby Hills and
Pigeon River Areas, 1993-1998, expressed as frequency and percentage of total
birds. ........................................................................................................................ 51

Table 21. Number of ruffed grouse classiﬁed as Type I non-dispersers in the Maltby
Hills area, 1993 - 1997, by sex and age class. ......................................................... 65

Table 22. Number of ruffed grouse classiﬁed as Type I non-dispersers in the Pigeon
River area, 1993 - 1997, by sex and age class ......................................................... 65

Table 23. Number of ruffed grouse used in the analysis of movements by sex and age
class. ........................................................................................................................ 66

Table 24. Number of ruffed grouse classiﬁed as Type I non-dispersers by sex and age
class. ........................................................................................................................ 66

Table 25. Mean activity area (ha) of ruffed grouse classiﬁed as Type I non-diSpersers in
the Maltby Hills open site, 1993 - 1997, by sex and age class ................................ 67

Table 26. Mean activity area (ha) of ruffed grouse classiﬁed as Type I non-dispersers in
the Maltby Hills closed site, 1993 - 1997, by sex and age class. ............................ 67

Table 27. Mean activity area (ha) of ruffed grouse classiﬁed as Type I non-dispersers in
the Pigeon River open site, 1993 - 1997, by sex and age class. .............................. 68

Table 28. Mean activity area (ha) of ruffed grouse classiﬁed as Type I non—dispersers in
the Pigeon River closed site, 1993 - 1997, by sex and age class ............................. 68

Table 29. Number of ruffed grouse classiﬁed as Type II non-dispersers in the Maltby
Hills area, 1993 - 1997, by sex and age class. ......................................................... 71

ix

Table 30. Number of ruffed grouse classiﬁed as Type II non-dispersers in the Pigeon
River area, 1993 - 1997, by sex and age class ......................................................... 71

Table 31. Number of ruffed grouse classiﬁed as Type II non—dispersers by sex and age
class. ........................................................................................................................ 72

Table 32. Mean activity area (ha) of ruffed grouse classiﬁed as Type II non-dispersers in
the Maltby Hills open site, 1993 - 1997, by sex and age class ................................ 73

Table 33. Mean activity area (ha) of ruffed grouse classiﬁed as Type II non-dispersers in
the Maltby Hills closed site, 1993 - 1997, by sex and age class. ............................ 73

Table 34. Mean activity area (ha) of ruffed grouse classiﬁed as Type II non-dispersers in
the Pigeon River open site, 1993 - 1997, by sex and age class. .............................. 74

Table 35. Mean activity area (ha) of ruffed grouse classiﬁed as Type II non-dispersers in
the Pigeon River closed site, 1993 - 1997, by sex and age class ............................. 74

Table 36. Number of ruffed grouse classiﬁed as Type III dispersers in the Maltby Hills
area, 1993 - 1997, by sex and age class ................................................................... 77

Table 37. Number of ruffed grouse classiﬁed as Type III dispersers in the Pigeon River
area, 1993 - 1997, by sex and age class ................................................................... 77

Table 38. Number of ruffed grouse classiﬁed as Type III dispersers by sex and age class.
................................................................................................................................. 78

Table 39. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Maltby Hills open site, 1993 - 1997, by sex and age class ............................ 79

Table 40. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Maltby Hills closed site, 1993 - 1997, by sex and age class. ........................ 79

Table 41. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Pigeon River open site, 1993 - 1997, by sex and age class. .......................... 80

Table 42. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Pigeon River closed site, 1993 - 1997, by sex and age class. ....................... 80

Table 43. Area and percentage of each forest type in the Maltby Hills open site, 1993 —
1997. ........................................................................................................................ 93

Table 44. Area and percentage of each forest type in the Maltby Hills closed site, 1993 —
1997. ........................................................................................................................ 93

Table 45. Area and percentage of each forest type in the Pigeon River open site, 1993 —
1997. ........................................................................................................................ 94

Table 46. Area and percentage of each forest type in the Pigeon River closed site, 1993 —
1997. ........................................................................................................................ 94

Table 47. Sample size, mean, and standard error of SI-ﬁeld values (HSI calculated
without distance to aspen variable) in the open and closed sites in Maltby Hills... 96

Table 48. Sample size, mean, and standard error of SI-ﬁeld values (HSI calculated
without distance to aspen variable) in the open and closed sites in the Pigeon River
Country State Forest. ............................................................................................... 96

Table 49. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of non-dispersing ruffed grouse in the Maltby Hills
open site.a .............................................................................................................. 115

Table 50. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of non-dispersing ruffed grouse in the Maltby Hills
closed site.a ............................................................................................................ 116

Table 51. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of non-dispersing ruffed grouse in the Pigeon River
open site.‘ .............................................................................................................. 117

Table 52. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of non-dispersing ruffed grouse in the Pigeon River
closed site.ll ............................................................................................................ 118

Table 53. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Maltby Hills open
site.‘| ....................................................................................................................... 120

Table 54. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Maltby Hills closed
site.‘ ........................................................................................................... ' ............ 121

Table 55. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Pigeon River open
site.‘ ....................................................................................................................... 122

Table 56. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Pigeon River closed
site.a ....................................................................................................................... 123

Table 57. Mean percent of activity range (standard error in parentheses) in each forest

type for ruffed grouse that survived less (short-lived) than and greater than (long-
lived) the median number of days in the Maltby Hills open site. ......................... 135

xi

Table 58. Mean percent of activity range (standard error in parentheses) in each forest
type for ruffed grouse that survived less than (short-lived) and greater than (long-
lived) the median number of days in the Maltby Hills closed site. ....................... 135

Table 59. Mean percent of activity range (standard error in parentheses) in each forest
type for ruffed grouse that survived less than (short-lived) and greater than (long-
lived) the median number of days in the Pigeon River open site. ......................... 136

Table 60. Mean percent of activity range (standard error in parentheses) in each forest
type for ruffed grouse that survived less than (short-lived) and greater than (long-
lived) the median number of days in the Pigeon River closed site. ...................... 136

Table 61. Risk ratios and P values generated using the Cox proportional hazards model
for variables related to survival in the Maltby Hills open and closed study sites. 145

Table 62. Risk ratios and P values generated using the Cox proportional hazards model
for variables related to survival in the Pigeon River open and closed study sites. 145

Table 63. Risk ratios and P values for explanatory values found in a stepwise regression
through the Cox proportional hazards model for grouse in the Maltby Hills and
Pigeon River open and closed sites. ...................................................................... I46

xii

LIST OF FIGURES

Figure 1. Location of Pigeon River and Maltby Hills open and closed study sites. ........... 7
Figure 2. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1993 - 1994, by sex and age class. ...................................................... 26
Figure 3. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1994 - 1995, by sex and age class. ..................................................... 27
Figure 4. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1995 - 1996, by sex and age class. ..................................................... 28
Figure 5. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1996 - 1997, by sex and age class. ...................................................... 29
Figure 6. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1997 - 1998, by sex and age class. ...................................................... 30
Figure 7. Ruffed grouse survival in the Maltby Hills sites, 1993 — 1994. ....................... 31
Figure 8. Ruffed grouse survival in the Maltby Hills sites, 1994 — 1995. ....................... 31
Figure 9. Ruffed grouse survival in the Maltby Hills sites, 1995 - 1996. ....................... 32
Figure 10. Ruffed grouse survival in the Maltby Hills sites, 1996 -— 1997. ..................... 32
Figure 11. Ruffed grouse survival in the Maltby Hills sites, 1997 -— 1998. ..................... 33
Figure 12. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1993 - 1994, by sex and age class .............. 38
Figure 13. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1994 — 1995, by sex and age class. ..................................................... 39
Figure 14. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1995 — 1996, by sex and age class. ..................................................... 40
Figure 15. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1996 - 1997, by sex and age class. ..................................................... 41
Figure 16. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1997 — 1998, by sex and age class. ..................................................... 42
Figure 17. Ruffed grouse survival in the Pigeon River sites, 1993 - 1994. ..................... 43

xiii

Figure 18. Ruffed grouse survival in the Pigeon River sites, 1994 — 1995. .................... 43

Figure 19. Ruffed grouse survival in the Pigeon River sites, 1995 — 1996. .................... 44
Figure 20. Ruffed grouse survival in the Pigeon River sites, 1996 - 1997. .................... 44
Figure 21. Ruffed grouse survival in the Pigeon River sites, 1997 — 1998. .................... 45

Figure 22. Frequency of occurrence of HSI values in the Maltby Hills open site, 1993,
and the Maltby Hills closed site, 1993 and 1997. ................................................... 99

Figure 23. Frequency of occurrence of HSI values in Pigeon River sites, 1993 and 1997.
............................................................................................................................... 100

Figure 24. F all habitat preference for ruffed grouse in the Maltby Hills open and closed
sites. Grouse habitat preference was not different (Waller-Duncan multiple
comparison procedure, P > 0.10) between underlined habitats. Mean HSI score for
cover variables for each forest type are shown for reference. ............................... 111

Figure 25. Fall habitat preference for ruffed grouse in the Pigeon River open and closed
sites. Grouse habitat preference was not different (Waller-Duncan multiple
comparison procedure, P > 0.10) between underlined habitats. Mean HSI score for
cover variables for each forest type are shown for reference. ............................... 112

Figure 26. Percent of HSI values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Maltby Hills open site. ...... 115

Figure 27. Percent of HSI values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Maltby Hills closed site. 116

Figure 28. Percent of HSI values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Pigeon River open site ....... 117

Figure 29. Percent of HSI values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Pigeon River closed site. 118

Figure 30. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Maltby Hills open site ............... 120

Figure 31. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Maltby Hills closed site. ........... 121

Figure 32. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Pigeon River open site. ............. 122

Figure 33. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Pigeon River closed site ............ 123

xiv

Figure 34. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Maltby Hills open site ...................................................................... 139

Figure 35. Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Maltby Hills closed site. ....................................................... 139

Figure 36. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Maltby Hills closed site. .................................................................. 140

Figure 37. Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Maltby Hills closed site. ....................................................... 140

Figure 38. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Pigeon River open site. .................................................................... 141

Figure 39. Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Pigeon River open site. ......................................................... 141

Figure 40. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Pigeon River closed site ................................................................... 142

Figure 41. Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Pigeon River closed site. ....................................................... 142

Figure 42. Lowland conifers within activity range (ha) vs. days survived past August 1

for ruffed grouse in the Maltby Hills closed site. .................................................. 151

Figure A 1. Distance moved from the harmonic center of the week after capture to the
harmonic center of each following week for: a) Type I non-disperser, b) Type II
non-disperser, and c) Type III disperser ................................................................ 161

Figure B 1. Index value for equivalent stem density in the Michigan model. ............... 162

Figure B 2. Index value for average height of deciduous trees in the Michigan model. 162

Figure B 3. Index value for height above ground of lower coniferous branches in the
Michigan model ..................................................................................................... 163

Figure B 4. Index value for average height of deciduous shrubs in the Michigan model.
............................................................................................................................... 163

Figure B 5. Index value for distance between life requisites in the Michigan model... 164

Figure C 1. Habitat suitability of the Pigeon River open site, 1993. ............................. 165

Figure C 2. Habitat suitability of the Pigeon River closed site, 1993 ............................ 166

XV

Figure C 3. Habitat suitability of the Maltby Hills open site, 1993. .............................. 167
Figure C 4. Habitat suitability of the Maltby Hills closed site, 1993. ........................... 168

xvi

INTRODUCTION

The ruffed grouse (Bonasa umbellus) is a popular game bird in the northern
Lower Peninsula of Michigan. Each year, approximately 130,000 small game hunters
pursue grouse in Michigan, and harvest an average of 250,000 birds annually (Karasek
1998). Since the start of regulated hunting seasons, the impact of hunting on grouse
populations has been disputed. In the early 19905, there was a call by Michigan grouse
hunters to investigate this question because of dwindling grouse populations. There was
additional interest in investigating future management strategies to ensure the
sustainability of a huntable population of this species in Michigan.

Hunting regulations have varied dramatically over the years, as have the opinions
on how many grouse can be safely harvested each season. Palmer and Bennett (1963)
suggested that as much as 50% of the fall population could be harvested in Michigan with
little risk of reduced productivity. More conservative estimates call for the taking of only
20 — 30% to ensure a healthy spring population (Bump et al. 1947). Confounding this
disagreement, it has been shown that ruffed grouse populations cycle over a period of
approximately 10 years. Many past studies on hunting mortality did not take the cycle
into account; several years of data were combined and averaged, across times of high and
low population numbers. There has not yet been a study that observed the effects of
hunting year by year, considering that effects may differ if population demographics
differ.

The decline in quality ruffed grouse habitat in northern Michigan is also a concern
in maintaining this species. In the northern United States, healthy ruffed grouse

populations are assumed to be closely linked with dense aspen (Populus spp.) forest types

(Brewer et al. 1991). The most limiting factors in the northern states for grouse are
winter food and fall/winter cover, both of which can be provided by a mixture of age
classes of aspen. Young stands provide dense horizontal and vertical cover that allow
birds to escape from predators such as hawks and owls. Additionally, mature aspen
catkins are a valuable winter food source that can be eaten quickly, thereby conserving
energy needed to survive winter temperatures (Vispo et al. 1994). Traditional ruffed
grouse management calls for a 40—year rotation of small (10 acres) cuts, arranged so that
various age classes are in close proximity (Gullion 1972).

Although young aspen was prevalent in this state in the recent past, the present
and future management practices stand to change the forest composition (Harnmill and
Visser 1984). Mature forests of pine or northern hardwoods are replacing areas once
dominated by aspen, and it is unknown what effects this will have on grouse populations.
Management of aspen resources is considered necessary for the continued success of the
ruffed grouse in the Great Lakes states, but there is limited information on the potential of
other forest types to provide the requirements for survival.

During the fall, a portion of the grouse population (generally assumed to be the
juvenile birds) will disperse from their summer range to a winter range. Other birds will
stay in an established area. Because they are considered the most important habitat
characteristics for grouse, food supply and cover are assumed to dictate their fall
movements (Chambers and Sharp 1980). There may be other contributing factors,
however, such as population density and hunting pressures. As the population of a
wildlife species increases, the movements of individuals may change. During times of

high density, birds may compete for territory, forcing some to move farther to ﬁnd

unoccupied high-quality habitat. Birds also may ﬂush more frequently in hunted areas,
and may move greater distances as a result. Knowing which factors inﬂuence the birds to
move in the fall may be important in improving habitat management.

In addition, movements and choice of fall/winter range may be indirectly related
to survival. As birds search for acceptable, unoccupied habitat, they may be forced to
travel through or settle in poor-quality habitat. If that habitat does not provide the
necessary components, the survival of that bird may be compromised. In a telemetry
study in Wisconsin, Small et al. (1991) showed that grouse that had dispersed during the
autumn season had a higher winter mortality than those that remained on an established
winter range. Understanding the relationship between hunting, movements, habitat
quality and availability, and survival has the potential to greatly enhance management for

ruffed grouse.

Recently, the increased use of Geographic Information Systems (GIS), has
improved the accuracy of habitat studies in the ﬁeld of natural resources. Large, digital
databases can now be used to study spatial relationships and can effectively automate the
application of habitat models (Donovan et al. 1987). Such databases have been
developed for two study areas in the northern Lower Peninsula of Michigan, where ruffed
grouse telemetry research took place from 1993 to 1998. This dissertation takes a
comprehensive ﬁve-year database of fall grouse locations in sites open and closed to
hunting and combines it with a habitat quality analysis in an attempt to analyze the

characteristics of grouse behavior and survival in northern Michigan.

OBJECTIVES

The speciﬁc objectives of this study are to:

1)

2)

3)

4)

5)

Analyze fall and winter survival rates and causes of mortality of ruffed grouse on
sites open and closed to hunting;

Analyze and compare habitat quality and habitat composition of open and closed
sites using GIS;

Analyze and compare movements, habitat use, and habitat preference of ruffed
grouse on sites open and closed to hunting using GIS;

Determine to what extent movements, habitat use, and habitat preference are
related to survival; and

Make recommendations for future ruffed grouse habitat management and hunting

regulations.

STUDY AREA

There are two study areas where ﬁeldwork took place in 1993 - 1998, both
located in the northern part of Michigan's Lower Peninsula (Figure 1). One portion of
this study was conducted in the Pigeon River Country State Forest (PRCSF), in
Cheboygan, Otsego and Montrnorency counties. Dominant forest types in the PRCSF
include quaking (Populus tremuloides) and bigtooth aspen (P. grandidentata), maple
(Acer saccharum), red pine (Pinus resinosa), white pine (Pinus strobus), and Northern
white cedar (Thuja occidentalis). Following the lumbering and post-harvest wildﬁres in
the early 19005, the Civilian Conservation Corps took on a reforesting effort resulting in
a substantial number of pine plantations in this area. The area also has a unique resource,
a managed elk herd, which sets it apart from other forested areas in northern Michigan.
Rocky Mountain elk (Cervus elaphus nelsoni) were reintroduced to the area in 1918, and
have grown to a population of 800-1300 in the 19903.

The Pigeon River study area is a mixture of moraines, till plains, and outwash
plains. Dominant soil types include Detour-Brevort, Rubicon-East Lake, and
Cheboygan-Blue Lake associations. The Pigeon, Sturgeon, and Black Rivers flow
through the area, and several small lakes are present. The average winter temperature in
the Pigeon River is —7.9 C, with an average minimum of —l 3.6 C. The average summer
temperature is 18.7 C, with an average maximum of 25.8 C. Total annual rainfall
averages 77.4 cm, of which 60% falls in April through September. Seasonal snowfall
averages 291 cm (Tardy 1991).

The second study area is in the Maltby Hills part of the Huron National Forest

(HNF) (approximately 80 km southeast of the PRCSF) in Oscoda, Alcona, and Ogemaw

counties. Because of the close proximity of the areas, forest types are similar in the HNF,
except for the existence of a substantial oak (Quercus spp.) component that does not exist
in the PRCSF. There was not as much planting in the HNF as there was in the Pigeon,
but a large ﬁre in the 19805 has left a continuous stand of dense jack pine in the area.

The summer weather in the HNF is similar to the PRCSF, with mean summer
temperature of 18.4 C and an average maximum of 25.6 C. The winter is a little less
severe in the HNF, with average temperature of —7.2 C and an average minimum of —12.6
C. Mean annual rainfall is 69.1 cm, and mean annual snowfall is 160 cm (Johnson 1990).
The area is dotted with small ponds and lakes, and is crossed by offshoots of the Ausable
River. The primary topographic divisions found in the HN F are moraines, till plains, and
deltas. Soil types include Grayling sand and Alﬁc Haplothods associations (Williams
1998)

Within each forest area, two study sites were established, one open to all hunting
under normal statewide regulations and one closed to grouse and woodcock hunting for
the duration of the study. Open and closed sites were separated by a buffer of about 4
km. Before the start of the study, the closed sites were posted at approximately each
tenth of a mile (0.16 km), at all county road intersections, and at frequently used parking
places. Maps of the closed sites were published in the yearly hunting regulations
published by the Michigan Department of Natural Resources. No special actions were
taken to post or publicize the sites open to hunting. Each site measured approximately 80

- 100 kmz, with the closed sites being slightly smaller than the open sites.

 

 

 

 

f“-

, Closed Pigeon
River

I.
o._
.._
..
uh
I.
.....
--.
.
u.
o._
u"
no
.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L l l T L Closed Maltby

 

 

 

 

 

 

 

Z
_’
\
Ll
r—
E?“
\
E

 

 

Figure 1. Location of Pigeon River and Maltby Hills open and closed study sites.

CHAPTER 1: Survival and causes of mortality

Introduction

One question that is often asked about hunted species is whether hunting is
additive or compensatory to natural mortality. The answer to such a question can aid
managers in understanding harvest limits and setting hunting regulations. The additive
theory assumes that hunting results in the deaths of more animals than would die
naturally. If hunting was additive to natural mortality, then survival probability would be
lower in hunted populations than in non-hunted populations (assuming that other
variables are generally equal). If survival probabilities are comparable between hunted
and non-hunted populations, then hunting mortality can be considered compensatory to
natural mortality. The way to test such an idea is to compare survival and causes of
mortality between wildlife populations in sites open to hunting to those in populations in
sites closed to hunting.

Various studies have come to different conclusions as to the effects of hunting in
ruffed grouse populations. Fischer and Keith (1974) compared mortality rates. in grouse
populations near roads (high hunting pressure) to those away from roads (low hunting
pressure). They found a 77% total annual mortality in the high-pressure area, with a
harvest of 23% of the banded birds. In the low-pressure area, total mortality was 64%,
with only 1% of birds harvested. They concluded, however, that fall hunting had no
measurable effect on spring population levels. Palmer and Bennett’s (1963) banding
study measured a take of almost 30% of one hunted population. Spring populations,

though, were comparable between areas of various hunting pressures. In contrast, in his

banding study in 1982, Gullion declared that in his study site the grouse population
depression was the result of excessive hunter harvest.

With the advent of telemetry methods that were appropriate for medium-sized
birds came more sophisticated studies on the survival and causes of mortality of ruffed
grouse. In a large-scale study in Wisconsin, researchers investigated collared birds over
8 years, and determined that avian predators were the largest source of mortality (Small
et al. 1991). Small et al. (1991) also concluded that hunters took a “signiﬁcant” number
of birds in the fall, and that hunting was at least partially additive to mortality.
Researchers warned that grouse numbers could be substantially reduced where areas of
high hunting mortality are combined with low rates of immigration. Although they
contain valuable information, all of the studies mentioned above lacked a site closed to
hunting. A site with no hunting pressure at all is needed as a control to study the above
mentioned questions.

Aside from comparing overall survival between hunted and non-hunted areas,
management processes may be improved by understanding the differences in risk to
different sex and age classes. Banding and harvesting studies have shown that by fall,
there are differing ratios of males to females and adults to juveniles in the grouse
population. Because of differing reproductive behaviors and associated risks, there are
slightly more males in the fall population than females (Bergerud and Gratson 1988).
There are generally more juveniles in the fall population than adults, but Domey (1963)
showed that winter juvenile mortality rates were generally 15% greater than those for
adults. It is assumed that juveniles will be more at risk during the hunting season, due to

lack of learned response behaviors.

Understanding how fall survival differs between areas open and closed to hunting
and between sex and age classes can help managers model populations across time.
Detailed survival information from a 5-year telemetry study of hunted and non-hunted

ruffed grouse in Northern Michigan is presented below.

10

Methods

Ruffed grouse trapping took place in August -— October 1993 — 1997 in both study
areas. Initial trapping areas were selected based on actual bird sightings by
Michigan Department of Natural Resources (MDNR) or project personnel, or by their
apparent quality as grouse habitat. Young aspen stands (approximately 5 — 20 years old)
with dense understory vegetation were frequently selected because they were likely to
support the highest density of grouse in summer and early fall. Modiﬁed cloverleaf traps,
as described by Domey and Mattison (1956), were placed in selected stands with an
attempt to orient the traps at right angles to suspected bird movements. Fifty birds was
considered the minimum sample size needed to detect differences in survival between
sites (Winterstein et al. in press). Length of the trapping season varied by site and by
year in an attempt to radio-collar 60 birds (10 above the minimum) per site per year.

Beginning in 1995, birds that survived with working collars from the previous
years were sometimes recollared to increase sample size. This allowed researchers to
reach their goal of 60 or more collared birds per season. Recapturing was accomplished
using alternative methods such as night-lighting and netting females on nest. All trapping
and handling procedures were approved by the Michigan State University Committee on
Research Involving Animal Subjects (AUF # 10/93-400-03).

Grouse were sexed according to tail and rump feather size and coloration as
described by Domey (1966), Roussel and Ouellet (1975), and Larsen and Taber (1980).
Birds were categorized as either juveniles or adults by observing body size until
September 1 and by wear on the 9th and 10th primary feathers aﬁer September 1 (Godin

1960). Weight and wing length were measured to aid in age classiﬁcation and to

II

determine the potential of the bird for holding a radio-collar. Birds that weighed over
350 g were banded and ﬁtted with a 10 g necklace-style radio collar before they were
released. Radios were obtained through LOTEK Engineering, Ontario, Canada or
Advanced Telemetry Systems, lsanti, MN (inclusion of company name does not imply
endorsement). Birds that weighed less than 350 g were considered too small to support
the weight of the collar, and were banded and released immediately. Birds that had
trapping injuries were released without processing.

Birds were located 1 — 4 times per week to monitor survival from late summer
through May 15 of the following year on each of the study sites. Radios were equipped
with 8-hour mortality sensors, and remains were collected as soon as possible after the
mortality sensor was known to be active. Cause of death was determined whenever
possible by examining the area for predator signs such as raptor whitewash, scat,
perching trees, or dens. All carcasses were collected for necropsy and sent to the
MDNR's Rose Lake facility.

Although every effort was made to locate all missing signals, each year in each
site there were birds that were lost and labeled “censored”. A bird was considered
censored if a signal was not found after 2 — 3 weeks of intense ground searching, and if
aerial surveys yielded no results. Some collars were lost for unknown reasons, but some
censored collars had obviously malfunctioned. In some cases, researchers observed that
collars emitted a mortality signal when the bird was actually alive, thereby shortening the
life of the battery. Additionally, some were lost after being in use for well over the 9-

month guaranteed battery life. Collars that were censored because of a supposed

12

malfunction (including age > 9 months) were included in the analysis but separated in the
presentation of data.

If a bird survived for at least 5 days after capture with a working collar it was
used in the analysis of survival. Birds that survived for at least 5 days after capture but
later died due to trapping or collaring were considered censored. Survival probabilities
for each site in each year of the study were calculated and survival curves were generated
using the Kaplan-Meier Product Limit estimator. This technique provides an estimate of
the survival probability at any point in the study period (Kaplan and Meier 1958).
Survival probability curves were also generated for birds by sex and age class. Survival
of grouse from 1 August through 1 May of each year was compared between hunted and
non-hunted sites using the Log-rank test (Lee 1992), provided by the Excel add-in
SurvanXL (S. G. Shering, 1995 - 1997). Survival of grouse between sex and age class
within years and within sites were also compared using the Log-rank test. Note that a
starting date of August 1 of each year was used for comparability of survival curves
between sites and years. Knowing that the populations cycle, I did not combine data
across many years (as previous studies have done), as survival may differ at different

population levels due to density dependent factors.

13

Results
Trapping

In the Maltby Hills open site, 452 birds were caught in modiﬁed cloverleaf traps
over the course of the study, and 267 were collared in 6264 trap-nights (Table 1). Note
that summary tables (Tables 1 — 4) do not include recaptures. Birds recaptured and
recollared from previous years are also excluded from the trapping summary, but are
included later in the summary of birds used for survival analysis. In all sites, the majority
of birds caught but not collared were young birds that did not meet the weight
requirement of 350 g. In the ﬁrst season, 1993, only 20 birds were collared (1/3 of the
desired number) in 1008 trap-nights in the open site. In 1994, though, the success rate on
this site was the highest recorded in the study, 0.058 grouse collared per trap-night (or
17.3 trap-nights per bird collared). The trapping success rate then decreased slightly in
1995, but was still much higher than in 1993. Trapping success was high in 1996 and
1997, and the overall success rate for the site was 0.072 birds and 0.043 collars per trap-
night.

Slightly fewer birds were caught in the Maltby Hills closed site (420) (Table 2),
but more of them were collared (301 ). Like the open site, trapping success was low in
1993, when only 36 birds were collared. More trap-nights and a higher success rate
resulted in 70 birds being collared in 1994. By the 1997 season, fewer trap-nights were
needed, and trapping success rose to 0.117 birds per trap-night, the highest recorded in
any site during the study. This site recorded the fewest trap-nights overall, and had a
success rate of 0.071 birds and 0.051 collars per trap-night. After 1993, over 60 birds

were collared in this site in each season.

14

In the Pigeon River sites, fewer birds were caught and collared than in the Maltby
Hills area, with slightly higher effort in both sites. In the open site, 347 birds were caught
in 7592 trap-nights (Table 3). Trapping success started at a low of 0.038 birds per trap
night in 1993. A high of 0.052 birds per trap night was reached in both 1996 and 1997.
Success in placing collars never surpassed 0.038 collars per trap night. The goal of 60
birds collared was only reached in 1997.

The fewest birds were collared in the closed Pigeon River site, 244 of 31 I
captured. As with the other sites, researchers collared few birds in the Pigeon River
closed site in 1993 (Table 4), when only 28 birds were collared of 33 caught. Trapping
success increased steadily from 1993 — 1997. We recorded similar trapping success in
the open and closed Pigeon sites in 1993 — 1995, but success in the closed site surpassed
that in the open site in 1996 and 1997. The highest trapping success during the study in
the Pigeon sites was in the closed site in 1997, when researchers caught 0.081 birds per

trap night. In that year, fewer than 1000 trap-nights were needed to collar over 60 birds.

15

Table 1. Summary of trapping effort and success in the Maltby Hills open site, 1993 —

 

 

1997.

1993 1994 1995 1996 1997 Total
Birds caught 37 101 87 98 129 452
Birds collared 20 74 49 58 66 267
Trap nights 1008 1283 1289 1226 1458 6264
Birds/trap night 0.037 0.079 0.067 0.080 0.088 0.072
Collars/trap night 0.020 0.058 0.038 0.047 0.045 0.043

 

Table 2. Summary of trapping effort and success in the Maltby Hills closed site, 1993 —

 

 

1997.

1993 1 994 1995 1996 1997 Total
Birds caught 51 85 106 95 83 420
Birds collared 36 70 69 63 63 301
Trap nights 1343 1529 1147 1182 710 5911
Birds/trap night 0.038 0.056 0.092 0.080 0.1 17 0.071
Collars/trap night 0.027 0.046 0.060 0.053 0.089 0.051

 

l6

Table 3. Summary of trapping effort and success in the Pigeon River open site, 1993 —

 

 

1997.

1993 1994 1995 1996 1997 Total
Birds 47 62 66 86 86 347
Collars 36 47 54 58 63 258
Trap nights 1239 1632 1423 1650 1648 7592
Birds/trap night 0.038 0.038 0.046 0.052 0.052 0.046
Collars/trap night 0.029 0.029 0.038 0.035 0.038 0.034

 

Table 4. Summary of trapping effort and success in the Pigeon River closed site, 1993 -

 

 

1997.

1993 1 994 1995 1996 1997 Total
Birds caught 33 51 67 86 74 311
Birds collared 28 40 54 60 62 244
Trap nights 1266 1755 1569 1098 913 6601
Birds/trap night 0.026 0.029 0.043 0.078 0.081 0.047
Collars/trap night 0.022 0.023 0.034 0.055 0.068 0.037

 

l7

Sample population used for survival analysis

Over 5 seasons in the 4 study sites, 135 birds did not survive or could not be
located within 5 days after capture, the time-frame required for use in the survival
analysis (Table 5). These birds constituted 12.6% of all birds collared. A slightly higher
percentage of collared birds were omitted from the open sites than from the closed sites.
During each trapping season, some ruffed grouse were lost due to stress or trapping
injuries. Fifty—four collared birds, or approximately 5% of all birds collared, were lost
presumably to trapping causes. These were distributed evenly across study sites, with a
range of 13 — 15 grouse lost per site. Twenty-four birds were censored before the 5 day
mark, including some that had malfunctioning collars, and possibly some that moved too
far out of the study area to be located. This left 935 birds to be used in the analysis of
survival, in addition to 115 birds that survived from previous years with working collars
(total = 1045 birds).

In the Maltby Hills open site, 65.4% of the birds used in the analysis of survival
were juveniles (Table 6). In 1993 and 1995 a higher proportion of birds used were adults,
possibly due to a very low sample size in 1993 and a high number of birds alive from the
previous year in 1995. Only 7 birds were classiﬁed as unknown age. The number of
males and females were nearly equal during the study, although within season differences
existed. More females than males were collared in 1993, possibly again due to small
sample size. Twenty-one birds were classiﬁed as unknown sex.

The proportion of adult birds used in the Maltby Hills closed site was 64.2%,
nearly equal to the proportion in the open site (Table 7). As in the open site, in 1993

more adults were used than juveniles. In each year except 1993, there were more males

18

used in the survival analysis than females. In the total sample, 55.6% of the grouse used
were male. Five birds were of unknown age and 8 of unknown sex.

In the Pigeon River open site (Table 8) more juveniles were used in the analysis
than adults (53.9%), but the difference was not as extreme as in the Maltby study sites.
The most pronounced differences were in 1995 and 1997, when juveniles made up more
than 60% of the sample population. In 1996 there were more adults used in the analysis
than juveniles, possibly because of the number of adult birds surviving from previous
years. Males and females comprised nearly equal proportions of the sample population,
but differed within the years 1993 and 1994. Seven birds were of unknown age and 5 of
unknown sex.

In the Pigeon River closed site, more adults were used than juveniles (Table 9).
The highest disparities were in the 1993 and 1994 seasons, when 71% and 60% of birds
used were adults. In 1997, 61% of birds used in the analysis were juveniles. Numbers of
males and females collared were similar overall, although many more males were
collared in 1996 and 1997. Twelve birds were of unknown age (7 of them in 1995) and 5

were of unknown sex.

19

Table 5. Summary of grouse excluded from analyses because mortality or loss of signal
before the 5-day mark by site.

 

Censored Stress/ Predator Other/ Total Birds %

 

trapping Unknown omitted collared omitted
Maltby Hills
Open 3 15 15 3 36 267 13.5
Closed 4 13 11 7 35 301 11.6
Pigeon River
Open 11 13 6 6 36 258 14.0
Closed 6 13 7 2 28 244 1 1.5
Totals 24 54 39 18 135 1070 12.6

 

20

 

 

 

 

 

 

 

OMN we mm mm av N 280,—.
_.N m ed o w.m N od o m.\. M ed o x3 :26:ch
Yew :_ N.Nv 5N edm _N m.mm _m mNm _N «.3 I 0380;
m. _ m N. ”um um 06m om 93c 5N odv 3 Ném 2 2n:
Qm N. od o ad N _.N_ h mg. M od 9 owe 555?:
v.3 2: ode on 03 .N adv 0N mNm 2 NdN N. 0:52;.
Row M: _ mam mN mdm 5N ﬁne. mN ado VN wdn N.— ~_:u<
.X. a o\.. a .x. u o\o A“ .x. a ..\o a
.58. 25— coo _ moo _ vaa _ moo _
ﬂag I 33 .8? memo—o Lozm :OOwE on. E mix—ac“ _a>_>.=.m of 5 new: 830% wot—c 38:8 Co Emma—o own van xom .o 035.
NmN Nm hm Nm Nv mN 280,—.
0N c a; _ o6 o o._ _ md v od o x8 =32.ch
«NV :— o. _ m N v.3 5N Név «N :3 S 9% C DEED..—
cdv m. _ Név VN QNm on o. _ m N va NN v. 3. N _ 0.32
9m 5 o._ _ a; _ ad m 5v N :6 0 Own 5505...:
mém eN_ v.3 vm v.3 \N m. G Nm 9.; oN ”.3. m_ 0:52;.
ENv o0 ﬁNm C 2% oN ﬁNm C 0.: cN N.mm 2 u_:v<
_.\o a. .X. a o\.. m .x. m. .X. a .x. k
.88. baa _ cg _ mac _ voa _ moo _

 

.32 I 32 .23 some cozy— :oowE 05 5 $3.28 3223 2: 5 new: 839% REE 38:8 mo momma—o own van xom .w 039—.

21

 

 

 

 

 

 

Ncm as me E. No N 280,—.

..N w o... c c. . 5m v 0.. . a6 N x3. 565::

N. . v N. N...m N 5.3 N 3% .m a. .v N h. . m m. 0.28...

9mm no. ..N.. 3 5% ..m o.om mm Wow mm «.3 N. 0..N.).

.2. m a... o ed o QN N o... o m... m own 5505...:

New g. Wow Nm Now we mg 3 5% ..m v.3 N. 2.5.5..

..3 mo. m.mm 0N «.mN m. on MN 2? N m: v. ~_:.u<
.x. a .x. a. o\° u .x. u .x. u .x. a.

.50...

.So. I mg. .27. .583 m5: 3:32 2.. a. mix—28 .9123. 2.. a. wow: 88% not... 33:8 .0 8320 own .28 xom N via...

NNN co m m .3 on o . 280...

n... .N m.» m ..N. .1 ..c. .. m. . m6 . x3 :32.ch

Q? N. Q? N v.9» ..N ..om ..N odm mm N60 N. 0380...

"I: N. 23 N v...V VN ..dm mm 0.3 Nm 9.; c o.3].

9N N o... o od o «N N 9m N .3. m own 5505.5.

«.3 a: 3K ..v 0.2. Q v.3 Nm N. 2. Se ”SM n 0:52;.

on 5 n. .N m. vNN m. 5% mm ”.mN N. v.3 o :32
.x. a .x. g. .x. u .x. m. o\.. u .x. a

 

.80...

Nam. I 8a. .8.» some 2....— .EHBS. 05 5 £92.23 .a>.>._:m o... :. now: 8.53 not...— 33:00 me momma? owe USN New 6 03a...

22

Survival - Maltby Hills

No ruffed grouse collared in the Maltby Hills open site in 1993 survived until 15
May 1994 (Figure 2, Table 10). Survival probability for both sexes and age classes was
0.00 (Table 11), but there was a signiﬁcant difference detected in the juvenile and adult
survival curves (P = 0.083). There was a sharp drop in survival in mid-late September in
all classes, coinciding with the ﬁrst 2 weeks of hunting season. Juvenile survival dropped
rapidly through late fall, and no collared juveniles were alive after 1 January 1994.

Grouse fared better in the closed site in 1993 (Figure 2), with a ﬁnal survival
probability of 0.23 (Table 10). The survival curve for the closed site was signiﬁcantly
different from the open site (Figure 7) (P = 0.098). Final survival (S) probabilities ranged
ﬁ'om 0.13 for males to 0.30 for females (Table 12). No signiﬁcant differences were
detected between sex or age classes.

In 1994, overall survival in the Maltby Hills open site was 0.41 (Table 10). Again
there was a sharp decline in survival during the beginning of hunting season (Figure 3),
but very few birds were lost in the time between October 1, 1993 and April 1, 1994.
Males had a higher survival probability than did females (Table 11), but no signiﬁcant
difference was found. Adults (S = 0.69) had signiﬁcantly higher survival than did
juveniles (S = 0.34, P = 0.092).

Survival of birds in the closed site was lower than in the open site (Figure 3) in
the 1994 — 1995 season but the difference was not signiﬁcant (Table 10). The only sharp
decline during the fall was in survival of males (Figure 3), but all birds exhibited a steady

mortality through winter and spring. Females had a survival probability of 0.54, versus

23

0.19 for males (Table 12). This difference was signiﬁcant (P = 0.042). Juveniles
survived slightly better than adults did, but the difference was not signiﬁcant.

In the 1995 -— 1996 season, grouse in the open Maltby Hills site had much lower
overall survival than in the previous year, at 0.21 (Table 10). There was no sharp drop in
survival, but a steady decline from winter to spring (Figure 4). Females had a higher
survival probability than males but the difference was not signiﬁcant. More juveniles
than adults were lost during the fall, and a drop in juvenile survival in January and
February resulted in a signiﬁcantly lower overall survival (Table 11) (P = 0.015).

Grouse in the Maltby Hills closed site in 1995 had a signiﬁcantly lower overall
survival probability than those in the open site (P = 0.093) (Table 10). The survival
curves for the open and closed sites were almost identical until January, when there was a
steep drop in survival in the closed site (Figure 9). Survival probability in the closed site
was only 0.10, and no collared females survived past the third week in February 1996
(Figure 4). Males had a signiﬁcantly higher survival probability than females (P =
0.039), and adults survived signiﬁcantly better than juveniles did (P = 0.090) (Table 12).

In the 1996 - 1997 season, grouse in the open Maltby Site had a survival
probability of 0.36 (Table 11). There was a signiﬁcant difference in survival between
males and females (P = 0.030). Although rates were similar through early fall, curves
separated by September (Figure 5), and males ended the season with a survival
probability of 0.19, while females were at 0.48. Adults survived better than juveniles did,
but not signiﬁcantly.

Grouse in the closed site had a survival probability of 0.33, similar to that in the

open site and much higher than in 1995 (Table 10). The curve for the closed site drops

24

early, but converges with the open site in October. Survival curves in the closed site
separate early by sex and age class (Figure 5), but are tightly grouped at the end, with
males and adults at around 0.32 and females and juveniles at around 0.42 (Table 12). No
signiﬁcant differences between sexes or age classes were detected.

Overall ruffed grouse survival in the Maltby Hills open site in 1997 - 1998 was
0.28 (Table 10). Survival curves for sex and age classes were tightly grouped until late
winter (Figure 6), when adult survival (S = 0.46) stayed well above juvenile survival (S
= 0.26). Males and females had comparable survival probabilities (Table 11). There
were no signiﬁcant differences between sexes or age classes.

Birds in the closed site (S = 0.33) survived slightly better than did those in the
open site (Table 10). The survival curve for the closed site was slightly higher but
parallel to the open site curve from winter through spring (Figure 11). Although adults
had much higher survival through February (Figure 6), they ﬁnished the season with
lower probability of survival than juveniles did (Table 12). All birds had a decrease in
survival in late February through early April. Males had a higher survival probability

than females, but the difference was not signiﬁcant.

25

 

 

 

 

 

 

 

 

 

1 .
0.8 -
g —Total
.9
,8 0-6 t - - - Male
E- —Female
E 0-4 s” ----- Adult
:tr Juvenile
m
0.2 .
0
8/1/93 9/20/93 11/9/93 12/29/93 2/17/94 4/8/94 5/28/94
Date
1
Closed
0.8 ~ -
3.3:
"" —T
E 0.6 —- ma]
3 - - - Male
°‘ I
E 0.4 Femae
S ------ Adult
a: 0.2 l~ — Juvenile
0

 

 

 

8/1/93 9/20/93 11/9/93 12/29/93 2/17/94 4/8/94 5/28/94

Date

Figure 2. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1993 - 1994, by sex and age class.

26

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

1 .
0.8 ~
.3!
LE —Total
g 0'6 ' - - - Male
E. —Female
:6
E 0'4 ‘ ----- Adult
Juvenile
m 0.2 «l» —~ ~
0
8/1/94 9/20/94 11/9/94 12/29/94 2/17/95 4/8/95 5/28/95
1
Closed
0.8 J
g. —Total
'3 - - - Male
<6 0.6 « - _ - L
E \— ——-Female
g ------ Adult
> 0.4 « .
E ~ ‘ . Juvenile
Us) ..u.- \n
0.2 -- ‘"'"'"~--- ,,
0
8/1/94 9/20/94 12/29/94 2/17/95 4/8/95 5/28/95

Date

Figure 3. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1994 — 1995, by sex and age class.

27

 

 

 

 

 

 

 

 

 

 

 

 

0.8
29
E": —Total
73 0.6
9. - - - Male
3 ——Female
g 0.4
g ------ Adult
VJ Juvenile
0.2
0
8/1/95 9/20/95 11/9/95 12/29/95 2/l7/96 4/7/96 5/27/96
Date
1
Closed
0.8 «H
3*
g 0.6 ._ <
e - - - Male
a.
'8 —Female
> 0.4 ~~
g ------ Adult
m Juvenile
0.2 -_., ~
0
8/1/95 9/20/95 ll/9/95 12/29/95 2/17/96 4/7/96 5/27/96

Date

Figure 4. Survival probability for ruffed grouse collared in the Maltby Hills open and

closed sites, 1995 - 1996, by sex and age class.

28

 

 

 

 

 

 

 

 

 

 

 

 

 

1
0.8 l
.é‘
a —Total
-8 0.6 —» ~
.g - - - Male
.3 0 4 -——Female
.3 ' ‘ ----- Adult
m Juvenile
0.2 «~
0
8/1/96 9/20/96 ll/9/96 12/29/96 2/17/97 4/8/97 5/28/97
Date
Closed
3‘
3;: _Total
.39: ' " " Male
g -—_Female
E ------ Adult
(2 Juvenile
0.2 4» -
0
8/1/96 9/20/96 11/9/96 12/29/96 2/17/97 4/8/97 5/28/97

Date

Figure 5. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1996 - 1997, by sex and age class.

29

 

 

 

 

 

 

 

    

 

l
0.8
.4?
E 0.6 —Total
2 - - - Male
o.
g 0.4 . _, —Female
E ----- Adult
m 0.2 «~— # Juvenile
0
8/1/97 9/20/97 11/9/97 12/29/97 2/17/98 4/8/98 5/28/98
Date
1
Closed
0.8 ~~ s _,
2:
E _Total
.0
g 0'6 4 - - - Male
.2: —Female
.2 0'4 A” ' ” l --‘---'Adult
cg . Juvenile
0.2 a !_ g
0

 

 

 

8/1/97 9/20/97 ll/9/97 l2/29/97 2/l7/98 4/8/98 5/28/98

Date

Figure 6. Survival probability for ruffed grouse collared in the Maltby Hills open and
closed sites, 1997 - 1998, by sex and age class.

30

 

 

 

 

 

 

 

 

 

1 -
0.8
:2?
'1': 0.6
",3; Open
75 - - - Closed
.3 0.4
S
m
0.2 J
0
8/1/93 9/20/93 11/9/93 12/29/93 2/17/94 4/8/94 5/28/94
Date
Figure 7. Ruffed grouse survival in the Maltby Hills sites, 1993 — 1994.
1 -
5‘ 0.8
2.5 06
7% ' P —0pen
a.
E 04 H " " ' CIOSCd
t
:3
m 0.2
0

 

 

8/1/94 9/20/94 11/9/94 12/29/94 2/17/95 4/8/95 5/28/95
Date

Figure 8. Ruffed grouse survival in the Maltby Hills sites, 1994 — 1995.

31

Survival probability
.9 .O .O
A Ox co

9
N

 

 

 

- - - Closed

 

 

0

8/1/95 9/20/95 11/9/95 12/29/95 2/17/96 4/7/96 5/27/96

Figure 9.

Survival probability
p
on

.o
N

O

P
ox

.9
4:.

 

Date

Ruffed grouse survival in the Maltby Hills sites, 1995 — 1996.

 

 

Open
- '- - Closed

 

 

 

8/1/96 9/20/96 11/9/96 12/29/96 2/17/97 4/8/97 5/28/97

Date

Figure 10. Ruffed grouse survival in the Maltby Hills sites, 1996 — 1997.

32

 

0.8 .

0.6 1

Survival probability

0.4 q

.——._._. -_. .A if

 

0.2

 

 

 

0

8/1/97 9/20/97 11/9/97 12/29/97 2/17/98 4/8/98 5/28/98

Figure 11. Ruffed grouse survival in the Maltby Hills sites, 1997 — 1998.

Date

33

 

Open
- - - Closed

Table 10. Overall survival of ruffed grouse in the Maltby Hills study sites for each year
(August 1 through May 15 of the following year).

 

 

1993 1994 1995 1996 1997
Open 0.00 a 0.41 0.21 a 0.36 0.28
Closed 0.23 0.33 0.10 0.33 0.33

 

a = Signiﬁcant difference in survival curves between open and closed sites (Log-rank
test, P < 0.10)

Table 11. Overall survival for ruffed grouse in the Maltby Hills open study site for each
year (August 1 through May 15 of the following year) by sex and age class.

 

 

1993 1994 1995 1996 1997
Male 0.00 0.59 0.12 0.19 a 0.28
Female 0.00 0.39 0.27 0.48 0.25
Adult 0.00 b 0.69 b 0.23 b 0.46 0.46
Juvenile 0.00 0.34 0.12 0.33 0.26

 

a = Signiﬁcant difference between survival curves for male and female ruffed grouse
(Log-rank test, P < 0.10)

b = Signiﬁcant difference between survival curves for adult and juvenile ruffed grouse
(Log-rank test, P < 0.10)

Table 12. Overall survival for ruffed grouse in the Maltby Hills closed study site for each
year (August 1 through May 15 of the following year) by sex and age class.

 

 

1993 1994 1995 1996 1997
Male 0.30 0.19 a 0.14 a 0.31 0.36
Female 0.13 0.54 0.00 0.43 0.28
Adult 0.27 0.30 0.13 b 0.33 0.19
Juvenile 0.20 0.38 0.07 0.41 0.36

 

a = Signiﬁcant difference between survival curves for male and female ruffed grouse
(Log-rank test, P < 0.10)

b = Signiﬁcant difference between survival curves for adult and juvenile ruffed grouse
(Log-rank test, P < 0.10)

34

Survival - Pigeon River

As in the Maltby Hills sites, there was a very low sample size for the Pigeon
River survival analysis in 1993. Birds in the Pigeon River open study site had a survival
probability of 0.19 in the 1993 — 1994 season (Table 13). As seen in the Maltby Hills
open site, there was a sharp drop in survival through the end of September (Figure 12),
but there was another sharp drop from mid-December through early January. No juvenile
birds survived past January 1994. Females survived better than males, but the difference
was not signiﬁcant.

In the Pigeon River study area, 1993 — 1994 was the only season where survival
curves for the open and closed sites were signiﬁcantly different (P = 0.006) (Table 13).
The closed site survival curve had a sharp drop later in September (Figure 12), but
survival then remained constant and high through the spring. The closed site had an
overall survival of 0.63, the highest recorded in any site during the study. Females and
adults survived better than did males and juveniles, but the differences were not
signiﬁcant.

In the Pigeon River open site in 1994, grouse had a survival probability of 0.33,
much higher than that for the previous season (Table 13). The drop in survival during
September is more pronounced in males and juveniles (Figure 13). Females ended the
season with a higher survival probability than males, but the difference was not
signiﬁcant (Table 14). Adult grouse had a signiﬁcantly higher survival probability (S =
0.51) than juvenile grouse (S = 0.22) (P = 0.041).

Survival in the closed site (S = 0.37) was comparable to open site survival (Table

13), but lower than in 1993. The survival curves for the open and closed site were nearly

35

identical (Figure 18). The survival curves for adults and juveniles in the closed site were
also similar (Figure 13), with ﬁnal survival at about 40%. The curves for male and

female survival separated in late September, and females had a ﬁnal survival probability
of 0.56. The survival curve for males is signiﬁcantly different than that for females (P =
0.057), and males ﬁnished the season with a survival probability of only 0.17 (Table 15).

As in the Maltby Hills study sites, both Pigeon River sites had lower survival in
1995 - 1996 than in the previous season. The open site had a very pronounced drop in
survival in the months of September - November (Figure 14), and ended with a ﬁnal
survival probability of 0.23 (Table 13). Females had a higher survival probability than
males, but no signiﬁcant differences were detected. Adults and juveniles both ﬁnished
the season with survival probabilities of 0.25 (Table 14).

During this season, the closed site had its lowest overall survival probability of
the entire study, 0.12 (Table 13). As in the open site, a sharp decrease in survival
occurred from mid-September through mid-December (Figure 19). Adults ended with a
signiﬁcantly higher survival probability (0.16) than juveniles (0.07) (P = 0.065) (Table
15). There was no signiﬁcant difference in survival between males and females.

In the 1996 — 1997 season, survival in the Pigeon River open site improved only
slightly from the previous year. Overall survival probability was 0.27 (Table 13).
Survival probability decreased steadily through January, and then fell sharply through
February (Figure 15). Juvenile survival was lower than adult survival, and males and
females had comparable survival probabilities. There were no signiﬁcant differences

detected (Table 14).

36

In 1996, the Pigeon River closed site had a comparable survival probability to the
open site (Table 13), also better than the 1995 — 1996 season. The open and closed
survival curves are generally parallel, but converge in spring (Figure 20). Very few birds
were lost in the ﬁrst month of this season (Figure 15), but survival probability decreased
steadily aﬁer mid-September. Females survived better than males did, but no signiﬁcant
differences were detected between sexes or age classes (Table 15).

The highest overall survival in the Pigeon River open site was in the 1997 — 1998
season. Final survival probability was 0.54 (Table 13). The dip in survival in September
was present, but not as extreme as in previous years (Figure 16). Survival was fairly
constant from fall to spring. Adults and juveniles had a higher overall survival
probability by May 1998, but no signiﬁcant differences were detected (Table 14).

The closed site also had a high survival probability in 1997 - 1998 (S = 0.47), the
highest since the ﬁrst year of the study (Table 13). The open and closed curves cross in
February (Figure 21), but aren’t signiﬁcantly different. Adults and males had slightly
higher survival probabilities than did females and juveniles, but the survival curves did
not differ signiﬁcantly (Table 15). There were two sharp drops in survival, one in

October and one in February, but the survival curves were otherwise relatively level.

37

 

0.8-.

0.6

 

0.4 -

Survival probability

0.2 .

 

0

 

Open

 

 

 

 

8/1/93 9/20/93

11/9/93 12/29/93 2/17/94 4/8/94 5/28/94
Date

 

0.8 ._- _. _ ,-.

Closed

 

0.6 -a _

0.4 -2 .

Survival probability

0.2 -.._ _,

 

0

 

 

 

 

 

8/1/93 9/20/93 11/9/93 12/29/93 2/17/94 4/8/94 5/28/94

Date

—Open
. . . Male
__ Female
...... Adult
Juvenile

 

——Clo sed
. . . Male
_. Female
...... Adult
Juvenile

 

Figure 12. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1993 - 1994, by sex and age class.

38

 

 

 

 

 

 

 

 

 

 

 

 

 

0.8
E
{B
D 0.6 -
‘8; __Open
2 004‘ ....Closed
m 0.2
0
8/1/95 9/20/95 11/9/95 12/29/95 2/17/96 4f7/96 5/27/96
Date
1
0.8
$3 —Closed
3 0.6 . - - Male
tea. —Female
3 0.4 ,___2__-- .. - - - - . ,, I - H - ,. g ...... Adult
S -""-.,_ ' Juvenile
W ..q‘
0.2 ._,__ 2 - - , X; ..
0
8/1/94 9/20/94 11/9/94 12/29/94 2/17/95 4/8/95 5/28/95

Date

Figure 13. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1994 - 1995, by sex and age class.

39

 

 

Survival probability

 

 

 

O

 

8/1/95 9/20/95 11/9/95 12/29/95 2/17/96 4/7/96 5/27/96

Date

 

1 Closed

Survival probability

 

 

 

 

 

8/1/95 9/20/95 11/9/95 12/29/95 2/17/96 4/7/96 5/27/96
Date

—Open
. . . Male
—_Female
...... Adult
Juvenile

 

—Closed
. . . Male
__.Female
...... Adult
Juvenile

 

Figure 14. Survival probability for ruffed grouse collared in the Pigeon River open and

closed sites, 1995 — 1996, by sex and age class.

40

Survival probability

Survival probability

 

 

   
 

 

 

 

 

    

 

Open
0.8 - _ a
_Open
0.6 . . . . Male
—Female
0.4 l J ...... Adult
Juvenile
0.2 -___._ , D_ i g 4
0
8/1/96 9/20/96 11/9/96 12/29/96 2/17/97 4/8/97 5/28/97
1
Closed
0.8 ._.— -
_Closed
0.6 .._ - . . Male
__Female
0.4 w k f ...... Adult
Juvenile
0.2 ..._ _ “~..
0

 

 

 

8/1/96 9/20/96 11/9/96 12/29/96 2/17/97 4/8/97 5/28/97

Figure 15. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1996 — 1997, by sex and age class.

41

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2;: —Open
3 I I I Male
g. ——Female
3 ...... Adult
S Juvenile
(I)
0.2 MW. -
0
8/1/97 9/20/97 11/9/97 12/29/97 2/17/98 4/8/98 5/28/98
Date
1
Closed
0.8 .t
2? _Closed
3 0.6 1» I I I Male
g —Fen'lale
E 0-4 w --.-...Adult
E Juvenile
m
0.2 -. _
0
8/1/97 9/20/97 11/9/97 12/29/97 2/17/98 4/8/98 5/28/98

Date

Figure 16. Survival probability for ruffed grouse collared in the Pigeon River open and
closed sites, 1997 — 1998, by sex and age class.

42

 

.0
co

9
as

Survival probability
.9
A

 

p
N

 

 

 

0

8/1/93 11/9/93 12/29/93 2/17/94 4/8/94 5/28/94

9/20/93

Date

Figure 17. Ruffed grouse survival in the Pigeon River sites, 1993 - 1994.

 

 

Survival probability

P
N

 

 

 

O

8/ l/94 9/20/94 5/28/95

ll/9/94 12/29/94 2/l7/95 4/8/95

Date

Figure 18. Ruffed grouse survival in the Pigeon River sites, 1994 — 1995.

43

 

Open
- - - Closed

- - - Closed

 

0.8 ~
0.6 ~ Open

04 - - - - Closed

Survival probability

0.2 --

 

 

 

 

0
8/1/95 9/20/95 ll/9/95 12/29/95 2/17/96 4/7/96 5/27/96

Date

Figure 19. Ruffed grouse survival in the Pigeon River sites, 1995 — 1996.

 

0.8 «F A a

0.6 «

 

Open

04 __ _ _ -‘- -Closed

 

Survival probability

0.2 a- ,

 

 

 

0
8/1/96 9/20/96 11/9/96 12/29/96 2/17/97 4/8/97 5/28/97

Date

Figure 20. Ruffed grouse survival in the Pigeon River sites, 1996 — 1997.

44

 

 

 

 

 

b 0.8 _ i
a 06 —* \'_ M
“g \‘ —Open
a. M - - _ - . - l d
E 0.4 q__ f , - Cose
m 0.2 w

0

 

8/1/97 9/20/97 11/9/97 12/29/97 2/17/98 4/8/98 5/28/98

Date

Figure 21. Ruffed grouse survival in the Pigeon River sites, 1997 — 1998.

45

Table 13. Overall survival of ruffed grouse in the Pigeon River study sites for each year
(August 1 through May 15 of the following year).

 

 

1993 1994 1995 1996 1997
Open 0.19 a 0.33 0.23 0.27 0.54
Closed 0.63 0.37 0.12 0.25 0.47

 

a = Signiﬁcant difference in survival curves between open and closed sites (Log rank test,
P < 0.10)

Table 14. Overall survival for ruffed grouse in the Pigeon River open study site for each
year (August 1 through May 15 of the following year) by sex and age class.

 

 

1993 1994 1995 1996 1997
Male 0.10 0.23 0.14 0.26 0.50
Female 0.28 0.37 0.34 0.29 0.65
Adult 0.19 0.51 b 0.25 0.32 0.70
Juvenile 0.00 0.22 0.25 0.20 0.47

 

b = Significant difference between survival curves for male and female ruffed grouse
(Log-rank test, P < 0.10)

Table 15. Overall survival for ruffed grouse in the Pigeon River closed study site for each
year (August 1 through May 15 of the following year) by sex and age class.

 

 

1993 1994 1995 1996 A 1997
Male 0.54 0.17 a 0.04 0.22 0.48
Female 0.73 0.56 0.19 0.35 0.44
Adult 0.68 0.38 0.16 b 0.24 0.51
Juvenile 0.51 0.41 0.07 0.28 0.43

 

a = Signiﬁcant difference between survival curves for male and female ruffed grouse
(Log-rank test, P < 0.10)

b = Signiﬁcant difference between survival curves for adult and juvenile ruffed grouse
(Log-rank test, P < 0.10)

46

Causes of mortality

In the Maltby Hills open site, avian predation represented 31% to 61% of known
mortalities each year (Table 16). [Note that in tables 16 -— 18, totals by sex and age class
may not equal the total number of birds; for some birds, sex, age, or both were unknown.
These birds were excluded from the respective analyses] Mammals accounted for 28%
of known mortalities in 1994, but varied from 8 - 15% in other years. In the ﬁrst two
seasons, 1993 and 1994, hunting accounted for 18% and 21% of known deaths. In the
last three seasons, the percentage of hunting mortality was 12 — 14%. The category
“other” included such causes as being hit by a car or ﬂying into an object, and accounted
for no more than 1 bird per year. In each year, the cause of death could not be
determined in 6 — 15% of the total mortalities. Two to 13 birds per year were classiﬁed
as censored.

In the Maltby Hills closed site, avian mortality ranged from 53 — 71% of total
mortalities (Table 17), slightly higher than the percentages for the open site. Mammals
accounted for 8 — 22% of mortalities, and no birds were known to have been shot in this
site. There was no diagnosis for up to 22% of known mortalities, a higher percentage
than in the open site. As in the open site, disease (such as malnutrition and pulmonary
congestion) took relatively few birds, and only 2 birds out of all years were classiﬁed as
“other”. Five to 22 birds per year were censored, with large numbers occurring in 1995
and 1997.

As in the Maltby sites, avian predation was generally the highest source of
mortality in the Pigeon River open site, ranging from 24 — 60% (Table 18). Mammals

accounted for 7 - 18% of all deaths, and a high number of birds were not diagnosed in

47

1995 and 1997. The percentage of birds shot varied from 13% in 1996 to 35% in 1997
surpassing the percentage of deaths due to avian predation in that year. Values ranged
from 17 — 28% of known mortalities in 1993 — 1995. Only 1 bird was lost due to disease
(in 1995), and one was classiﬁed as “other”. Nine to 20 birds were censored per year.

Avian predation in the closed site took 38 — 77% of mortalities (Table 19), and
was higher than the percentages in the open site in all years but 1993. Mammals were
responsible for 25% of deaths in 1993 (2 out of 10), but that number dropped to 6 — 12%
in the remaining years. No birds were known to have been shot in the Pigeon closed site.
Again there was a high number of birds not diagnosed in 1995, and a high number of
censored birds in 1997. Four birds died of disease across the 5 years in this site, and 2
were categorized as “other” (both in 1997).

Over the 5 study seasons, both open study sites, more males were shot than
females (Table 20). In the Maltby site, the ratio of males to females shot (16:7) was
signiﬁcantly different than the ratio of males to females used in the analysis of survival
(126:125) (Chi-square, P = 0.078). In the Pigeon River site, the difference in ratios was
not signiﬁcant. In all, 12% of males collared and 8% of females collared were
attributable to hunting mortality.

In the Maltby open site, 12 juveniles, and 10 adults were shot, which is not
signiﬁcantly different than the age structure of the birds used in the analysis of survival.
In the Pigeon site, over 3 times more juveniles were shot than adults: 5 adults and 16
juveniles. This ratio was signiﬁcantly different than the ratio of adults to juveniles used
in the analysis (99: 126) (Chi-square, P = 0.067). Hunters shot 8% of collared adults and

9% of collared juveniles in the 5 years of the study.

48

Table 16. Fate of all birds used in survival analysis and percentage of known mortalities
for each year in the Maltby Hills Open study site, as of May 15 of the following year.

1993 1994 1995 1996 1997

 

 

# % # % # % # % # %

 

Alive 0 30 11 18 1 1
Avian 9 52.9 9 31.0 29 60.4 20 58.8 22 61.1
Mammal 2 11.8 8 27.6 7 14.6 4 11.8 3 8.3
Shot 3 17.6 6 20.7 6 12.5 4 11.8 5 13.9
No diagnosis 1 5.9 4 13.8 5 10.4 5 14.7 5 13.9
Disease 1 5.9 1 3.4 0 0.0 0 0.0 0 0.0
Other 1 5.9 1 3.4 1 2.1 1 2.9 1 2.8
Total mortality 17 29 48 34 36
Total censored 2 7 10 6 13
Radio failure 1 4 5 1 8
Other 1 3 5 5 5

 

Table 17. Fate of all birds used in survival analysis and percentage of known mortalities
for each year in the Maltby Hills closed study site, as of May 15 of the following year.

1 993 1994 l 995 1 996 1997
# % # % # % # % # %

 

 

 

Alive 5 15 5 21 16
Avian 13 68.4 17 53.1 29 63.0 22 71.0 26 63.4
Mammal 3 15.8 7 21.9 7 15.2 3 9.7 '5 12.2
Shot 0 0.0 O 0.0 0 0.0 0 0.0 0 0.0
No diagnosis 3 15.8 6 18.8 10 21.7 4 12.9 9 22.0
Disease 0 0.0 2 6.3 0 0.0 1 3.2 0 0.0
Other 0 0.0 0 0.0 0 0.0 1 3.2 1 2.4
Total mortality 19 32 46 31 41
Total censored 5 15 19 1 1 21
Radio failure 3 3 5 4 10
Other 2 12 14 7 11

 

49

Table 18. F ate of all birds used in survival analysis and percentage of known mortalities
for each year in the Pigeon River open study site, as of May 15 of the following year.

1993 1994 1995 1996 1997
# % # % # % # % # %

 

 

 

Alive 3 6 7 7 17
Avian 10 58.8 8 44.4 12 38.7 18 60.0 4 23.5
Mammal 2 11.8 3 16.7 2 6.4 5 16.7 3 17.6
Shot 4 23.5 5 27.8 5 16.1 4 13.3 6 35.3
No diagnosis 0 0.0 2 11.1 10 32.3 3 10.0 4 23.5
Disease 1 5.9 0 0.0 1 3.2 0 0.0 0 0.0
Other 0 0.0 0 0.0 1 3.2 0 0.0 O 0.0
Total mortality l7 18 31 30 17
Total censored 9 18 14 20 18
Radio failure 5 4 3 7 4
Other 4 14 11 13 14

 

Table 19. Fate of all birds used in survival analysis and percentage of known mortalities
for each year in the Pigeon River closed study site, as of May 15 of the following year.

1993 1994 1995 1996 1997
# % # % # % # % # %

 

 

 

Alive 10 10 5 10 16
Avian 3 37.5 13 76.5 25 60.1 20 62.5 13 54.2
Mammal 2 25.0 1 5.9 4 9.8 3 9.4 2 8.3
Shot 0 0.0 O 0.0 0 0.0 O 0.0 0 0.0
No diagnosis 2 25.0 3 17.6 12 29.3 7 21.9 6 25.0
Disease 1 12.5 0 0.0 0 0.0 2 6.3 1 4.2
Other 0 0.0 0 0.0 0 0.0 0 0.0 2 8.3
Total mortality 8 17 41 32 24
Total censored 6 13 12 11 24
Radio failure 2 5 8 6 10
Other 4 8 4 5 14

 

50

Table 20. Sex and age classes of collared ruffed grouse shot in the Maltby Hills and
Pigeon River Areas, 1993-1998, expressed as frequency and percentage of total birds.

 

 

 

Maltby Hills Pigeon River Total
Freq. % Freq. % Freq. %
Male 16 12.7 13 11.3 29 12.0
Female 7 5.6 11 9.9 18 7.6
Adult 10 11.5 5 5.1 15 8.1
Juvenile 12 6.7 16 12.7 28 9.2
Totals 24 8.8 24 10.3 48 9.5

 

51

Discussion

Although our trapping methods were not randomized, I expected that the sample
caught would reﬂect the sex and age structure of the population. Thus it was expected
that there would be a higher number of juveniles than adults in our sample population,
because they probably represent a higher proportion of the population during the fall
distribution. In some cases, however, we caught and collared more adults than juveniles
in a given year. This could be due to extremely high juvenile mortality, or may be an
artifact of the sampling methods used.

One possible reason for such a result is a bias in the trapping method. The ﬁmnel
of the cloverleaf trap is sized to prevent birds from exiting, but in August, juvenile birds
are still much smaller than adults. Some smaller birds may have escaped through the
funnel (evidence included feathers on or near the trap entrance, but an empty trap). In
addition, weight restrictions on collar placement probably resulted in juvenile birds being
released because they were too small (< 350 g) to be collared.

Grouse populations are known to cycle over a period of around 10 years. This has
been documented in the northern states regardless of differences in hunting regulations
(Bump et al. 1947). In Michigan, the low point in the cycle occurred around 1992-1993,
and according to drumming and ﬂushing counts, the population increased as the study
progressed (DNR report — unpublished data). Assuming that trapping success is at least
loosely correlated with population number, we expected to catch more birds later in the
study. This result is supported by the general increase in trapping success across years in
both the Maltby Hills and Pigeon River study areas. Although some of the increase could

be due to an increase in trapping experience by the researchers, experience is not likely to

52

be the only explanation. Trapping success continued to increase even through the last
year of the study, after lead researchers had several years of trapping experience.

The results presented provide little evidence that hunting is additive to natural
mortality in Michigan. If hunting mortality was additive, I would expect to see a
difference in fall to spring survival between sites open and closed to hunting. In the ﬁrst
year of the study (1993-1994), there were signiﬁcant differences in survival between sites
in both areas, with closed sites having higher survival probabilities. In subsequent years,
however, survival probabilities were statistically comparable in the open sites. In fact,
birds in the Maltby Hills open site survived better than those in the closed site in 1995.
Although adult birds dominated the open site sample in that year, both adults and
juveniles had higher age speciﬁc survival rates than closed site birds. Such data suggest
that differences may show up in years of low population numbers (or low sample size)
but that harvest is not a factor in fall to spring survival in years of rising populations.

Other evidence of hunting mortality being additive would be equal losses due to
natural mortality in open and closed sites, with additional hunting losses in the open sites.
Using ﬂushing and drumming counts, Palmer and Bennett (1963) measured a 60 - 70%
fall to spring loss of birds regardless of hunting pressure. Total mortality in my study, as
a percentage of birds used in the analysis was comparable between open and closed sites
in most years. With the exception of Pigeon River 1993, a higher percentage of
mortalities in the closed sites were attributable to avian predation. This may suggest a
more compensatory role of hunting, where a lower number of birds are lost to natural

causes than would be in an non-hunted site.

53

Rusch and Keith (1971b) showed that grouse populations declined rapidly
through November, which coincided with the hunting season. Then populations
remained stationary through the following spring. I recorded a similar drop in survival in
most cases, but in some years there were similar drops in the sites closed to hunting.
Survival curves then appear to equalize over time, resulting in similar survival rates.
There may be other factors that inﬂuence that drop in survival. Some authors suggest
that migrating hawks have some effect on survival during the fall in the northern United
States (Rusch and Keith 1971b; Small et al. 1991).

Small et al. ( 1991) found no signiﬁcant difference in hunting mortality between
adults and juvenile grouse. Our study, however did show a difference in survival in some
cases, and signiﬁcantly more juveniles shot than adults in the Pigeon River open site. In
both Open sites adults always survived better than did juveniles, with some signiﬁcant
differences detected. In the closed sites, though, there were no noticeable trends.
Survival was comparable for adults and juveniles in the closed sites (except for 1995),
suggesting that hunting may have an impact on juvenile survival.

In nearly every case, avian predation was the major source of fall to spring
mortality for ruffed grouse, which is consistent with the research of Small et al. (1991).
Overall, hunting took only about 10% of collared birds, below that reported by other
researchers, and well below the recommended 20-50% safe harvest levels. Recent
studies in Ohio show a mean harvest rate of 18% of the fall population, with annual rates
ranging from 13-24% (Stoll and Culbertson 1995). My data are closer to that reported by

Small et al. (1991) for private hunting lands in central Wisconsin.

54

High censoring levels in some years could be due to lack of reported hunting
mortality, but counting all censored as hunted still wouldn’t nearly reach the level of 60%
reported by Small et al. (1991) for public hunting areas. There is a concern, though,
about differential censoring between sites open and closed to hunting. In the Maltby
Hills closed site, the censoring rate was nearly double that in the open site. Although
there was no direct evidence of hunting in the posted area, some of the censored birds
may have been shot illegally or shot just outside the closed site borders. The effect is the
reverse in the Pigeon River sites, however, with more birds being censored in the site
open to hunting. The basic assumption of this project is that non-hunting areas are not
hunted, and that collars from shot birds are returned. It is unknown to what extent lack of
reporting and illegal hunting affected the results. Such problems will exist in any such
study, but may be minimized by public education. outreach, and a fair and prompt reward
system.

Ideally, the Maltby Hills and Pigeon River sites could be combined and used to
generalize over a broad range. Although the Maltby Hills and Pigeon River differ in
many ways, there were some effects that were similar. Trapping trends were consistent,
indicating that the population was increasing across northern Michigan. Both sites had
signiﬁcantly higher survival in the closed sites in 1993, but not in any other years. In
1995, both sites had lower survival than other years, with open sites showing higher
survival than in the closed sites. Such indications suggest that some effects are broad and
results may be generalized over a wide geographic area. In the next 3 chapters, I will

investigate ruffed grouse movements and habitat use. These results may begin to explain

55

and model some of the effects that take place on a smaller scale and cause differences

between our study areas.

56

CHAPTER 2: Fall activity range and movements

Introduction

To aid in the management of habitat for wildlife species, information about
movement patterns must be known. Through the year, ruffed grouse exhibit movements
that can be quantiﬁed and used to plan habitat management on a landscape level. Ruffed
grouse are known to disperse in the fall, when broods break up, to search for suitable
winter habitat. It is generally thought that birds will move to and settle in winter habitat
that provides them with the resources (food and cover) needed to enhance survival
through the spring (Bergerud and Gratson 1988). Understanding the characteristics of
movement, activity range size, and distances traveled by grouse in the fall can aid in the
prediction of the effects of habitat management decisions.

Movements (or lack of movements) are assumed to be related to the presence or
absence of suitable winter habitat. For some grouse, their winter range will contain the
resources needed to survive through the winter, and they will not disperse. The size of
seasonal activity areas varies by region, and thus probably by habitat, food availability, or
density. In a Missouri study, fall-winter home range size for male grouse was 83 ha
(Neher 1993). Thompson (1987, also in Missouri) combined males and females, and
found a fall to winter activity range of 78 ha. Fall activity areas for grouse in Virginia
measured around 20 ha (Fearer et al. 1999), a much smaller area than in the Missouri

studies.

57

Birds that do not ﬁnd adequate winter food and cover on their fall ranges will
need to move to ﬁnd habitat of sufﬁcient quality. This is supported by results in Virginia
that showed an increase in daily movement as the percent of lower quality food
(rhododendron/laurel habitat) within a bird’s home range increased (F earer et a1. 1999).
According to a compilation of telemetry studies, juveniles usually travel farther than
adults and have larger ranges, possibly because of competition for resources and breeding
sites. Adult males are the most sedentary, followed by adult females, juvenile males, and
juvenile females (Bergerud and Gratson 1988). The Virginia study showed that females
moved farther on a daily basis than did males in all seasons (Fearer et al. 1999).
Juveniles females in Wisconsin dispersed an average of 4.8 km, while juvenile males
dispersed 2.1 km (Small and Rusch 1989). This is valuable information to know in
Michigan, because birds that move longer distances may be at an increased risk for
predation.

Density dependent factors are those that change in direct response to changes in
population numbers. Behavior and productivity are among possible factors that can be
inﬂuenced by proximity of members of the same species. If habitat is a limiting factor
for a species, then as population numbers climb, increased competition may force some
animals to move farther to ﬁnd suitable and unoccupied habitat. If ruffed grouse
movements are related to density, then it is presumed that movements will increase when
populations increase.

During the study, the grouse population was increasing across the northern lower
peninsula of Michigan. In fact, the number of birds taken by hunters rose from 1.7

grouse/hunter in 1994 to 2.8 grouse/hunter in 1996 (Karasek 1998). Assuming that fall

58

movements were associated with density, I hypothesized that movements of grouse
would increase from 1992 - 1997. The Wisconsin study mentioned above measured
mean values for dispersal over several years of research. The possibility of density
dependence, however, precludes the combination of data across years.

In addition to the direct loss of birds discussed in Chapter 1, hunting of grouse in
Michigan probably has indirect effects on the ability of birds to survive through the
winter. As hunters walk through occupied habitat, they may ﬂush the same birds
repeatedly. Assuming that hunters choose habitat that appears “high quality”, grouse
may be forced to move out of good habitat into poor habitat. I assumed that birds would
move more in hunted areas, because of increased ﬂushing. I also expected both larger
home ranges for non-dispersers and farther dispersal distances in hunted areas.

This chapter will quantify grouse movements during fall and winter in two study

areas in northern Michigan.

59

Methods

Radio locations for each bird were taken a minimum of 2 times per week ﬁ'om the
time of capture through late December. The time of location for each bird was varied
daily to provide maximum habitat use information. Bird locations were either manually
estimated on topographic maps or calculated by the computer program TELEM88
(Virginia Polytechnic Institute and State University, Dept. of Fisheries and Wildlife,
Blacksburg, VA) using two compass bearings from known locations, usually on county
roads. Because of the high density of roads in the study areas, compass bearings were
generally taken from < 1 km from the bird’s location. Field trials by research personnel
indicated an average bearing error of approximately 2°. Locations were recorded in
Universal Transverse Mercator (UTM) coordinates to the nearest 50 m.

Because birds were located for a variable number of days per week, a weekly
harmonic center (Dixon and Chapman 1980) was calculated for each bird. Two criteria
were used to determine if a bird would be used for the analysis of fall movements. The
bird must have been located during at least 5 different weeks between time of capture and
December 31. In 1989, Small and Rusch found that ruffed grouse in Wisconsin dispersed
beginning in late September to mid-October. To ensure that dispersal was not overlooked
by recording movements too early or too late in the fall, birds were eliminated from the
analysis if their analysis period did not include the week of October 3. It is important to
note, however, that the inclusion of birds that did not survive through the end of the study
season means that some dispersal patterns may have been miscategorized.

For birds included in the analysis, the distances between the weekly harmonic

center of the week after capture and the weekly harmonic centers of all subsequent weeks

60

were calculated and plotted (Appendix A). Grouse were separated into dispersal

categories using the following criteria:

1) Birds which never moved more than 500 m from the harmonic center of the ﬁrst
week after capture were considered Type I non-dispersers;

2) Birds which moved more than 500 m from the harmonic center of the ﬁrst week after
capture for 1 or 2 weeks at a time, but then returned to within 500 m before
November 14, were considered Type II non-dispersers;

3) Birds which moved more than 500 m from the harmonic center of the ﬁrst week after
capture and stayed 500 m or farther away for 3 or more consecutive weeks before

November 14 were considered dispersers, or Type III.

The November 14 cutoff was used to distinguish fall movements from late season
movements. Late season movements have been documented in other telemetry studies
(Small and Rusch 1989), and are considered to occur after a bird has settled into a winter
range.

Fall activity ranges were calculated for non-dispersers using the CALHOME
program (U. S. Forest Service, Albany, CA). The adaptive Kernel home range method,
using 95% contours, was used to calculate areas of grouse activity (Worton 1989).
Dispersal distance was considered the linear distance between the harmonic center of the

locations in the week captured and the harmonic center of all locations the week after

dispersal has ended.

61

Sex and age ratios of birds used in the analysis were compiled. and compared to
the sex and age ratios of birds within each dispersal type. For each year, activity ranges
and dispersal distances of each movement type were summarized by sex and age class.
They were compared within sites using a Kruskal-Wallis one-way ANOVA (Siegel and
Castellan 1988) with a level of signiﬁcance set at 0.10. Paired sites (open and closed to
hunting) were also compared within each year.

I tested for linear trends in activity ranges and dispersal distances among years
using the linear contrast method. Using the ProcGLM procedure in SAS (SAS Institute,
Cary, North Carolina), orthogonal weights of -2, -1, 0, 1 and 2 were assigned to years

1993 - 1997. Signiﬁcance was set at 0.10.

62

Results
Type I non-dispersers

Over the course of the study, 665 birds were classiﬁed into movement types
(Table 23). Of the 174 birds used in the analysis in the Maltby Hills open site, 35 were
Type I non-dispersers (20.1% of those used in the analysis). The closed site had 50 Type
I birds out of 213 used in the analysis (23.5%). With the exception of the open site in
1993 and 1996, more males were classiﬁed as Type I birds than females (Table 21). The
sex ratio of Type I birds was signiﬁcantly different than the sex ratio of birds used in the
movement analysis in the closed site (Table 24, P < 0.01). In most cases, more adults
were classiﬁed as Type 1 birds than juveniles. In both the open and closed sites, the age
ratio was signiﬁcantly different than the ratio of birds used in the analysis (both P values
< 0.01).

In the Pigeon River study area, 64 birds were classiﬁed as Type I, 26 out of 123 in
the open site (21.1%), and 38 out of 155 in the closed site (24.5%). As in the Maltby
Hills sites, more males were classiﬁed as Type I birds than females in most years (Table
22) and overall (Table 24). No signiﬁcant differences were detected. In the Pigeon River
closed site, the age ratio was signiﬁcantly different than the age ratio of those birds used
in the analysis (P = 0.01), with more adults being classiﬁed as Type I than juveniles.

In 1993, the mean fall activity range size for Type I grouse in the open Maltby
site was 71.5 ha (Table 25). In the years following, mean range size was much smaller,
and consistently around 40 ha. Female birds tended to have slightly larger ranges than

did male birds, but no signiﬁcant differences were detected between sex or age classes.

63

In 1994, 1995, and 1997 juveniles had larger activity ranges than adults. Among years, a
linear trend was detected, with range size decreasing across time (P = 0.01).

In the closed site, activity range size varied yearly, ranging from 29.0 ha in 1995
to 67.3 ha in 1996 (Table 26). In 1993 and 1995, home ranges were signiﬁcantly smaller
in the closed site than the open site (P = 0.08, P = 0.05, respectively). There were no
signiﬁcant differences between sex or age classes. No linear trend in activity range size
among years was detected.

As in the Maltby Hills open site, birds in the Pigeon River open site in 1993 had
larger home range sizes than the years following (Table 27). All Type I birds in that year
were adult males. In 1994, only one bird was classiﬁed as Type I, also an adult male. In
1995 - 1997, females had larger home ranges than males, but the difference was only
signiﬁcant in 1997 (P < 0.01). Juveniles had larger home ranges than adults, but no
signiﬁcant differences were detected. No linear trend in activity range size among years
was detected.

Mean activity range size for grouse in the Pigeon River closed site differed from
the open site in 1997 only, when it was signiﬁcantly smaller (Table 28, P = 0.05). There
was a linear trend detected, with a decrease in mean activity range size from 1993 to
1997 (P < 0.01). In 1993, males had larger ranges than females, but the difference was
not signiﬁcant. In 1994 and 1996, females had larger ranges and the difference was
signiﬁcant in 1996 (P = 0.09). There was no trend apparent between age classes. Adults
had signiﬁcantly larger ranges than juveniles in 1995 (P = 0.05), and juveniles had larger

ranges in 1997 (P = 0.01).

64

 

v. . . a N c w m . w v 230...

m N m N m m . . o . o 0:32...
o o m o m m w . v v :33.
v o N . o N . o N o 2.33...
o . m o . o o v . n v 2a.).

 

38.0 3&0 38.0 22.0 38.0 :80 38.0 3&0 38.0 22.0
woo. woo. 38. v3. m3.

 

.mmﬁo owe 98 x3 .3 a30. - 30. .33 33y. :oowr. 2.. :. whomaoamEéo: . 0%... mm 3.07.8.0 088m 3b.: .3 35.832 .NN v.3...

 

m . o m N m . m. o. N. v v 280...

o c v o . n v . N o 0:32....
o m v N N. o. o o N m :23.
. m o . . v o N N m 0.83“.

v. c w . . . w o. m N . 0.3).

 

38.0 :80 38.0 3&0 38.0 :30 38.0 3&0 38.0 :30
38. 03. mac. «3. m3.

 

.mmﬂo own use x8 .3 K3. - moo. .33 2.... .33.). 2.. a. maomaommﬁco: . on»... we 3.0.320 088m 3b.: ..o 39:52 ..N 0.3...

65

Table 23. Number of ruffed grouse used in the analysis of movements by sex and age
class.

 

Maltby Hills Pigeon River
Open Closed Open Closed

 

 

Male 83 123 66 82
Female 80 85 56 71

Adult 63 83 64 79
Juvenile 109 128 59 68

Totals 174 213 123 155

 

Table 24. Number of ruffed grouse classiﬁed as Type I non-dispersers by sex and age
class.

 

Maltby Hills Pigeon River
Open Closed Open Closed

 

Male 21 45‘I 17 29
Female 13 4 9 9

Adult 24b 30b 17 22b
Juvenile 10 20 9 l6

 

I Signiﬁcant difference between sex distribution of birds classiﬁed as Type I and birds
used in movement analysis (Chi-square, P < 0.10)
b Signiﬁcant difference between age distribution of birds classiﬁed as Type I and birds
used in movement analysis (Chi-square, P < 0.10)

66

Table 25. Mean activity area (ha) of ruffed grouse classiﬁed as Type I non-dispersers in
the Maltby Hills open site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)
Male 74.0 38.2 (4.9) 39.9 (7.6) 37.0 37.0 (10.3)
Female 70.7 (15.6) 48.0 (6.0) 44.3 (10.3) 42.0 39.7 (5.5)
Adult 81.7 (6.2) 39.5 (4.4) 38.3 (4.8) 39.5 (2.5) 28.0 (10.5)
Juvenile 50.0 52.3 (18.7) 42.8 (8.7)
All birds 71.5 (11.1) 41.0 (4.0) 41.5 (5.4) 39.5 (2.5) 37.9 (6.9)

 

Table 26. Mean activity area (ha) of ruffed grouse classiﬁed as Type I non-dispersers in
the Maltby Hills closed site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)
Male 36.0 (3.0) 41.8 (9.1) 30.6 (5.8) 67.3 (19.9) 51.1 (8.1)
Female 43.0 (1.0) 26.0 60.0
Adult 40.5 (1.5) 31.3 (5.6) 30.3 (5.3) 46.5 (11.8) 44.0 (14.8)
Juvenile 38.5 (5.5) 57.5 (20.1) 14.0 88.0 (37.7) 56.8 (82)
All birds 39.5 (2.4)b 41.8 (9.1) 29.0 (5.0)b 67.25 (19.9) 51.7 (7.5)

 

u Signiﬁcant difference in mean home range between open and closed sites (Kruskal-
Wallis one-way ANOVA, P < 0.10)

67

Table 27 . Mean activity area (ha) of ruffed grouse classiﬁed as Type I non-dispersers in
the Pigeon River open site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997

Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)

Male 73.8 (24.8) 35.0 44.3 (6.7) 52.0 24.6 (4.8)'
Female 101.5 (61.5) 70.0 57.7 (4.7)
Adult 73.8 (24.8) 35.0 34.3 (5.2) 21.7 (7.2)
Juvenile 73.2 (23.3) 61.0 (9.0) 54.5 (0.5)

All birds 73.8 (24.8) 35.0 58.6 (15.7) 61.0 (9.0) 42.6 (6.12)

 

' Signiﬁcant difference detected in mean home range between sex or age classes
(Kruskal-Wallis one-way AN OVA, P < 0.10)

Table 28. Mean activity area (ha) of ruffed grouse classiﬁed as Type I non-dispersers in
the Pigeon River closed site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)
Male 110.7 (47.7) 51.0 (11.42) 59.7 (10.3) 35.7 (9.6)' 25.7 (5.1)
Female 43.5 (11.5) 103.0 104.5 (51.5) 28.8 (10.2)
Adult 96.8 (36.5) 51.0 (11.4) 76.7 (13.1)' 61.0 (26.1) 15.0 (1.7)'
Juvenile 32.0 42.7 (8.4) 39.3 (10.8) 35.25 (6.1)

Allbirds 83.8(31.l) 61.4(13.7) 59.7(10.3) 52.9(16.5) 26.6(4.4)b

 

' Signiﬁcant difference detected in mean home range between sex or age classes

(Kruskal-Wallis one way ANOVA, P < 0.10)
b Signiﬁcant difference in mean home range between open and closed sites (Kruskal-

Wallis one—way AN OVA, P < 0.10)

68

Type II non-dispersers
In the Maltby Hills study sites, 75 birds exhibited Type II movements, 30 in the

open site and 45 in the closed site. These birds represented 17.2% and 21.1% of total
birds used in the analysis for those sites. In both sites, more males were classiﬁed as
Type 11 than females (T able 29), but the distribution was not signiﬁcantly different than
the sex distribution of those used in the analysis (Table 31). More adults were classiﬁed
as Type II than juveniles, and in the closed site, the age distribution was signiﬁcantly
different than the distribution of those used in the analysis (P < 0.01).

In the Pigeon River sites, 74 birds were Type II, with 35 in the open site (28.5%)
and 39 in the closed site (25.2%). In the open site slightly more males were classiﬁed
Type 11 than females, but in the closed site more females exhibited Type II movements
(Table 30). In both the open and closed sites more adults were classiﬁed as Type II than
juveniles, but there were no signiﬁcant differences between the sex or age distribution of
the Type II birds and the distributions of those birds used in the analysis (Table 31).

The mean activity area for Type II ruffed grouse in the Maltby Hills open study
site ranged from 63.3 ha in 1995 to 122.2 ha in 1994 (Table 32). There was no trend
evident between sex or age classes, and no linear trend detected. In 1996, mean juvenile
activity area was signiﬁcantly larger than mean adult activity area (P = 0.06).

In the Maltby Hills closed site, the smallest mean activity area was recorded in
1995 (60.9 ha, Table 33). No linear trends were detected and there were no consistent
differences between sex or age classes. Juvenile ranges were signiﬁcantly larger than

adult ranges in 1997 (P = 0.09), but adult ranges were larger in 1994 (P = 0.09). In all

69

years except 1993, activity ranges were smaller in the site closed to hunting. In 1994, the
difference was signiﬁcant (P = 0.05).

Mean activity areas for Type II birds in the Pigeon River open site ranged from
80.3 ha in 1996 to 147.3 ha in 1995 (Table 34). Differences between sex and age classes
varied widely and inconsistently by year. There were no signiﬁcant differences between
sex or age classes and no linear trend was detected.

A signiﬁcant linear downward trend was detected among years in the mean
activity range of ruffed grouse in the Pigeon River closed site (Table 35, P < 0.01). Birds
had a mean range of212.3 ha in 1993, and a mean range of35.2 ha in 1997. Female
birds had larger ranges than did males, and the difference was signiﬁcant in 1993 (P =
0.06) and 1995 (P = 0.01). No signiﬁcant differences were detected between adult and
juvenile ranges. In 1997, the mean range size in the closed site was signiﬁcantly smaller
than the mean range size in the open site (P = 0.07). In 1993 — 1996 there was no clear

trend between open and closed sites.

70

 

 

 

 

 

c v A. N. N. a o N o m £50...

.4 o v m N m N . N . o..3>3

N v m N w v v o .4 N 233‘

m N m o N m o m N N 0:23...

m N m o m e o m e . 0.3).

38.0 3&0 38.0 3&0 38.0 3&0 38.0 3&0 38.0 3&0
Nam. 0&2 mg. .43. m3.

.820 own .38 3m .2. .33 - mg. .33 33y. 309A. 2.. a. m3m3&m..o-:o: .. 0&3. 8 3.0.3.3 088w 3&8 ..o 3937. .9.“ 0.3...
m. N e N a c. c m m 28¢...

w .8 .4 N N m .4 o . 0:32...

o. m N m o N N N N .33.

v m N N N m N N N 0.23...

v. .4 m m N N v . o 0.3).

 

38.0 3&0 38.0 3&0 38.0 3&0 38.0 3&0 38.0 3&0

N3.

 

03.

m8.

3.. .

m3.

8% as e5 as B .82 - 32 ans 2:: 3%: as e. 88.2.3.8: = 6.5. a 83.83 638m Bea .e 8.52 .3 2.3

71

Table 31 . Number of ruffed grouse classiﬁed as Type II non-dispersers by sex and age
class.

 

Maltby Hills Pigeon River
Open Closed Open Closed

Male 17 31 18 15
Female 1 1 l4 16 23
Adult 15 27b 23 ‘ 20
Juvenile 14 17 12 19

 

1’ Signiﬁcant difference between age distribution of birds classiﬁed as Type II and birds
used in movement analysis (Chi-square, P < 0.10)

72

Table 32. Mean activity area (ha) of ruffed grouse classiﬁed as Type II non-dispersers in
the Maltby Hills open site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)
Male 115.5 (29.5) 61.0 (14.2) 133.7 (82.4) 115.0 (29.8)
Female 111.5 (51.5) 135.5 (6.0) 71.5 (12.5) 64.0 (13.0) 87.7 (11.1)
Adult 111.5 (51.5) 129.5 (29.5) 52.3 (6.8) 42.0 (9.0)' 87.7 (11.1)
Juvenile 94.0 118.5 (29.2) 98.5 (46.5) 160.8 (54.0) 115.0 (29.8)
All birds 105.7 (30.3) 122.2 (20.1) 63.3 (11.1) 121.2 (42.4) 101.3 (15.5)

 

' Signiﬁcant difference detected in mean home range between sex or age classes
(Kruskal-Wallis one-way AN OVA, P < 0.10)

Table 33. Mean activity area (ha) of ruffed grouse classiﬁed as Type II non-dispersers in
the Maltby Hills closed site, 1993 - 1997, by sex and age class.

 

 

 

1993 1994 1995 1996 1997
Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)

Male 96.0 86.1 (12.0) 52.4 (1 1.1) 86.0 (19.2) 89.4 (9.9)
Female 158.5 (109.5) 72.0 (13.9) 82.0 (21.0) 106.7 (49.1) 73.5 (15.8)
Adult 182.0 (86.0) 90.7 (10.9)' 67.4 (12.9) 77.0 (23.4) 74.8 (10.8)'
Juvenile 61.3 (11.8) 44.5 (16.5) 108.3 (34.9) 99.6 (12.3)

All birds 137.7 (66.6) 81.9 (9.2)” 60.9 (10.5) 94.9 (21.6) 85.8 (8.4)
' Signiﬁcant difference detected in mean home range between sex or age classes

(Kruskal-Wallis one-way ANOVA, P < 0.10)
b Signiﬁcant difference in mean home range between open and closed sites (Kruskal-
Wallis one-way ANOVA, P < 0.10)

73

Table 34. Mean activity area (ha) of ruﬁ‘ed grouse classiﬁed as Type II non-dispersers in
the Pigeon River open site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)
Male 175.0 99.0 (21.1) 175.3 (43.2) 77.3 (7.7) 95.5 (22.5)
Female 103.0 (57.0) 83.7 (11.2) 91.3 (30.5) 83.2 (11.1) 106.5 (68.5)
Adult 110.5 (64.5) 104.2 (19.1) 150.5 (22.7) 77.9 (5.3)
Juvenile 160.0 106.0 144.8 (58.9) 83.6 (14.7) 101.0 (29.6)
All birds 127.0 (40.7) 104.4 (16.2) 147.3 (32.4) 80.3 (6.5) 101.0 (29.6)

 

Table 35. Mean activity area (ha) of ruffed grouse classiﬁed as Type II non-dispersers in
the Pigeon River closed site, 1993 - 1997, by sex and age class.

 

 

 

1993 1994 1995 1996 1997
Area (SE) Area (SE) Area (SE) Area (SE) Area (SE)
Male 147.3 (17.6)' 74.2 (16.9)3 81.7 (17.7) 27.3 (5.8)
Female 342.5 (77.5) 1 16.0 (20.2) 189.7 (35.6) 91.4 (35.3) 43.0 (3.6)
Adult 168.3 (35.0) 114.0 (25.8) 180.4 (53.9) 98.0 (33.1) 30.0 (9.0)
Juvenile 300.5 (119.5) 120.0 (46.0) 113.9 (25.3) 106.3 (39.7) 37.8 (5.8)
All birds 212.3 (47.1) 116.0 (20.2) 141.6 (27.2) 101.7 (23.9) 35.2 (4.7)”
' Signiﬁcant difference detected in mean home range between sex or age classes

(Kruskal-Wallis one way ANOVA, P < 0.10)
b Signiﬁcant difference in mean home range between Open and closed sites (Kruskal-
Wallis one way ANOVA, P < 0.10)

74

Dispersers

In the Maltby Hills open site, 110 birds (63.2%) were classiﬁed as dispersers
during the 5 years of the study. In the closed site, 118 grouse dispersed (55.4%). In the
Pigeon River sites, 62 birds (50.4%) dispersed in the open site and 78 dispersed in the
closed site (50.3%). In most years in both sites, more females dispersed than males and
more juveniles dispersed than adults (Table 36). The age class distribution of dispersers
differed signiﬁcantly from those used in the analysis in the Maltby open and closed sites
(Table 38, all P values < 0.01), and the sex distribution differed in the Maltby closed site
(P < 0.01). In the Pigeon River sites (Table 37), the sex and age distribution of dispersers
did not differ ﬁom the distributions of birds used in the analysis.

In 1993 in the Maltby Hills open site, only one grouse was classiﬁed as a
disperser, a juvenile male that moved 599 m (Table 39). Mean dispersal in the following
years ranged from 2250 m to 3599 m. In 1994, 1996, and 1997, females moved
signiﬁcantly farther than males (P = 0.09, P = 0.03, P = 0.06, respectively). In 1995, the
large dispersal distance in males was largely due to one bird that moved over 15 km. The
mean dispersal distance of males calculated after omitting that individual was 2348 m,
which was not signiﬁcantly different from the dispersal distance of females. In all years
(1994 — 1997) juveniles moved farther than adults and the difference was signiﬁcant in
1995 and 1996 (P = 0.06, P = 0.02). No linear trend was detected in dispersal distances
among years.

The mean activity ranges in the Maltby Hills closed site did not differ from the
open site in any of the study seasons (Table 40). In 1993, 1994, 1996, and 1997 female

grouse moved farther than did males and the difference was signiﬁcant in 1996 (P =

75

0.04). Juveniles moved signiﬁcantly farther than adults in 1994, 1996, and 1997 (P <
0.01, P = 0.03, P = 0.03).

Mean dispersal distance for birds in the Pigeon River open site ranged from a low
of 990 m in 1995 to a high of 2640 m in 1997 (Table 41). Males and females moved
comparable distances, and there were no differences detected. In 1993 and 1994 adults
moved farther than juveniles, and in 1995 - 1997 juveniles moved farther than adults.
None of the differences were not signiﬁcant. There was a signiﬁcant downward linear
trend detected among years (P = 0.03).

In the Pigeon River closed site, mean dispersal distances never exceeded 2 km
(Table 42). Grouse dispersed the farthest in 1996, when the mean distance was 1941 m.
In 1993 and 1997, female grouse dispersed signiﬁcantly farther than males (P = 0.06, P =
0.01). In 1994, however, males moved signiﬁcantly farther than females (P = 0.08). In
all years except 1994 juveniles dispersed farther than adults and the difference was
signiﬁcant in 1997 (P = 0.08). No linear trend was detected in dispersal distances among
years. In 1995 birds moved signiﬁcantly shorter distances in the open site than in the

closed site (P = 0.07), while the reverse was detected in 1997 (P < 0.01).

76

 

 

 

 

 

 

N3.

 

08.

m3.

#3 .

m3.

.N N. cN m. 8 .. .. m. c 3 _So...
.. a a N o a v m N v 0:32;
o. m o. w a N m o v m :33.
a m m o v. e e e N m 0.23".
N. N N. w o m v a v v 0.3).
38.0 3&0 38.0 3&0 33.0 3&0 38.0 3&0 38.0 3&0
N3. 08. 3a. «8. mg.

.33 owe 23 3a .3 .N&&. .. mam. .33 3.5. coowr. 2.. a. ﬁoﬁamB E 09.... 3 3593.0 892w 3......— ..o .337. .Nm 23...
NN 3N mm ..N NN NN «N on m . .50...
w. mN Nm N N m. m. MN N . 0:32:

N v m m m a c. w m o :34.
m. m. a. o. o. N. e. m. m o 0.323...
c. m. c. a a N .. m. N . 0.3).

38.0 3&0 38.0 3&0 38.0 3&0 38.0 3&0 38.0 3&0

.336 as as. as... B .32 - 82 .88 2.... 3%: 2.. a £28.... E 95 a 333.6 8:8» Bee .6 895.2 .8 2.3

77

Table 38. Number of ruffed grouse classiﬁed as Type III dispersers by sex and age class.

 

Maltby Hills Pigeon River
Open Closed Open Closed

 

Male 45 47‘ 30 38
Female 56 67 3 l 39

Adult 24" 26b 27 37
Juvenile 85 91 34 34

 

IrSigniﬁcant difference between sex distribution of birds classiﬁed as Type III and birds
used in movement analysis (Chi-square, P < 0.10)
b Signiﬁcant difference between age distribution of birds classiﬁed as Type III and birds
used in movement analysis (Chi-square, P < 0.10)

78

Table 39. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Maltby Hills open site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997"" —
Dist. (SE) Dist. (SE) Dist (SE) Dist. (SE) Dist. (SE)
Male 599.0 1749.3 (215)' 4491.0 (2265) 1525.9 (579)' 1701.3 (377)'
Female 3368.3 (604) 3308.1 (945) 3252.6 (672) 2603.0 (407)
Adult 599.0 2240.8 (848) 2738.0 (1366)' 856.6 (238)' 2100.8 (1051)
Juvenile 2694.3 (402) 4128.3 (1033) 2876.7 (518) 2273.8 (294)
A11 birds 599.0 2558.8 (349) 3598.6 (817) 2516.0 (451) 2249.9 (284)

 

' Signiﬁcant diﬁ'erence detected in mean home range between sex or age classes

(Kruskal-Wallis one-way AN OVA, P < 0.10)
b Signiﬁcant difference in mean home range between open and closed sites (Kruskal-

Wallis one-way AN OVA, P < 0.10)

Table 40. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Maltby Hills closed site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Dist. (SB) Dist. (SE) Dist (SE) Dist. (SE) Dist. (SE)
Male 712.0 (9) 3751.0 (958) 3694.4 (1629) 1966.1 (3 72)‘ 1496.5 (279)
Female 915.7 (185) 3831.1 (913) 2992.1 (540) 3173.5 (399) 2674.8 (622)
Adult 888.0 (195) 1645.5 (514)‘I 3624.0 (1417) 874.3 (228)' 1250.1 (568)'I
Juvenile 753.5 (51) 5122.6 (818) 3242.6 (644) 2785.3 (302) 2561.6 (471)
All birds 834.2 (113) 3880.8 (637) 3186.1 (572) 2621.5 (291) 2162.5 (385)

O

' Signiﬁcant difference detected in mean home range between sex or age classes
(Kruskal-Wallis one-way ANOVA, P < 0.10)

79

Table 41. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Pigeon River open site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Dist. (SE) Dist. (SE) Dist (SE) Dist. (SE) Dist. (SE)
Male 1392.3 (169) 1747.0 (521) 951.0 (235) 1810.7 (467) 2507.6 (436)

Female 1402.6 (809) 1606.2 (441) 1022.7 (354) 1169.6 (183) 2824.2 (666)

Adult 1698.4 (758) 1682.1 (461) 655.5 (54) 1286.3 (292) 1826.7 (490)
Juvenile 1022.5 (279) 1595.0 (562) 1064.4 (253) 1585.7 (364) 2910.4 (426)

All birds 1398.0 (432) 1632.4 (322) 990.1 (121) 1426.0 (226) 2639.5 (360)

 

Table 42. Mean dispersal distance (m) of ruffed grouse classiﬁed as Type III dispersers
in the Pigeon River closed site, 1993 - 1997, by sex and age class.

 

 

1993 1994 1995 1996 1997
Dist. (SE) Dist. (SE) Dist (SE) Dist. (SE) Dist. (SE)
Male 778.3 (52)“ 1679.3 (393)“ 1516.2 (603) 1970.5 (392) 866.1 (74)“

Female 1944.5 (50) 1019.2 (154) 1526.6 (288) 1895.4 (449) 2288.0 (641)

Adult 1089.3 (306) 1401.6 (352) 141 1.8 (448) 1504.3 (388) 1040.0 (236)a
Juvenile 1322.5 (573) 1 142.3 (296) 1714.4 (425) 2238.8 (490) 1871.4 (540)

All birds 1167.0 (248) 1315.5 (184) 1523.5 (262)” 1940.5 (289) 1475.5 (311)h

 

' Signiﬁcant difference detected in mean home range between sex or age classes
(Kruskal-Wallis one-way ANOVA, P < 0.10)
b Signiﬁcant difference in mean home range between open and closed sites (Kruskal-

Wallis one-way ANOVA, P < 0.10)

80

Discussion

Studies on the movements of grouse are relatively new, being dependent on
current radio-technology for data collection. This chapter adds to the general knowledge
base with the added dimensions of density-dependent trends and effects of hunting
pressure. Although most movement results obtained in this research were consistent with
previous studies, there were some inconsistencies and unexpected trends observed.

It was assumed that activity ranges and dispersal distances would increase as
density of grouse increased from 1993 — 1997. Linear trends in activity ranges were
either not detected, or detected in a downward direction across time. As ruffed grouse
density increased, home range size (in 3 of 8 cases) showed a signiﬁcant linear decline in
area. The effect of increased density may be a tightening of territories as grouse ﬁll in
space at the fringes of other birds’ ranges. Bergerud and Gratson (1988) put forth the
argument that aggressive phenotypes of ruffed grouse appear more frequently at high
densities. At high grouse numbers, populations may be able to maximize ﬁtness by
aggressively defending small territories that contain necessary habitat components. At
low numbers, grouse may beneﬁt by keeping more loosely deﬁned ranges, without
wasting needed energy on territory defense. Ranges then would not have tight borders,
but birds would use whatever acceptable habitat was available. If a bird is not challenged
as it searches the forest for food, larger activity ranges may result.

Such results may not hold true in all habitats, however. Ruffed grouse densities
(and thus presumably range) seem to remain constant in areas of high quality habitat in
Ontario through the cycle (Theberge and Gauthier 1982). Poor or marginal habitat

(sparsely populated or unoccupied at low density) may ﬁll up with birds as population

81

numbers increase. This deserves to be investigated, as it would drive down the mean
activity range.

Type I birds (non-dispersers that did not move far from point of capture) were
mostly males and mostly adults. This is consistent with the idea that successful adults
will stake claim to an area that previously provided an advertising area with adequate
cover. Males tend to show ﬁdelity towards selected sites and may even display in the fall
(Bergerud and Gratson 1988), many months before the breeding season. They are not
competing for winter food, rather competing to occupy habitat that will be beneﬁcial in
the spring and provides adequate winter habitat (both food and cover). F all activity range
size was comparable to that measured in Missouri studies (Thompson 1987; Neher
1993)

Instead of categorizing birds only as dispersers or non-dispersers, a third category
was described to take into account birds that showed an increase in non—directional
movement in the fall (Type II non-dispersers). Type 11 birds had larger home ranges
than Type I birds by deﬁnition, as they were categorized by the their distance moved
from point of capture. Small et al. (1989) described a transient phase of dispersal in the
fall, when birds were exhibiting increased daily movements. This was followed by a
settlement phase, where the birds’ movements level off and the birds settled into their
winter range. In the case of Type II dispersers, the transient phase occurred as described
by Small et al., but was followed by settlement within the pre-movement area.

The type of non-directional (“shuffle”) movement described may be explained by
the predisposition to disperse without an acceptable corridor for movement. Quantifying

habitat quality within activity ranges may shed some light on why birds would expend the

82

energy required to move, and then settle back where they began. Type II birds were
mostly adults, represented equally by males and females. Where signiﬁcant differences
were detected, females and juveniles had larger ranges than did males and adults (similar
to Type I birds). The importance in understanding such movements lies in the possibility
that the risks incurred from energy expenditure and vulnerability to predators may be
more pronounced than the risks to a bird that stayed in a settled area through the fall and
winter.

In all sites, approximately half of the grouse used in the movement analysis were
classiﬁed as dispersers. In general, more females dispersed and they dispersed longer
distances than did males. Even in the fall, females may be avoiding each other (and
spacing themselves) in their search for suitable spring nesting sites. In the Wisconsin
study (Small and Rusch 1989), 6 of 6 females followed eventually nested in their
wintering areas, whereas males wandered through the spring. These younger males may
want to advertise near other successful males to increase their chance of mating, and thus
are not as concerned with spacing themselves into unoccupied territories. Also, more
juveniles dispersed and they moved longer distances than adults. This was expected as
broods are known to break up in the fall as young birds attempt to ﬁnd acceptable winter
habitat.

Grouse in the Maltby Hills study sites moved farther than those in the Pigeon
River sites. Herein lies a result that deserves further investigation. In the Maltby Hills
open site, 13 birds dispersed more than 5 km, while 21 birds dispersed more than 5 km in
the Maltby Hills closed site. The net distance traveled in the Maltby Hills sites by

juvenile grouse was generally similar to the mean distance recorded in the Wisconsin

83

study. In contrast, only one bird in the Pigeon River area closed site dispersed farther
than 5 km, and none traveled that far in the open site. Dispersal distances for juvenile
grouse in the Pigeon River area were substantially lower than those recorded by Small
and Rusch (1989). Although populations were on an upward climb from 1993 — 1997,
our study sites may never have reached a density so high that birds had to move very long
distances to ﬁnd unoccupied habitat. This hypothesis is supported by the lack of linear
trends detected in dispersal distances across time.

It was predicted that birds would show increased movements in the sites open to
hunting because of increased ﬂushing in high quality habitat. Of the signiﬁcant
differences found in activity ranges between open and closed study sites, the closed site
birds did have smaller ranges. Differences in dispersal distances, however, were not
consistent. Signiﬁcant differences were only found in the Pigeon River sites; in 1995 the
open site had a larger mean dispersal distance, and in 1997 the closed site had a larger
mean. Disturbance by hunting may only have an effect on those birds that are attempting
to stay sedentary through the fall and winter, while ﬂushing birds that are already in the
dispersal phase produces no effect.

An important addition to the above information will be habitat quality of the study
areas and habitat quality of ranges selected by ruffed grouse. The next two chapters will

explore such issues through a detailed analysis of all four study sites and preferred

habitat.

84

CHAPTER 3: Habitat quality and composition

Introduction

Availability and quality of habitat is key in the management of any wildlife
species. Population management tools such as hunting regulations and season length go
hand-in-hand with a comprehensive, long-term habitat management plan. Management
for ruffed grouse has generally taken the shape of aspen clear-cutting, with a goal of
creating coverts of multiple age classes. Although there is evidence that some aspen
stands have grown from seed, reproduction from suckers is the common way in which
stands are regenerated (Graham et al. 1963). In Michigan, forest practices that stimulate
the grth of aspen suckers have been declining in acreage in recent years (Hammill and
Visser 1984; Leatherberry and Spencer 1996). There may be a need to expand the vision
of grouse management to other combinations of forest types and timber management
strategies.

It is generally accepted that aspen is important for grouse in northern Michigan,
providing both food and cover from predators (Gullion 1970; Hammill and Moran 1986).
Before European settlement, though, the forests of Michigan were probably largely
dominated by pines and upland hardwood communities, interrupted at a small scale by
burns and windfalls. Aspen, a shade intolerant species, was probably in short supply.
The landscape was changed drastically in the early 19005: widespread harvesting
followed by wildﬁres led to large, continuous tracts of young regenerating forests across
the state (Hammill and Visser 1984). During the ‘30s and ‘405, wildlife that depended

on such habitat thrived and hunters enjoyed a surplus of game. The ‘50s through the ‘70s

85

saw a continuation of clearcutting practices on a smaller scale to satisfy the demand for
pulpwood.

Because of the suppression of ﬁre, clearcutting has been the primary way in
which succession is set back to create wildlife habitat. The decline in the value of
pulpwood in recent decades has led to fewer clearcuts (Hammill and Visser 1984), and
thus less acreage of regenerating forest. If this trend continues, the amount of traditional
grouse habitat can only decrease.

Existing habitat models weight aspen types heavily, considering it the only source
of winter food (Cade and Sousa 1985; Hammill and Moran 1986). This theory needs
further testing because grouse are known to be omnivores, seeming to vary their food
input by its availability. Bump et al. (1947) recorded that ruffed grouse winter foods in
New York included buds or catkins of cherry (Prunus spp.), aspen (Populus spp.), birch
(Betula spp.), and hop-hornbeam (Ostrya virginiana). Ruffed grouse also exist outside of
the range of aspen in the US, relying on other food sources such as acorns (Korschgen
1966). Although grouse may be closely tied with areas that have mature aspen as a food
source, other food sources may ﬁll in the gaps where mature aspen is missing in the
northern states.

Although the issue of food is debatable, the issue of cover is not. Ruffed grouse
need dense forested stands to escape predators, both avian and terrestrial (Cade and Sousa
1985). If the quality of cover used by grouse is related to their survival, especially

through the hunting season, then it is important to know whether the habitat that is

available (aspen or other forest types) provides the necessary components. An

86

assessment of the quantity of each forest type present, and the value of each as cover is
necessary to make educated management decisions.

Because of the decline of aspen it may beneﬁt managers to consider the
possibility of managing other forest types for quality cover and food. The ruffed grouse
HSI model for Michigan has been tested both by its developers (Hammill and Moran
1986) and by Roloff (1994), and has been shown to predict ruffed grouse density
adequately. This chapter summarizes composition and uses the Michigan model to assess

the quality of various habitat types for ruffed grouse in four study sites.

87

Methods
Habitat quality for ruffed grouse in all study sites was measured in the ﬁeld in

1994 and 1995 (Clark 1996; Gormley 1996) using the HSI model for ruffed grouse in
Michigan (Hammill and Moran 1986). Vegetation variables were measured in 3
randomly placed rectangular plots per stand in 3 — 20 randomly selected stands in 10
different forest types. The forest types were: 3 age classes of aspen, 2 age classes of
pine, upland hardwoods, lowland hardwoods, lowland conifers, jack pine, and oak. There
was no substantial oak component in the Pigeon River area, so it was eliminated from the
analysis there. The size of plots varied by forest type (50 m2 - 250 m2) with sparsely
stocked types requiring larger plot size. Variables measured included equivalent stem
density, height of deciduous trees and shrubs, and low branch height of coniferous trees.
The 3 plots were combined to reﬂect the mean value for each variable for each stand.

For each stand and for each variable a suitability index (SI) score was calculated
on a scale of 0 -— 1 (0 = poor habitat, 1 = high quality habitat) according to the linear
production functions provided by the model (Appendix B). These values were then
combined using the model equations to give an overall suitability value for the stand, and
stands of like forest type were averaged. I will refer to the resulting mean SI as SI-ﬁeld.

Classiﬁed LANDSAT images of the areas were obtained from the University of
Michigan. The images were taken in 1993 and have a 30 m resolution. Maps were
reclassiﬁed in ARC/GRID to correspond to the ten forest types that were sampled for
quality as ruffed grouse habitat. Because of the drastic change caused by clearcutting and
the assumed importance of aspen to ruffed grouse, it was necessary to alter the images for

each year of the study. Maps of each aspen clearcut in each year were obtained ﬁ'om

88

MDNR and United States Forest Service (U SFS) ﬁles. The maps were digitized and
overlaid onto the images in ARC/INFO, creating a more accurate image for each site in
each of 5 years. Using a topographic map, UTM coordinates for the study area borders
were estimated and polygon coverages were created. Borders were cut from each forest
type grid, and the amount of each forest type in each area was summarized.

Once the appropriate maps were created, each pixel was assigned the mean SI-
ﬁeld value for the forest type in that pixel. Then the spatial aspect of the model, distance
to an aspen food source, was added. A distance value was calculated for each pixel in
each area by measuring the distance from that pixel to a pixel classiﬁed as mature aspen
(30+ years). The distance was then translated to a value between 0 and 1 using the
equation of the line forming the production function in the model (Appendix B). I will
refer to this variable as SI-food.

The overall value for each pixel was modiﬁed by multiplying the SI-ﬁeld value
by the SI-food value, as called for in the model. This produced a habitat suitability map,
one of which was created for each site for the ﬁrst year of the study, 1993. This process
was repeated for each year if the availability of forest types was different than the
previous year due to clearcutting of aspen. The distribution of forest types in 1993 was
compared with the distribution of forest types in 1997, using a Chi-square test with
signiﬁcance set at 0.10.

To compare habitat quality between areas, an Arc Macro Language (AML)
program was written to randomly sample the sites for quality. Circles with a radius of
500 m (area = 53.5 ha, similar in size to a grouse activity range) were placed on the

habitat suitability map at random UTM coordinates, and a mean HSI value was calculated

89

for that circle’s area. This was repeated at 200 sets of random UTM coordinates, creating
a set of “possible available ranges”. A set of available ranges was generated for each
study site for 1993 and 1997 (if changes in classiﬁcation took place by the addition of
clearcut areas), using the same random coordinates. A distribution of the ﬁequency of
possible available ranges was created by assigning each circle to an HSI value bin.
Frequencies were compared between sites and years using a Chi-square test with

signiﬁcance set at 0.10.

90

Results
Habitat composition

Upland hardwoods covered the largest area of the Maltby Hills open site, 23%
(Table 43). Pine and aspen over 30 years old were also well represented, with 16 and
17% of the area, respectively. Medium aged (11 — 29 years old) aspen made up 13% of
the site, but young aspen only made up 3% of the site (the lowest percentage of young
aspen of all sites). The category “undeﬁned” represents those areas with water or
openings as landcover and those not classiﬁed because of cloud cover or technical
difﬁculties, and represented less than 2% of the grid. There were no clearcuts in the open
site between 1993 and 1997.

The Maltby Hills closed site was also dominated by upland hardwoods, which
comprised 32% of the site (Table 44). Old and medium aged aspen each made up 15% of
the site, while young aspen made up 6%. There was less older pine (5%) than in the open
site. Small aspen stands were clearcut each year (totaling 136 ha), and acreages of old
aspen, medium aspen, and upland hardwoods decreased slightly because these cuts. Two
percent of the area was undeﬁned.

The Pigeon River open site was dominated by lowland hardwoods (19%), upland
hardwoods (20%), and older pine (19%) (Table 45). Aspen covered 17% with all age
classes, evenly distributed with 5 — 6% in each class. Clearcuts occurred in 1994, 1995,
and 1996, which increased the acreage classiﬁed as young aspen by 244 ha These cuts
led to a loss of old and medium aged aspen, and small losses of upland hardwoods and

pine. Undeﬁned areas made up 5% of this site, slightly higher than the other sites.

91

Lowland conifers (19%), pine over 30 years (18%), and lowland hardwoods
(17%) dominated the Pigeon River closed site (Table 46). Compared to the open site,
there was slightly more aspen in the closed site: young aspen comprised 8% of the site,
medium aspen 5%, and old aspen 9%. In 1994 and 1997 there were clearcuts in the
closed Pigeon River site, creating 130 ha of new aspen. The cuts led to a small decrease

of old and medium aged aspen. Less than 2% of the areas was classiﬁed as undeﬁned.

92

Table 43. Area and percentage of each forest type in the Maltby Hills open site, 1993 —

 

 

1997.

Area %
Forest type (ha)
Undeﬁned 133 1 .6
AspenO- 10 years 217 2.6
Aspen ll -29 years 1126 13.4
Aspen 30+years 1418 16.9
Upland hardwoods 1945 23.2
Lowland conifers 775 9.2
Pine 0 - 29 years 388 4.6
Pine 30 + years 1352 16.1
Jack pine 411 4.9
Oak 620 7.4

 

Table 44. Area and percentage of each forest type in the Maltby Hills closed site, 1993 —

 

 

 

1997.
1 993 1994 1995 1996 1997

Area % Area % Area % Area % Area %
Forest type (ha) (ha) (ha) (ha) (ha)
Undeﬁned 234 2.3 234 2.3 234 2.3 233 2.3 233 2.3
Aspen 0 — 10 years 599 6.0 659 6.6 702 7.0 705 7.0 735 7.3
Aspen 11 —29 years 1516 15.1 1507 15.1 1493 14.9 1493 14.9 1485 14.8
Aspen 30+years 1535 15.3 1531 15.3 1518 15.2 1517 15.2 1506 15.1
Upland hardwoods 3194 31.9 3189 31.9 3179 31.8 3178 31.8 3168 31.7
Lowland conifers 105 1.0 103 1.0 103 1.0 103 1.0 103 1.0
Pine 0 — 29 years 504 5.0 493 4.9 492 4.9 492 4.9 492 4.9
Pine 30 + years 512 5.1 500 5.0 499 5.0 499 5.0 499 5.0
Jack pine 923 9.2 911 9.1 911 9.1 911 9.1 911 9.1
Oak 885 8.8 878 8.8 876 8.8 876 8.8 875 8.7

 

93

Table 45. Area and percentage of each forest type in the Pigeon River open site, 1993 —
1997.

 

 

 

1993 1994 1995 1996-97
Area % Area % Area % Area %
Forest type (ha) (ha) (ha) (ha)
Undeﬁned 534 4.9 531 4.9 530 4.8 530 4.8
Aspen 0 — 10 years 657 6.0 802 7.3 901 8.2 901 8.2
Aspen 11 - 29 years 651 6.0 646 5.9 643 5.9 643 5.9
Aspen 30 + years 562 5.1 555 5.1 551 5.0 551 5.0
Upland hardwoods 2173 19.9 2159 19.7 2157 19.7 2157 19.7
Lowland hardwoods 2113 19.3 2068 18.9 2015 18.4 2015 18.4
Lowland conifers 1478 13.5 1455 13.3 1443 13.2 1443 13.2
Pine0—29years 185 1.7 181 1.7 180 1.6 180 1.6
Pine 30 + years 2028 18.5 1997 18.2 1975 18.0 1975 18.0
Jack pine 560 5.1 551 5.0 549 5.0 549 5.0

 

Table 46. Area and percentage of each forest type in the Pigeon River closed site, 1993 -—
1997.

 

 

 

1 993 1 994-96 1 997
Area % Area % Area %

Forest type (ha) (ha) (ha)

Undeﬁned 139 1.6 129 1.4 129 1.4
Aspen 0 — 10 years 686 7.7 762 8.5 816 9.1
Aspen 1 1 — 29 years 424 4.8 408 4.6 407 4.6
Aspen 30 + years 756 8.5 739 8.3 730 8.2
Upland hardwoods 1235 13.8 1234 13.8 1218 13.7
Lowland hardwoods 1483 16.6 1469 16.5 1444 16.2
Lowland conifers 1714 19.2 171 1 19.2 171 1 19.2
Pine 0 — 29 years 436 4.9 428 4.8 427 4.8
Pine 30 + years 1586 17.8 1582 17.7 1580 17.7
Jack pine 463 5.2 460 5.2 460 5.2

 

94

Habitat quality — ﬁeld variables

In Maltby Hills open site, medium-aged aspen (1 1-29 years) had the highest mean
SI-ﬁeld value (0.62), followed by lowland conifers (0.39) (T able 47). Young pine (0.25)
and young aspen (0.22) were of moderate quality. Older aspen, older pine and oak both
all ranked low in quality, with SI-ﬁeld values less than 0.10.

Medium-aged aspen (mean SI-ﬁeld value = 0.50) also ranked highest in the
closed Maltby Hills site (Table 47), but its value was lower than in the open site.
Lowland conifers, jack pine, and young pine all had high SI-ﬁeld values (0.31 - 0.41).

As in the open site, older aspen and oak ranked low, but in addition, upland hardwoods
ranked much lower than in the open site (0.06 vs 0.19). The young aspen category also
ranked lower than the open site.

In the Pigeon River area, values also showed some consistencies between the
open and closed sites. In the open site, lowland conifers ranked highest with a mean
value of 0.6 (Table 48). Medium aspen was second, with a mean value of 0.52. Unlike in
the Maltby sites, older aged aspen also ranked high (0.49). Upland hardwoods had the
lowest SI-ﬁeld value, 0.07.

In the closed Pigeon River site, no forest types scored above 0.50 (Table 48). The
highest HSI value was 0.47 for lowland hardwoods, and lowland conifers and young pine
also had high values (0.42, 0.43, respectively). Older aspen had medium quality, 0.16.

As in the open site, upland hardwoods had the lowest SI-ﬁeld value at 0.03.

95

Table 47. Sample size, mean, and standard error of SI-ﬁeld values (HSI calculated
without distance to aspen variable) in the open and closed sites in Maltby Hills.

 

 

 

Open site Closed site

Forest type N Mean SE N Mean SE

Aspen 0 — 10 years 10 0.22 0.10 16 0.14 0.06
Aspen 11 — 29 years 18 0.62 0.08 21 0.50 0.08
Aspen 30 + years 13 0.06 0.03 20 0.07 0.05
Upland hardwoods 9 0.19 0.07 10 0.06 0.05
Lowland hardwoods 4 0.07 0.05 3 0.22 0.17
Lowland conifers 5 0.39 0.23 7 0.41 0.11
Pine 0 — 29 years 7 0.25 0.12 8 0.31 0.11
Pine 30 + years 8 0.03 0.01 7 0.10 0.09
Jack pine 13 0.19 0.10 20 0.34 0.10
Oak 8 0.05 0.04 8 0.07 0.06

 

Table 48. Sample size, mean, and standard error of SI-ﬁeld values (HSI calculated
without distance to aspen variable) in the Open and closed sites in the Pigeon River
Country State Forest.

 

 

 

Open site Closed site

Forest type N Mean SE N Mean SE

Aspen 0 — 10 years 19 0.19 0.07 16 0.20 0.06
Aspen ll — 29 years 18 0.52 0.08 22 0.33 0.07
Aspen 30 + years 18 0.49 0.09 16 0.16 0.07
Upland hardwoods 8 0.07 0.04 10 0.03 0.02
Lowland hardwoods 4 0.25 0.13 4 0.47 0.23
Lowland conifers 4 0.6 0.22 8 0.42 0.13
Pine 0 - 29 years 2 0.12 0.12 8 0.43 0.15
Pine 30 + years 10 0.19 0.06 9 0.19 0.11
Jack pine 7 0.28 0.11 10 0.31 0.13

 

96

Habitat quality — random circles

In the Maltby Hills open site, the center of the distribution of mean HSI values
within 200 random circles was 0.2 — 0.24 (Figure 22). The distribution was skewed left
with a strong peak, and there were no values over 0.36. The mean value was 0.182, with
a standard error of 0.004.

The centers of the distributions in the closed site in 1993 and 1997 were not as
obvious (Figure 22). These distributions had 2 peaks, one at 0 and one at 0.12. The
mean HSI value in 1993 was 0.099, with a standard error of 0.005. There were no circles
with HSI values over 0.28. Ten of the 200 circles were categorized into different bins in
1997 due to clearcuts, but the distribution in 1997 did not differ signiﬁcantly from the
distribution in 1993 (P = 0.997). The closed site distribution in 1993 did differ
signiﬁcantly from the open site distribution (P < 0.001).

The peak at 0 in the closed site made the distribution incomparable in shape to the
open site. The open and closed sites were therefore compared again to see whether the
distribution of mean values outside of the low quality area differed from the open site.
The ﬁrst bin (0 — 0.04) was removed ﬁ'om the analysis and open and closed sites and the
Chi-square test was redone. There was still a signiﬁcant difference detected (P < 0.001).
The Pigeon River open site in 1993 had no 0 HSI values, and peaked at 0.16 (Figure 23).
This distribution was skewed slightly right and was ﬂatter than the Maltby Hills
histograms. As in the Maltby Hills open sites, no circles had a mean value over 0.36.
The mean HSI value for 1993 was 0.153, and the standard error was 0.005. There was no

signiﬁcant difference due to clearcutting between 1993 and 1997 (P = 0.999).

97

The distribution for the Pigeon River closed site had even less of a peak than the
open site, but maximum frequencies occurred at 0.20 and 0.28 (Figure 23). The mean
HSI score was 0.160, and the standard error was 0.006. Clearcuts caused no signiﬁcant
differences between 1993 and 1997 (P = 0.999). The closed site distribution was skewed

to the left, and differed signiﬁcantly from the open site distribution (P = 0.039).

98

 

Maltby open

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36 0.4
HSI value

 

Maltby Hills closed

 

 

 

 

 

 

 

 

 

 

 

 

 

01 ﬁ| . p e; r. T I T .
O 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36 0.4

HSI value

.1993

.1993
.1997

Figure 22. Frequency of occurrence of HSI values in the Maltby Hills open site, 1993,

and the Maltby Hills closed site, 1993 and 1997.

99

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PigeonRiveropen
50
.1993
.1997
0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36 0.4
HSI value
60
PigeonRiverclosed
50 _._. _ _ __. mun.-- _.
3 4° "
t:
8 30 - I1993
g .1997
u. 20 ..
10« — - e
()1

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36 0.4
HSI value

Figure 23. Frequency of occurrence of HSI values in Pigeon River sites, 1993 and 1997.

100

Discussion

It appears that management for aspen in northern Michigan is not of high priority,
with only 510 ha (or 1.3%) of the study areas cut between 1993 and 1997. Such a small
amount suggests that the priority of cutting follows the trends of most northern states. In
fact, no clearcuts took place in the Maltby Hills open site during the study. Upland
hardwood and mature pine together represented 30 — 50% of the total area of the sites.
Managers need to acknowledge that if this continues, the early successional species
important for wildlife habitat will diminish to a very small percentage of public lands.
Such information makes an analysis of quality even more important, because managers
may need to consider alternatives to traditional grouse management.

As many HSI models do, the ruffed grouse model evaluates habitat on a scale
from 0 — 1, with 0 being poor habitat and 1 being optimum habitat. Habitat suitability
values (SI-ﬁeld) were generally low, ranging from only 0.03 — 0.62 in the Pigeon and
0.07 to 0.52 in the Maltby Hills. These values were decreased again during the spatial
analysis, when they were modiﬁed by the model requirement of aspen within 300 meters.
Because of the structure of the model, any large areas without mature aspen will have low
HSI values. The model requires that the SI value for mature aspen proximity be
multiplied by the SI-ﬁeld value. The result of this mathematical set-up is that if an area is
far from aspen, and has an SI-food value of 0, then the area gets an HSI value of 0.

In the Pigeon River study sites, continuous high quality habitat was interrupted by
areas of low (or 0) value (Appendix C, Figures C 1 and C 2). In the open site, upland
hardwoods comprised much of the low quality areas, with no mature aspen interspersed.

The low quality areas in the closed site consisted of upland hardwoods, pines and

101

lowland conifers. The Maltby Hills open site is much different ﬁom the Pigeon River,
with almost continuous coverage of medium quality habitat (Appendix C, Figure C 3)
and mature aspen scattered evenly throughout. The closed site had only a narrow band of
medium to high quality habitat (Appendix C, Figure C 4), the only area where mature
aspen was present. In all study sites, either much of the area was low quality, or the
heavy weighting of aspen was inappropriate. This weighting of aspen as food is
understandable in northern forests, but it is unknown what other food sources may be
substituted. This question will be addressed in the next chapter, when I test whether
grouse actually favor areas that score high according to this model.

Shifting focus to the SI-ﬁeld variable, the cover requirements for grouse were best
met by medium aged aspen and lowland conifers. The fact that 11-29 year old aspen
topped the list was consistent with traditional management, which calls for such habitat in
quantity. This result may also reﬂect the fact that the model was developed with
medium-aged aspen as the ideal, high quality habitat. In all sites, lowland conifers also
topped the list with high SI-ﬁeld values, and had some of the highest equivalent stem
densities recorded. The lowland conifer types (especially cedar) are recognized to be
important to other species for thermal cover, sheltering understories from wind and snow
(Blouch 1984).

This leads into the debate of whether conifers are of value as ruffed grouse cover.
For many years, grouse management followed the tenet that conifer stands were sinks for
ruffed grouse, providing better cover for predators. Gullion (1972) acknowledged that
some conifers as understory were beneﬁcial, but that stands should be avoided. Some

models suggest a penalty for conifers (Cade and Sousa 1985), decreasing the value of

102

deciduous stands with conifer components. The HSI model I used, however, weights
conifers as important cover for ruffed grouse, and weights them all equally. In fact, one
conifer stem is worth 4 times one deciduous stem. This structure caused lowland conifers
to be high in SI-ﬁeld values, even though the lowest branch height may have been high
enough to provide roosting sites for raptors. A valuable addition to the model may be the
differential weighting of conifer by type. Lowland conifers such as white cedar have a
much different growth structure than red or white pine, and thus may affect grouse
survival differently.

The Maltby Hills sites generally had less conifer cover than the Pigeon River,
though one contiguous jack pine stand comprised a large portion of the closed Maltby
Hills site. The jack pine, while adequate in cover, contributed greatly to the low HSI
values in the closed site distribution. It had an overall HSI value of 0 because of its
distance from aspen. The Maltby sites are also lacking a conifer component in the aspen
understory, leading to lower SI-ﬁeld values for aspen over 30 years of age. In both areas,
the number of conifers in the understory made the difference between areas of high
quality cover and low quality cover.

If management for grouse is a priority, then the long-term presence of adequate
habitat must be addressed and planned for. The steady loss of aspen types will continue
to be a problem and other management tools must be considered. A Wisconsin study
calls for more research into the improvement of pine plantations for grouse habitat by
increasing stem densities in the understory (McCaffery et al. 1996). In the following
chapters, I will address the value of such areas for grouse in our study, and assess

whether such management strategies are viable.

103

CHAPTER 4: Habitat use and preference

Introduction

In the previous chapter, habitat composition and quality were measured for the
study sites in the Pigeon River and Maltby Hills areas. To use such information to
improve management for grouse, we need to address two additional questions. To fully
understand how to manage the composition of habitat in our sites, we need to know
whether birds exhibit patterns of preference and avoidance of forest types. Additionally
we need to verify/refute our use of the model by testing whether birds are settling in areas
that are considered high quality.

A forest type is considered preferred if it is used in higher proportion than its
availability, and avoided if it is used in lower proportion than its availability. Past
banding and ﬂushing studies combined with anecdotal evidence have built some
generalizations about grouse habitat. Aspen is considered important for ruffed grouse in
the northern climates (Bump et a1. 1947; Cade and Sousa 1983), and is expected to be
highly selected for. In addition, upland hardwoods and older pines provide little cover
for grouse (Gullion 1972; 1982), and are expected to be avoided. Other types that may
fulfill the habitat requirements for grouse (or conversely, be sinks) have not been
extensively studied. If the goal of management is to provide the best habitat possible, it
is important to know what forest types (including which age classes of aspen) are used
more or less than available.

In one of the most extensive telemetry studies done on ruffed grouse, radio-

collared birds in Wisconsin selected lowland conifers in summer and fall, and open oak

104

in winter (Small and Rusch 1989). In all seasons, birds avoided older pine stands. These
grouse were probably selecting areas that had suitable winter cover in addition to a food
source. The fact that they were preferring oak may not have been predicted, because
existing models (and popular opinion) emphasize the need for aspen for food. Ruffed
grouse, as stated in Chapter 3, are known to be opportunists that may choose many
different foods (including acorns) depending on their availability (Bump et al. 1947).
Habitat selection in the fall and early winter may be driven by factors other than those
accounted for in the models, such as predator density, winter food other than aspen, or

disturbance by hunters.

The habitat suitability index (HSI) model by Hammill and Moran (1986) outlines
the speciﬁc requirements for food and cover for ruffed grouse in Michigan. It takes the
classic view that the two major components needed to sustain grouse are a high stem
density and the presence of mature aspen. One way to test the model is to see if birds are
selecting those forest types that ﬁt the model’s description of good habitat. If a forest
type is shown to be high quality by the model but birds are avoiding it, then the model
may be missing a component that detracts from its quality for grouse.

The next step is to test whether the model is an effective management tool when
used at the landscape scale. In Chapter 3, I assessed and compared the habitat suitability
(according to the model) of the study areas used in the Michigan Ruffed Grouse Project.
I took a model that was designed to analyze habitat on a stand by stand basis and applied
it to large study sites using a geographic information system. I was able to separately
calculate the HSI value of individual 30x30 m pixels by generalizing the cover variables

and measuring the spatial variable (distance to mature aspen) over a very large scale.

105

Before I recommend this type of use of the model it is important to carry out a spatial
analysis to test whether ruffed grouse are using the areas that rank high according to the
model. If so, then I can recommend the use of the model as a management guide for
landscape planning.

In this chapter, I will use GIS to intersect ruffed grouse activity ranges obtained
over the course of the Michigan Ruffed Grouse Project (1993 — 1997) with the habitat
and HSI grids created for Chapter 3. In this manner I will address the questions of habitat
preference/avoidance and test the landscape level use of the Ruffed Grouse Suitability

Model for Michigan (Hammill and Moran 1986).

106

 

Methods

Grouse locations through fall and early winter were used to quantify habitat use
and preference. If a bird was located on at least 10 different days, it was used in the
analysis of habitat use and preference. Habitat use was determined for each non-
disperser (Types I and II) by intersecting ruffed grouse activity ranges (95% kernel
contours in UTM coordinates) with forest cover maps in ARC/GRID. After the
intersections were completed, birds were omitted from the analysis if their ranges
contained a majority of unclassiﬁed pixels, or if their ranges went outside the borders of
the study site. In each study area, 9 forest types were differentiated: 3 age classes of
aspen and 2 age classes of pine, upland hardwoods, lowland hardwoods (absent in the
Maltby Hills sites), lowland conifers, jack pine, and oak (absent in the Pigeon River
sites). Results were summarized by number of pixels in each forest type using an AML
program, and exported into a text ﬁle for analysis.

Habitat availability was measured for each bird by surrounding activity areas
(95% kernel contours) by a 500 m buffer and generating new polygons in ARC/INFO.
Based on previous research (Clark 1996), the 500 m buffer was considered a reasonable
estimation of the possible movement of each bird toward acceptable habitat. These
polygons were then intersected with forest cover maps in ARC/GRID, and the number of
pixels of each forest type within the “available polygon” was exported to a text ﬁle.

Habitat use was compared to availability in each site to determine preference or
avoidance using the population of collared birds in each site (years combined) using the
program PREFER (Great Lakes Fishery Laboratory, U. S. Fish and Wildlife Service, Ann

Arbor, Michigan). This program uses a ranking procedure that compares habitat use of

107

each individual to the habitat available to that individual. A value is then produced for
each forest type, corresponding to preferred, neutral, or avoided. Within this program,
the Waller-Duncan procedure is used to test for signiﬁcant differences between the
preference of forest types, with signiﬁcance set at 0.10.

An additional analysis was performed to discern whether individuals chose habitat
that ranked higher than average by the grouse HSI model. A weighted HSI value was
calculated for each activity area for non-dispersers by intersecting the activity area with
HSI grids in ARC/GRID. For dispersers, there were not enough points to calculate an
activity range after dispersal ended. An activity range was therefore estimated by
creating a 500 m radius circle centered around the harmonic center of the week after
dispersal ended. The distribution of HSI values in the areas used was compared to the
distribution of values within the random circles generated in the habitat quality analysis
discussed in Chapter 3. Comparisons were made using a Chi-square test with

signiﬁcance set at 0.10.

108

Results
Forest 0pc preference

In the Maltby Hills open site, 61 non-dispersers (Type I and Type II) were used in
the analysis of preference. Young aspen, young pine, and oak were chosen by ruffed
grouse in higher proportion than their availability (Figure 24a). Although 11 - 29 year
old aspen rated high in cover requirements (0.62, Chapter 3), it was not strongly selected
for. Aspen over 30 years, jack pine, and lowland conifers were also used in proportion to
their availability. Upland hardwoods and pine over 30 years were avoided.

The non-dispersing grouse in the Maltby Hills closed site (N = 94) selected
strongly for lowland conifers, using them signiﬁcantly more than their availability
(Figure 24b). As in the open site, young pine was selected for, but not as strongly.
Again, 11 — 29 year old aspen rated highest in cover requirements, but was not selected
for. All three age classes of aspen, oak and jack pine were used in proportion to their
availability. As in the open site, pine > 30 years and upland hardwoods were avoided.

In the Pigeon River open site, 61 non-dispersers were used in the analysis. These
grouse preferred aspen ll - 29 years old, jack pine, and aspen over 30 years old (Figure
25a). Older aspen ranked high (0.49) in cover suitability value in this site when
compared to the closed site (0.20). Young aspen and young pine, lowland conifers, old
pine, and lowland hardwoods were used in proportion to their availability, and upland
hardwoods were strongly avoided.

The 77 grouse used in the Pigeon River closed site analysis also preferred aspen
11 — 29 years and jack pine, but they also selected for young pine. Young and old aspen

and pine were not selected for nor avoided. Lowland conifers, lowland hardwoods, and

109

upland hardwoods were avoided, although both lowland types ranked high in cover

suitability.

110

.ooaouomou 8m 8527. 93 Kb 528 :08 8m
moietg “goo Sm 808 “mm 302 .3852 35:35. c8253 2: .o A m dun—cocoa 83388 03:38 gouache—33v acouobmv

we: we? 3:232“ «853 8:80 .mozm 388 98 none .53”— coowE 05 E 8:on vow?— uom ooeouomoa «83a: =am .mm Esmﬁ

 

 

 

 

M52 wcmmmouooa
A
86 23 Q3 a H .o o _ .o end 23 _md mmd
30032: mamazoo i + om i + on i S I o .3. am - e SE a a u :
mQOOBgm 9724304 9793304 mam ZmEm< 2mmm< mZE MUS. ZmEm<
8mm tome—o 83M aoowE 3
Ed mad Ed cod .26 2d 36 wad de

 

 

 

 

800354.: a + an mamazou a» a - c a S I o a + on mza

a a .. :
800392: 972.53 mza @2333 BE 22? 2mm?

MU<~ mem<

36 zone 53¢ :oomE Q

111

.3e20m8 com 56% 8a 25 “no.8.“ some 8m

833:3 “260 .8.“ 88m 5: :82 .3833 3513:: 50253 Sﬁo A m .2358:— c8maa88 033:8 gouache—33v avenge
8a ma? 3583a “53¢: 3:95 .mozm come—o Ea come m5: 3:“: 05 E 859% cube 8% ooeﬁomoa “SEE zam .vm v.53..—

 

 

 

 

 

 

 

xqﬁ mammogooa
A
8.0 25 So «no So one Ed _2 :3
.3 + 2 a + om mza a a .. : an 2 n o i a - o 3328
883% mza 2%? mos 55 2mm? 2%? BE az<§3
as 883 3%: 3
85 26 one 25 3o 86 85 NS «S
a + 2 95.328 mza .: + on a a I : .s a - o i 2 n o
mza 883% 2433 53 2mm? 2mm? m<o sz 2mm?

8% some >832 Am

112

Habitat quality preference — non-dispersers

In the Maltby Hills open site, the distribution of HSI values within activity ranges
(Figure 26) differed signiﬁcantly from the distribution of HSI values within random
circles (Chi-square, P < 0.01). The most extreme difference was within the HSI range of
0.28 — 0.32, which is the peak of the distribution of activity ranges. Thirteen percent of
the random circles (available ranges) fell within this category, while 31% of the activity
ranges fell within this category (Table 49). Nearly 10% of activity ranges fell into the
range of 0.32 — 0.36 (the highest measured HSI range). while only 2% of random circles
did.

The habitat quality difference between activity ranges used and random circles in
the Maltby Hills closed site was even more extreme (Figure 27, P < 0.01). The peak of
the HSI distribution for the birds used in the analysis was between 0.20 and 0.24. Over
38% of all activity range HSI values fell into this range (Table 50). The two peaks of the
distribution for random circles were at 0.00 — 0.04 and 0.12 — 0.16. The distributions
were also tested with the elimination of the 0.00 — 0.04 category to see if avoidance of the
extensive jack pine area was driving the difference. In this case, a signiﬁcant difference
was still detected (P < 0.01).

In the Pigeon River open site, the distribution of HSI values in the random circles
differed signiﬁcantly from the distribution of HSI values in the activity ranges of the
birds used in the analysis (Figure 28, P < 0.01). The distribution of the random circles is
ﬂat, while the distribution of the activity ranges shows a clear peak at 0.20 — 0.24. A
higher percentage of the random circles (6%) fell into the 0.32 — 0.36 category (Table

51), but a higher percentage also fell into the 3 lowest categories between 0.04 and 0.16.

113

The activity range distribution in the Pigeon River closed site had two peaks, at
ranges 0.12 - 0.16 and 0.28 — 0.32 (Figure 29). This differed signiﬁcantly ﬁ'om the
random circles distribution (Chi-square, P = 0.018). which did not have a prominent
peak. Over 40% of the activity ranges had HSI values between 0.24 and 0.32, relatively

high quality (Table 52). Only 30% of the random circles fell between these values, while

8% fell below 0.08.

114

 

 

 

 

 

 

 

4o
30
g 20 lAvailable
m IUsed
10
O .J

 

   

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36
HSI value

Figure 26. Percent of HSI values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Maltby Hills open site.

Table 49. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of non-dispersing ruffed grouse in the Maltby Hills open
site.a

 

Available Available Used Used

 

HSI range Freq. % Freq. %
0.00 - 0.04 2 1.00 l 1.64
0.04 - 0.08 7 3.50 5 8.20
0.08 - 0.12 10 5.00 O 0.00
0.12-0.16 8 4.00 1 1.64
0.16 - 0.20 29 14.50 2 3.28
0.20 - 0.24 57 28.50 11 18.03
0.24 - 0.28 57 28.50 16 26.23
0.28 - 0.32 26 13.00 19 31.15
0.32 - 0.36 4 2.00 6 9.84
0.36 - 0.40 0 0.00 0 0.00

 

TDistribution of HSI values within random circles differed signiﬁcantly from distribution

of HSI values within activity areas (Chi-square, P < 0.01)

115

 

 

 

 

 

 

4O
30
a .
{3: 20 _ IAvarlable
d: IUsed
10 -
o 1

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36
HSI value

Figure 27. Percent of HSI values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Maltby Hills closed site.

Table 50. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of non-dispersing ruffed grouse in the Maltby Hills closed
site.‘

 

Available Available Used Used

 

HSI range Freq. % Freq. %
0.00 - 0.04 41 20.50 1 1.06
0.04 - 0.08 5 2.50 3 3.19
0.08 - 0.12 21 10.50 1 1.06
0.12 - 0.16 57 28.50 14 14.89
0.16 - 0.20 35 17.50 26 27.66
0.20 - 0.24 35 17.50 36 38.30
0.24 - 0.28 6 3.00 10 10.64
0.28 - 0.32 O 0.00 3 3.19
0.32 - 0.36 0 0.00 O 0.00
0.36 - 0.40 0 0.00 0 0.00

 

1rBistribution of HSI values within random circles differed signiﬁcantly from distribution
of HSI values within activity areas (Chi-square, P < 0.01)

116

 

 

 

 

 

 

40
30 _.- ._ __
E Avaihbb
§ 20 -..._--_.__w---.s '
a: .USCd
10
O ‘ 17

 

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36 0.4
HSIvalue

Figure 28. Percent of HS] values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Pigeon River open site.

Table 51. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of non—dispersing ruffed grouse in the Pigeon River open
site.a

 

Available Available Used Used

 

HSI range Freq. % Freq. %
0.00 - 0.04 O 0.00 O 0.00
0.04 - 0.08 16 8.00 1 1.64
0.08 - 0.12 22 11.00 2 3.28
0.12 - 0.16 29 14.50 3 4.92
0.16 - 0.20 38 19.00 14 22.95
0.20 - 0.24 35 17.50 18 29.51
0.24 - 0.28 36 18.00 10 16.39
0.28 - 0.32 12 6.00 12 19.67
0.32 - 0.36 12 6.00 1 1.64
0.36 - 0.40 0 0.00 0 0.00

 

1|-ff)istribution of HSI values within random circles differed signiﬁcantly from distribution
of HSI values within activity areas (Chi-square, P < 0.01)

117

 

 

 

Percent

 

 

 

3O
20
I Avaihbb
- Used
10
o 1 ,

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36
HSIvahe

Figure 29. Percent of HSI values within random circles (available) and within activity
ranges (used) of non-dispersing ruffed grouse in the Pigeon River closed site.

Table 52. Frequency and percent of HSI values within random circles (available) and

within activity ranges (used) of non-dispersing ruffed grouse in the Pigeon River closed
- a

 

 

Slte.
Available Available Used Used
HSI range Freq. % Freq. %

0.00 - 0.04 1 0.50 0 0.00
0.04 - 0.08 15 7.50 O 0.00
0.08 - 0.12 24 12.00 8 10.39
0.12-0.16 25 12.50 14 18.18
0.16 - 0.20 30 15.00 10 12.99
0.20 - 0.24 37 18.50 5 6.49
0.24 - 0.28 29 14.50 16 20.78
0.28 - 0.32 33 16.50 18 23.38
0.32 - 0.36 6 3.00 6 7.79
0.36 - 0.40 0 0.00 0 0.00

 

1rDistribution of H81 values within random circles differed signiﬁcantly from distribution
of HSI values within activity areas (Chi—square, P = 0.018)

118

Habitat quality preference - dispersers

The distribution of HSI values within random circles did not differ ﬁom the
distribution of HSI values within estimated activity ranges of dispersers (N = 87) in the
Maltby Hills open site (Figure 30). Peaks in the histograms were similar, near the 0.20 —
0.24 category, with around 60% of activity areas and random circles falling within that

range (Table 53). Both histograms were skewed to the left.
1 In the Maltby Hills closed site, a signiﬁcant difference was detected between the
HSI values of the activity areas of the 89 grouse used in the analysis and the HSI values
of the random circles (Figure 31). The peak of the activity range distribution was 0.16 —
0.20. Although this peak occurred at lower quality than the open site ranges, it was
higher than the peak of the random circle distribution (0.12 — 0.16). While almost 54%
of activity areas had values over 0.16, only 38% of random circles had values that high
(Table 54).

Of the activity ranges of the 57 dispersers used in the analysis in the Pigeon River
open site, 58% had HSI values between 0.20 and 0.24 (Table 55). This shows up as a
clear peak in the activity range distribution, whereas the random circle distribution has
no peak (Figure 32). A signiﬁcant difference in distribution was detected (Chi-square, P
< 0.01).

The random circles and activityranges for dispersers in the Pigeon River closed
site did not differ signiﬁcantly. Both histograms have a shallow peak between 0.20 and
0.28 (Figure 33). Three percent of the random circles, however, fell in the range of 0.32

— 0.36, while none of the activity ranges did (Table 56).

119

 

 

I Available
I Used

 

 

Percent

 

 

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36

HSI value

Figure 30. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Maltby Hills open site.

Table 53. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Maltby Hills open site.‘I

 

Available Available Used Used

 

HSI range Freq. % Freq. %
0.00 - 0.04 2 1.00 1 1.15
0.04 - 0.08 7 3.50 8 9.20
0.08 - 0.12 10 5.00 4 4.60
0.12-0.16 8 4.00 1 1.15
0.16 - 0.20 29 14.50 6 6.90
0.20 - 0.24 57 28.50 33 37.93
0.24 - 0.28 57 28.50 22 25.29
0.28 - 0.32 26 13.00 12 13.79
0.32 - 0.36 4 2.00 O 0.00
0.36 - 0.40 0 0.00 0 0.00

 

Tﬁistribution of HSI values within random circles did not differ signiﬁcantly from
distribution of HSI values within activity areas (Chi-square, P = 0.158)

120

Percent

 

 

 

I Available
I Used

 

 

 

 

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36
HSI value

Figure 31. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Maltby Hills closed site.

Table 54. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Maltby Hills closed site.II

 

Available Available Used Used

 

HSI range Freq. % Freq. %
0.00 - 0.04 41 20.50 5 5.62
0.04 - 0.08 5 2.50 3 3.37
0.08 - 0.12 21 10.50 5 5.62
0.12 - 0.16 57 28.50 28 31.46
0.16 - 0.20 35 17.50 32 35.96
0.20 - 0.24 35 17.50 14 15.73
0.24 - 0.28 6 3.00 2 2.25
0.28 - 0.32 0 0.00 O 0.00
0.32 - 0.36 0 0.00 0 0.00
0.36 - 0.40 O 0.00 0 0.00

 

1rDistribution of HSI values within random circles differed signiﬁcantly from distribution
of HSI values within activity areas (Chi-square, P < 0.01)

121

 

 

 

 

 

 

 

40

30 my L
g 20 IAvarlable
a. IUsed

10

o .

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36
HSI value

Figure 32. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Pigeon River open site.

Table 55. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Pigeon River open site.a

 

Available Available Used Used

 

HSI range Freq. % Freq. %
0.00 - 0.04 0 0.00 0 0.00
0.04 - 0.08 16 8.00 1 1.75
0.08 - 0.12 22 11.00 1 1.75
0.12 - 0.16 29 14.50 6 10.53
0.16 - 0.20 38 19.00 8 14.04
0.20 - 0.24 35 17.50 17 29.82
0.24 - 0.28 36 18.00 16 28.07
0.28 - 0.32 12 6.00 7 12.28
0.32 - 0.36 12 6.00 1 1.75
0.36 - 0.40 0 0.00 0 0.00

 

TDistribution of HSI values within random circles differed signiﬁcantly from distribution
of HSI values within activity areas (Chi-square, P < 0.01)

122

 

 

 

Percent

 

 

30
20
I Available
I Used
1 0
o 1

 

 

 

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36
HSI value

Figure 33. Percent of HSI values within random circles (available) and within activity
ranges (used) of dispersing ruffed grouse in the Pigeon River closed site.

Table 56. Frequency and percent of HSI values within random circles (available) and
within activity ranges (used) of dispersing ruffed grouse in the Pigeon River closed site.ll

 

Available Available Used Used

 

HSI range Freq. % Freq. %
0.00 - 0.04 1 0.5 O 0.00
0.04 - 0.08 15 7.5 5 6.94
0.08 - 0.12 24 12 5 6.94
0.12-0.16 25 12.5 11 15.28
0.16 - 0.20 30 15 10 13.89
0.20 - 0.24 37 18.5 16 22.22
0.24 - 0.28 29 14.5 16 22.22
0.28 - 0.32 33 16.5 9 12.50
0.32 - 0.36 6 3 0 0.00
0.36 - 0.40 0 0 0 0.00

 

‘ Distribution of HSI values within random circles did not differ signiﬁcantly from
distribution of HSI values within activity areas (Chi-square, P =0.54)

123

Discussion

As expected, grouse selected for medium-aged aspen habitat in the Pigeon River
sites. In Alberta, Rusch and Keith (1971a) recorded 85% of grouse observations within
aspen woods. This forest type is known to provide good hiding cover and is usually close
to an area of mature aspen (a food source). A surprising result, however, was that this
age class of aspen wasn’t preferred over other types in the Maltby Hills sites. Instead,
Maltby Hills grouse chose young aspen and young pine (predominantly red or white pine,
0 — 29 yr.), both of which ranked high in preference in the open and closed sites.
Medium-aged aspen had the highest quality cover, yet was used in direct proportion to its
availability. The Maltby Hills sites, however, had more medium-aged aspen than the
Pigeon River sites, so it was more likely that the area surrounding the activity range (the
available habitat) contained it.

In many cases, forest types that rank high in fall cover according to the model
were not preferred. Birds may not select areas with good cover because of the lack of a
food source, increased competition, or presence of predators. There is no evidence that
hunter disturbance caused birds to avoid aspen, as its use was similar in areas open and
closed to hunting. Birds in the Maltby Hills closed site, though, preferred lowland
conifers which were not used much in the open site.

A possible explanation for the disparity in use and cover quality is the method I
used to assign ﬁeld variable suitability values. To use the model on a landscape scale, I
assigned the mean value for each forest type to each pixel of that type. Perhaps some
areas are of very high quality cover and some are medium or low, and grouse are using

these differently. This could result in a forest type appearing to be used in proportion to

124

its availability when in reality grouse are choosing high quality cover and avoiding low
quality cover. Averaging over a smaller area or correlating stem density data to forest
cover grids would overcome this problem.

In all study sites, upland hardwoods (which comprised a large percentage of 3 of
the four sites) were avoided by ruffed grouse. This was an expected result, because of
their general lack of understory cover and low HSI score. In the two study areas, the
upland hardwoods category generally consisted of maple and beech, thin stands with high
basal area and low stem density. In some cases these stands were managed for timber
production, and were thinned to decrease competition. Where oaks were present,
however, they were used to a greater extent. There was very little oak present in the
Pigeon River sites (not enough to be accurately classiﬁed and mapped), and the oak
present in the Maltby Hills sites had low cover value. But McDonald et al. (1998)
showed that mixed oak stands could be managed to provide food and cover for grouse in
Pennsylvania. The possibility of improving understory cover in oak stands (possibly
using group selection cuts) should be investigated further in the Lakes States. Any such
management would have to be carefully regulated, though, to prevent the conversion of
oaks to other forest types.

f In the Maltby sites, the older pines were also avoided. These types of stands
probably provide good roosting places for raptors (Mosher et al. 1986). In the Pigeon
sites, however, older pines were used more frequently or used in proportion to their
availability. Possible keys to this difference may be greater understory cover (older pines
in the Pigeon River had higher quality cover than in the Maltby sites), a difference in

structure, or the presence of some food source that replaces the need for aspen. Grouse in

125

the Pigeon River also used jack pine frequently, whereas grouse in the Maltby Hills sites
did not. In both areas jack pine cover quality was similar. Although small to medium
sized pines are thought to provide good cover as understory (Gullion 1984; McCaffery et
al. 1996), they have never been shown to be favored by grouse.

In all study sites, the distributions of HSI values within activity ranges of non-
dispersing ruffed grouse were signiﬁcantly different from the distributions of HSI values
within random circles. Non-dispersers generally chose higher quality habitats, with very
few choosing habitats with HSI values lower than 0.10. The difference was most
obviously expressed in the Maltby Hills closed site, where many of the random circles
sampled had values of O - 0.04. These facts lend credit to the predictive value of the
model on a landscape scale. Model applicability could be improved, though, by using
grids that categorize forest type within smaller pixels. By using 30 x 30 m pixels, our
sampling may not have picked up small pockets of high or low quality habitats within
other homogeneous stands. The model states that a clone as small as 0.04 ha would
qualify as adequate food for a grouse through the winter, and this was too small for us to
pick up and categorize. As acknowledged by Donovan et al. (1987), accurate usage of a
GIS based model relies on the compatibility of the scales used in model development and
those that are available in the GIS data.

Habitat quality within the calculated activity areas of dispersers was different
from random in 2 of the 4 sites, the closed Maltby Hills site and the open Pigeon River
site. In both cases, areas surrounding activity ranges ranked higher in HSI value than
HSI values in random circles. In the Maltby Hills open site and the Pigeon River closed

site, activity areas did not differ signiﬁcantly from random circles. It was expected that

126

the activity ranges in all areas would be higher in quality than random, but dispersing
birds may be forced to settle in areas that are not of the highest quality. Additionally,
birds may move to areas of higher quality after their primary fall dispersal, as the need
for adequate winter cover increases.

The ultimate goal of modeling habitat quality and preference for grouse is to use
the information gained to improve or create habitat. Although I have shown that most
grouse settle in habitat that is considered high quality, it would be useful to know whether
such choices affect survival. In the ﬁnal chapter, I will use the information obtained thus
far to see whether habitat use and habitat quality have an effect of the survival of ruffed

grouse in northern Michigan.

127

CHAPTER 5: Relating habitat factors to survival

Introduction

The last step in this project is to link the factors discussed in the previous
chapters. As a step in improving existing ruffed grouse management, it will be useful to
relate survival to various parameters that are within the control of managers. Although I
have shown preference and avoidance of habitats by grouse, I cannot make the
assumption that preference (or use of areas with a high HSI value) is predictive of ruffed
grouse persistence in an area. Such assumptions must be tested by relating habitat choice
and quality to some measure of ruffed grouse success.

The development and testing of habitat models in the past has taken the approach
of correlating habitat quality with species abundance, occurrence, or other such response
(Scharnberger and O’Neil 1986). Building models by sampling for animal density,
however, can lead to errors of commission and bias towards certain habitats. Most
wildlife sampling methods have associated error, such as different sighting functions for
different habitats. In that case, calculated densities can vary by habitat even though
actual densities are equal. Additionally, the density of a species and habitat quality may
not be positively correlated (V anHome 1983).

Researchers often make the assumption that the measure of abundance is a
reﬂection of productivity or survival, which may be misleading. Natural populations are
rarely closed to immigration and emigration, and animal movements could cause habitats
of low quality to appear good. This type of error is illustrated in the study by Paradis and

Croset (1995) of voles in meadows and orchards. Although measured densities were

128

similar between the sites, no juveniles survived in the meadows. If the measure of
quality had been based solely on density then the habitats would have rated as equal in
quality. The meadows, however, were a sink; the demographic parameters of the
population measured in that area were not sufficient for the population to persist.

A preferable means of evaluation would relate quality directly to survival,
fecundity, or other demographic parameter. These types of studies are rare and badly
needed (Morrison et a1. 1988). Suitable habitat should be deﬁned as an area which
allows the population to persist over time, not necessarily an area with a consistent
presence of a certain species. Jones (1996) tested the HSI model for the bald eagle by
relating habitat quality to productivity in northern Michigan. He found that the existing
model underestimated the affect of human activity on nesting success. Although eagles
were present in areas of human disturbance, their success rate was substantially lower
than in undisturbed areas.

Models of ruffed grouse habitat quality have been developed and tested by
relating spring drumming counts with the surrounding habitat (Roloff 1994; Hammill and
Moran 1986; Cade and Sousa 1983). Drumming counts are repeatable, cheap, and easy
to perform, and are probably in some way related to the quality of habitat present. But
ruffed grouse have been thought to move from “sources” (such as areas far from roads
where they are safe from hunting pressure) to “sinks” (areas where hunting pressure is
high) when density within the “sinks” decreases (Small et a1. 1991). Thus, densities may
be equal even though some areas are better for long-term survival.

In this chapter, I will begin to describe the differences between activity ranges of

grouse that live through fall and winter and those that do not. I will examine how

129

demographic and environmental variables predict long-term survival, and test the
relationship of the existing model to ruffed grouse survival. My intent is to make steps
toward developing a model (or improving the existing model) that can be used on a large

scale to evaluate present and future quality of habitat for ruffed grouse in Michigan.

130

 

Methods

A comprehensive database was compiled using the data gathered on movements,
habitat use, habitat quality, and survival for non-dispersing ruffed grouse. For each bird,
identiﬁcation data included area, site, band number, year, sex, and age. The additional
variables were the date the bird entered the study, the date the bird left the study, fate
(alive, dead, or censored), activity range size, HSI value within the activity range, and
frequency and percent of each forest type within the activity range. The last date out was
limited to May 15 of the following year. Past that date there was a high chance of
censoring due to radio failure (radio-collar batteries had a 9 month expected life-span).
The maximum number of days survived was therefore 287.

The median HSI value within activity ranges was calculated for each site. Birds
were categorized as low-HSI (HSI score < median) or high-HSI (HSI score > = median).
Censored birds were eliminated from this analysis. The number of days survived after
August 1 was compared between low-HSI and high-HSI birds within each site using a
Student’s t-test (Rosner 1990) with signiﬁcance set at 0.10.

The median number of days survived was also calculated for each site, also using
only birds whose exact fate was known. Birds were categorized as short-lived (days
survived < median) or long-lived (days survived > median). The percent of each habitat
within activity ranges were compared between short-lived and long-lived birds using a
Kruskal-Wallis one-way ANOVA (Siegel and Castellan 1988). Scatter plots were
generated using days survived past August 1 versus HSI value and versus cover value

(suitability of habitat without the spatial variable included) within the activity range of

131

each bird. Linear regressions were used to test whether a direct relationship existed
between quality of habitat and days survived.

The Cox proportional hazards model (Cox 1972) was used to evaluate potential
risks based on vegetation parameters for non-dispersers (including censored birds). The
variables used were: sex, age, HSI value, and the number of hectares in activity range of
young aspen (aspen 0 — 29 years old), old aspen (aspen 30 years or older), upland
hardwoods (upland hardwoods and oak in the Maltby Hills sites), upland conifers (young
pine, old pine, and jack pine), and lowlands. The model was ﬁrst run using proc PHREG
in SAS (SAS Institute, Cary, North Carolina) with all variables included. Then the most
probable proportional hazards model was found through stepwise regression. For a
covariate to enter and stay in the model, P was set at 0.2. For the purposes of describing
model results, the percent of increase or decrease in the hazard of a covariate was

determined using the following equation: % difference in hazard = 100*(risk ratio-1).

132

Results
Activity range composition

In the Maltby Hills open site, 53 non-dispersers had data on habitat use, habitat
quality, and a known fate. The median number of days survived past August 1 was 245.
Long-lived and short-lived grouse had similar mean percentages of most forest types
within their activity ranges (Table 57). The one signiﬁcant difference detected was in
aspen over 30 years of age. Short-lived birds had a signiﬁcantly higher mean percentage
of this type in their activity ranges (20.4%) when compared to long-lived birds (17.7%).

The median number of days survived for Maltby Hills closed site birds was 225
(N = 79). Long-lived birds had signiﬁcantly more young aspen within their activity
ranges (11.6%) than did short-lived birds (8.8%) (Table 58, P = 0.076). Activity ranges
of long-lived birds had signiﬁcantly lower percentages of lowland conifers, pine 0 - 30
years old, and pine over 30 years old than did short-lived birds (P = 0.004, P = 0.056, P =
0.019, respectively).

In the Pigeon River open site, the median number of days survived was 216 (N =
48). Long-lived birds had a signiﬁcantly greater young aspen component within their
home ranges (13.2%) than did short-lived birds (8.1%) (Table 59). Short-lived birds had
a signiﬁcantly higher mean percentage of lowland hardwoods (30.5%) than long-lived
birds (23.8%, P = 0.018).

The median number of days survived for the 61 birds in the Pigeon River closed
site was 207. The signiﬁcant differences detected within activity ranges between long-
and short-lived birds were the same as in the open site (Table 60). Activity ranges for

long-lived birds contained a higher proportion of young aspen (13.4%) than short-lived

133

birds (10.0, P = 0.069) and a lower proportion of lowland hardwoods (11.9% vs. 19.1%,

P = 0.017).

134

Table 57. Mean percent of activity range (standard error in parentheses) in each forest
type for ruffed grouse that survived less (short-lived) than and greater than (long-lived)
the median number of days in the Maltby Hills open site.

Short-lived Long-lived
% (SE) % (SE) P

 

Aspen 0 — 10 years 3.3 (0.6) 3.6 (0.5) 0.482
Aspen 11 — 29 years 16.7 (2.0) 21.5 (2.4) 0.145
Aspen 30 + years 20.4 (1.6) 17.7 (1.5) 0.084.
Upland hardwoods 28.5 (4.0) 23.1 (3.3) 0.413
Lowland conifers 8.1 (1.7) 7.9 (1.0) 0.413
Pine 0 — 29 years 4.4 (0.7) 6.5 (2.2) 0.337
Pine 30 + years 9.5 (2.8) 9.1 (2.5) 0.669
Jack pine 1.1 (0.6) 2.4 (1.6) 0.605
Oak 8.0 (1.2) 8.2 (0.8) 0.444

 

ISigniﬁcant difference in mean percent of forest type within activity range between short-
lived and long-lived grouse (Kruskal-Wallis one-way AN OVA)

Table 58. Mean percent of activity range (standard error in parentheses) in each forest
type for ruffed grouse that survived less than (short-lived) and greater than (long-lived)
the median number of days in the Maltby Hills closed site.

 

Short-lived Long-lived
% (SE) % (SE) P

Aspen 0 — 10 years 8.8 (1.1) 11.6 (1.2) 0.076'’
Aspen 11 - 29 years 26.0 (1.8) 30.1 (2.1) 0.114
Aspen 30 + years 14.5 (1.2) 14.1 (1.1) 0.829
Upland hardwoods 19.0 (3.1) 18.2 (3.0) 0.632
Lowland conifers 1.6 (0.1) 1.1 (0.1) 0.004”I
Pine 0 — 29 years 6.3 (0.8) 4.4 (0.7) 0.056a
Pine 30 + years 5.8 (0.9) 2.6 (0.5) 0.019’I
Jack pine 2.1 (0.5) 2.8 (1.4) 0.500
Oak 15.9 (1.5) 15.1 (1.7) 0.624

 

'Signiﬁcant difference in mean percent of forest type within activity range between short-
lived and long-lived grouse (Kruskal-Wallis one-way ANOVA)

135

Table 59. Mean percent of activity range (standard error in parentheses) in each forest
type for ruffed grouse that survived less than (short-lived) and greater than (long—lived)
the median number of days in the Pigeon River open site.

 

 

 

Short-lived Long-lived

% (SE) % (SE) P
AspenO— 10 yr. 8.1 (1.4) 13.2 (2.1) 0.081“
Aspen 11 —29 yr. 10.2 (1.1) 9.4 (1.0) 0.951
Aspen 30 + yr. 5.9 (0.8) 5.9 (0.9) 0.821
Upland hard. 14.0 (3.0) 12.3 (3.0) 0.665
Lowland hard. 30.5 (2.3) 23.8 (2.7) 0.018'l
Lowland con. 9.5 (1.7) 12.0 (2.7) 0.773
Pine 0 — 29 yr. 1.0 (0.3) 1.1 (0.3) 0.821
Pine 30 + yr. 17.9 (2.2) 20.0 (2.3) 0.606
Jack pine 2.8 (0.7) 2.4 (0.5) 0.673

 

ISigniﬁcant difference in mean percent of forest type within activity range between short-
lived and long-lived grouse (Kruskal-Wallis one-way ANOVA)

Table 60. Mean percent of activity range (standard error in parentheses) in each forest
type for ruffed grouse that survived less than (short-lived) and greater than (long-lived)
the median number of days in the Pigeon River closed site.

 

 

 

Short-lived Long-lived

% (SE) % (SE) P
Aspen 0 — 10 yr. 10.0 (1.8) 13.4 (1.6) 0.069"
Aspen 11 — 29 yr. 5.2 (0.7) 6.2 (0.7) 0.341
Aspen 30+yr. 13.5 (1.8) 10.4 (1.7) 0.161
Upland hard. 5.7 (2.4) 13.1 (9.4) 0.918
Lowland hard. 19.1 (2.3) 11.9 (1.3) 0.017“
Lowland con. 16.7 (2.3) 12.3 (1.6) 0.193
Pine 0 - 29 yr. 5.1 (0.8) 4.9 (0.6) 0.812

' Pine 30 + yr. 19.6 (2.8) 21.9 (3.3) 0.931

Jack pine 4.8 (0.5) 5.9 (0.8) 0.461

 

ISigniﬁcant difference in mean percent of forest type within activity range between short-
lived and long-lived grouse (Kruskal-Wallis one-way AN OVA)

136

HSI value and days survived
In the open Maltby Hills site, the median HSI value was 0.234. Those birds that

had values lower than the median (low-HSI birds) survived an average of 197 days after
1 August. High-HSI birds survived an average of 241 days. This difference was
signiﬁcant (T-test, P = 0.014). There was no signiﬁcant linear relationship, however,
between days survived past August 1 and HSI value (Figure 34). Two birds whose
habitat ranked lowest in overall quality survived the maximum number of days. There
was a weak positive linear relationship between cover suitability and days survived
(Figure 35) (P = 0.07, R2 = 0.063).

The median HSI value in the Maltby Hills closed site was 0.165. Low-HSI birds
survived an average of 207 days while high-HSI birds survived 230 days. Again the
difference was signiﬁcant (T-test, P = 0.08). No linear relationship existed between days
survived and HSI value (Figure 36), or between days survived and cover suitability
(Figure 37). Birds that survived the maximum number of days had a wide range of
values for both overall HSI and cover suitability.

In the Pigeon River open site the median HSI value was 0.177. Low-HSI birds
survived an average of 193 days, and high-HSI birds survived 219 days. The difference
was not signiﬁcant (T-test, P = 0.13). No relationships were detected between HSI value
(Figure 38) or cover suitability (Figure 39) and days survived. Seven of the lowest
ranking cover suitability values were for birds that survived the full study period.

Low-HSI birds in the Pigeon River closed site (median = 0.19) survived an
average of 219 days, longer than the average for high-HSI birds (193 days). The

difference was signiﬁcant (T-test, P = 0.07). There was no relationship detected between

137

HSI value and days survived (Figure 40), and a negative linear relationship between
cover suitability and days survived (Figure 41) (P = 0.009, R2 = 0.111). Of the 7 birds

that had activity ranges with the lowest values for cover suitability, 6 of them lived more

than 200 days.

138

 

0.4 w Lu“-..

 

P = 0.07 R2 = 0.0625 '

0 o
0.3—
O 9 . o .0 ." i
0 ~ __4_.W o

 

 

 

 

 

 

8
E
————'""’

"‘ 0.2 a o o 8
E; : .. O O O O
8 o’

0.1 - o O

0.0 T . -- m-.. s ,

0 50 100 150 200 250 300 350
Days survived past August 1

Figure 34. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Maltby Hills open site.

 

 

 

 

 

0_4 _ -____WW._-.__, - ,,,,, _. W- ' . -.WW, . .Wi
P=0.45 R2=0.0112
0.3 . . . o
O . . . . O O.
O
o o 0’ 9’ . ’ J
s z 0 . . t
'55 e
O
m 0.1 ,
O
0 —~-——--~ —~ ——»--~ ..- --»~~—-.—— ~~-—A——~— -————.+
0 50 100 150 200 250 300 350

Days survived past August 1

Figure 35. Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Maltby Hills closed site.

139

 

 

 

 

 

 

 

0.3 -~ w~—A—————-—-- »-
. P = 0.55 R2 = 0.0036
0 O
0 2 — ’ O 0 o 0 o’
1:5 9’ ’ o
g 00 o O
.— . ; . 9
g 0 ° 3 , ° 0
01 0 ’
. z. 9. O .
’ o
O T f _ ‘——— —_ T‘ ﬂaw _ l T ﬁ
0 50 100 150 200 250 300 350
Days survived past August 1

Figure 36. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Maltby Hills closed site.

 

 

 

Cover SI value

 

 

 

 

0.4 i__ __ ---__-__.-. W. -._-_- _-
P = 0.46 R2 =0.0071
0.3 4 . ‘
° 00 ° ° ’
O 9 o o
O“ . O .0 Q
0.2 - W
O . .
o o o. .
O
o . O
0.1 — O .0 . . .
0.0 —..___——._—_ «Tm .- _.____ _. r
.0 50 100 150 200 250 300 350

Days survived past August 1

Figure 37. Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Maltby Hills closed site.

140

 

 

 

 

 

 

 

0.3
0 P = 0.75 R2 = 0.0023 0
O O O
0.2 q , ’ 3
3 WOW
'3 O O. . O O
> . .
('3 . O . . . O
:1:
0.1— t
’0
o T j .7—"_—_ “__ , V -_ T T
0 50 100 150 200 250 300 350

Days survived past August 1

Figure 38. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Pigeon River open site.

 

 

 

 

 

 

0.5 “WWWWW.-. -- .- .‘ _-- W__W E
P=0.99 R2=.OOO6 ,
0.4 - , .
o O
:3 0
"g 0.3 _ e o3 ’
a O O O 9 :
§ 0.2 — '0 0 ° ° 3 T
° 9
U
0.1 -
0.0 , WW v . - W._ .W__T W
0 50 100 150 200 250 300 350

Days survived past August 1

Figure 39. Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Pigeon River open site.

141

 

 

 

 

 

 

0.4
P = 0.15 R2 = 0.0051
0.3 1 0 , .
o o o
g 0 . o . . ‘ 9 ’3
3 0,2 1 o , o o 0
F3 . + o
m . o e
o
0.1 d .0 . . . 3
o. , ’
Q 0
0.0 1 T -——~—--—1————h——- j 1 1
0 50 100 150 200 250 300 350

Days survived past August 1

Figure 40. HSI value within activity range vs. days survived past August 1 for ruffed
grouse in the Pigeon River closed site.

 

 

 

 

0 4 _— _—_— _4____ _,-,_ - _- mm ,, -_._ ,-_,_,__---_______-___..___
P = 0.009 R2 = 0.1112
0.3 ~ . O “Q. . .094
_3 o ” o . ’ o
a
> 0
E3 0.2 — o
’2‘ t .
8 . i
0.1 _ . o 3
0.0 . ,-_ —. _-__- . . j
0 50 100 150 200 250 300 350

Days survived past August 1

Figure 41 . Cover SI value within activity range vs. days survived past August 1 for
ruffed grouse in the Pigeon River closed site.

142

 

Cox model results

When the Cox proportional hazards model was run with all variables for birds in
the Maltby Hills open site, age, size of activity range, and number of ha within aspen
over 30 years old were signiﬁcant high risk factors (Table 61). The number of ha within
young aspen and upland hardwoods were covariates that were positively related to
survival. The best ﬁt model used these variables plus the number of ha within upland
conifers to predict survival probability (Table 63). Juveniles had a risk of mortality that
was 5 times that of adults. Every ha increase in activity range size increased mortality by
1.5%. Additionally, with every 1 ha increase in the number of acres of old aspen within
the activity range, the risk of mortality increased 16%. Increases in the number of ha of
young aspen, upland hardwoods, and upland conifers decreased risk by 14.1%, 4.1%, and
6.5%, respectively.

The two covariates that had signiﬁcant effects in the Maltby Hills closed site were
number of ha in young aspen and lowland conifers (Table 61). These were also the two
variables picked as the most likely ﬁt for the model (Table 63). For every ha of young
aspen present in the activity range, risk of mortality decreased 5.4%. But every increase
in ha of lowland conifers increased the risk by over 3 times.

When all variables were used, age, number of ha in aspen, and number of ha in
lowland hardwoods were signiﬁcant covariates in the Pigeon River open site (Table 62).
When the stepwise regression was performed, old aspen was no longer signiﬁcant (Table
63). Age and number of ha in lowland hardwoods were associated with an increased risk

of mortality. Juveniles had over twice the risk of mortality than did adults. For every ha

143

 

A—Hy- .—

increase in lowland hardwoods within a bird’s activity area, the risk of mortality
increased by 2.5%.

Age and number of ha in lowland conifers were signiﬁcant covariates when all
the variables were used in the Pigeon River closed site (Table 62), but the most likely
model included age and number of ha within young aspen (Table 63). Juveniles were
again at a higher risk for mortality, over half again as high as for adults. With every one

ha increase of young aspen within activity ranges, risk of mortality decreased by 3.1%.

144

Table 61. Risk ratios and P values generated using the Cox proportional hazards model
for variables related to survival in the Maltby Hills open and closed study sites.

 

 

 

Open site Closed site
Risk ratio P Risk ratio P
Age 5.456 0.002 1.033 0.920
Sex 1.486 0.390 1.254 0.566
HSI 4.359 0.696 10.938 0.534
Range 1.020 0.061 0.999 0.915
Aspen 0 — 29 yr. 0.858 0.001 0.943 0.008
Aspen 30+ yr. 1.145 0.007 0.997 0.949
Upland hardwoods 0.961 0.021 1.008 0.726
Upland conifers 0.948 0.135 1.015 0.507
Lowland conifers 0.935 0.291 2.598 0.063
N = 56 N = 92

 

Table 62. Risk ratios and P values generated using the Cox proportional hazards model
for variables related to survival in the Pigeon River open and closed study sites.

 

 

 

Open site Closed site
Risk ratio P Risk ratio P
Age 3.152 0.013 2.001 0.046
Sex 1.418 0.365 1.690 0.201
HSI 1951.161 0.159 12.396 0.497
Range 1.004 0.252 0.996 0.598
Aspen 0 — 29 yr. 1.008 0.814 0.964 0.254
Aspen 30+ yr. 0.739 0.037 1.014 0.809
Upland hardwoods 0.975 0.286 1.007 0.484
Upland conifers 1.029 0.460 0.994 0.721
Lowland conifers 0.913 0.228 1.064 0.073
Lowland hardwoods 1.090 0.071 0.989 0.705
N = 60 = 75

 

14S

Table 63. Risk ratios and P values for explanatory values found in a stepwise regression
through the Cox proportional hazards model for grouse in the Maltby Hills and Pigeon

River open and closed sites.

 

 

 

 

Maltby Hills Pigeon River
Open Closed Open Closed

Risk P Risk P Risk P Risk P
Age 4.880 0.003 -° - 2.182 0.035 1.632 0.120
Sex - - - - - - - ..
HSI - - - - - - - -
Range 1.015 0.076 - - - - - -
Aspen 0 — 29 yr. 0.859 < 0.001 0.946 0.002 - - 0.969 0.055
Aspen 30+ yr. 1.160 0.003 - - - - - -
Upland hardwoods 0.959 0.016 - - - - - -
Upland conifers 0.935 0.049 - - - - - -
Lowland conifers - - 3.260 0.001 - - - -
Lowland hardwoods nab na 1.025 0.129 - -

 

I’Variable not selected
t’Not applicable

146

Discussion

The fact that ruffed grouse chose habitat that ranked higher than random
according to the model (Chapter 4) is evidence that the model is at least an adequate
predictor of use. The ruﬂed grouse model as a predictor of survival was only loosely
veriﬁed by the results in the Maltby Hills study sites. Although the Cox proportional
hazards model did not ﬁnd a signiﬁcant relationship between HSI value and survival,
birds that used habitat that ranked highest in quality survived longer than birds that used
lower-quality habitat. There was no linear relationship detected between HSI value and
survival in either site, and the weak relationship in the closed site was found after the
removal of the spatial variable (distance to mature aspen). This suggests that the use of
the aspen variable as a quality modiﬁer and predictor for high quality habitat may be
inappropriate.

In the Maltby Hills open site, old aspen was in fact a high risk component of
activity ranges. In the Maltby Hills, the aspen over 30 years ranked very low in quality
of cover when compared to the Pigeon River. According to Bergerud and Gratson
(1988), ruffed grouse should use aspen as a winter food source when they can optimally
forage; that is, when there is reduced predation risk. If a food source is present but
contains very little cover, then the bird does not maximize its energy reserves by feeding
there, and it may increase its risk of predation. These facts may negate the value of
mature aspen as a winter food source for grouse in some cases.

The effect of young to medium aged (0 — 29 years old) aspen within grouse
activity ranges was an increased probability of survival in 3 of the 4 study sites. Young

aspen was among the highest in quality for cover and had high stem densities in all sites.

147

Vispo et al. (1994) used the Cox model to relate grouse habitat and survival in the
northwest United States. They found that the presence of aspen in the overstory was
associated with reduced mortality. Rickers et a1. (1995) presented a model based on the
arrangement of aspen age classes to predict the affects of timber management decisions
on habitat. My results show that presence of aspen is not a good predictor in and of itself.
The cover quality of aspen within age classes can vary widely (Roloff et al. 1994), and
should be considered when planning management and measuring quality. Vispo et al.
also noted that the presence of conifers was associated with decreased winter mortality. I
found that jack pine and young pine (both of which had high cover values) tended to be
used equally in higher proportion to their availability, but did not seem to have any affect
on survival.

Relationships detected between cover and HSI value and survival were weak. R2
values were low even where signiﬁcant linear relationships were found. Additionally,
survival and cover value were not always positively related. In the Pigeon River closed
site, there was a negative relationship between HSI value (and HSI cover value) and clays
survived. Birds that used habitat that ranked high in quality according to the model
survived fewer days than those that chose low quality habitat. In that site, the lowland
hardwoods were very high in cover value, but there was a risk associated with number of
ha of a bird’s activity range in lowland hardwoods (detected by both the T-test and the
Cox model). This suggests that the value of cover is variable, perhaps by the amount
present and forest type. In my study, birds that had a high component of lowland areas
within their activity ranges were at a higher risk of mortality. Most of the lowland

hardwoods in the Pigeon River study sites (none were classiﬁed in the Maltby sites) were

148

thick with speckled alder (A Inus rugosa). Nutritional analyses of alder catkins in a
captive ruffed grouse population showed a negative effect due to rapid mass and water
loss (Gulielmo and Karasov 1995).

The number of hectares of lowland conifers within an activity range had a very
strong negative effect on survival in the Maltby Hills closed site. This type, while
ranking high in cover value (see Chapter 3) may have provided good cover for predators.
Rusch and Keith (1971a) found that although Alberta grouse spent the majority of their
time in aspen, 38% of mortalities were located in spruce woods. There was no linear
relationship between lowland conifers and days survived (Figure 42) in that site,
suggesting that some interaction between covariates was taking effect.

Further analyses need to be done before speciﬁc modiﬁcations of the existing
model are proposed. Age was a factor in survival in 3 of the 4 sites. Juvenile birds had a
much higher risk of mortality than adult birds, conﬁrmation of the results discussed in
Chapter 1. In the Maltby Hills open site, range size was also a factor in survival.
Juveniles tended to have larger home ranges than adults, so these covariates are probably
not independent. It may be useful to stratify the birds by age and run the analysis again.
Then it can be determined whether young birds are using different habitat than adults or
if they are surviving differently in similar habitats, and if range has an effect independent
of age. It might also be beneﬁcial to combine the open and closed study sites for the Cox
analysis, adding “hunted” or “not hunted” as a variable.

Although more exploration is needed, the results of this chapter are a good

starting point to aid in the improvement of ruffed grouse habitat modeling and

149

 

management. In the ﬁnal section, I will discuss these implications, management options,

and modeling possibilities.

150

 

3.5

 

 

 

 

 

 

 

 

 

 

 

 

 

P=0.548 R2=0.0039
3 o
O
2.5 ,
.SE 2
“v 0 O
1.4 9
91,53) 15 O 00 : O . O
O a... 0“
g}? 1 H ’ 0’ £
.2.
1%: 0.5 m
0 O.
A 0 , . ”“4 .". ”01—. .
0 50 100 150 200. 250 300

Days survived past August 1

350

Figure 42. Lowland conifers within activity range (ha) vs. days survived past August 1
for ruffed grouse in the Maltby Hills closed site.

151

 

 

MANAGEMENT IMPLICATIONS

The initial question addressed in this study was whether hunting has an effect on
the fall to spring survival of ruffed grouse in Michigan. Translated into management
terms, the question becomes whether the existing hunting regulations are adequate or too
liberal. My results pertaining to this question were mixed; in the ﬁrst year of the study,
when grouse populations were at their lowest, survival rates in the sites closed to hunting
were signiﬁcantly higher than in those open to hunting. In all years except 1993, though,
the survival curves converged, and the study seasons ended with comparable or higher
survival in the open study sites. This would suggest that changes to the hunting
regulations at the low point of the cycle could be beneﬁcial.

One variable that may have confounded the comparison of survival between the
open and closed study sites was the difference in habitat quality. The difference is most
pronounced in the Maltby sites, where the closed site has fewer areas of continuous
habitat that ranks high in quality. If habitat quality affects survival, the low survival in
the closed site may have been caused by poor habitat. This is preferable, however, to the
problem encountered by Kubisiak (1994). His area of high hunting pressure was of lower
quality than his area of low hunting pressure, which meant he was unable to discern
whether hunting had an effect on population density. In my sites, the difference in
survival found in 1993 was despite the lower quality habitat.

The possible options for changing the ruffed grouse hunting season (during or
preceding cyclic lows) in Michigan are changing the bag limits (currently 5 per hunter
per day in the Lower Peninsula), shortening the season (currently September 15 —

November 14 and December 1 — December 31), or eliminating the season. It is arguable

152

whether any of these options would make a measurable difference in the huntable
population over the course of a 10-year cycle.

Changing bag limits is an option that the Michigan Department of Natural
Resources (MDNR) has investigated in recent years. Existing MDNR survey data should
be studied to see if hunters take their bag limits during low population years. If they do
not, then decreasing the bag limit wouldn’t have any effect on the number of birds taken.
Hunters may already be shooting fewer birds at the low point of the cycle because there
are fewer to shoot. If, however, they are compensating for the low population by
spending more time in the ﬁeld until they ﬁll their limits (as per Small et al. 1991), then
lowering the bag limit may allow for more birds to live through the fall and winter.

Season length is another factor that has caused some controversy among hunters
in recent years. Michigan has a 12 week season, including the month of December.
There has been concern voiced by early fall hunters that the December hunt causes undue
pressure on the population. They feel that eliminating the winter hunt would increase fall
to spring survival. But of the 20 birds shot in 1993 and 1994 (when grouse populations
were at their lowest), only 1 was shot in December (Gorrnley 1996). In northern
Michigan, the snow accumulation probably deters many hunters ﬁom the ﬁeld (Palmer
and Bennett 1963). Thus, a change in regulations would probably only affect the
southern half of the Lower Peninsula, where snowfall is less of an issue in the early
winter.

Changing the season by altering the opening day is another possibility. However
research shows that most of the kill probably takes place in the ﬁrst few weeks of the

hunting season, regardless of when it starts or how long it lasts. In Wisconsin, Rusch et

153

'

 

 

al. (1983) reported that of 3600 grouse killed in 2 seasons, 55% were killed in October,
26% in November, 13% in December, and 6% in January. In Riﬂe River, Michigan, 75%
of the kill was made in the ﬁrst 15 days of a 41 day season, leading researchers to believe
that overall season length had little to do with total harvest (Palmer and Bennett 1963).
An intensive study of black grouse (T etrao tetrix) regulations in Switzerland found that
season date changes had no effect on the overall harvest (Zbinden and Salvioni 1997).
Starting the season a few weeks later in the fall may reduce the harvest, because broods
would be more likely to have broken up. Thus, hunters would not be as likely to shoot
multiple birds from the same brood, and fall to winter juvenile survival would be
increased.

Eliminating the harvest altogether in years of low population would undoubtedly
increase survival to some extent. Slather et al. (1996) examined a model of maximizing
the sustained yield of a game species when that species has a ﬂuctuating population.
They suggest that if harvesting is stopped when population is at the lowest level, then
over time harvest is maximized. But the pertinent question is, would there be a
noticeable difference? Maximizing harvest over the long term may mean that over 10
years, a given hunter will shoot 30 grouse instead of 29. Does that 1 bird change whether
the hunter is satisﬁed with her take? Wisconsin and Minnesota instituted a ban on
hunting in the 19403 when the grouse population was approaching a low point (Bergerud
and Gratson 1988). At the end of the time, there didn’t seem to be any more grouse
present than in Michigan, a state that kept its hunting season intact. Time has shown that

cyclic declines will happen regardless and that grouse populations will recover. Perhaps

154

the more important management focus should be on making sure that there is adequate
habitat for birds when the population is at its highest.

There is a need for habitat models that are usable on a large scale, with ease of
application and accuracy. I suggest the creation of a GIS-based model that can be used to
evaluate current habitat quality, to project the effects of management decisions, and to
plan habitat management activities for grouse and other species. The current trend in the
Midwestern United States is a reduction in the amount of aspen clearcuts (Hammill and
Visser 1984). Red and white pine and northern hardwoods, types that may be of little
value if not accompanied by areas of good fall/winter cover, are replacing mature aspen.
Management plans and models for ruffed grouse need to take this fact into account, and
look beyond aspen. The placement and landscape interspersion will be key to make the
most of the limited aspen cuts that are carried out.

Over the scale used in this project, the existing model (Hammill and Moran 1986)
was only loosely veriﬁed, and could be improved using my results. In its present form
the HSI model for ruffed grouse in Michigan overemphasizes the need for mature aspen
as a food source when used on a landscape scale. Because of the method used to
compute the ﬁnal value (multiplying the aspen suitability value by the cover suitability
value), a stand far away from mature aspen is assumed to have no value to grouse. This
is probably not the case. In fact, mature aspen was a high risk factor in the Maltby Hills
open site, probably because of its limited understory cover. Although aspen may be

preferable as winter food, other foods combined with adequate cover may be of similar

value.

155

 

An ideal model would take a remotely collected landscape cover, and
automatically run an evaluation to give an output of quality. Donovan et al. (1987) warns
researchers to consider the feasibility of generalizing habitat requirements so that
variables measured remotely accurately represent the species’ requirements. An
individual based model developed by Cary et al. (1992), for example, used land cover as
the sole measurement of quality. My results show that this generalization that may not be
appropriate as a predictor of the persistence of grouse. In different areas of the state (and
even within our study areas), though, the cover value of similar forest types can vary
greatly. The landscape cover analysis would have to be accompanied by ﬁeld
measurements like those in the existing model. A stand by stand evaluation is not
optimal for a large-scale project, but some ﬁeld measure of cover is needed for an
accurate model. If groundwork is not possible, perhaps existing compartment data could
be used to assess stem density, height, and presence of a winter food source.

These cover data could then be combined with a measure of interspersion, as in
the aspen model put forth by Rickers et al. (1995). This entails using a moving window
within a grid based GIS to measure and qualify the components of the area surrounding a
pixel. For example if a given pixel contained a forest type with high cover and a food
source and was surrounded by a mix of components that would enhance ruffed grouse
survival, it would receive a high quality value. Possible variables for inclusion in the
interspersion model are percent composition of aspen types, pines, upland hardwoods,
and lowlands.

Areas of young and medium aged aspen have been shown to support high

densities of grouse, and my results supported the continuation of traditional aspen

156

management for grouse. Birds with activity ranges primarily in aspen had a high
probability of survival. When qualifying interspersion, as the proportion of the
evaluation area in young and medium-aged aspen increased, so would the suitability
value. Mature aspen would only increase the quality value if it reached a certain
threshold of cover, in which case its value as a winter food resource would be
maximized.

Although old pine and upland hardwoods were avoided, they did not appear to
have a consistent negative effect on survival. Many grouse in the Pigeon River closed
site used mature pine stands regularly. The recommendations that pines can be improved
as grouse habitat should be explored. Pennsylvania researchers McCaffery et al. (1996)
suggest that stocking be decreased and herbicide use (or manual understory cutting) be
reduced to increase stem density and improve cover. In the Pigeon River, the removal of
groups of overstory trees may increase understory cover within pine stands. Oak stands
also have the potential for providing the life requirements for grouse, especially in the
Maltby Hills sites. Improvements in understory cover should be added to the already
present food source to increase their overall quality. The quality value of the presence of
certain hardwoods and pine in the interspersion evaluation would depend on their
quantity and cover quality.

Jack pine and young pine were preferred in many cases, their cover ranked high
by the existing model, and they did not seem to have a negative effect on survival. If a
winter food source was present, these types would earn a high interspersion suitability
value. Lowland hardwoods and conifers, though, were generally avoided and at times

were associated with an increased risk of mortality. Therefore if an area contained more

157

than some threshold of lowland hardwoods or conifers, it would be substantially
decreased in interspersion suitability value.

Two other spatial characteristics that may be considered important in the model
are continuity of quality habitat and distance from roads. Thus far I have proposed to
manage habitats that allow for the survival of non-dispersers. The model will assume
that after fall movements are done, dispersers will beneﬁt from the same habitat as non-
dispersers. But too often we forget that birds in the fall may move long distances (up to
15 km in my study). More and more, our forest management schemes create, isolated
cuts, instead of connecting smaller cuts across the landscape. Habitat management needs
to stretch beyond creating islands, and focus on the layout of forest types over a large
area. If fragmentation is a problem, it may be necessary to create corridors of continuous
high quality habitat that protect dispersers from predation.

Increasing the survival rate of juveniles may also be beneﬁcial. I found that
adults generally survived better than juveniles, and a disproportionate number of
juveniles were shot in the Pigeon River site. Fisher and Keith (1974) found that at a
greater distance from roads, birds had a higher probability of surviving the hunting
season. Small et al. (1991) added that birds in low hunting pressure areas actually
provide sources for other areas through emigration. Providing quality habitat away from
roads and closing logging trails to trafﬁc should be considered as measures to reduce
hunting pressure in some areas, thus increasing the survival of juvenile birds through the
fall. These variables may not be part of the evaluation process per se, but observed

during the planning process to ensure that placement of high quality habitat is of

maximum beneﬁt.

158

The ruffed grouse population will cycle even if hunting is eliminated (Bergerud
and Gratson 1988), and adequate habitat is provided (Gullion 1972). But the range of the
ruffed grouse has changed with the changing habitat in the northern United States
(Aldrich 1963). If ruffed grouse are a priority in Michigan, then the types of habitat
monitoring and improvements proposed are keys to long term persistence and continued

harvest of the species.

159

APPENDICES

160

 

APPENDIX A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a) ’g‘ 2000 _ - W_.___
E, 1500 — ~— ~—---—
8 1000 L- _-_.-
a?
500
g 0 T M
6/30 8/19 10/8 11/27 1/16
Date
b) g 2000 T—MHTUWWM’L
g 1500 _ _ __
D
E 1000
43
500 _- -
.2 O T"——-_"’ " 1" —— ”_' i 5" ”T 1‘7”.- _' ___.- _._____.T.._ 7" “_—_—J:
6/30 8/19 10/8 11/27 1/16
Date
°’ g 2000 MN.
g 1500 -____A__-----.-_- f __
a 1000 _____.___ ........ -- “--N--” -.-_------.__-._.
3
500 _-. “Ev . -21-"? _
g 0 J1 1 V T 1
6/30 8/19 10/8 11/27 1/16

Date

Figure A 1. Distance moved from the harmonic center of the week after capture
to the harmonic center of each following week for: a) Type I non-disperser, b) Type

II non-disperser, and c) Type III disperser.

161

APPENDIX B

 

 

 

1 .

0.75 "
E
a 0.5
35
u h
.5 0.25

0 i v i v i
0 4000 8000 12000 16000 20000

Equivalent stem density (stems/ha)

Figure B 1. Index value for equivalent stem density in the Michigan
model.

 

 

 

1 1
0.75 t

8
'3 0.5
>
x
g {D
5 0.25

O l t ‘r 1 ¥ . ‘v i 1

0 1 2 3 4 5 6 7 8 9
Deciduous stem height (m)

Figure B 2. Index value for average height of deciduous trees in
the Michigan model.

162

APPENDIX B

0.75 ‘-

Index value
.°
Ut

0.25 .

 

 

 

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Low conifer branch height (m)

Figure B 3. Index value for height above ground of lower coniferous
branches in the Michigan model.

 

 

 

1 W
0.751-
8
'5
x 0.5 ~
0
“U
.5
0.2511
0 .L : : : .L . ﬂ
0 1 2 3 4 5 6 7

Shrub height (m)

Figure B 4. Index value for average height of deciduous shrubs in
the Michigan model.

163

 

APPENDIX B

0.75 11

Index value
9
u.

0.25 1*

 

 

 

0 4 f 4 : .L
0 50 100 150 200 250 300 350

Distance to food source (m)

Figure B 5. Index value for distance between life requisites in the
Michigan model.

164

APPENDIX C

.82 an 5% coax doom: 2: co gases 3%: ._ o 2:5

 

165

8% 02 I
seesaw 2E: I

8.8 $.36 33 D

 

hov—

 

APPENDIX C

.Emmﬁo

 

83 oz -
Sega emf I

m

8.8 gone 33 D

 

hov—

QE .3; B83 5,2 oooma 2: Co 338% 5%: .N o 3:5

 

166

 

APPENDIX C

3km.

2

Key
Low quality (0.0)

I High quality (0.7)

D

 

 

 

- 1,; 'x ‘.
j l.- r

_J

/ .

' .

    

~41 £43521?! .
wars-2.:

        
      
  
 
  
   

  
   
  

‘3 I

4' -.
a r
‘ jli '. .
.—.- ~14“. *4 '~
.4. {Ail-.1 '

    

   
 
 
 
 

'5353‘1“ "i‘fu." .1» .
'Qﬁ.’:”~."¢ " '

      

- :1) , . . ,2
‘ 2&5? ’

 
 

167

Figure C 3. Habitat suitability of the Maltby Hills open si , 1993.

APPENDIX C

.Exm~_o

 

8% 02 I
Eagooo .3: I

E

8.8 £95 33 U

 

mov—

ée so. Bozo ma: 3%: 2: Co 338% “as“: .o o 2%;

 

 

168

LITERATURE CITED

169

LITERATURE CITED

Aldrich, J. W. 1963. Geographic orientation of American Tatraonidae. Journal of
Wildlife Management. 4:529 — 545.

Bergerud, A. T., and M. W. Gratson. 1988. Adaptive strategies and population ecology
of northern grouse. Vol. II. University of Minnesota Press, Minneapolis,
Minnesota.

Blouch, R. L. 1984. Northern great lakes states and Ontario forests. p. 391-410 Q L. K.
Halls (ed.) White-tailed Deer Ecology and Management. Stackpole Books,
Harrisburg, Pennsylvania.

Brewer, R., G. A. McPeek, and R. J. Adams, Jr. 1991 . The atlas of breeding birds of
Michigan. Michigan State University Press, East Lansing, Michigan.

Bump, G., R. W. Darrow, F. C. Edminster, and W. F. Crissey. 1947. The ruﬁ'ed grouse:
life history, propagation, management. The Holling Press, Inc., Buffalo, New
York.

Cade, B. S., and P. J. Sousa. 1985. Habitat Suitability Index Models: Ruffed Grouse.
U. S. Department of the Interior, Fish and Wildlife Service, Biological Report 82
(10.86). Washington DC.

Cary, J. R., R. J. Small, and D. H. Rusch. 1992. Dispersal of ruffed grouse: a large-
scale individual-based model. i_n D. R. McCullough and R. H. Barrett (eds.)
Wildlife 2001: populations. Elsevier Applied Science, Essex, England.

Clark, M. E. 1996. Movements, habitat use, and survival of ruffed grouse in northern
Michigan. Master’s Thesis, Michigan State University, East Lansing.

Chambers, R. E., and W. M. Sharp. 1980. Movement and dispersal within a population
of ruffed grouse. Journal of Wildlife Management 22:231-239.

Cox, D. R. 1972. Regression models and life tables (with discussion). Journal of the
Royal Statistical Society, Series B 34:187-220.

Dixon, K., and J. Chapman. 1980. Harmonic mean measures of animal activity areas.
Ecology 61:1040-1044.

Donovan, M. L., D. L. Rabe, and C. E. Olson, Jr. 1987. Use of geographic information
systems to develop habitat suitability models. Wildlife Society Bulletin 15:574-

579.

Domey, R. S. 1963. Sex and age structure of Wisconsin ruffed grouse populations.
Journal of Wildlife Management. 27:599-603.

170

Domey, R. S. 1966. A new method for sexing ruffed grouse in late summer. Journal of
Wildlife Management 30:623-625.

, and H. M. Mattison. 1956. Trapping techniques for ruffed grouse. Journal of
Wildlife Management 20:47-50.

Fearer, T. M., D. F. Stauffer, and R. L. Kirkpatrick. 1999. Relationship of ruffed grouse
home range size and movement to landscape characteristics in southwestern
Virginia. Report to the Virginia Department of Game and Inland Fisheries. 95pp.

Fisher, C. A., and L. B. Keith. 1974. Population responses of central Alberta ruffed
grouse to hunting. Journal of Wildlife Management 38:585-600.

Godin, A. J. 1960. A compilation of diagnostic characteristics used in aging and sexing
game birds and mammals. Page 156 iii S.C. Schemnitz (ed.) Wildlife management
techniques manual, 4th ed. The Wildlife Society, Washington, DC.

Gorrnley, A. 1996. Causes of mortality and factors affecting survival of ruffed grouse in
northern Michigan. Master’s Thesis, Michigan State University, East Lansing,
Michigan.

Graham, S. A., R. P. Harrison, and C. E. Westell. 1963. Aspens: Phoenix trees of the
Great Lakes region. The University of Michigan Press, Ann Arbor, Michigan.

Guglielmo, C. G., and W. H. Karasov. 1995. Nutritional quality of winter browse for
ruffed grouse. Journal of Wildlife Management 59:427 — 436.

Gullion, G. W. 1970. Factors inﬂuencing ruffed grouse populations. Transactions of
the North American Wildlife Natural Resources Conference 35:93-105.

. 1972. Improving your forested lands for ruffed grouse. Biddon Press, Inc.,
Rochester, New York.

. 1982. Rejuvenation and maintenance of forest habitats for the American ruffed
grouse. Pages 11-29 i_n T. W. Lovel, ed. Proceedings of the Second International
Symposium on Grouse. World Pheasant Association, Exning, England.

. 1984. Managing northern forests for wildlife. University of Minnesota
Agricultural Experiment Station, Saint Paul, Minnesota.

Hammill, J ., and R. Moran. 1986. A habitat model for ruffed grouse in Michigan. Pages

15-18 in J. Vernor, ed. Wildlife 2000. The University of Wisconsin Press,
Madison, Wisconsin.

171

Hammill, J ., and L. Visser. 1984. Status of aspen in northern Michigan as ruffed grouse
habitat. Pages 123-136 in W. L. Robinson (ed.) Ruffed grouse management: state
of the art in the early 1980’s. Northcentral Section, The Wildlife Society,

Bathesda, Maryland.

Johnson, E. P. 1990. Soil Survey of Ogemaw County, Michigan. U. S. Department of
Agriculture, East Lansing, Michigan.

Jones, K. C. 1996. GIS approach to evaluate a habitat suitability index (HSI) model
(breeding season) for bald eagles around Tippy Darn Pond, Manistee River,
Michigan. M. S. Thesis, Michigan State University, East Lansing, Michigan.

Kaplan E. L., and P. Meier. 1958. Nonparametric estimation from incomplete
observations. Journal of the American Statistical Association 53:457-481.

Karasek, G. L. 1998. Hunting results: Michigan small game seasons, 1992 — 1996.
MDNR Wildlife Division Report No. 3285.

Korschgen, L. J. 1966. Foods and nutrition of ruffed grouse in Missouri. Journal of
Wildlife Management 30:86-100.

Kubisiak, J. 1994. Ruffed grouse densities and population trends in central Wisconsin.
in Wisconsin Grouse Symposium Proceedings. University of Wisconsin bulletin.

Larsen, J. S., and R. D. Taber. 1980. Criteria of sex and age. Pages 143-202 1;) S. D.
Schemnitz (ed.) Wildlife management techniques manual, 4th ed. The Wildlife

Society, Washington, D. C.

Leatherberry, E. C., and J. S. Spencer. 1996. Michigan forest statistics, 1993. U. S.
Department of Agriculture, Forest Service Resource Bulletin NC-67.

Lee, E. T. 1992. Statistical methods for survival data analysis. John Wiley and Sons,
Inc., New York, New York.

McCaffery, K. R., J. E. Ashbrenner, W. A. Creed, and B. E. Kohn. 1996. Integrating
forest and ruffed grouse management: a case study at the Stone Lake Area.
Wisconsin Department of Natural Resources Technical Bulletin 189. Madison,

Wisconsin.

McDonald, J. E., G. L. Storm, and W. L. Palmer. 1998. Home range and habitat use of
male ruffed grouse in managed mixed oak and aspen forests. Forest Ecology and

Management. 109:271 - 278.

Morrison, M. L., B. B. Marcot, and R. W. Mannan. 1988. Wildlife-habitat relationships:
concepts and applications. University of Wisconsin Press, Madison, Wisconsin.

172

Mosher, J. A., K. Titus, and M. R. Fuller. 1986. Developing a practical model to predict
nesting habitat of woodland hawks. Pages 31 — 36 'm J. Verner, M. L. Morrison,
and C. J. Ralph (eds) Wildlife 2000. University of Wisconsin Press, Madison,

Wisconsin.

Neher, Lynn N. 1993. Winter movements and macrohabitat selection of ruffed grouse in
central Missouri. MS Thesis. University of Missouri, Columbia, Missouri, USA.

Palmer, W. L., and C. L. Bennett. 1963. Relation of season length to hunting harvest of
ruffed grouse. Journal of Wildlife Management 27:634-639.

Paradis, E., and H. Croset. 1995. Assessment of habitat quality in the Mediterranean
pine vole (Microtus duodecimcostatus) by the study of survival rates. Canadian
Journal onoology 73:1511-1518.

Rickers, J. R., L. P. Queen, and G. J. Arthaud. 1995. A proxirnity-bascd approach to
assessing habitat. Landscape Ecology 10:309-321.

Roloff, G. J. 1994. Using an ecological classiﬁcation system and wildlife habitat models
in forest planning. PhD Dissertation, Michigan State University, East Lansing,
Michigan.

., J. B. Hauﬂer, and M. C. Reynolds. 1994. Variability in the structure and
composition of regenerating aspen with implications for ruffed grouse
management. Michigan Academician. XXVI:75-82.

Rosner, B. 1990. Fundamentals of biostatistics. Wadsworth, Inc., Boston.

Roussel, Y., and R. Ouellet. 1975. A new criterion for sexing Quebec ruffed grouse.
Journal of Wildlife Management 39:443-445.

Rusch, D. H., S. DeStefano, and R. J. Small. 1983. Seasonal harvest and mortality of
ruffed grouse in Wisconsin. in W. L. Robinson, ed. Ruffed grouse management;
State of the art in the early 1980’s. BookCrafters, Chelsea, Michigan.

Rusch, D. H., and L. B. Keith. 1971a. Ruffed grouse-vegetation relationships in central
Alberta. Journal of Wildlife Management 35:417-429.

Rusch, D. H., and L. B. Keith. 1971b. Seasonal and annual trends in numbers of Alberta
ruffed grouse. Journal of Wildlife Management 35:803-822.

Schamberger, M. L. and L. J. O’Neil. 1986. Concepts and constraints of habitat-model
testing. Pages 5 - 10 i_n J. Verner, M. L. Morrison, and C. J. Ralph eds. Wildlife
2000. University of Wisconsin Press, Madison, Wisconsin.

173

Siegel, S., and N. J. Castellan, Jr. 1988. Nonparametric statistics for the behavioral
sciences. Second ed. McGraw-Hill, Inc., New York, NY.

Slather, B., S. Engen, and R. Lande. 1996. Density dependence and optimal harvesting
of ﬂuctuating populations. Oikos 76:40-46.

Small, R. J ., J. C. Holzwart and D. H. Rusch. 1991. Predation and hunting mortality of
ruffed grouse in central Wisconsin. Journal of Wildlife Management 55:512-520.

Small, R. J ., and D. H. Rusch. 1989. The natal dispersal of ruffed grouse. Auk. 106:72-
79.

Stoll, R. J ., and W. L. Culbertson. 1995. Ruffed grouse hunting pressure and harvest on
an Ohio public hunting area. Ohio Department of Natural Resources Technical
Bulletin No. 12. Columbus, Ohio.

Tardy, S. W. 1991. Soil Survey of Cheboygan County, Michigan. U. S. Department of
Agriculture, East Lansing, Michigan.

Theberge, J. B., and D. A. Gauthier. 1982. Factors inﬂuencing densities of territorial
male ruffed grouse, Algonquin Park, Ontario. Journal of Wildlife Management.
46:263-268.

Thompson, F. R. 111. 1987. The ecology of the ruffed grouse in central Missouri. Ph. D.
Dissertation. University of Missouri, Columbia, Missouri.

VanHorne, B. 1983. Density as a misleading indicator of habitat quality. Journal of
Wildlife Management 47:893-901.

Vispo, C. R., J. R. Cary, D. J. Lauten, and C. Balzer. 1994. Assessment of habitat
quality in a habitat mosaic: an example with ruffed grouse. Pages 470-474 i_n J.
Bissonette and P. Krausman, eds. Integrating people and Wildlife for a
sustainable future: Proceedings of the International Wildlife Management
Congress. The Wildlife Society, Bethesda, Maryland.

Williams, T. E. 1998. Soil Survey of Alcona County, Michigan. Natural Resources
Conservation Service, Washington, D. C.

Winterstein, S. R., K. H. Pollock, and C. M. Bunck. in press. Analysis of Survival data
from radio telemetry studies. '

Worton, B. J. 1989. Kernel methods for estimating the utilization distribution in home-
range studies. Ecology 70:164-168.

Zbinden, N., and M. Salvioni. 1997. Hunting of black grouse Tetrao tetrix in the Canton
of Ticino. Ornithologische Beobachter 94:331-346.

174