LIBR \RY
Michigan State

University

 

 

'This is to certify that the

thesis entitled

Are There "Hot Spots" ochovine Tuberculosis
In The Free-Ranging White-Tailed Deer

(Odocoileus virginianus) Herd of Northeastern
Michigan?

presented by

Brandi Daneille Hughey

has been accepted towards fulfillment
of the requirements for

M.S. degree inFisheries & Wildlife

S (4‘)“ V'VI a ”C \ WE .I/L'x \AA/x

Major professor

0.7639 MSU is an Affirmative Action/Equal Opportunity Institution

H" fi' '
,__..__— ——~——-

 

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

 

UMWZ '05

 

 

 

 

 

 

 

 

 

 

 

 

 

 

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

ARE THERE “HOT spore" OF BOVINE TUBERCULOSIS IN THE FREE-
RANGING WHITE-TAILED DEER (ODOCOILEUS VIRGINIANUS) HERD OF
NORTHEASTERN MICHIGAN?

By

Brandi Danielle Hughey

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
For the degree of

MASTERS OF SCIENCE
Department of Fisheries and Wildlife

2003

ABSTRACT
ARE THERE “HOT SPOTS” OF BOVINE TUBERCULOSIS IN THE FREE-
RANGING WHITE-TAILED DEER (ODOCOILEUS VIRGINIANUS) HERD OF
NORTHEASTERN MICHIGAN?
By
Brandi Danielle Hughey
This project looks at whether high case frequency areas or “hot spots” of
bovine tuberculosis (tb) exist in free ranging white-tailed deer in northeastern
Michigan, and examines the factors associated with them. Michigan Department
of Natural Resources researchers have been collecting deer heads annually
since 1996 in the five county area of Alcona, Alpena, Montmorency, Oscoda, and
Presque Isle as part of an effort to manage an outbreak of bovine tb in the wild
white-tailed deer population. From these heads a database containing age, sex,
and harvest or collection location information was constructed. The townships
within these five counties were categorized as having zero case frequency,
medium case frequency, or high case frequency. These categories are based on
the number of years at least one tb infected deer was detected in that township.
Each case frequency category was then examined individually looking at the
yearly sample size distributions, yearly case frequency, and cumulative sample
size distributions. Using GlS these areas of varying case frequency were
compared to five deer use categories, Summer Use (high quality summer
habitat), Summer Other (poor summer habitat), Winter Use (high quality winter

habitat), Winter Other (poor winter habitat), and Rare Use (includes areas rarely

or never used by deer, such as water, urban and industrial areas).

ACKNOWLEDGEMENTS

Funding was provided by the MDNR Wildlife Division and carried out in
conjunction with the Partnership for Ecosystem Research and Management
(PERM) between the MDNR and MSU, The Michigan Agricultural Experiment
Station, The Native American Institute at Michigan State University, and AISES
(American Indian Science and Engineering Society).

I would especially like to thank Dr. Scott Winterstein, my major professor
for his patience, understanding and guidance, particulariy during my first
semester of graduate school. Without it, I would have never made it this far.

Thanks to Drs Joseph Messina and Kelly Millenbah (committee members)
for answering my endless GIS questions, and giving encouragement when I
needed it most.

I would also like to thank the MDNR for providing the data for this research
and technical assistance. More specifically Drs Steve Schmitt and Daniel
O’Brien for agreeing to be on my committee, and Jean Fierke and Stephanie
Hogle for answering database and GIS questions.

Thanks to Erin Kelley my research assistant, who provided much
appreciated assistance with literature searches, checking the databases, and
numerous other tasks I assigned her.

A very special thanks to all my family and friends for their love,

encouragement and understanding. Mom, Dad, and Brother for understanding

when I could not visit as often as I would have liked. Thanks to Corina for not
letting the fame and fortune of being a graduate student go to my head.
Special thanks to Devine intervention and inspiration, without it, this thesis

would not have been written.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

LIST OF TABLES ................................................................................. vii

LIST OF FIGURES ................................... viii

CHAPTER 1

INTRODUCTION ................................................................................... 1

OBJECTIVES ....................................................................................... 6

STUDY SITE ....................................................................................... 6

CHAPTER 2

METHODS ............................................................................................ 8
Databases ................................................................................... 8
GIS .......................................................................................... 12
Case F reggencv Cgverages ........................................................... 12
Original Case Frequency Coverage ................................................. 12
Tb+ Deer Point Coverage ............................................................. 1 5
Home Range Case Frequency Coverage .......................................... 15
New Case Frequency Coverage ..................................................... 19
Deer Use Grid/Coveraga ............................................................... 20
Kernel Core Am ........................................................................ 22
MCP Core Area .......................................................................... 22
_Seasonal Home Ranges & Habitat Use ............................................ 22
Statistical Analyses ...................................................................... 24

' Databases ................................................................................. 24

9g .......................................................................................... 24

CHAPTER 3

RESULTS .......................................................................................... 27
Databases ................................................................................. 27
GIS ........................................................................................... 32
Case Freguencv Coverage Analyses .......................................... ' ..... 32
Core Area Analyses ..................................................................... 36
_Sflsonal Home Ranges & Habitat Use ............................................ 38
Rare Use ................................................................................... 38
Summer Use .............................................................................. 39
Winter Use ................................................................................. 41
Summer Other ............................................................................ 42
Winter Other .............................................................................. 43

CHAPTER 4

DISCUSSION ...................................................................................... 45
Case Frequency vs. Prevalence ..................................................... 45
Databases ................................................................................. 45
GIS .......................................................................................... 47

Case Frequency Coverages ........................................................... 47

 

 

 

 

 

 

 

 

 

Moran’s I ................................................................................... 47
Core Area .................................................................................. 48
Habitat Use ................................................................................ 48
Case Freggencv and Scale ............................................................ 50
CONCLUSIONS AND RECOMMENDATIONS TO THE MDNR ..................... 53
APPENDICES
CHANGES MADE TO TB DATABASE ...................................................... 55
METADATA ........................................................................................ 56
_O_rjginal Cfie Freggencv Coveraga ................................................. 56
New Case Frequency Coverage ..................................................... 60
Home Range CaseFredgencv Coverage ......................................... 64
REGIONS .......................................................................................... 68
RECLASSIFICATIONS OF LAND USE TO DEER USE ................................ 69
CHI-SQUARE ANALYSES ..................................................................... 70
Home Range Case F reggencv ....................................................... 70
Kernel Core Area ........................................................................ 71
MCP Core Area .......................................................................... 72
LITERATURE CITED ........................................................................... 73

vi

LIST OF TABLES

Table 1. Record of missing (Q) or incorrect data ......................................... 9
Table 2. Composition of TB Database .................. . ................................... 10
Table 3. Composition of Suspicious Tissue Database ................................. 10

Table 4. Estimated Home Ranges and Their Averages Based on Age and Sex
Categories .......................................................................................... 19

Table 5. Chi-Square Tests Comparing the Number of Deer Tested for Tb by Sex
among the 3 Case Frequency Areas ....................................................... 30 -

Table 6. Chi-Square Tests Comparing Sample Sizes by Sex Between Hot Spots

vs. Not Hot Spots ................................................................................ 31
Table 7. Chi-Square Tests for Biases by Sex in the Number of TB+ Carcasses
vs. Heads Turned in by Hunters ........................................................... 32
Table 8. Summary of MANOVA and Tukey's HSD Analyses for Differences of %
Deer Use Among Case frequency Categories ........................................... 33
Table 9. Summary of ANOVA Testing for Differences in Habitat Use .............. 40
Appendix 4. Reclassifications of Land Use to Deer Use ....................... 69

Appendix 5a. Chi-Square Tests Comparing Deer Use Categories in the 3 Case
Frequency Areas for the Home Range Case Frequency Coverage ........ 70

Appendix 5b. Chi-Square Tests Comparing Deer Use Categories in the 3 Case
Frequency Areas for the Kernel Core Area ....................................... 71

Appendix 50. Chi-Square Tests Comparing Deer Use Categories in the 3 Case
Frequency Areas for the Minimum Convex Polygon Core Area ............. 72

vii

LIST OF FIGURES

Figure 1. Map Showing Focal Areas of Baiting and Feeding Regulations .......... 4
Figure 2. Study Area: Alcona, Alpena, Montmorency, Oscoda, and Presque Isle
Counties .............................................................................................. 7
Figure 3. Data Flow Diagram for Original Case Frequency Coverage ............. 13
Figure 4. Original Case Frequency Coverage for the Five County Area ........... 14
Figure 5. Data Flow Diagram for Tb+ Deer Point Coverage .......................... 15

Figure 6. Data Flow Diagram for Home Range Case Frequency Coverage......16
Figure 7. Home Range Case Frequency Coverage for the Five County Area....17
Figure 8. Data Flow Diagram for New Case Frequency Coverage .................. 20
Figure 9. New Case Frequency Coverage for the Five County Area ............... 21

Figure 10. Kernel and MCP Core Areas Overlayed onto the Home Range Case
Frequency Coverage ............................................................................ 23

Figure 11. Examples of how habitat use was calculated .............................. 26

Figure 12. Example of How Percent Points, Percent Area, and Index Were
Calculated .......................................................................................... 26

Figure 13. Percent Sample Size Categories for All Case Frequency Distributions
1996-2000 .......................................................................................... 28

Figure 14. Sample Size Distribution of Deer Tested for TB in All Townships of
the Five County Area 1996-2000 ............................................................. 28

Figure 15. Percent Sample Size Categories for High Case Frequency
Townships .......................................................................................... 29

Figure 16. Distribution of Sample Sizes for High Case Frequency
Townships .......................................................................................... 29

Figure 17. Distribution of Sample Sizes of Sections within High Case Frequency
Townships .......................................................................................... 30

Figure 18. Percentage of Deer Use Categories for Original Case Frequency
Coverage ........................................................................................... 34

viii

Figure 19. Percentage of Deer Use Categories for New Case Frequency

Coverage ........................................................................................... 34
Figure 20. Percentage of Deer Use Categories for Home Range Case

Frequency Coverage ............................................................................ 35
Figure 21. Percentage of Deer Use Categories for Kernel Core Area ............. 37
Figure 22. Percentage of Deer Use Categories for MCP Core Area ............... 38
Figure 23. Rare Use Interaction Between TB Areas & Migratory Status .......... 39

Figure 24. Summer Use Interactions between TB Area and Migratory Status ..41
Figure 25. Winter Use Interaction Between TB Areas 8 Migratory Status ........ 42
Figure 26. Summer Other Interaction Between TB Areas & Migratory Status...42
Figure 27. Winter Other Interaction Between TB Areas & Migratory Status ...... 43
Figure 28. Percent Deer Use in the No TB, TB, and 5 County Areas .............. 44

Appendix Figure 1. Example of the Overlapping Feature for Regions Data
Model ................................................................................................ 68

Appendix Figure 2A. Example of the Home Range Case Frequency Coverage
With Polygons. 2B. Example of the Home Range Case Frequency Coverage
With Regions ....................................................................................... 68

CHAPTER 1
INTRODUCTION

Bovine tuberculosis (tb) (Mycobacterium bovis) is a bacterium that
generally attacks the respiratory system of mammals. It is most easily
transmitted through aerosol created by coughing or sneezing of an infected
individual (Michigan Department of Natural Resources, 2000). Historically, tb in
Michigan has been associated with cattle. However, in 1975, a hunter-harvested
9.5 year old white-tailed doe (Odocoileus virginianus) in Alcona County tested
positive for tb. This was thought to be an isolated incident, and no further action
was taken. In 1994, a 4.5 year old buck was harvested in Alpena County, and
tested tb positive (Schmitt et al., 1997). Since Michigan had been considered a
tb free state since 1979 this caused concern. In the spring of 1995, the Michigan
Department of Agriculture (MDA) surveyed all livestock within a 10 mile radius of
the tb positive deer. None of the livestock were found to be tb positive. In the fall
of 1995, the Michigan Department of Natural Resources (MDNR) surveyed
hunter-harvested deer within a 10 mile radius of where the tb positive deer was
harvested. Of the 354 deer harvested and tested, 18 cultured positive for tb
(Fitzgerald et al., 1997).

Tb posses a threat to animal health, as well as the economic success of
agriculture and wildlife industries of Michigan. For example costs to the
agricultural industry, due to the loss of Michigan’s tb Accredited-Free State
status, are anticipated to be around $16 million a year (Whitcomb, 1999). In an

effort to better understand and manage this outbreak of bovine tb, MDNR

beginning in 1996 systematically expanded their surveillance area, eventually
including the entire state of Michigan, with a focus on the five counties of Alcona,
Alpena, Montmorency, Oscoda, and Presque Isle. As of February 12, 2003, 449
white-tailed deer, 41 carnivores, 1 captive deer herd, 28 beef and dairy herds, 2
elk, 1 feral cat and 1 human have tested positive for tb within the state. The vast
majority of these animals were located in the “core” area of Alcona, Alpena,
Montmorency, and Oscoda Counties (Michigan Department of Natural
Resources, 2003). Because of the nature of the findings, it is believed that
bovine tb is endemic to the white-tailed deer in this region. This is the first
occurrence of tb maintained within a free-ranging white-tailed deer population
(Schmitt et al., 1997). As a result, Michigan lost its tb Accredited-Free State
status in 2000 (Michigan Department of Natural Resources, 2000).

Bovine tb is hypothesized to be prevalent in the tb core area is because of
the combined effects of high deer densities, baiting and feeding. Deer densities
are beyond that which the natural environment can support increasing the
contact rate between deer. This area is known as “club country” with numerous
privately owned clubs ranging from 40 - 28,000 acres in size (Fitzgerald et al.,
1997). Extensive baiting and feeding and mild winters have helped to lower
winter mortality, while the reluctance to harvest antlerless deer has led to a
higher ratio of does to bucks. Consequently, this has contributed to the high deer
densities in this particular area of Michigan. Increased numbers of deer
concentrated into smaller areas, increases the probability of transmitting tb

(O’Brien et al., 2002).

Historically the amount of “good” habitat in the tb core area has been
limited by the poor quality of the soil for traditional agricultural row crops and
vegetation in this area. Therefore, many Iandownersresorted to feeding and
baiting (a cheaper alternative to habitat improvement) to make up for habitat
deficiencies, and to maintain deer on their property (Peyton, 2000). Probable
pathways of transmitting tb are through face-to-face contacts (Garner, 2001 ), and
consumption of contaminated feed (Michigan Department of Natural Resources,
2001). To decrease the number of face-to-face contacts and the consumption of
contaminated feed the Michigan Department of Agriculture banned feeding and
the Natural Resource Commission restricted baiting to a 5-gallon bucket
maximum in 1998 (Michigan Department of Natural Resources, 1999). The ban
was applied to an area east of I-75 and north of M-55, which consists of the
aforementioned five county area and a portion of the surrounding counties
(Figure 1). Garner (2001) showed that decreasing the amount of bait actually
increased the number of face-to-face contacts, because deer were crowding into
a smaller area. It is thought that deer crowd into these smaller areas because
they have learned that they need to be the first ones to feed, or they would miss
out on a free meal. In 1999-2000 both baiting and feeding were banned in this
area. In 2001, the MDNR changed its regulations to allow 1 gallon of grain or
corn per day to be used October 1 to November 30 in Deer Management Unit
(DMU) 452. According to published guidelines, the bait had to be spread over an

area at least 100 square feet. The MDNR changed their ban on baiting because

 

 

 

 

- Core Area

 

/\/ Highways .

[:3 Counties ‘

  

 

 

Charlevoix

  

Antrim

  

   

 

 

 

Presque Isle

   
 

 

 

 

Montmorency

  

 

Ifi

  

  

 

Crawford

 

   

 

 

 

 

 

 

 

Kalkaska I 75 Oscoda Alcona
Ogemaw
L. M'55
Missaukee Roscommon i i l f
40 o 40 80 Kilometers

 

 

Figure 1. Map Showing Focal Areas of Baiting and Feeding Regulations

  
        

 

there was a decrease in bow hunter participation, and some managers were
concerned that the ban was negatively impacting the efforts to eradicate tb
(Michigan Department of Natural Resources, 2001). In 2002-2003 baiting was
banned in the counties of Alcona, Alpena, Crawford, Montmorency, Oscoda,
Otsego, and Presque Isle (Figure 1). In the remainder of the state baiting was
restricted to a maximum of two gallons in a 100 square feet area at a hunting
site. Baiting is allowed in Michigan (except for the previously mentioned 7
counties) October 1 through January 1(MDNR, 2002).

In an effort to manage this outbreak of tb in the wild deer herd, MDNR
researchers have been collecting data annually, since 1996, on the deer heads
collected for tb monitoring in the five county area of Alcona, Alpena,
Montmorency, Oscoda, and Presque Isle. This study will look at whether high
case frequency areas (defined by both political and biological boundaries) or “hot
spots” of tb exist in free-ranging white-tailed deer in northeastern Michigan, and

examine the ecological factors associated with these hot spots.

OBJECTIVES
The main objectives of this research project are to:
1) Determine if “hot spots” exist by examining the tb database for
accuracy and biases,
2) Determine whether a correlation exists between these high case
frequency areas of tb, and selected ecological factors, and
3) Make recommendations to the MDNR.
STUDY SITE
The study area includes the five counties of Alcona, Alpena,
Montmorency, Oscoda, and Presque Isle (Figure 2). The average minimum
temperature ranges from —11.9 C° in the north to -9.6 C° in the south. The
average maximum temperature ranges from 23.4 C° in the north to 25.7 C° in the
south (Knapp 1988, Williams 1992). The average snowfall is about 175 cm and
rainfall is approximately 72.5 cm (Sitar 1996). The topography of the area was
formed by glacial deposits, and is characterized by level and undulating plains
and rolling to hiIIic moraines. The region is dominated by wooded/forested land,
with farmland being the second most prevalent land type. The elevation ranges

from 150 m to 390 m above sea level (Williams 1992).

 

 

 

 

 

Presque Isle

Montmorency

 

Oscoda

 

20 0 20 40 Kilometers Source: MDNR

 

 

Figure 2. Study Site: Northeastern Lower Peninsula of Michigan

CHAPTER 2
METHODS

Databases

MDNR researchers have been collecting data annually on the deer heads
collected for tb monitoring in the five county area of Alcona, Alpena,
Montmorency, Oscoda, and Presque Isle from 1996-2002. This study examined
data collected from 1996-2000. These data were organized into two databases,
the tb database and suspicious tissue database. The tb database contains test
records for individual deer, consisting of survey type (deer obtained from hunt or
non-hunt origin), tb identification number, tb test result (negative or positive), age
(aged on the half year), sex, and harvest location (county, township, and section)
information. The suspicious tissue database contains data fields for tissue type
(head or carcass), survey type (hunt or non-hunt), tb identification number, tb test
result (negative or positive), age (aged on the half year), sex, and harvest
location information. A Q in any of the data fields indicated no data present.

The first step in determining whether “hot spots” exist is to examine the
database and check for errors and detectable biases. Error checking includes
examination of data consistency, and correctness of data. For example some
sections were listed as _2, 2_, or 02; for data consistency all numbers were
changed to 02. Location errors include counties that have townships listed,
which are not located in that county. The MDNR was contacted to see if there

were any corrections for the location errors, and changes were made to the

database. All changes to the tb database are listed in Appendix 1. If there were
no corrections available then the record was deleted (Table 1).

Table 1. Record of MlsslngLQ) or Incorrect Data

 

Date Category Data Type Number % of Total
Section Incorrect 2 0.01 %
Town Incorrect 401 1 .07%
Range Incorrect 407 1 .09%
Sex Q 88 0.24%
Age Q 90 0.24%
Section Q 764 2.04%
Town Q 459 1.23%
Range Q 453 1.21%

 

860 2.30%

Total Deleted Records
Bias checking was done by looking at sample sizes of each township by

sex and age (1 1/2, 21/2, 3‘/2, 4V2”) categories, and comparing the distributions of
each of these categories. Sample sizes of at least 300 are required to detect 2
1% prevalence with a 95% confidence interval (Schmitt et al., 1997). In addition,
to determine if there was a positive bias in the number of carcasses turned in due
to visible lesions (i.e. are hunters more likely to turn a deer into a check station if
there were lesions visible in the chest cavity?), the number of tb+ carcasses and
heads turned in during the hunt season were examined. For the analyses, 860
records were deleted from the tb and suspicious tissue databases because of
incorrect or missing location information. For the tb database, this is 2.30% of
the total records in the tb database (Table 1). No tb+ animals were deleted from
either database. This left 29,078 usable records for the five county area. There
are 356 records from 1994-1995, or 1.22% of the 29,078 records, which were
excluded. Fawns comprise 5.32% of the total usable records and were also

excluded from the current analysis. Records from 1994-1995 were excluded

because sampling was not over the entire five county area. Fawns were
excluded because they were thought too young to exhibit signs (visible lesions)
of tuberculosis, and were only collected by mistake or if the hunter insisted.
There were 27,085 records that contained all the correct location, sex and age

information from 1996-2000. This will be the basis for the sex/age analysis

 

 

 

discussed herein (Table 2).
Table 2. Composition of TB Database
Data Category Number % of Total
1994-2000 records w/correct inforrnatlon 29078 100.00%
1994-1995 Records 356 1.22%
1975 Record 1 0.00%
1996-2000 Missing Sex/Age 90 0.31%
Fawns 1546 5.32%
1996-2000 records w/correct information 27085 93.15%

For the suspicious tissue database, there are 29,071 records (Table 3).
There are 27,066 records that contain all the correct information from 1996-2000.
There are 355 records from 1995, which comprises 1.22% of the database.

Fawns consist of 5.30% of the total suspicious tissue database.

Table 3. Composition of Suspicious Tissue Database

 

 

 

Data Category Number % of Total
Usable Head (1996-2000) 26540 91.29
Usable Carcass (1996-2000) 526 1.81
Unknown Head/Carcass 15 0.05
Unknown Sex/Age (1996-2000) 93 0.32
Fawns(1996-2000) 1542 5.30
1995 Head/Carcass 355 1.22
Total 29071 100.00

 

10

Females comprise 78.43% of the non-hunt data, and 46.6% of the hunt
data, while males make up 21.25% of the non-hunt data, and 53.17% of the hunt
data.

A new column, case frequency, was added to the tb database, which
denoted the case frequency category of each township. Case frequency
categories are defined by the number of years a tb positive animal has been
harvested in that township. Zero case frequency townships are defined as never
having a tb+ animal in that township. Medium case frequency is defined as
having at least one tb+ animal in up to each of two years in that township. High
case frequency is defined as having at least one tb+ animal in each of three or
more years in that township. There are 59 townships classified as zero case
frequency townships, 21 townships are medium case frequency, and 20
townships are high case frequency. Case frequency differs from prevalence.
Prevalence is calculated as the number of tb+ deer present in any variously
defined area divided by the total number of deer tested in that area.

For each case frequency category, sample size distributions were
determined yearly and cumulatively using 1996-2000 data. Yearly data were
broken into five distribution categories: n = 0, 1-25, 26-50, 51-100, or > 100 deer
checked. Cumulative data were also broken into five distribution categories, n =
0, 1-100, 101-200, 201-300, or > 300. This is based on a sample size of at least
300 being required to detect a 2 1% prevalence with a 95% confidence interval
(Schmitt et al., 1997). Sex and age distributions were classified into their

respective case frequency category.

11

GIS

County, township, section coverages, and Northern Lower Peninsula 1993
land cover grid were downloaded from the Michigan Geographic Data Library
(http://www.state.mi.us/webapp/cgi/quI/).

Case FreggencLCoveragfi

Three case frequency coverages were created: original case frequency
(using politically defined boundaries), home range case frequency (using
biologically defined boundaries), and new case frequency (using a combination
of political and biological boundaries) (Appendix 2).

Original Case Frequency Coverage

The tb database provides town and range information for dead deer. In
ArcViewTM a shapefile for zero, medium and high case frequency townships was
created by linking the tb database to the township coverage by. the town and
range column. ArclnfoTM was used to clean and build the coverage (Figures 3

and 4).

12

 

 

Obtain Daatabases
Coverages from DNR

TB Township
Database
Bias/Error
Checking feline TgwnsD

 

 

 

 

 

 

 

 

 

 

l Cleaned Coverage
l
Final I Join Database to
TB Database Township Coverage

 

 

 

 

 

Task performed in ArcView
Task performed in Arclnfo

 

 

 

Figure 3. Data Flow Diagram for Original Case Frequency Coverage

13

 

 

 

 

LEGEND
Case Freguency # Townships

High 20
Medium 21

- Zero 56

Source: Michigan Department at Natural Resources

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Z) 0 Z) 4) Klaraers

 

Figure 4. Original Case Frequency Coverage for the Five County Area

14

 

Tb+ Deer Point Coverage

The tb database gives locations of deer to the section (this is not a point
location). To have actual point locations for which to ”generate the home range
case frequency coverage, the tb+ deer point coverage was constructed. To
create a coverage for tb+ deer point locations (Figure 5) several steps were
taken. First, a table of tb+ deer locations was created, and linked to the section
coverage. Point locations were created in ArclnfoTM by using the create/abel,
centroidlabel (places the location of the tb+ deer in the center of the section in

which it was harvested), and build (creates topology) commands.

Final Jorned 4— Section
TB Database—T add9d& column Coverage

V

 

 

 

 

 

 

 

New
Coverage

/\/

1 -2000
Graded, centered, , -996

Pornt Coverage
built point coverage V

Task performed in ArcView
Task performed in Arclnfo

 

 

 

 

 

 

Figure 5. Data Flow Diagram for TB+ Deer Point Coverage

Home Range Case Frequency Coverage
Home range data for deer in this area, provided by previous
radiotelemetry studies (Garner 2001, Muzo unpublished data, Sitar 1996), were

incorporated to create the home range case frequency coverage (Figures 6 and

15

7). The reasoning behind this methodology is that deer are not stationary

objects. Point locations by themselves take into account only one instance in

 

 

 

Queried Out
Sex/Age
Cateogon'es

J ' ed B ff C ' .
0'" u 2' overage 4- R :g'gount <- regroncount" <—
Regioncount Table‘
Unioned All Sex/Age Added Column,
_, Buffer Coverages Dissolved
w/Townshrp Coverage w/Patch Analyst

I
Task performed in ArcView ‘_ C 9&3"
regionpoly

Task performed in Arclnfo
Figure 6. Data Flow Diagram of Home Range Case frequency Coverage

 

 

 

 

 

 

 

 

 

 

     

       

 

 

Buffer

 

Case
Frequency
Coverag
‘Task performed on each individual sex/age cover\/

 

 

 

time,’and for example if that deer had been harvested a day earlier, it may have
been in an entirely different township. Using home range estimates takes into
account that deer are mobile and that one fixed location (i.e. harvest site) does
not accurately reflect the ecology of deer. This study does not take into account
that approximately 19% of the deer in this region are migratory (Garner 2001),
however a conservative estimate can be deduced. Migratory deer are defined as
having separate summer and winter home ranges, that are at least 1 km apart
(Sitar 1996). The complexity of incorporating the direction and distance of

migration into the spatial modeling is beyond the scope of this project.

16

 

LEGEND
Case Fregyuenc

, High
Medium

   

Source: Michigan Department of Natural Resources

 

 

 

 

 

 

20 0 20 40 Kilometers

 

 

 

 

Figure 7. Home Range Case Frequency Coverage for the Five County Area

17

Home range sizes were calculated and averaged within age and sex
categories. In ArcViewTM, using the animal movement extension (Hooge 2000),
home ranges were determined by the kernel method with a 95% probability
contour and a least-squares cross-validation (LSCV) choice of h (smoothing
parameter). Age categories were females 1‘/2, 2% and 31/2" and males 1%: and
272*; each category was a separate coverage. The age categories differed
between the tb database and GIS evaluations. This difference was due to
inadequate sample sizes from radio-collared deer. For this study, non-migratory
home range estimates were used for the females, and both migratory and non-
migratory season-specific home range estimates were used for the males. This
difference was again due to the small sample size of radio-collared male deer.
This gives a conservative approximation of area utilized by deer in this region.

The average home range area was used to calculate a radius, which was
usedto create a sex/age specific buffer around the tb positive deer locations
(Table 4). For each point coverage a new buffer coverage was created with the
regionbuffer (for more information on regions see Appendix 3) command in
Arcinfom. Then using regioncount, a table was created and joined to each
coverage. In ArcViewTM, each coverage was combined using union, and an
additional case frequency column was placed in the attribute table, which added
all polygons that intersected. Using this new column, case frequency was
determined for each polygon using the same criteria as were used to determine
the original case frequency. Zero case frequency is defined as never having a

tb+ animal in that polygon. Medium case frequency is defined as having one or

18

two tb+ animals in that polygon. High case frequency is defined as having three
or more tb+ animals in that polygon. Once this was completed the shapefile was
dissolved by the case frequency column using the patch analyst extension (Elkie
et al. 1999), and was clipped to fit the five county area using xtools extension
(DeLaune 2001). Finally, the home range case frequency coverage was cleaned

in Arcinfom.

Table 4. Estimated Home Ranges and Their Averages Based on Aggand Sex Categories

 

Home Range

 

 

Sex Age Radizus Sample )2 "will? Mtgmgm
(m ) 8120 (m2) (m2) (m2)
Female 1.5 1,706 10 9,141,251 1,406,601 23,399,887
Female 2.5 1,347 20 5.700.450 1,119,122 25,541,237
Female 3.5+ 1,884 35 11,143,857 668,693 141,231,996
Male 1.5 2,064 19 13,380,289 641,953 89,628,439

Male 2.5+ 2,813 5 24,853,722 6,523,754 31,109,171

New Case Frequency Coverage

In ArcViewTM the new case frequency coverage was created by overlaying
the home range case frequency coverage over a copy of the original case
frequency coverage. A new case frequency column was added to the original
case frequency attribute table, which categorized townships based on the
number of individual home ranges. Zero case frequency townships are defined
as never having a tb+ home range in that township. Medium case frequency is
defined as having one or two tb+ home ranges in that township. High case

frequency is defined as three or more tb+ home ranges in that township.

19

ArclnfoTM was used to clean and build the New Case frequency Coverage

(Figures 8 and 9).

 

 

Original
Case

Frequency

Add
New Column

 

 

  

 

Case Frequency

Task performed in Arc View
Task performed in Arclnfo

 

 

 

 

 

 

Figure 8. Data Flow Diagram for New Case frequency Coverage

' Deer Use Grid/Coverage

In ArcViewTM, the NLP 1993 land cover grid was reclassified into five deer
use categories (see Appendix 4), Summer Use (high quality summer habitat),
Summer Other (poor summer habitat), Winter Use (high quality winter habitat),
Winter Other (poor winter habitat), and Rare Use (includes areas rarely or never
used by deer, such as water, urban and industrial areas). These deer use
categories are based on habitat use literature (Kohn et al., 1971, McCaffery et
al., 1974, Rogers et al., 1981, Stonner et al., 1980, Van Deelen et al., 1996, and
Davenport, 1941 ). In cases where land cover types overlapped in seasonal use,
the more frequently used season was applied. For the habitat use analysis, the

deer use grid was converted to a vector coverage using ArclnfoTM.

20

 

 

 

 

 

 

 

 

 

 

 

 

LEGEND
Case Freguency # Townshigs

7 7 High 32
Medium

- Zero 40

Source: Michigan Department of Natural Resources

 

it"s

 

20 0 20 40 IGIomaers

 

 

 

Figure 9. New Case frequency Coverage for the Five County Area

 

Kernel Core A_re_a_

The kernel (Worton 1989) and minimum convex polygon (MCP) (Mohr
1947) core area coverages (Figure 10) were created‘as a way to compare the
individual case frequency values, and to determine the most significant polygons.

In ArcviewTM using the animal movement extension and the tb+ deer point
locations 3 fixed kernel range estimate with a 95% probability contour was
calculated. The resulting shapefile was clipped from the home range case
frequency coverage. This new shapefile was used to compare case frequencies
and deer use within the kernel core area.

MCP Core Area

In ArcviewTM using the animal movement extension a minimum convex
polygon using 95% of the tb+ deer points was calculated. The resulting shapefile
was clipped from the home range case frequency coverage. This new shapefile
was used to compare case frequencies and deer use within the MCP core area.

_Se_asonal Home Ranges & Habitat lg

To better understand the patterns observed in the three case frequency
coverages (original, new, and home range), movement patterns and habitat use
of radio-collared adult does in the study area were examined. Use is defined as
any time a deer is present in a specified habitat type. Deer were placed into
each of three categories: range season (summer or winter), migratory status
(non-migratory or migratory), and tb area (y=yes in tb area, n=not in tb area, or
p=partial tb area). The tb area was defined by the kernel core area. Those deer

whose home range was inside the kernel core area were delineated as y, those

22

 

 

8
E

9
a}
N

medium

20 re

 

40 ”bum

40 Iflbrrners

 

 

 

 

23

Figure 10. Kernel and MCP Core Areas Overlayed onto the Home Range Case Frequency Coverage

outside the kernel core area were delineated as n, and those that had a home
range that were partially inside the kernel core area were delineated as p. These
three categories were used in statistical comparisons.

Habitat use was determined by calculating summer and winter home
ranges for 29 does, 15 non-migratory and 14 migratory. Summer and winter
home ranges were defined using the average migration dates of radio-collared
deer in the area, March 29 and October 28 (Garner 2001). In ArcViewTM, using
the animal movement extension, home ranges were determined by the kernel
method with a 95% probability contour and a least-squares cross-validation
(LSCV) (Seaman et al. 1996) choice of h (smoothing parameter). Each home
range was used to clip an area from the deer use coverage. The percent area of
each deer use category was calculated from the clipped coverages. The point
coverage used to estimate the home ranges were overlayed on the clipped
coverages, and the percent points contained in each deer use category was
calculated (Figures 11 and 12).

Statistical Analyses

Databases

The tb database and suspicious tissue database were evaluated for
biases by comparing samples sizes by sex and age categories for statistical
differences using chi-square analyses.

_G_l_$_

For both the original and new case frequency coverages, tabulate areas in

Spatial Analyst was used to calculate the area of each deer use category per

24

township. This was converted to percentages per township. MANOVA
(multivariate analysis of variance) and Tukey’s HSD analyses were performed
using the GLM procedure in the srtsTM statistical package (SAS, 1999) to
determine if there were correlations or differences in deer use categories among
case frequency categories. MANOVA is a technique used for assessing group
differences across multiple metric dependent variables (deer use categories) at
the same time, based on a set of categorical independent variables (case
frequency) (Johnson et al., 1982). Tabulate areas in Spatial Analyst was used to
calculate the area of each deer use category per area of case frequency for the
home range, kernel 8 MCP core area coverages. The data were evaluated using
chi-square analysis to determine differences among deer use categories by case
frequencies. ArclnfoTM was used to calculate Moran’s Ifor the home range case
frequency coverage. Moran’s I (Moran 1950) is a weighted correlation coefficient
that identifies deviations from spatial randomness.

To determine differences in habitat use a usage index was calculated as

% points

 

(<1 no to low use, 1 moderate use, >1 high use). For each deer use
% area

category the index was ranked, starting with one as the lowest number for that
category (Figures 11 and 12). An ANOVA was run, in SASTM using the GLM
procedure, on the ranks for each deer use category and migratory status, range
season, and tb area sub-categories (Conover et al., 1981 ). Interactions between

these sub-categories were also tested and identified.

25

 

 
   
 

     

Estimated Deer Use

Tabulate Area «— <—
. Task Performed in ArcView
Output _, 05,2232? 2016825 Task Performed in Excel/SAS
Table Run. Stats

 

 

 

 

 

 

 

Figure 11. Data Flow Diagram for Seasonal Home Range and Habitat Use
Analyses

 

a

    

o _ :
Point Locations C .3 HomeRange ‘ .
:31; fig ' '
Deer Use Coverage
(not to scale)

 

Rare Summer Winter Summer Winter

Use Use Use Other Other
'I. . fi .

 

 

 

Points 0.00 18.82 4.71 0.00 70.47 - 7
96
C Ii

Area 031 27.01 5.57 0.12 66.99 ‘ p
[:1 Rare Use

Index 000 0.70 0.85 0.00 1.14 I s“"‘"‘" u”

. I Summer Other

mwmter Use
[:1 Winter Other

 

 

Figure 12. Example of How Percent Points, Percent Area, and Index Were
Calculated

CHAPTER 3
RESULTS

Databases

Bias checking was done by looking at sample size distributions, and
performing chi-square analyses comparing sample sizes of sex and age
categories among the different case frequency groups. Comparing the case
frequency categories cumulatively, high case frequency areas are well sampled,
medium case frequency will have adequate sample sizes in a year or two, and
zero case frequency areas will likely require at least two years to reach adequate
sample sizes (Figures 13 and 14). This is based on Schmitt et al. (1997) finding
that sample sizes of at least 300 are required to detect 2 1% prevalence with a
95% confidence interval. Comparing the yeany sample size distributions within
the high case frequency category shows a marked decrease in sampling in 1999
and 2000 (Figures 15 and 16). Figure 17 illustrates high case frequency
townships broken down into sections. This shows that sampling is inadequate
for statistically significant analysis at the section level.

To determine if “hot spots” really existed or if they were an artifact of over
sampling in the high case frequency areas and under sampling in the medium
and zero case frequency categories, chi-square analysis was done. Another
concern was that since older males (25 years) are more likely to be infected with
bovine tb (O’Brien et al., 2002), over sampling of older males would overestimate
tb case frequency. Chi-square analyses show no differences (x2=2.64,

p=o,3524) among age groups in the female sample sizes (Table 5).

27

30% -

% of Total Townships
S
a?

10% 1

 

 

{DO I1-100 I101-200 I201-300 I>300

 

 

0%

Zero

Medium

Degree of Case Frequency

 

High

 

Figure 13. Percent Sample Size Categories for All Case Frequency Distributions

 

 

 

 

 

 

 

 

 

 

 

201 -300

1996-2000
CIZero IMedIum IHigh
20

17 17

'—l
0 16 .2 7 4
.9
fi

12
g 12 «7 7 7 W m“
[-
'6
L- 8 ‘l‘ _v, i v 7%
3
E
z 4‘ >Ai ""l
2
1
0 0 0 0 - 0 0
0 1-100 101-200
Sample Size

>300

 

Figure 14. Sample Size Distribution of Deer Tested for TB in All Townships of the
Five County Area 1996-2000

28

 

E10 01-25 I26-50 I51- 100 I>100J

III

Year

 

 

% of Total Townships

 

 

 

1996 1997

Figure 15. Percent Sample Size Categories for High Case Frequency Townships

 

[31996 .1997 31998 31999 .2000]

 

 

 

 

Number of Townships

 

 

 

 

0 1-25 26-50 51-100 > 100
Sample Size

Figure 16. Distribution of Sample Sizes for High Case Frequency Townships

29

 

 

Number of Sections

 

E31996 E31997 .1998 51999 .2000

 

 

 

 

 

 

   
 

 

'//////////////////////////2

1-25

 

 

 

 

  

I

III

\ Y

    

EKFSO

Sample Size

 

51-100

>.1oo

 

 

 

Figure 17. Distribution of Sample Sizes of Sections within High Case frequency
Townships

Table 5. Chi-Square Tests Comparing the Number of Deer Tested for Tb by Sex among the

 

 

 

 

 

 

 

3 Case Frequency Areas
Females Observed Expected

1.5 2.5 3.5 4.5+ 1.5 2.5 3.5 4.5+ Total P x1
Zero 1209 1136 1125 1946 1181 1162 1113 1960 5416 0.8524 2.64
Medium 741 752 709 1245 752 739 709 1247 3447 .
_H_igh 913 928 865 1559 930 915 877 1543 4265
Total 2863 2816 2699 4750 2863 2816 2699 4750 13128
Males Observed Expected

1.5 2.5 3.5 4.5+ 1.5 2.5 3.5 4.5+ Total P x2
Zero 3680 1204 423 141 3516 1321 488 124 5448 so.ooo1 141.69
Medium 2570 809 316 77 2434 914 338 86 3772
flgh 2757 1370 511 99 3057 1148 424 108 4737
Total 9007 3383 1250 317 9007 3383 1250 317 13957

 

In the males, the sample sizes differed (x2=141.69, p>0.0005) among age

categories, with the number of males tested decreasing with age. Table 6

compares “hot spots” (high case frequency townships) with ”non-hot spots” (zero

and medium case frequency townships). The same trends were exhibited here;

females were not different, with the number of males tested decreasing with age.

30

Table 6. Chi-Square Tests Comparing Sample Sizes by Sex Between Hot Spots vs. Not Hot

 

 

 

 

 

 

 

Spots
Females Observed Expected
1.5 2.5 3.5 4.5+ 1.5 2.5 3.5 4.5+ Total P x‘
“Non-
Hot
spots“ 1950 1888 1834 3191 1933 1901 1822 3207 8863 0.7472 1.22
”Hot
spots“ 913 928 865 1559 930 915 877 1543 4265
Total 2863 2816 2699 4750 2863 2816 2699 4750 13128
Males Observed Expected
1.5 2.5 3.5 4.5+ 1.5 2.5 3.5 4.5+ Total P x’
“Non-
Hot
spots" 6250 2013 739 218 5950 2235 826 209 9220 >0.0001 137.31
”Hot
spots" 2757 1370 511 99 3057 1148 424 108 4737
Total 9007 3383 1250 317 9007 3383 1250 317 13957

 

Due to the difference in age groups among the males, a chi-square test

was used to determine if there was a positive bias in the number of carcasses

turned in due to visible lesions. Older bucks have the potential of being infected

longer, and therefore may be more likely to exhibit visible symptoms (eg. lesions

in the chest cavity). Chi-square analyses of the suspicious tissue database

showed that the higher case frequency of tb in older bucks is not due to a greater

number of carcasses being turned in (x2 =2.15, p=0.5426) (Table 7). Based on

these results, hot spots of tb do exist and are not an artifact of sampling.

31

Table 7. Chi-Square Tests for Biases by Sex in the Number of TB+ Carcasses vs. Heads
Turned In by Hunters

 

 

 

 

 

 

 

Females Observed Expected

1.5 2.5 3.5 4.5+ 1.5 2.5 3.5 4.5+ Total P x’
Carcass 4 14 14 29 6 13 13‘ 29 61 0.7698 1.13
Head 6 10 1o 23 4 11 11 23 49
Total 10 24 24 52 10 24 24 52 1 10
Males Observed Expected

1.5 2.5 3.5 4.5+ 1.5 2.5 3.5 4.5+ Total P x’
Carcass 21 47 23 4 22 47 20 6 95 0.5426 2.15
Head 19 38 13 6 18 38 16 4 76
Total 40 85 36 10 40 85 36 10 171

GIS

Qse Freggencv Coverage Analyses

Once it was determined that “hot spots” did exist, potential correlations
between case frequency areas and deer use categories were established.
MANOVA analyses of the two township coverages (original case frequency
coverage and new case frequency coverage) compared percent deer use
categories per township among each case frequency category. Tukey’s HSD
was used to determine where these differences lie. Results for the original case
frequency coverage show rare use (F2,98=2.16, p=0.1207), winter use (F2,98=0.61,
p=0.5480), summer other (F2,93=0.83, p=0.4399), and winter other (F2.93=1.56,

=0.2155) exhibited no differences among case frequency categories. However,
summer use (F2_98=4.08, p=0.0198) showed differences between high and zero
case frequency categories (Table 8). Results for the new case frequency
coverage showed winter use (F2,93=0.72, p=0.4897) and summer other
(F2,ga=0.46, p=0.6346) revealed no differences among case frequency

categories. Conversely, for the new case frequency coverage, rare use

32

(F2,gg=5.82, p=0.0041), summer use (F2,98=5.96, p=0.0036), and winter other

(F2,ga=3.24, p=0.0435) showed differences. Rare use differed between high and

zero, and medium and zero case frequency categories. Summer use differed

between high and zero case frequency categories. Winter other differed

between high and medium case frequency categories. Figures 18 and 19, which

show the percentages of deer use categories for each case frequency category

and the entire five county area, help illustrate why these differences are present.

Table 8. Summary of MANOVA and Tukey's HSD Analyses for Differences of % Deer Use

Among Case frequency Categories

 

Overall Model For Original Case frequency Coverage

 

 

 

 

 

 

 

Rare Use Summer Winter Summer Winter
Use Use Other Other
F(p) 2.16(0.1207) 4.08(0.0198) 0.61 (0.5480) 0.83(0.4399) 1 .56(0.2155)
Tukey’s HSD"
Case Rare Use Summer Winter Summer Winter
frequency Use Use Other Other
Zero a a a a a
Medium a ab a a a
High a b a a a
Overall Model For New Case frequency Coverage
Rare Use Summer Winter Summer Winter
Use Use Other Other
F(pL 5.82(0.0041) 5.96(0.0036) 0.72(0.4897) 046106346) 3.24(0.0435)
Tukey's HSD‘
Case Rare Use Summer Winter Summer Winter
frequency Use Use Other Other
Zero a a a a a
Medium b ab 3 a a
High b b a a b

'Within a deer use category case frequency categories with the same letter do not differ (a=0.05).

33

 

 

DZero I Medium I High I 5 Counties

 

 

a
.8

 

as
.o
8

l
l
l

Percentage Deer Use
{3 8
8 8

 

'8

 

  

 

 

 

 

 

 

.0
.8

 

Summer Use Winter Use Summer Other Winter Other
Deer Use Categories

 

 

 

Figure 18. Percentage of Deer Use Categories for Original Case Frequency
Coverage

 

 

DZero sMedium IHigh I5 Counties]

 

 

 

 

 

Rare Use Summer Use Winter Use Summer Other Winter Other
Deer Use Categories

 

 

 

Figure 19. Percentage of Deer Use Categories for New Case Frequency Coverage
Chi-square analysis was used to compare deer use categories among

zero, medium, and high case frequency areas for the home range case
frequency coverage (Appendix 58). Results of the analysis show differences

among deer use categories (x2=83.63, p>0.0001). Figure 20 shows high case

34

frequency has the largest amount of summer use, and the least amount of

summer other and winter other use.

 

 

[CIZero .Medium .High .5 Counties

 

 

 

 

 

 

 

 

 

 

Rare Use Summer Use Winter Use Summer Winter Other
Other

Deer Use Categories

 

 

 

Figure 20. Percentage of Deer Use Categories for Home Range Case Frequency
Coverage

Trends observed over all three case frequency coverage analyses were
as follows:

1. Rare use was evenly distributed among high, medium and zero case
frequency for the original case frequency coverage and home range case
frequency coverage. However, there was a significantly larger portion of rare use
for zero case frequency (10.36%), than high and medium case frequency (7.70%
and 7.33%, respectively) for the new case frequency coverage.

2. Summer use increased as case frequency level increased, with the
most dramatic increase exhibited by the home range case frequency coverage.

3. Winter use was evenly distributed among high, medium and zero case
frequency for all coverages.

4. Summer other use was evenly distributed among high, medium and

35

zero case frequency for original and new case frequency coverages. The home
range case frequency coverage showed a significantly smaller amount of
summer other use (5.57%) for high case frequency compared to the medium and
zero case frequencys (11.10% and 10.41%, respectively).

5. Winter other use decreased as case frequency increased. This trend
was consistent for all coverages.

Moran’s Iwas performed on the home range case frequency coverage,
which was used to determine the spatial distribution (random vs. clustered
distribution) of bovine tb in the area. The Moran’s I = 0.9958, which shows that
case frequency categories are clustered (Figure 7). For example, high case
frequency areas have a higher probablility of being next to another high case
frequency area (i.e. positive spatial autocorrelation).

Core A_rea Analysefis

‘ Core area analyses were used to compare the individual case frequency
values, and to determine the most significant polygons, significant polygons were
those that are correctly categorized into their respective case frequency
category, and are not an artifact of how the home range coverage was
calculated. For example if there is a 25 km2 area categorized as medium case
frequency surrounding a 1ha high case frequency area, is this area truly high
case frequency or an artifact of overlapping home ranges? This analysis did not
succeed in answering the question of significance. However, it did provide a
useful means of comparing central areas of high case frequency.

Chi-Square analyses were used to compare deer use categories in the

36

kernel core area and the MCP core area (Appendix 5b and 5c). While this does
not tell which polygons are most significant, it does demonstrate areas of high tb
activity, and excludes areas (polygons) of low activity. Both kernel and MCP
analyses showed differences of deer use among case frequency categories
(x2=34.36, p>0.0001;x2=44.78, p>0.0001, respectively). Figures 21 and 22
show similar trends for both analyses. Rare use and summer other use
categories decrease as case frequency increases. Winter use category
increases as case frequency increases. In the MCP core area analysis, summer
use follows the same trends as the three case frequency analyses. The winter
other use category also exhibits similar trends to the case frequency analyses.
Zero and medium case frequencies are similar, with high case frequency

significantly smaller.

 

 

DZero I Medium I High]

 

S
8

 

8

 

.3 .8
8
m

as
.0
8
l

 

 

8
8

 

 

.3
8
I
I

Percentage of Deer Use

8

 

 

 

 

 

 

 

_o
8

 

Rare Use Summer Use Winter Use Summer Other Winter Other
Deer Use Categories

 

 

 

Figure 21. Percentage of Deer Use Categories for Kernel Core Area

37

 

 

UZero IMedium IHigh

 

 

 

Percentage of Deer Use
0)
.°
8

£5
8

 

 

 

 

 

0.00 —

Rare Use Summer Use Winter Use Summer Other Winter Other
Deer Use Categories

 

 

 

Figure 22. Percentage of Deer Use Categories for MCP Core Area
Seasonal Home Ranges 8. Habitat Use

To better understand the relationships observed in the previous analyses,
movement patterns and habitat use of radio-collared deer in the study area were

% points _ <1

 

examined. To determine differences in habitat use a usage index ( o/
0 area

no to low use, 1 moderate use, >1 high use) was calculated. For each deer use
category the index was ranked starting with one for the smallest index value. An
ANOVA was run on the ranks for each deer use category and migratory status,
range season, and tb area sub-categories. Interactions among these sub-
categories (SR*MS, SR*TB, MS*TB, SR*MS*TB) were also tested.

Rare Use

Results from the ANOVA show a difference (F=2.15, p=0.0390) for the

overall model, and for seasonal range (F=14.52, p=0.0004) for rare use. It is the

38

only category that shows significant differences in seasonal ranges. Figure 23
shows the average index value plotted for migratory status and tb area sub-
categories. Average index value varies substantially by migratory status
between tb and non-tb areas. However, it did not show up as being statistically
different (Table 9). In the tb area non-migratory deer not only use rare use more
often than did migratory deer in the tb area, but they also have the highest use of

all of the four groups.

 

 

[-e—TB +No TB

 

 

1.80
1.601
1.40 J»
1.20 1 —
1.00 .. -

0.80 1 - —
0.40 «e -, — ——- ———___- _._-.
0.20 _ _ _-.___..______ _
0.00

 

 

 

 

 

 

 

Average Index Value

 

 

Non-migratory Migratory
Migratory Status

 

 

 

Figure 23. Rare Use Interaction Between TB Areas 8- Migratory Status

Summer Use

Results from the ANOVA show that the overall model is not significant
(F=1.31, p=0.2569). Figure 24 illustrates high quality summer habitat use of deer
in the five county area. Deer in the no tb area are consistent in summer habitat
use regardless of migratory status. Non-migratory deer in the tb area show a

substantially larger amount of high quality summer habitat use than any other

group.

39

moduo Kv drum—O.

 

 

550
288.98.. 288.98.. 8.5.8.935 c.8353 238.965? 288.985 258.98., 858.555 .353
550
853.39.? 88.935 68:32.. 848.935 668.955 @8333 282.83 38.925 SEE...»
so:
648.925 68098.5 38.8.55? 285.985 885.995 885.985 @3858 82598.0 .353
8:
283.985 58.225 $80388 8.3.285 commence 658.928 $235.0 883:3 555%
theorem 883:8 £8353 £53943 88455.5 653.28; 238.98.: 8855K on: cam
2.5% 39.3. .80!
Eta-2.x» are: are» make 83 2. been... .238»... :85 8: coon
mzozoémbz. whomuum 22:
4.32:6."—

 

8: 53...: s 885559 .o. 9:62. <>oz< so seesaw .e 633

40

 

 

[—e—TB +No TBl

 

 

1.20

1.15

 

 

 

1.05 «la—2-.—

 

 

 

1.00 .E..- ,

 

0.95 .r 4- -

Average Index Value

0.90 ~~_~~— —— sew—4 s - e _-_--.- —~-— --—-——~

 

 

0.85

 

Migratory Status

 

 

 

Figure 24. Summer Use Interactions between Migratory Status and TB Area

Winter Use

The outcome of the statistical analysis for winter use shows no differences
(F=0.63, p=0.7793) or interactions (Table 9). Figure 25 demonstrates the use of
high quality winter habitat. Again, non-migratory and migratory deer in the no tb
area have similar winter use patterns. Migratory deer in the tb area show a
considerably larger difference of high quality winter habitat use than any other

group.

41

 

 

L—e— TB + NO TB

 

 

4.00

2.00 i

 

Average Index
Value

3.00 «m

 

1.00 _-

 

 

 

 

0.00

 

Non-migratory Migratory

Migratory Status

 

 

Figure 25. Winter Use Interaction Between TB Areas 8. Migratory Status

Summer Other

Results of the ANOVA show an interaction (F=8.78, p=0.0006) between

migratory status and lb area. Poor summer habitat use is illustrated in Figure 26.

Non-migratory deer in both the no tb and tb areas exhibit low to moderate use of

poor summer habitat. While migratory deer in the tb area have very low use, and

non-migratory deer in the no tb area have high use of poor summer habitat.

 

 

I-e—TB +No TB]

 

 

1 .40

0.60 .2-

Average Index Value

 

1.20 —— - -

 

1.00 4t“ -
0.80 lm

0,4o._._----.- - - --
0.20-4-

 

 

 

0.00

 

Non-migratory Migratory

Migratory Status

 

 

Figure 26. Summer Other Interaction Between TB Areas 81 Migratory Status

42

Winter Other

Results of the ANOVA show differences (F=3.77, p=0.0009) in the overall
model, and differences (F=13.78, p=<0.0001) between no tb and tb areas.
Figure 27 shows low use of poor winter habitat by deer in the no tb area, with no
differences between non-migratory and migratory deer. While deer in the tb area
also show no difference between non-migratory and migratory deer, they use
poor winter habitat by a substantially larger amount (Figure 28). This is even
more unusual considering that there is a considerably smaller amount of winter

other habitat available in the tb area.

 

 

[-e—TB +No TB

 

1.80
1.60 .
1.40 + - ,
1.20 + 7 .- — --

1.00 1~m~ - - -- — — -— ———~—~ ,.
0.80~W- —W— -~ - - -- u-----—-——4W_W—-
060 _._.--. . ., ..__ .._ “____.___ _ . _. . ._ __. 2, H, _

0.40 -——-WWWF -- - ‘U.--

0.20 W ~ - - _-_.-“m -- W “a--. 2-2.2.. W
0.00

 

 

 

 

 

 

Average Index Value

 

Non-migratory Migratory
Migratory Status

 

 

 

Figure 27. Winter Other Interaction Between TB Areas 8: Migratory Status

43

 

 

 

[o No TB Area I TB Area I 5 County Area ]

 

 

8

40

Percent Deer Use
8 8

.3
O

 

 

 

 

 

 

Rare Use Summer Use Winter Use Summer Other Winter Other
Deer Use Categorles

 

Figure 28. Percent Deer Use In the No TB, TB, and 5 County Areas

44

 

CHAPTER 4
DISCUSSION

Case Frequency Vs. Prevalence

Why use case frequency (number of infected animals in a given area)
instead of prevalence (% of infected animals in a given area)? Prevalence (in
this case apparent prevalence) has the potential to greatly over estimate actual
prevalence, depending on how effectively areas are sampled. For example if 3
tb+ deer were found in a township, and that was also the total number of deer
sampled for that township, that township would have a prevalence of 100%. At
the beginning of this project affects of sampling on detected levels of tb were
unknown. In an attempt to deal with this uncertainty and potential bias, case
frequency levels were created. While the above mentioned issues with
calculating prevalence are not completely eliminated, it is believed that this is a
more objective way of looking at the intensity of disease. The Original case
frequency coverage map was compared, visually, to an apparent prevalence
map created by the MDNR, and similar patterns were observed. Attempts at
categorizing disease are generally subjective, and are based on the researchers
knowledge and perceptions of what is high or low (whether it be prevalence or
case frequency). This study attempts to move away from what has already been
done, while still effectively measuring case frequency.

Databases

“Hot spots” are not an artifact of sampling. Results of analyses show that
sampling is not overestimating case frequency in females. Sample sizes did not

differ among case frequency categories and age groups. A recent study by

45

O’Brien et al. (2002), shows that the risk of being infected with tb, was not
different between sexes among the fawn and yearling age groups. There was a
small increase in tb risk for females from age 2-2.5 years, which leveled off at 24
years (Odd ratio (O.R.)=1], 95% CI 0.9, 3.1 for 2-2.5 yrs; O.R.=2.5, 95% CI 1.4,
4.7 for 44.5 yrs; O.R.=2.5, 95% CI 1.4, 4.3 for 25 yrs; O’Brien et al., 2002).
Males, however, showed a n increase in risk at age 2-2.5 years which continued
to increase with age (O.R.=4.5, 95% CI 2.7, 7.4 for 2-2.5 yrs; O.R.=11.3, 95% Cl
3.2, 40.3 for 25 yrs; O’Brien et al., 2002). Even though (for this study) a
difference was found for males among the different age categories, sample sizes
were shown to decrease as age increased. This reduced the probability of over
estimating case frequency due to over sampling animals with higher rates of
infection.

Another factor that was examined was whether this increased risk of
infection was due to the sampling method. Deer samples are obtained,
voluntarily, from hunters, and therefore what is sampled is based on hunter skills
and their willingness to submit samples for testing. One potential way hunters
could influence sampling is that their perception of tb will make them more likely
to have a deer tested if it has visible lesions. This would mean that there would
be differences in the proportion of carcasses turned in relative to heads. There
was, however, no difference found between the proportion of carcasses and
heads by age class, which means the sampling method, in this instance did not

influence sampling. Other issues of sampling such as; “Do hunter harvest

46

surveys accurately reflect wild deer p0pulations?”, have been addressed by
O’Brien et al. (2002), who suggest that the bias is negligible.

GIS

Case Freguency Coverages

Several correlations were observed among the three case frequency
coverages, the most prominent being that summer use of quality summer habit
increases as case frequency increases. While interesting, these observed
correlations alone do not demonstrate cause and effect. These results coupled
with habitat use of deer in the study area, which will be discussed later, present
interesting insights into what may be driving levels of tb in this area.

Moran’s I

The Moran’s I shows that case frequency categories for the home range
case frequency coverage are clustered . These results are similar to those of
O’Brien et al. (2002), who found central areas of high prevalence surrounded by
larger areas of low prevalence. The region known as the core area by the MDNR
(essentially the portions of Montmorency, Alpena, Oscoda and Alcona counties
where their boundaries intersect) demonstrates this clumped pattern, and the
peak concentration of high case frequency in the home range case frequency
coverage resembles the shape of the MDNR’s core area. This helps to support
other analyses done on tb in this area, which show the core area as having the

highest levels of tb (Hickling, 2002; O’Brien et al., 2002; Schmitt et al., 1997).

47

Core Area

Core area (defined as kernel and MCP), while not fulfilling their original
purpose, provided more support that the main area oftb activity is centralized in
the five county area. The kernel core area analysis results show that summer
use category did not differ among case frequency categories. This is explained
by the distribution of the summer use category within the five county area.
Summer use category is most densely concentrated in the central areas of
Montmorency, Alpena, Alcona, and Oscoda, which are almost completely
encompassed by the kernel core area.

Habitat Use

The seasonal range and habitat use analyses proved to be one of the
more telling analyses of this study. A study done by Felix and Hughey
(unpublished data) also examined seasonal home ranges and habitat use of
does in this area. Habitat was broken down into three coverages: Spring and
Summer Food Potential (SSF P), Fall and Winter Food Potential (FWFP), and
Thermal Cover Potential (T CP) (Felix, 2003). Each of these coverages had five
different levels of habitat quality; low, medium low, medium, medium high, and
high. Habitat quality levels were based on an index created by Felix
(unpublished). Results of analyses showed that migratory deer tend to use
poorer habitat, and that all deer regardless of migratory status showed a shift in
seasonal habitat use (eg. higher quality winter habitat in the winter, and higher

quality summer habitat in the summer) . Distributions of habitat potential (quality

48

of the habitat) among these three coverages (SSF P, FWF P, and TCP) were
similar to those in the deer use coverage.

The tendency for non-migratory deer to use higher quality habitat is
illustrated by both tb and no—tb area groups only in the summer use category
(Figure 22). In contrast, non-migratory deer in the tb area showed moderate-high
use (index 21) to the highest use of poor quality habitat (rare use, summer
other, and winter other). Also, deer in the tb area show high use of poor quality
winter habitat, even though it makes up a smaller percentage of habitat than in
the no-tb area. This lends support to the belief that baiting and supplemental
feeding influence deer behavior in this area.

What does it all mean? The correlations observed among the summer
use category and case frequency levels for all three coverages (original case
frequency, new case frequency, and home range case frequency) are most likely
due to the distribution of summer use present in each case frequency area
(meaning the percent of that area composed of that habitat type). Correlations
between rare use, summer other, and winter other could also be influenced by
distribution. However, those correlations, coupled with the increase in deer
activity in these areas, specifically non-migratory deer in the tb area, may be
showing two things. One, deer behavior in the tb area has been influenced by
practices such as baiting and supplemental feeding, and thus deer are being
lured into these areas with prospects of an easy meal. Two, since the majority of
this use is by non-migratory deer it may explain why tb is centralized and not as

wide spread. Of course this is only speculation and should be tested further.

49

One of the limitations to this analysis was that bucks were not included,
due to small sample sizes. Since males have much different movement patterns,
and life histories it would have been interesting to see the differences in habitat
use.

Q55 Freggencvand Scale

Several studies have examined the effect of changing scale on pattern
and results, or the Modifiable Areal Unit Problem (MAUP) (Openshaw et al. 1979;
Turner et al., 1989; Malanson et al., 1997). Openshaw and Taylor (1979),
termed MAUP as being two distinct, but interrelated issues. The first issue is
looking at an area of study at the same scale, but aggregating the units of
measure differently (e.g. township case frequency is defined by number of
infected deer or the number of infected deer home ranges). The second issue
being, when the same area of study (extent) is looked at with increasingly larger
areal units of measurement (e.g. looking at the same five county area starting at
the section level and moving to the county level).

This study also attempts to address MAUP, with the three case frequency
coverages. Both the original and new case frequency coverage are at the same
scale. However, they differ in how township case frequency is defined. The
statistical analyses show differing results dependant on how township case
frequency was defined. lnforrnation is gained or lost depending on which
coverage is chosen. Which scale is correct? Levin (1992) states that this
(changing scale) is the “principal technique of science", moving from one scale to

another helps us shift from variable, unrepeatable phenomena to a compilation of

50

information for which general statements can be made. He also states that there
is no one correct scale, and that several scales should be examined in order to
truly understand an ecological system.

Looking at disease on a township level has several advantages:

1) Data collection is easier.

2) Comparing data is easier, because larger sample sizes facilitate
statistical analyses.

3) Boundaries are easily identifiable, landowners (as a general rule) know
what township they are in, which makes implementing and enforcing
regulations easier.

Despite these advantages examining data on only a township level has

two major limitations:

1) The smallest area that can be examined is a township, so information
on biological phenomena that occur at smaller scales is lost.

2) Looking at an ecological issue on a political scale can mask what is
actually happening in the real world.

The home range case frequency coverage addresses these issues by
using smaller units of measure and using a “biological scale” to look at tb. This
helps identify specific areas of clustering that are not evident with townships. It
also provides a new perspective, which helps support previous findings, and
uncovers new ones.

One main problem with applying the home range case frequency

coverage in a management situation are tiny polygons that may be a result of

51

how the coverage was created, and do not represent what is actually occurring in
the real world. A possible solution to this problem would be to set the minimum
mapping unit (mmu)(smallest polygon) to that of the smallest area of
management. The smallest area of management is determined by the managers
goals and knowledge of the area. Once this is determined, the polygons smaller
than the mmu could be dissolved into the largest adjacent polygon or according

to a set of dissolve criteria (decision tree).

52

1)

2)

3)

4)

CONCLUSIONS AND RECOMMENDATIONS TO MDNR
Attempts to create a data entry system for the tb database should be
made to standardize data, and eliminate entry of incorrect information.
This would facilitate the use of such data, and increase confidence in
the quality of the data collected.

Geographic “hot spots” of tb do exist, and are not merely an artifact of
sampling.

The correlations observed among the summer use category are most
likely due to it’s distribution in each case frequency area. Correlations
between rare use, summer other, and winter other could also be
influenced by their distributions. Deer behavior in the tb area is
influenced by practices such as baiting and supplemental feeding, and
since the majority of increased activity is by non-migratory deer it may
explain why tb appears to be centralized.

Kernel & MCP core area analyses provide more support that the main

area of tb activity is centralized in the five county area.

5) Patterns of bovine tb should be examined at multiple scales to get a

better understanding of the mechanisms driving the disease.

53

APPENDICES

54

APPENDIX 1
CHANGES MADE TO THE TB DATABASE

. Section numbers were corrected from single digits to two digits, eg. “2 ”

and “ 2” are now “02"

. Deleted incorrect town and range values; a file of incorrect town and range

values was created; the file name is incorrectTwang.xls

. Changed the following location information as per instructions from Jean

Fierke, Lab Scientist, Rose Lake Wildlife Disease Lab, MDNR

a. Alcona 28N 04E
b. Alpena 28N 04E
c. Alpena 28N 04E

d. Montmorency 31N 05E

24 to Oscoda 28N 04E
07 to Montmorency 29N 04E
25 to Montmorency 29N 04E
21 to Alpena 31N 05E

. Changed ages 1, 2, 3, 4, to 1.5, 2.5, 3.5, 4.5, respectively

55

24
O7
25
21

APPENDIX 2
METADATA

Original Case Freguency Coverage
Identification Information

Citation:

Citation Information:
Originator.

Brandi D. Hughey, Graduate Research Assistant, Michigan State
University, Department of Fisheries and Wildlife, 13 Natural
Resources, East Lansing, MI 48824-1222

Scott R. Winterstein, Professor, Michigan State University, Department of

Fisheries and Wildlife, 13 Natural Resources, East Lansing, MI

48824-1222

Date: 2003

Title: Original Case Frequency Coverage

Title of File: prevalence

Format Arclnfo coverage and all associated files

Description:

Abstract This digital map and associated database describes and projects
case frequency of bovine tuberculosis based on township boundaries in Alcona,
Alpena, Montmorency, Oscoda, and Presque Isle counties. Townships are
categorized into one of 3 tb case frequency categories zero, medium, and high.
Zero case frequency townships are defined as never having a tb+ animal in that
township. Medium case frequency is defined as having at least one tb+ animal in
up to each of two years in that township. High case frequency is defined as
having at least one tb+ animal in each of three or more years in that township.

Purpose: Provides spatial distribution of bovine tuberculosis for five county
area (Alcona, Alpena, Montmorency, Oscoda, and Presque Isle).

Status:
Progress: Complete

Keywords:
Theme:
Theme Keyword: bovine tuberculosis
Theme Keyword: township
Place:
Place Keyword: Michigan
Place Keyword: Northern Lower Peninsula
Point of Contact:
Contact lnfonnation:

Contact Organization: Michigan State University,
Department of Fisheries and Wildlife

56

Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: MI
Postal Code: 48824-1222
Contact Voice Telephone: 517-353-2042
Contact Facsimile Telephone: 517-432-1699

Native Data Set Environment. Windows NT version 4.0; ESRI ArcView 3.2

Cross Reference:
Citation Infonnation:
Originator:
Michigan DNR - Resource Mapping and Aerial Photography
Publication Date: Unpublished Material
Title: Tr324 - PLSS Coverage of Michigan
Series lnfonnation:
Publication Information:

Publisher:
Data Quality Information
Accuracy Report:
Lineage:

Source lnforrnation:
Source Citation:
Originator: Michigan DNR - Resource Mapping and Aerial

Photography.

Publication Date: Unpublished Material

Title: sectionsup.eOO

Geospatial Data Presentation Format: Vector Digital
Database

Geospatial Data Presentation Format: Digital Database
Publication Infonnation:
Publication Place: Lansing, Michigan
Publisher. Michigan Natural Features Inventory

Source Infonnation:
Source Citation:
Originator: Michgan Department of Natural Resources
Publication Date: Unpublished
Title: State tb data base
Geospatial Data Presentation Fonnat:
Publication Information:
Publication Place:
Publisher:
Source Contribution:

57

This was used to identify case frequency level for each polygon.

Spatial Data Organization Information
Direct Spatial Reference Method: Vector
Spatial Reference Information
Spatial Reference: Michigan GeoRef from Oblique Mercator projection
Scale factor at center- - 0. 9996
Azimuthal angle = 337.25556
False easting = 2546731496
False northing = 4354009816
Horizontal datum name = North American Datum 1983 (NAD83)
Planar Distance Units: meters

Entity and Attribute lnforrnation

Attribute Description:
Attribute:
Attribute Label: Town
Attribute Definition: Township
Attribute:
Attribute Label: Range
Attribute Definition: Range
Attribute:
Attribute Label: Town_range
Attribute Definition: Township and Range
Attribute:
Attribute Label: Prevalence
Attribute Definition: Prevalence refers to number of tb+ deer found
in that township

Distribution Information
Distributor:
Contact Information:
Contact Person: Brandi D. Hughey
Contact Organization: Michigan State University, Department of Fisheries
and Wildlife
Contact Address:
Address: 13 Natural Resources
City. East Lansing
State: MI
Postal Code: 48824-1222
Contact Voice Telephone: 517-432-4959
Contact Electronic Mail: hugheybr@msu.edu

Contact Information:
Contact Person: Scott R. Winterstein

58

Contact Organization: Michigan State University, Department of
Fisheries and Wildlife

Contact Address:
Address: 13 Natural Resources
City. East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-353-2042
Contact Electronic Mail: winterst@msu.edu

Metadata Reference lnforrnation
Metadata Date: 20030125

Metadata Contact:
Contact Infonnation:
Contact Organization: Michigan State University, Department of
Fisheries and Wildlife
Contact: Brandi D. Hughey
Contact Address:
Address: 13 Natural Resources
City. East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-432-4959
Contact Electronic Mail: hugheybr@msu.edu

Contact Scott R. Winterstein
Contact Address:
Address: 13 Natural Resources
City. East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-353-2042
Contact Electronic Mail: winterst@msu.edu

59

New Case Freggencvgrveragg
Identification Information

Citation:

Citation Information:
Originator:

Brandi D. Hughey, Graduate Research Assistant, Michigan State
University, Department of Fisheries and Wildlife, 13 Natural
Resources, East Lansing, MI 48824-1222

Scott R. Winterstein, Professor, Michigan State University, Department of

Fisheries and Wildlife, 13 Natural Resources, East Lansing, MI

48824-1222

Date: 2003

Title: New Case Frequency Coverage

Title of File: new_prev

Fonnat: Arclnfo coverage and all associated files

Description:

Abstract: This digital map and associated database describes and projects
case frequency of bovine tuberculosis based on township boundaries and
estimated average home ranges for white-tailed deer in Alcona, Alpena,
Montmorency, Oscoda, and Presque Isle counties. Townships are categorized
into one of 3 tb case frequency categories zero, medium, and high. Zero case
frequency townships are defined as never having a tb+ animal home range in
that township. Medium case frequency is defined as having at least one tb+
animal home range in up to each of two years in that township. High case
frequency is defined as having at least one tb+ animal home range in each of
three or more years in that township.

Purpose: Provides spatial distribution of bovine tuberculosis for five county
area (Alcona, Alpena, Montmorency, Oscoda, and Presque Isle)

Status:
Progress: Complete
Keywords:
Theme:
Theme Keyword: bovine tuberculosis
Theme Keyword: township
Place:
Place Keyword: Michigan
Place Keyword: Northern Lower Peninsula
Point of Contact:
Contact Infonnation:
Contact Organization: Michigan State University, Department of
Fisheries and Wildlife
Contact Address:

60

Address: 13 Natural Resources
City: East Lansing
State: Ml
Postal Code: 48824-1 222
Contact Voice Telephone: 517-353-2042
Contact Facsimile Telephone: 517-432-1699

Native Data Set Environment: Windows NT version 4.0; ESRI ArcView 3.2

Cross Reference:
Citation lnfonnation:
Originator:
Michigan DNR - Resource Mapping and Aerial Photography
Publication Date: Unpublished Material
Title: TrsZ4 - PLSS Coverage of Michigan
Series Information:
Publication Information:

Publisher:
Data Quality Information
Accuracy Report:
Lineage:

Source lnfonnation:

Source Citation:
Originator: Michigan DNR - Resource Mapping and Aerial
Photography. ‘
Publication Date: Unpublished Material
Title: sectionsup.eOO
Geospatial Data Presentation Format: Vector Dig ital
Database

Geospatial Data Presentation Format: Digital Database
Publication Information:
Publication Place: Lansing, Michigan
Publisher. Michigan Natural Features Inventory

Source lnfonnation:
Source Citation:
Originator: Michgan Department of Natural Resources
Publication Date: Unpublished
Title: State tb data base
Geospatial Data Presentation Format:
Publication lnfonnation:
Publication Place:
Publisher:
Source Contribution:

61

This was used to identify case frequency level for each polygon.

Spatial Data Organization Information
Direct Spatial Reference Method: Vector

Spatial Reference Infonnation

Spatial Reference: Michigan GeoRef from Oblique Mercator projection
Scale factor at center = 0.9996
Azimuthal angle = 337.25556
False easting = 2546731496
False northing = 4354009816
Horizontal datum name = North American Datum 1983 (NAD83)

Planar Distance Units: meters
Entity and Attribute Information
Attribute Description:
Attribute:
Attribute Label: Town
Attribute Definition: Township
Attribute:
Attribute Label: Range
Attribute Definition: Range
Attribute:
Attribute Label: Town_range
Attribute Definition: Township and Range
Attribute:
Attribute Label: New_Prev
Attribute Definition: Prevalence refers to number of tb+ deer home
ranges found in that township

Distribution Information
Distributor:
Contact lnfonnation:
Contact Person: Brandi D. Hughey
Contact Organization: Michigan State University, Department of Fisheries
and Wildlife
Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-432-4959
Contact Electronic Mail: hugheybr@msu.edu

Contact Information:

62

Contact Person: Scott R. Winterstein
Contact Organization: Michigan State University, Department of
Fisheries and Wildlife
Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-353-2042
Contact Electronic Mail: winterst@msu.edu

Metadata Reference lnforrnation
Metadata Date: 20030125

Metadata Contact:
Contact lnfonnation:
Contact Organization: Michigan State University, Department of
Fisheries and Wildlife
Contact: Brandi D. Hughey
Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: MI
Postal Code: 48824-1222
Contact Voice Telephone: 517-432-4959
Contact Electronic Mail: hugheybr@msu.edu

Contact: Scott R. Winterstein
Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-353-2042
Contact Electronic Mail: winterst@msu.edu

63

Home Range Case Frquency Coverage
Identification lnfonnation

Citation:

Citation Information:
Originator:

Brandi D. Hughey, Graduate Research Assistant, Michigan State
University, Department of Fisheries and Wildlife, 13 Natural
Resources, East Lansing, MI 48824-1222

Scott R. Winterstein, Professor, Michigan State University, Department of

Fisheries and Wildlife, 13 Natural Resources, East Lansing, MI

48824-1222

Date: 2003

Title: Home Range Case Frequency Coverage
Title of File: hr_prev

Format: Arclnfo coverage and all associated files

Description:

Abstract: This digital map and associated database describes and projects
case frequency of bovine tuberculosis based on estimated average home ranges
for white-tailed deer in Alcona, Alpena, Montmorency, Oscoda, and Presque Isle
counties. Polygons are categorized into one of 3 tb case frequency categories
zero, medium, and high. Zero case frequency is defined as never having a tb+
animal home range for that polygon. Medium case frequency is defined as
having at least one tb+ animal home range in up to each of Mo years for that
polygon. High case frequency is defined as having at least one tb+ animal home
range in each of three or more years for that polygon.

Purpose: Provides spatial distribution of bovine tuberculosis for five county
area (Alcona, Alpena, Montmorency, Oscoda, and Presque Isle).

Status:
Progress: Complete

Keywords:
Theme:
Theme Keyword: bovine tuberculosis
Theme Keyword: home range
Place:
Place Keyword: Michigan
Place Keyword: Northern Lower Peninsula
Point of Contact:
Contact lnfonnation:
Contact Organization: Michigan State University, Department of
Fisheries and Wildlife
Contact Address:

Address: 13 Natural Resources
City: East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-353-2042
Contact Facsimile Telephone: 517-432-1699

Native Data Set Environment: Windows NT version 4.0; ESRI ArcView 3.2

Cross Reference:
Citation lnfonnation:
Originator:
Michigan DNR - Resource Mapping and Aerial Photography
Publication Date: Unpublished Material
Title: Tr524 - PLSS Coverage of Michigan
Series lnfonnation:
Publication lnfonnation:

Publisher:
Data Quality Information
Accuracy Report:
Lineage:

Source lnfonnation:

Source Citation:
Originator: Michigan DNR - Resource Mapping and Aerial
Photography. -
Publication Date: Unpublished Material
Title: sectionsup.eOO
Geospatial Data Presentation Format: Vector Digital
Database

Geospatial Data Presentation Format: Digital Database
Publication lnfonnation:
Publication Place: Lansing, Michigan
Publisher: Michigan Natural Features Inventory

Source lnfonnation:
Source Citation:
Originator: Michgan Department of Natural Resources
Publication Date: Unpublished
Title: State tb database
Geospatial Data Presentation Fonnat:
Publication lnfonnation:
Publication Place:
Publisher:
Source Contribution:

65

 

This was used to identify case frequency level for each polygon.

Spatial Data Organization Information
Direct Spatial Reference Method: Vector

Spatial Reference Information
Spatial Reference: Michigan GeoRef from Oblique Mercator projection
Scale factor at center = 0.9996
Azimuthal angle = 337.25556
False easting = 2546731496
False northing = 4354009816
Horizontal datum name = North American Datum 1983 (NA083)
Planar Distance Units: meters

Entity and Attribute Information

Attribute Description:
Attribute:
Attribute Label: Area
Attribute Definition: Area of poly/region in square coverage units
Attribute:
Attribute Label: Perimeter
Attribute Definition: perimeter of poly/region in coverage units
Attribute:
Attribute Label: Prevalence
Attribute Definition: Prevalence refers to number of tb+ deer home
ranges found in that township

Distribution lnforrnation
Distributor:
Contact lnfonnation:
Contact Person: Brandi D. Hughey
Contact Organization: Michigan State University, Department of Fisheries
and Wildlife
Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-432-4959
Contact Electronic Mail: hugheybr@msu.edu

Contact lnfonnation:
Contact Person: Scott R. Winterstein
Contact Organization: Michigan State University, Department of
Fisheries and Wildlife

66

Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: Ml
Postal Code: 48824-1222
Contact Voice Telephone: 517-353-2042
Contact Electronic Mail: winterst@msu.edu

Metadata Reference Information
Metadata Date: 20030125

Metadata Contact:
Contact lnfonnation:
Contact Organization: Michigan State University, Department of
Fisheries and Wildlife
Contact: Brandi D. Hughey
Contact Address:
Address: 13 Natural Resources
City: East Lansing
State: MI
Postal Code: 48824-1222
Contact Voice Telephone: 517-432-4959
Contact Electronic Mail: hugheybr@msu.edu

Contact: Scott R. Winterstein
Contact Address:
Address: 13 Natural Resources

City: East Lansing
State: Ml
Postal Code: 48824-1222

Contact Voice Telephone: 517-353-2042

Contact Electronic Mail: winterst@msu.edu

67

APPENDIX 3
REGIONS

Regions are vector data objects, and are used to represent areal
geographic features. Regions are comprised of one or more polygons, just as
polygons are made up of lines, and lines are made up points. There are three
main features of regions nested feature, associated feature, and overlapping
feature. The overlapping feature was used in this study, which is illustrated in

Figures 1 and 2.

 

 

 

Attribute Table
Region # Polygon List
101 1, 2
‘ \ 102 2, 3, 4
103 4, 5
104 6

 

 

 

 

Appendix Figure 1. Example of the Overlapping Feature for Regions Data
Model.

 

 

5"“
A

Appendix Figure 2A. Example of the Home Range Case frequency
Coverage With Polygons. 28. Example of the Home Range Case
frequency Coverage With Regions.

 

 

 

 

 

 

68

Reclassifications of Land Use to Deer Use

APPENDIX 4

 

 

New New Class Old Code Old Class Name
Code Name

0 Background 0 Background

1 Rare Use 1 ' High Intensity Urban

1 Rare Use 2 Low Intensity Urban

1 Rare Use 3 Extractive

1 Rare Use 44 Barren Land

1 Rare Use 45 Water

1 Rare Use 46 Urban Grassland

2 Summer Use 8 Herbaceous Openiand

2 Summer Use 9 Shrubland

2 Summer Use 14 Northern Hardwood

2 Summer Use 15 Northern Hardwood/Conifer

2 Summer Use 16 Aspen/Birch

2 Summer Use 19 Oak

2 Summer Use 31 Emergent Wetland/Wet Meadow

2 Summer Use 33 Lowland Broad-Leaved Deciduous
Shrubland

2 Summer Use 37 Mixed Lowland Hardwood

3 Winter Use 42 Mixed Lowland Conifer/Hardwood

3 Winter Use 35 Other Forested Wetland

3 Winter Use 39 Black Spruce

3 Winter Use 23 White Pine

3 Winter Use 29 Cedar/Spruce/Fir

3 Winter Use 34 Lowland Broad-Leaved Evergreeen
Shrubland

3 Winter Use 38 Lowland Jack Pine

3 Winter Use 41 Northern White Cedar

3 Winter Use 47 Lowland Needle-Leaved Evergreen
Shrubland

4 Summer Other 4 Agricultural Crops

4 Summer Other 6 OrchardNineyard

4 Summer Other 13 Other Broad-Leaved Deciduous Forest

4 Summer Other 32 Other Lowland Shrub

5 Winter Other 22 Other Coniferous Forest

5 Winter Other 25 Upland Jack Pine

5 Winter Other 24 Red Pine

69

CHISQUARE ANALYSES

APPENDIX 5

Appendix 5a. Chi-Square Tests Comparing Deer Use Categories in the 3 Case Frequency
Areas for the Home Range Case Frequency Coverage

 

 

 

 

 

 

 

 

 

 

 

Observed
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 544.20 2925.14 1151.08 657.87 1043.16 6321.46
Medium 93.08 547.67 208.50 120.84 118.55 1088.65
High 40.24 372.02 137.35 35.41 51.24 636.26
Total 677.52 3844.83 1496.93 814.13 1212.95 8046.36
Expected
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 532.28 3020.62 1 176.03 639.60 952.93 6321 .46
Medium 91.67 520.19 202.53 110.15 164.11 . 1088.65
High 53.57 304.03 118.37 64.38 95.91 636.26
Total 677.52 3844.83 1496.93 814.13 1212.95 8046.36
Chi-Square
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 0.27 3.02 0.53 0.52 8.54 12.88
Medium 0.02 1.45 0.18 1.04 12.65 15.33
High 3.32 15.21 3.05 13.03 20.81 55.41 p-value
Total 3.61 19.67 3.75 14.59 42.00 83.63 < 0.0001
Percent Deer Use
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 8.61 46.27 18.21 10.41 16.50 100.00
Medium 8.55 50.31 19.15 11.10 10.89 100.00
High 6.32 58.47 21.59 5.57 8.05 100.00

70

APPENDIX 5 (cont’d)

Appendix 5b. Chi-Square Tests Comparing Deer Use Categories in the 3 Case Frequency
Areas for the Kernel Core Area

 

 

 

 

 

 

 

 

 

 

 

Observed
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 50.48 273.58 72.94 56.75 51 .02 504.77
Medium 75.86 449.65 161 .64 85.80 104.53 877.48
HiLh 40.05 373.23 137.37 35.21 50.40 636.26
Total 166.39 1096.46 371.96 177.76 205.95 2018.51
Expected
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 41.61 274.19 93.02 44.45 51.50 504.77
Medium 72.33 476.65 161 .70 77.27 89.53 877.48
Hm 52.45 345.61 117.25 56.03 64.92 636.26
Total 166.39 1096.46 371.96 177.76 205.95 2018.51
Chi-Square
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 1.89 0.00 4.33 3.40 0.00 9.63
Medium 0.17 1.53 0.00 0.94 2.51 5.15
High 2.93 2.21 3.45 7.73 3.25 19.57 p-value
Total 5.00 3.74 7.79 12.08 5.76 34.36 < 0.0001
Percent Deer Use
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 10.00 54.20 14.45 11.24 10.11 100.00
Medium 8.65 51.24 18.42 9.78 11.91 100.00
High 6.29 58.66 21.59 5.53 7.92 100.00

71

 

APPENDIX 5 (cont’d)

Appendix 5c. Chi-Square Tests Comparing Deer Use Categories in the 3 Case Frequency
Areas for the Minimum Convex Polygon Core Area

 

 

 

 

 

 

 

 

 

 

 

Observed
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 102.73 574.19 194.67 143.41 136.68 1151.68
Medium 77.13 440.46 151.64 83.82 89.71 842.77
HigL 38.85 370.82 136.49 31.57 50.57 628.31
Total 218.72 1385.48 482.80 258.79 276.96 2622.75
Expected
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 96.04 608.38 212.00 113.64 121.62 1151.68
Medium 70.28 445.20 155.14 83.16 89.00 842.77
High 52.40 331.91 115.66 62.00 66.35 628.31
Total 218.72 1385.48 482.80 258.79 276.96 2622.75
Chi-Square
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 0.47 1 .92 1 .42 7.80 1 .86 13.47
Medium 0.67 0.05 0.08 0.01 0.01 0.81
High 3.50 4.56 3.75 14.93 3.75 30.50 p-value
Total 4.64 6.53 5.25 22.74 5.62 44.78 < 0.0001
Percent Deer Use .
Case Rare Summer Winter Summer Winter
Frequency Use Use Use Other Other Total
Zero 8.92 49.86 16.90 12.45 11.87 100.00
Medium 9.15 52.26 17.99 9.95 10.64 100.00
High 6.18 59.02 21.72 5.02 8.05 100.00

72

LITERATURE CITED

Conover, W. J. and R. L. iman. 1981. Rank transformations as a bridge
between parametric and nonparametric statistics. Amer. Stat. 35: 124-
129.

Davenport, L. 1941. Deer Nutrition Studies. MDNR unpublished data.
DeLaune, MG. (2001). XTools ArcView Extension (Version 6/1/2001).

Elkie, P., R. Rempel and A. Carr. 1999. Patch Analyst User’s Manual. Ont. Min.
Natur. Resour. Northwest Sci. & Technol. Thunder Bay, Ont. TM002.
16pp + Append.

Felix, A. 2003. Development of Landscape-Scale Models To Describe Habitat
Potential of White-Tailed Deer (Odocoileus Virginianus) In Michigan. MS.
Thesis, Michigan State University, East Lansing, Michigan, USA.

Fitzgerald, S. D., S. M. Schmitt, L. Sullivan. 1997. Annual update on
tuberculosis surveillance of white-tailed deer in Michigan. Anim. Health
Diagn. Lab. And Dep. Pathol., Coll. Vet. Med., Michigan State University,
East Lansing. Mich. Vet. Conf.: 249-252.

Garner, M. S. 2001. Movement Patterns and Behavior at Winter Feeding and
Fall Baiting Stations n a Population of White-Tailed Deer Infected with
Bovine Tuberculosis in the Northeastern Lower Peninsula of Michigan.
Ph.D. Thesis, Michigan State University, East Lansing, Michigan, USA.

Hickiing, G. March 2002. Dynamics of Bovine Tuberculosis in Wild White-tailed
Deer in Michigan. Wildlife Division Report No. 3363. MDNR, Lansing, MI,
USA.

Hooge, P. N. and B. Eichenlaub. 2000. Animal movement extension to Arcview.
ver. 2.0. Alaska Science Center - Biological Science Office, US.
Geological Survey, Anchorage, AK, USA.Kohn, B. E., J. J. Mooty. 1971 .
Summer Habitat of White-Tailed Deer in North-Central Minnesota.
Joumai of Wildlife Management. 35(3): 476-487.

Johnson, R. A., and D. W. Wichem. Applied Multivariate Statistical Analysis.
Englewood Cliffs, N.J., Prentice Hall. 1982. 594 pp.

Knapp, B. D. 1988. Soil survey of Presque Isle County, Michigan. U.S.D.A. Soil
Conservation Service. 252 pp.

Levin, S. A.1992. The Problem of Pattern and Scale in Ecology. Ecology. 73(6):
1943-1967.

73

Malanson, G. P., and M. P. Armstrong. 1997. Issues in Spatial Representation:
Effects of Number of Cells and Between-Cell Step Size on Models of
Environmental Processes. Geographical 8 Environmental Modeling,
1(1):47-64.

McCaffery, K. R., J. Tranetzki, J. Piechura Jr. 1974. Summer Foods of Deer in
Northern Wisconsin. Joumai of Wildlife Management. 38(2): 215-219.

Michigan Department of Natural Resources. 1999. Deer Baiting Issues in
Michigan. MDNR, Lansing, Mi, USA.

Michigan Department of Natural Resources. 2000. Bovine Tuberculosis in
Michigan. MDNR, Lansing, MI, USA.
<http://www.dnr.state.mi.us/wildlife/Division/RoseLake/bovinetb/Publicatio
ns/2OOOBrochure%20for°/020web.html>

Michigan Department of Natural Resources. 2001. DMU 452 Baiting
Regulations for 2001 - As Passed — 9/14/2001. MDNR, Lansing, MI,
USA. <http://www.tb.com/>

Michigan Department of Natural Resources. 2003. Wildlife Surveillance
Summary. MDNR, Lansing, MI, USA. <http://www.tb.com/>

Michigan Department of Natural Resources. 2002. Baiting and Feeding
Regulations. MDNR, Lansing, MI, USA. <www.michigan.gov/dnr>

Mohr, C. O. 1947. Table of equivalent populations of North American mammals.
Am. Midland Nat. 37:223-249.

Moran, P. A. P. 1950. Notes on continuous stochastic phenomena. Biometrika.
37:17-23.

O’Brien, D. J., S. M. Schmitt, J. S. Fierke, S. A. Hogle, S. R. Winterstein, T. M.
Cooley, W. E. Moritz, K. L. Diegel, S. D. Fitzgerald, D. E. Berry, J. B.
Kaneene. 2002. Epidemiology of Mycobacterium bovis in free-ranging
white-tailed deer, Michigan, USA, 1995-2000. Preventive Veterinary
Medicine. 54:47-63.

Openshaw, S., and P.J.Tayior. 1979. A Million or so Correlation Coefficients:
Three Experiments on the Modifiable Areal Unit Problem. In Statistical
Methods in The Spatial Sciences, Wrigley, N. (ed.), pp: 127-144, London,
Routledge & Kegan Paul.

Openshaw, S. 1984. The Modifiable Areal Unit Problem, Concepts and
Techniques in Modern Geography, 38, Norwich, Geo Books.

74

Peyton, B. R. 2000. Wildlife Management: cropping to manage or managing to
crop? Wildlife Society Bulletin. 28(4): 774-779.

Rogers, L. L., J. J. Mooty. D. Dawson. 1981. Foods of White-Tailed Deer in the
Upper Great Lakes Region-A Review. USDA. Forest Service. 24 pp.

SAS Institute Inc., SAS IINSIGHT User‘s Guide, Version 8, Cary, NC: SAS
Institute Inc., 1999. 752 pp.

Schmitt, S. M., S. D. Fitzgerald, T. M. Cooley, C. S. Bruning-Fann, L. Sullivan, D.
Berry, T. Carlson, R. B. Minnis, J. B. Payeur, and J. Sikarskie. 1997.
Bovine tuberculosis in free-ranging white-tailed deer from Michigan.
Joumai of Wildlife Diseases 33: 749-758.

Seaman, D.E. & Powell, RA 1996. An evaluation of the accuracy of kernel
density estimators for home range analysis. Ecology (Washington D C) 7722075-
2085.

Sitar, K. L. 1996. Seasonal Movements, Habitat Use Patterns, and Population
Dynamics of White-tailed Deer (Odocoileus virginianus) in an Agricultural
Region of Northern Lower Michigan. MS. Thesis, Michigan State
University, East Lansing, Michigan, USA.

Stonner, F. A., W. A. Bauer. 1980. Summer Forage Use by Tame Deer in
Northern Michigan. Joumai of Wildlife Management. 44(1): 98-106.

Turner, M. G., R. V. O’Neill, R. H. Gardner and B. T. Milne. 1989. Effects of
changing spatial scale on the analysis of landscape pattern. Landscape
Ecology. 3(3/4):1 53-162.

Van Deelen, T. R., K. S. Pregitzer, and J. B. Haufler. 1996. A Comparison of
Presettlement and Present day Forests in Two Northern Michigan Deer
Yards. The American Midland Naturalist. 135: 181-194.

Williams, T. E. 1992. Soil survey of Alcona County, Michigan. U.S.D.A. Soil
Conservation Service. 409 pp.

Whitcomb, S. D. 1999. Deer and Elk Feeding Issues in Michigan. Wildl. Div.
Briefing Paper. Michigan Department of Natural Resources, Wildlife
Division, Lansing, MI. 9 pp.

Worton, B.J. 1989. Kernel methods for estimating the utilization distribution in
home-range studies. Ecology 70:164-168.

75

76

 
 

A

IIgllzllglljlllilyinil

.lil