THESIS

llllllllllllllllllllllllllllllllllllllllllllllllll

23 02079 9262

This is to certify that the

thesis entitled

Assessing Social-ecological Feasibility of Land
Acquisition for Refuge Expansion

presented by

Amanda A. McDonald

has been accepted towards fulfillment
of the requirements for

M.S. degreein Fish. & Wildl.

l/QWA

Major professor

 

 

 

 

O-7639 MSUis an Affirmative Action/Equal Opportunity Institution

 

 

 

Mishigan State
University

 

 

ASSESSING SOCIAL—ECOLOGICAL FEASIBILITY OF LAND ACQUISITION
FOR REFUGE EXPANSION

By

Amanda A. McDonald

A THESIS

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

MASTER OF SCIENCE

Department of Fisheries and Wildlife

2000

ABSTRACT

ASSESSING SOCIAL-ECOLOGICAL FEASIBILITY OF LAND ACQUISITION
FOR REFUGE EXPANSION

By

Amanda A. McDonald

To better protect biodiversity, it is desirable to purchase land to expand
the boundaries of protected areas such as wildlife refuges. Unfortunately, an
acquisition project may not be able to achieve the stated objectives even after a
lot of time, energy, and money have been invested. Using Shiawassee National
Wildlife Refuge as a case-study, I conducted a feasibility assessment on an on-
going acquisition project. First, I conducted a survey of the private landowners
within the acquisition area. The survey indicated that only 49.5% of the
respondents would be willing to sell and only 38.9% would be willing to sell at an
appraised market value. In addition, I found that a higher percentage of
landowners who had owned their land for less than 21 years were willing to sell
than other landowners and that a higher percentage of wetland owners were
willing to sell than non-wetland owners. Non-sellers were also predisposed to
refusing to work with the refuge to manage their land or to enter into a
conservation easement. The most expressed reason for not selling was simply
that the landowner had no reason to sell. Furthermore, the non-sellers hold
50.4% of the land area interspersed within the total acquisition area disrupting

habitat connectivity. Next, | created a model that would simulate land purchases

within the acquisition boundary. Employing hypothetical incentive levels, I found
that increasing the incentive to sell increased the amount of land that was
purchased. I also found that increasing the number of landowners contacted with
purchase offers shortens the time required to complete the acquisition and if the
survey information is assumed to be true across time, the expansion of
Shiawassee National Wildlife Refuge will take approximately 250 years to
complete. In the end, I found that less than half of the acquisition area would be
purchasable without using incentives of some sort. To facilitate the acquisition
project, the United States Fish and Wildlife Service should increase the number
of annual purchase offers made and offer non-monetary incentives to the

landowners for selling their land.

Copyright by
Amanda Ann McDonald
2000

In loving memory of my grandfather,
LeRoy McDonald

who saw me as greater than I am.

ACKNOWLEDGMENTS

There were many people who deserve praise for their help with this
research. First, and foremost, I would like to thank my guidance committee: Dr.
Jianguo Liu, Dr. Harold Prince, and Dr. James Hart. Then, there are Christopher
Lepczyk, Daniel Rutledge, Marc Linderman, and An Li who taught me to
challenge everything. Doug Longpre, the ever intrepid intern who stared at the
computer everyday until the land coverages were completed, thank you. Much
appreciation goes to Drs. Ben Peyton and Robert Holsman for helping to
construct the survey questionnaire and for teaching me everything about
conducting surveys. Also, thanks go to David Peters of Shiawassee National
Wildlife Refuge for being kind enough to aid in data collection,and to the United
States Fish and Wildlife Service for allowing me to do the work and showing
interest. I am grateful to Kierstan Kress who helped with the formation of the
ideas for the project as well as with the initial data collection. Finally, I must
acknowledge my funding sources, Michigan Agricultural Experimentation Station
and Wildlife Division of the Michigan Department of Natural Resources as well as
the people who helped to get the funding: Drs. Ben Peyton, Scott Winterstein,

and Henry Campa, lll.

vi

TABLE OF CONTENTS

List of Tables ...................................................................................... ix

List of Figures ...................................................................................... x

Chapter 1

Introduction ......................................................................................... 1
Land Acquisition .............................................................................. 1
The Expansion of Shiawassee National Wildlife Refuge ........................... 4
Overview of Thesis ........................................................................... 6
Literature Cited ............................................................................... 7

Chapter 2

Attitude Assessment of Landowners within the Acquisition

Boundary of a National Wildlife Refuge in Southern Michigan ..................... 10
Abstract ....................................................................................... 10
Introduction ................................................................................... 1 1
Methods ....................................................................................... 14

Study Area ............................................................................... 14
Survey .................................................................................... 15

Results ........................................................................................ 19
Discussion .................................................................................... 30
Acknowledgements ...................................................................... 36
Literature Cited .............................................................................. 36

Chapter 3

A Socio-ecological Simulation Model of Land Acquisition

to Expand a National Wildlife Refuge ...................................................... 39
Abstract ....................................................................................... 39

vii

Introduction ................................................................................... 40

Methods ....................................................................................... 42
Study Area ............................................................................... 42
Simulation Model ....................................................................... 43
Landowner Attitudes ................................................................ 45
Acquisition Scenarios ................................................................. 48

Results ........................................................................................ 51
Sensitivity of Non-Respondent Probabilities of Sale ........................ 53
Hypothetical vs. Survey—Derived Probabilities of Sale ...................... 54
Incentives ................................................................................ 58

Effects of Attempted Purchases on Time Required for

Completing the Acquisition ......................................................... 60
Discussion .................................................................................... 63
Acknowledgements ........................................................................ 65
Literature Cited ............................................................................. 65

Chapter 4
Conclusions and Recommendations ....................................................... 68
Literature Cited ............................................................................ 71

viii

LIST OF TABLES

Table 2-1. Frequency and percentage of landowners that answered yes, no, or
undecided to selling their land to the USFWS based on age, occupation, or
length of ownership.

Table 2-2. Landowners’ willingness to sell in relation to the type of land they
own.

Table 2-3. Frequency and percentage of landowners responses to the
alternative questions of whether they would be willing to work with the USFWS to
manage their land or to enter into a conservation easement compared to their
responses to the original question of whether they would sell their land.

Table 2-4. Frequency and percentage of landowner responses compared
between the two alternative questions of working with the USFWS to manage the
land for wildlife or entering a conservation easement. The percentages within the
columns are of the column total, and likewise, the percentages within the rows
are of the row total.

Table 2-5. Frequency of high-score answers (4 or 5 on a scale of 1 to 5) and
overall mean scores (n= 35) of expressed reasons for not selling property to
USFWS.

Table 2-6. Quantification of the responses to the question of selling to the
USFWS by evaluating the areas of the parcels in question. These values are
compared to the total area accounted for by the survey responses and the total
proposed acquisition area.

Table 3-1. Examination of the sensitivity of the non-respondent probability of
sale. The numbers represent the average values (is.d.).

Table 3-2. Simulation results to examine survey importance. The numbers
represent the average values (is.d.).

Table 3-3. Simulation results using limited purchase attempts per year.

LIST OF FIGURES

Figure 1-1. Location of Shiawassee National Wildlife Refuge in Michigan.

Figure 1-2. Map of Shiawassee National Wildlife Refuge and the proposed
acquisition.

Figure 2-1. Frequency of landowner responses (yes, no, or undecided) to the
question of whether or not they would sell their land to the USFWS.

Figure 2-2. Frequency of landowner responses (yes, no, or undecided) to the
question of whether or not they would sell their land to the USFWS with an
adjustment for people who would sell, but not at appraised value.

Figure 2-3. Frequency distribution of landowner responses to the question of
when they would be willing to sell their land.

Figure 24. Map of Shiawassee National Wildlife refuge and the acquisition area
(from US. Fish and Wildlife Service, 1996) with the parcels owned by survey
respondents and responses highlighted.

Figure 3-1. Example of spatial distribution of purchased parcels resulting from
land acquisition simulation.

Figure 3-2. Amount of each land type purchased. The simulations were
executed using a low (0.333) and a high (0.500) hypothetical probability of sale
as well as probabilities of sale derived from a previously conducted survey.

Figure 3-3. Hydric versus non-hydric soil areas purchased. The simulations
were executed using a low (0.333) and a high (0.500) hypothetical probability of
sale as well as probabilities of sale derived from a previously conducted survey.

Figure 3-4. Purchased amounts of existing land types resulting from the land
purchase simulation investigating the effects of incentives. Three incentive levels
were used: no incentives, low incentives, and high incentives.

Figure 3-5. Hydric versus non-hydric soil areas purchased during the land
purchase simulations investigating the effects of incentives. Three incentive
levels were used: no incentives, low incentives, and high incentives.

Figure 36 Relationship between the number of attempted purchases per year
and the time required to purchase all of the land owned by willing sellers.

Chapter 1

Introduction

Land Acquisition

Land acquisition has historically been used by special interest groups and
conservation organizations to establish and expand wildlife reserves and
recreational areas. These groups include Ducks Unlimited (Ducks Unlimited
2000), The Nature Conservancy (Press et al. 1996), the United States Fish and
Wildlife Service (USFWS; United States Fish and Wildlife Service 1996), and the
United States Park Service (National Park Service 2000), as well as state
conservation organizations such as the Michigan Department of Natural
Resources (Michigan Department of Natural Resources 1996). Usually, these
reserves are created to counter habitat degradation and destruction (Reinecke et
al. 1989, Terbough and Schaik 1997, Liu et al. 1999) and offer protection for
wildlife and the natural systems that they require, as well as recreational
opportunities for the public (DellaSala et al. 1996, Turner and Rylander 1997).

In creating or expanding these reserves, 3 land acquisition project is
initiated. An acquisition project consists of two steps: the planning process and
the acquisition process. In the planning process, objectives for the reserve are
defined, an acquisition boundary is developed, and money is appropriated for the
purchases. After the planning is completed, the acquisition process begins
where the landowners are contacted with purchase offers for the management

rights to their land. The purchase offers can take any of several forms since

there are several ways that the management rights to the land can be acquired.
Two of the methods commonly employed are the direct purchase of the title fees
and the establishment of conservation easements (Meeks 1990, Press et al.
1996). Of the two methods, direct purchase of the title fees may be the oldest
method of land acquisition (Meeks 1990) and would be the preferable method
because it involves a single payment of a predetermined amount, which is
usually the lowest possible fair price (Kershow 1975). In addition, there are no
restrictions on the authority of the acquiring agency over the management of the
acquired land (Ramsey and Addison 1996, Press et al. 1996).

Time is an issue when dealing with objectives related to the conservation
of biodiversity, and an acquisition project can take several years to complete.
Obviously, it would be desirable to immediately acquire the land and begin
restoration and/or protection of the habitats contained within the area of interest.
Therefore, it would be beneficial to accelerate the acquisition project. Wright and
Tanimoto (1998) have suggested using geographic information systems (GIS) to
prioritize the land parcels within an area of interest based on habitat complexity.
This method would facilitate the planning process, but that is only part of an
acquisition project.

After the planning process is over and the acquisition process has begun,
the landowners are contacted with offers to buy their land. How the contacts are
done differs depending on which organization or agency is doing the acquiring. It
has been suggested by Ramsey and Addison (1996) that mass mailings of

purchase contracts to all the landowners and an assembly line method of

 

  

l Titabawassee s' k
. Q (.8 (f,
2 Shiawassee o /(({i'{
3 Cass R (r6 “'1/
4 Flint \ / \\ i
“ 3 \\\
j , \ \.
. , , 1 \\
- l,
. [L \.,_.
V ’ ' “ /
f \ ' f
o , /
L /
7
/

Figure 1-1. Location of Shiawassee National Wildlife Refuge in Michigan.

 

handling the contract responses would facilitate the acquisition process.
However, this can be labor intensive and financially inefficient, especially for
government agencies that have a lot of regulations that need to be met.

Finally, after years of execution, the acquisition project may or may not
have been successful in meeting the objectives that were originally defined for it.
If the acquisition project was not successful, then money has already been
invested and effort spent to no avail. Regardless of whether or not the
acquisition project will be accelerated, the first questions for an acquisition
project should always be about whether or not the acquisition is feasible and

whether or not it should be undertaken.

The Expansion of Shiawassee National Wildlife Refuge
In examining the feasibility of an acquisition project, I chose to use a
national wildlife refuge, Shiawassee National Wildlife Refuge (hearafter refuge),
which is currently undergoing an expansion. The refuge is located in Saginaw
County, Michigan, near Saginaw Bay (Figure 1-1) on the juncture of four rivers:
the Shiawassee, the Titabawassee, the Cass, and the Flint Rivers. The refuge
was initially established in 1953 by the USFWS to provide a migratory staging
area for waterfowl and to serve as a flood containment unit, but presently, the
goals also include maintaining the biodiversity of the area and providing habitat
for the native endangered species of the area (United States Fish and Wildlife

Service, 1996).

Currently, the refuge consists of 3680 contiguous hectares, a third of
which are classified as wetlands. The USFWS decided to attempt an expansion
of the refuge (United States Fish and Wildlife Service, 1996). An acquisition
boundary was drawn around 317 land parcels (3035 ha); and it extends outwards
on three of the rivers: the Shiawassee, the Titabawassee, and the Cass (Figure
1—2). Each parcel of land has been classified as high priority, medium priority or
low priority based on its evaluated habitat potential for both upland and wetland

species (United States Fish and Wildlife Service, 1996).

@ E Refuge

Proposed Acquisition Area

- Top Priority
Medium Priority

  

S Low Priority

 

Figure 1—2. Map of Shiawassee National Wildlife Refuge and the proposed
acquisition.

The USFWS followed the National Environmental Policy Act protocols for
planning this acquisition project (United States Fish and Wildlife Service 1996).
Public meetings were held to identify and discuss potential issues with the
proposed acquisition.

Now that the acquisition proposal was approved and some funding was
appropriated, landowners are being selectively contacted with offers to buy their
land. By selectively, I mean that the realty division contacts landowners based
on the refuge manager’s recommendation of who might be willing to sell (John
Connors, pers. comm.) It must be kept in mind that not all of the landowners
have been contacted or are even aware that the USFWS is planning this

expansion (John Connors, pers. comm; McDonald, unpublished data).

Overview of Thesis

It was my goal to create a simulation model that would project the results
(e.g., area purchased, habitat purchased, and distribution of purchases) of an
acquisition project. These results would then be used to assess the feasibility of
the acquisition project. To do this, I chose to use Shiawassee National Wildlife
Refuge as a case study since it was trying to expand. The work was divided into
two sections. The first was an assessment of the landowners’ attitudes within the
acquisition area towards selling to the USFWS. This could be considered a
preliminary feasibility assessment and is described in Chapter 2. The next step
was to incorporate the landowners’ attitudes into a simulation model that would

project the amount of land purchased, the amount of different habitat types

purchased (both existing and potential), and the distribution of the purchased
parcels. The model and the information that it needed are described in Chapter
3. Both Chapters 2 and 3 are written in manuscript format, so there is a little

repetition of information between these two chapters.

Literature Cited

Connors, John. 1999. Perscomm. United States Fish and Wildlife Service,
Realty Division. District 3 Offices, Minneapolis, MN.

DellaSala, D.A., J.R. Strittholt, R.F. N053, and BM. Olson. 1996. A critical role
for core reserves in managing Inland Northwest landscapes for natural
resources and biodiversity. Wild. Soc. Bull. 24: 209-221.

Ducks Unlimited. 2000. “Ducks Unlimited Conservation: How We Conserve”
Available: http://www.ducks.org/conservation/how_we_conserve.asp [Date
Visited: April 6, 2000]

Kershow, W.W. 1975. Land Acquisition. National Recreation and Park
Association, Arlington, VA. 26 pp.

Liu, J., Z. Ouyang, W. Taylor, R. Groop, Y. Tan, and H. Zhang. 1999. A
framework for evaluating effects of human factors on wildlife habitat: The
case of the giant pandas. Conservation Biology 13: 1360-1370.

Meeks, (3., Jr. 1990. State land conservation and growth management policy: A
Legislator’s Guide. National Conference of State Legislatures; Denver.

64 pp.

Michigan Department of Natural Resources. 1996. Shiawassee River State
Game Area Master Plan. Lansing, MI. 32 pp.

National Park Service. 2000. Strategic Plan: 2000-2005. US. Government
Printing Office, Washington, DC. 48 pp.

Press, D., D.F. Doak, and P. Steinberg. 1996. The role of local government in
the conservation of rare species. Conservation Biology 10:1538-1548.

Ramsey, OJ. and DS. Addison. 1996. Facilitating a multiparcel land acquisition
project in the Western Big Cypress Region of Collier County, Florida,
USA. Natural Areas Journal 16:36-40.

Reinecke, K.S., R.M. Kaminski, D.J. Moorhead, J.D. Hodges, and JR. Nassar.
1989. Mississippi Alluvial Valley. pp 203-247 i3 Habitat Management for
Migrating and Wintering Waterfowl in North America. Texas Tech
University Press, Lubbock. 560 pp.

Terbough, J. and GP van Schaik. 1997. Minimizing species loss: the
imperative of protection. pp 15 — 35 i_n Kramer, R., C. van Schaik, and J.
Johnson (ed.). Last Stand: Protected Areas & the Defense of Tropical
Biodiversity. Oxford University Press, New York. 242 pp.

Turner, J. and J. Rylander. 1997. Land use: the forgotten agenda. pp 60-75 i3
Chertow,M.R. and DC. Esty (ed). Thinking Ecologically: the Next
Generation of Environmental Policy. Yale University Press, New Haven.

271 pp.

United States Fish and Wildlife Service. 1996. Final Environmental Assessment
for Additions to Shiawassee National Wildlife Refuge. US. Government
Printing Office, Washington, DC. 103 pp.

Wright, GR. and PD. Tanimoto. 1998. Using GIS to prioritize land conservation
actions: integrating factors of habitat diversity, land ownership, and

development risk. Natural Areas Journal 18:38-44.

Chapter 2

Attitude Assessment of Landowners within the Acquisition
Boundary of a National Wildlife Refuge in

Southern Michigan

Abstract

To better protect biodiversity, it is desirable to purchase land to expand
the boundaries of protected areas such as wildlife refuges. Unfortunately, an
acquisition project may not be able to achieve the stated objectives if the
landowners are not willing to sell their land. Therefore, it is necessary to assess
the landowners’ willingness to sell their land as part of the feasibility assessment
effort. As a case study, I conducted a survey of landowners within the proposed
acquisition boundary of Shiawassee National Wildlife Refuge in Southern
Michigan. The United States Fish and Wildlife Service approved the proposal for
the expansion of the refuge, but heretofore, the feasibility of purchasing from the
adjacent private landowners was unclear. My survey results indicate that only
49.5% of the respondents would be willing to sell and only 38.9% would be
willing to sell at an appraised market value. Furthermore, I found that a higher
percentage of landowners who were younger than 60 years old were willing to
sell than the older landowners, that a higher percentage of landowners who had
owned their land for less than 21 years were willing to sell than other landowners,

that a higher percentage of wetland owners were willing to sell than non-wetland

10

owners, and that there were no large differences in willingness to sell among
occupations. Non-sellers were also predisposed to refusing to work with the
refuge to manage their land or to enter into a conservation easement. The most
expressed reason for not selling was simply that the landowner had no reason to
sell. Without additidnal incentives, only 13.7% of the 1215 ha accounted for in the
survey will be purchasable. Furthermore, the non-sellers hold 50.4% of the land
area interspersed within the total acquisition area disrupting habitat connectivity.
This study demonstrates how managers can assess the potential success of an
acquisition project before a lot of time and money is invested. I suggest that this
type of survey should also be useful to other land acquisition projects for wildlife

conservation.

Introduction

There is a continuing decline in biological diversity around the world
because wildlife habitats have become increasingly fragmented and degraded,
resulting in smaller and smaller populations of animals (DellaSala et al. 1996,
Terbough and Schaik 1997). To counter this habitat degradation and
destruction, many protected areas, such as wildlife refuges, have been created
(Reinecke et al. 1989, Terbough and Schaik 1997, Liu et al. 1999). These areas
offer protection for wildlife and for the natural systems that the animals require,
as well as recreational opportunities for the public (DellaSala et al. 1996, Turner
and Rylander 1997). Unfortunately, many of the protected areas are too small

(Fitter and Scott 1978). Generally speaking, the larger the area of protection, the

11

more the core area is buffered from human disturbances (DellaSala et al. 1996)
and the better the core area is for wildlife habitat, especially for species that are
very specialized in their habitat requirements (Saunders et al. 1991). To protect
these specialized species and biodiversity in general, many protected areas
would need to be expanded.

An expansion of protected areas could be accomplished through
purchasing the title fees, leasing, or contracting easements (Meeks 1990, Press
et al. 1996). Direct purchase of titles may be the oldest method of conservation
(Meeks 1990). It would be the most appealing approach since it concerns a
single payment of a predetermined amount, usually the lowest possible price
(Kershow 1975), and there are no restrictions on the authority of the agency over
the management of the purchased land (Ramsey and Addison 1996, Press et al.
1996). Also, a direct purchase can be the best way to safeguard the land from
the pressure of economic development (Press et al. 1996). Land purchasing is a
practice that has been employed by organizations such as the Nature
Conservancy (Press et al. 1996), the United States Fish and Wildlife Service
(USFWS) (United States Fish and Wildlife Service 1996), and state conservation
agencies (Michigan Department of Natural Resources 1996) to establish and
expand managed areas.

The first step of a land purchase project is usually to define the area of
interest and assess the ecological value of the parcels. The assessment enables
the acquiring agency to prioritize the land parcels within the acquisition boundary

for wildlife habitat (Wright and Tanimoto 1998). Then, the landowners have to

12

be contacted and given an offer for their land (Ramsey and Addison 1996). Land
parcel by land parcel, decisions are made regarding the purchase of the parcel
depending on its rated priority and the money available to complete the
acquisition. Sometimes, high priority land parcels are passed over to acquire
more acreage. Landowners usually have to be contacted several times before a
final decision is made. Because there are few predetermined guidelines,
deciding which parcels to purchase can make land acquisition a long, time-
consuming process. This process may continue until all the land within the
acquisition boundary is purchased or the funding for the project runs out.

One problem with the acquisition process that was realized through
conversations with area land managers is that while planning the project, the
willingness of the specific landowners to sell is often not assessed. Some land
addition projects are doomed to failure because some landowners will never sell
their land, some high value areas will not be attainable for social or economic
reasons, or not enough of the acquisition area would be attainable to produce
effective results. This type of project is not even worth attempting because the
applied effort would vastly outweigh the benefits. The feasibility of a project
should be assessed before lots of energy, money, and manpower are invested in
it.

One way to evaluate the feasibility of an acquisition project is to consider
the attitudes of the landowners towards selling. Previous research has indicated
that many people think there is a need for wildlife conservation and management

(Kellert 1978, Gill 1996), but how many individuals would give up their land for

13

such a purpose? To answer this question, I conducted a survey of the
landowners within the proposed boundary of an ongoing acquisition project
administered by the USFWS. The goal of this survey was to assess the attitudes
of the landowners towards selling to determine whether or not the acquisition
would be feasible. It is my hope that the approach used in this study can also
stimulate socio-economic feasibility assessment for other similar land purchase

projects.

Methods

Study Area

My study area was the Shiawassee National Wildlife Refuge (hereafter
refuge) in Saginaw County, Michigan that was designed to be a migratory bird
sanctuary and flood containment unit. The refuge was authorized in 1953 and
consists mostly of managed wetlands. It is located near Saginaw Bay of the
Lower Peninsula of Michigan at the junction of four rivers: the Shiawassee, the
Titabawassee, the Cass, and the Flint (Figure 1-1). Presently, the goals of the
refuge include maintaining the biodiversity of the area, providing habitat for the
native endangered species of the area, as well as the original goals of providing
a migratory staging area for waterfowl and flood containment (United States Fish
and Wildlife Service 1996).

Currently, the refuge consists of 3680 ha. The USFWS is interested in
purchasing land for an expansion of the refuge (United States Fish and Wildlife

Service 1996). The area of interest for the acquisition has been bounded and

14

consists of 3035 ha extending outwards from the refuge along three of the rivers:
the Shiawassee, the Titabawassee, and the Cass Rivers (Figure 1-2). The
majority of the lands in question either encompass wetlands or are within the
flood plain of the rivers since these lands are the most useful for the purposes of
the refuge. Each parcel of land has been classified as high priority, medium
priority or low priority based on its evaluated habitat potential for both upland and
wetland species (United States Fish and Wildlife Service 1996). The landowners
are being selectively contacted for the land acquisition project and decisions
about purchases are being made based upon the available monies for the

project.

Survey

To assess the attitudes of landowners towards the acquisition project, a
survey was conducted of the landowners within the acquisition boundary of the
refuge. A questionnaire was developed to assess the attitudes of the owners
towards selling their land to the USFWS and to gather some general information
about the owner and the status of the land. The primary question was whether or
not they would sell their land to the USFWS. Then, the owners were asked some
other information such as age, occupation, and how long they have owned the
land, to help analyze their attitudes in relation to the profile of respondents. If the
owner said that they were willing to sell their land, they were then asked to
describe how much of their land they would sell and were given the options of

selling all, all but the residential, or only a specified part of their land. Following

15

 

 

this line of questioning, they were asked if their property was currently on the
market, if they would sell for appraised value, and when in the next twenty years
they would be willing to sell. They were given six categories to choose from for
the time of sale: now, 1-2, 3-5, 6-10, 11-15, 16-20 years.

I wanted to know for what reasons people would not be willing to sell their
land, so I could assess the obstacles to the acquisition project. Thus, for the
people who responded that they would not sell their land or the people that would
not commit to an answer, several conditions were presented, and the respondent
was asked to rate each one as to its influence on the decision of whether or not
to sell. The conditions were: having children at home, farming the land, having
owned the land for generations, using the land for recreation purposes, not
agreeing with the USFWS practices, not being able to sell for the amount they
think the land is worth, and having no reason to sell. I offered a scale of 1 to 5
for the respondent to evaluate each condition with 1 being the least important
and 5 being the most important.

Some background information was collected to look at trends and profiles.
The owner’s year of birth and occupation were requested. Then, the owners
were asked how long they had owned the land in question and were given six
categories to choose from. The categories were $5 years, 6-10 years, 11-15
years, 16-20 years, 20-24 years, and 25+ years. Following the length of
ownership question, the owners were asked to describe their land. They were
given land type categories in which to fill out the acreage they own. The

categories were the total land owned, agricultural land, wetlands, forested lands,

16

open fields, residential land, and unclassified lands. This question was included
to identify any significant differences in response associated with owning a
certain type of land.

Two alternatives to selling were presented in the questionnaire. These
alternatives were to work with the USFWS in managing the land or to enter into a
conservation easement and allow the USFWS to manage the land while the
owner receives some form of remuneration (e.g., annual payments, tax write-offs,
or both). These will be important variables if the USFWS cannot appropriate
enough funding to complete the proposed acquisition. It was expected that some
of the people questioned would not know what a conservation easement entails
and would therefore answer “undecided.”

The survey began in May of 1998 and ended in August of 1998. A mailing
list of the 198 private owners of potential acquisition lands was obtained from the
USFWS. The list was a year old so some of the lands had already changed
hands. A letter was mailed to the recipients informing them that the survey would
be following. Thirteen of the letters were returned because the owners had
moved or the address was incorrect. In the two cases where the owners had
moved but the address was to the actual parcel of land, the address was
changed to “current owner.” However, for the 11 incorrect addresses or returned
letters that were not sent to the actual parcel, these addresses and names were
removed from the list. One person wrote back that the land in question had been
sold and gave me the name and address of the new owner. This new owner was

already on the list so I added that parcel to his holdings.

17

Of the 187 questionnaires that were sent out, 104 (56%) were returned.
Nine of these responses were notification that the recipient had already sold the
land either to the Michigan Department of Natural Resources or to someone else.
The remaining 95 (51% of the total questionnaires sent out) were reported for
this study. Unfortunately, the addresses were my only available method for
contacting the landowners, so no non-response follow-up was possible.

From the responses, I examined the number of respondents who
answered yes and no to the question of whether or not they would be willing to
sell. I also analyzed their willingness to sell at appraised value and how much
land they would sell. To search for response patterns, I examined the ages,
occupations, and lengths of ownership against their willingness to sell at
appraised value. The ages were grouped into five-year classes because of the
small population. Occupations were divided into four groups: laborer,
professional, self-employed, and retired. The laborer category consisted of
factory workers and other labor-intensive occupations, while the self-employed
category consisted of farmers and people who stated they were self—employed.
The professional category included businessmen, doctors, teachers, and any
other white-collarjobs. There were only three cases where the household had
two occupations, but for each, both occupations fell within the same category so
this was not a problem.

The reasons for not selling were analyzed in two ways. First, I compared
the average rated importance placed upon the condition. If a respondent left a

reason for not selling blank, but rated at least one of the others, I assumed that

18

the blank reason held no importance and inserted a value of 0 into the data set.
Secondly, using an answer of 4 or 5 to suggest that the reason held meaningful
importance, I examined the frequency of a meaningful rating.

The types of land people owned were compared to their willingness to sell
at appraised value. Since most of the parcels contain more than one land type,
the percentage of willing sellers was calculated for each land type independently
of the others, meaning that most landowners will be in more than one group. The
opposite of each group (e.g., agriculture vs. non-agriculture) was also calculated

for comparison so there were 10 groups total for the land type analysis.

Results
An overall profile of the respondents showed that the ages ranged from 32
to 91 years old with an average of 58.9 (s.d. = 14.1). For occupations, 65% of
the respondents were quite evenly distributed between the categories of laborer,
professional, and self-employed. The remaining 35% were retired. Regarding
the length of land ownership, half of the respondents have owned their land
parcels for more than 20 years. Furthermore, 38.2% of the respondents have

owned their parcels for over 25 years.

19

 

 

01
O
4

 

45-

 

(A) (A) -b
O (J1 O
L l L

N
1L4

_x_x
0010010
1

l

 

Frequency of response
N
01

l

 

 

 

 

 

 

 

 

Yes No Undecided

Type of response
Figure 2-1. Frequency of landowner responses (yes, no, or undecided) to the

question of whether or not they would sell their land to the USFWS.
Forty-seven of the respondents (49.5%) said that they would be willing to
sell at least a portion of their land to the USFWS (Figure 2-1). Thirty—six (37.9%)
replied that they would definitely not sell their land to the USFWS. Twelve
(12.6%) did not commit to one answer or the other. However, if the answers to
the initial sell question were analyzed in conjunction with whether or not they
would sell at appraised value, the distribution evened out (Figure 2-2). Ten of the
people who answered yes to whether or not they would sell also answered no to
selling for appraised value, which puts them into the conditional category
meaning that there was no commitment to a yes or no answer and incentives for

selling may be necessary.

20

 

 

 

 

 

 

Frequency of response
l\)
O
I

 

 

 

 

 

 

 

 

Yes No Undecided
Type of response

Figure 2-2. Frequency of landowner responses (yes, no, or undecided) to the
question of whether or not they would sell their land to the USFWS with an
adjustment for people who would sell, but not at appraised value.

Of the people that said they would be willing to sell their land, 20 (42.5%)
said that they would definitely sell all of their land, 4 (8.4%) said all they would
sell was the non—residential portion of their land, and 2 (4.3%) said that they
would only sell a specific part of their land, usually the wetland portion of their
property. The rest remained non-committal as to the amount of land that they
would sell. In reply to when they would be willing to sell, nearly fifty percent (17
of 35) of the respondents answered “immediately” (Figure 2-3). Only two
respondents answered that their property was currently on the market. One of
the two said that he or she would sell to the USFWS, while the other was

uncertain.

21

Frequency Of response
_x _x _\ _\ _x N
O N J}- O) 00 O N b 0) 00 O
l J 1 l I l 1 I 1 r

 

 

 

Wm H

I l I

now 1-2 3-5 6-10 11-15 16-20
Time (years from now)

 

Figure 2-3. Frequency distribution of landowner responses to the question of
when they would be willing to sell their land.

To examine response patterns, the age of the owner, occupation of the
owner, the length of ownership (Table 2-1), and the type of land owned were
analyzed (Table 2-2). The ages were grouped into two groups because of the
small sample size. Since the median age of 58 was close to the senior citizen
delineation of 60 years old, I chose 60 years old as a dividing point. The
respondents under the age of 60 showed a greater number of willing sellers and
only a few uncertain responses (Table 2-1). The respondents over the age of 60
showed similar numbers for each answer.

After grouping the occupation responses, I found that the occupation
category that received the most responses was the retired category (Table 2—1).

Of the 85 people who responded to the occupation question, 30 (35.3%) were

22

 

 

 

3. $8. 5 E Aswan; 2 388$ : 28> + 5
9:28:26
ms 38s: s @028 3 3.39 E 28> Ev .6 98>
om 20$.va s @00va 2 @003 2 88m
t 38.5 m Asa. E s 3%. E N 8288.65
2 Scream €8.st m @039 : 593
cm Asoms a $00.93 w Soon 3 m 85353.8 88880
ow @039 ms @058 2 $00.09 NF 5-8
h. Ashe: m Asses 2 Seems mm. 8-8 86 2838?
:38. 820mg: 02 mm> 0305 mwcoawom
owcoqmmm .6 83.

.QEmLmEso .6 £92 .5 6033380 due :0 Emma w>>.._m: 9: 9. 653 :9: 9.53

9 32026:: .0 .o: d?» voaosmcm RE wamcsoucfl u6 ommEmEma ncm Sconce... TN 2an

23

retired. The greatest difference in response was for the laborer group with 11
answering yes, 5 answering no, and 2 undecided. The self-employed group had
a low number of sellers but was split almost evenly between no’s and
undecided’s. The professional group was similar with the yes’s and no’s receiving
equal support. Occupation was not an indication of whether or not the
landowner would sell.

The responses to the question about the length of ownership (n=89) had a
median that fell within the 21 to 25 year category so two fairly even groups could
be formed using 21 years as the dividing line (Table 2-1). The number of willing
sellers who have owned their land for less than 21 years was twice the number of
willing sellers who have owned their land for more than 21 years. The lower
percentage of sellers within the people who owned their land for 21 or more
years could reflect sentimental value of the land to the owners.

After examining the different land types (Table 2-2), I found that the
wetland category had a higher percentage of willing sellers. Fifty-six percent
were willing to sell. The higher percentage of willing sellers for wetland owners
may be a result of legislation that limits their activities involving the wetland
portions of their land. The non-residential group had a higher percentage of
willing sellers, which might be related to the idea that the land is non-essential to
them if they do not live on it. The other land type groups were more even in their
distribution among the answers.

In the two alternative choices to selling (working with the USFWS or the

conservation easement), a total of 74 people responded, but not all 74 answered

24

Table 2-2. Landowners’ willingness to sell in relation to the type of land they

 

 

own.
Type of Response
Land type Yes No Undecided Total
Agriculture 11 (32.4%) 15 (44.1%) 8 (23.5%) 34
Non-agriculture 24 (44.4%) 19 (35.2%) 11 (20.4%) 54
Wetland 13 (56.5%) 4 (17.4%) 6 (26.1%) 23
Non-wetland 22 (33.8%) 30 (46.2%) 13 (20.0%) 65
Forested 13 (38.2%) 12 (35.3%) 9 (26.5%) 34
Non-forested 22 (40.7%) 22 (40.7%) 10 (18.6%) 54
Open fields 7 (35.0%) 8 (40.0%) 5 (25.0%) 20
Non-open 28 (41.2%) 26 (38.2%) 14 (20.6%) 68
Residential 23 (35.4%) 26 (40.0%) 16 (24.6%) 65
Non-residential 12 (52.2%) 8 (34.8%) 3 (13.0%) 23

 

25

Table 2-3. Frequency and percentage of landowners responses to the

alternative questions of whether they would be willing to work with the USFWS to

manage their land or to enter into a conservation easement compared to their
responses to the original question of whether they would sell their land.

Work with USFWS

Conservation easement

 

 

Answer to
original question Yes No Undecided Yes No Undecided

Yes 13 9 15 13 7 16
35.1% 24.3% 40.6% 35.7% 17.9% 46.4%

No 5 14 8 1 15 12
18.5% 51.9% 29.6% 3.6% 53.6% 42.8%

Undecided 2 2 5 1 3 4
22.2% 22.2% 55.6% 12.5% 37.5% 50.0%

Total 20 25 28 15 25 32
27.4% 34.2% 38.4% 20.8% 34.7% 44.5%

 

26

both questions (Table 2-3). The most frequent answer was “undecided" in either
case (Table 2-3). When asked whether they would work with the USFWS, 25
(34.2%) said no, 20 (27.4%) said yes, and 28 (38.4%) said they were undecided.
Comparing these answers to the original question of selling (Table 2-3), I found
that people who would not sell their land were more likely not to want to work
with the USFWS in managing their lands with 51.9% (14 out of 27) answering
negative. When the landowners were asked whether they would be interested in
entering a conservation easement, 25 (34.7%) said no, 15 (20.8%) said yes, and
32 (44.4%) said they were undecided (Table 2-3). The higher number of
undecided’s in the easement question may be the result of ignorance of what a
conservation easement entails. Again, the people who answered no to the
original question were predisposed to answering no to entering into a
conservation easement with over 50% answering no (Table 2-3).

Comparing the two alternative choices, I found that the same number of
individuals responded no to each question (Table 2-3). However, only 19 of
them responded no to both questions (Table 2-4). About half (9) of the people
that would be interested in working with the USFWS responded “no” or
“undecided” to the easement. Only five of the people who would not work or
were undecided about working with the USFWS would agree to the easement.
Three of the people who were undecided about working with the USFWS gave a
definitive negative to entering the conservation easement and allowing the
USFWS to manage their land. The only category that was not found to have a

large variation in the distribution was the group of people who answered yes to

27

Table 2-4. Frequency and percentage of landowner responses compared
between the two alternative questions of working with the USFWS to manage the
land for wildlife or entering a conservation easement. The percentages within the
columns are of the column total, and likewise, the percentages within the rows
are of the row total.

Response to the conservation easement

 

Response to
working with the

 

USFWS Yes No Undecided Total
Yes 10 (52.6%) 3 (15.8%) 6 (31.6%) 27
(26.7%) (12.0%) (64.5%)
No 1 (4.0%) 19 (76.0%) 5 (20.0%) 25
(6.7%) (76.0%) (16.1%)
Undecided 4 (14.8%) 3 (11.1%) 20 (74.1%) 19
(66.6%) (12.0%) (19.4%)
Total 15 25 31 71

 

28

Table 2-5. Frequency of high-score answers (4 or 5 on a scale of 1 to 5) and
overall mean scores (n= 35) of expressed reasons for not selling property to
USFWS.

Frequency of

 

high-score
answers Mean score (sd)
Having children at home 5 1.057 (1.679)
Land is being farmed 15 2.486 (2.228)
Land has been in family for generations 12 2.029 (2.203)
Land is used for recreation 4 2.286 (2.283)
Do not agree with the USFWS strategies 13 2.114 (2.259)
Cannot sell for perceived value 9 1.686 (2.153)
No reason to sell 28 4.171 (1.618)

 

29

working with the USFWS (Table 2-4). The people who responded no to either
question were more likely to respond no to the other question with more than
70% answering no (19 out of 25). People who responded yes to the
conservation easement were more likely to work with the USFWS with two-thirds
answering yes (10 out of 15).

Among the expressed reasons that landowners would not sell (Table 2-5),
having no reason to sell was the most frequent response (28 rated 4 or higher).
The reason that received the lowest response was that they had children living at
home. Since nearly half of the people who responded were over 60 years old
and past the typical child-rearing age, this answer was not surprising. There
were no noticeably large differences between the frequencies of meaningful
responses (4 or higher) among that the land was being farmed, that the land had
been in the family for generations, that the land was used for recreational
purposes, and that the respondent did not agree with the USFWS’s management
practices. Having no reason to sell had the highest mean (Table 2-5) and having
children at home had the lowest, corresponding with the results from the

frequency of meaningful responses.

Discussion
The results indicate that without extra measures (eg. incentive),
completion of the proposed acquisition project will not be feasible since only 50%
of the survey respondents said that they would be willing to sell their land and

only 38% would be willing to sell at the appraised value. It should be kept in

30

mind that these percentages refer to the number of landowners willing to sell, but
give no indication of the amount of land that would then be acquired. By linking
the survey response data with land parcel data and assuming that sellers will sell
all of their land, I found that of the 1215 hectares that can be accounted for with
the survey, at most 166 ha could be purchased at appraised value (Table 2-6).
This represents only 13.7% of the acreage accounted for by the survey (1215 ha)
and 5.5% of the total acquisition area (3035 ha). Adding in the people who would
positively be willing to sell for more than appraised value only brings the number

up to 21.4% of the surveyed acreage and 8.6% of the total. In addition, based

Table 2-6. Quantification of the responses to the question of selling to the
USFWS by evaluating the areas of the parcels in question. These values are
compared to the total area accounted for by the survey responses and the total
acquisition area.

 

Area % of the % of the total
accounted for surveyed acquisition
Response to question of selling by survey (ha) area area
Yes, at appraised value 165.9 13.7 5.5
Yes, but not at appraised 260.3 21.4 8.6
value
Total of the yes’s 426.2 35.1 14.0
Undecided 176.9 14.6 5.8
No 611.5 50.3 20.1

 

31

on my survey data, 612 ha will not be available for purchase because it is owned
by the non-willing sellers (Table 2-6).

Parcels that cannot be purchased are interspersed within the matrix
(Figure 2-4) in a way that is not helpful for habitat connectivity. Unfortunately,
while a survey like the one presented in this study allows wildlife managers to
assess the attitudes of the respondents towards selling, the survey gives no
indication of how much land and what kind of land will be available for purchase
from the non-respondents. If managers assume that the respondents are
representative of the non-respondents, they can calculate the percentage of
parcels available for purchase, but they cannot determine which parcels will be
purchased using the survey information alone. If more information is needed to
decide the feasibility of an acquisition project, the managers would either need a
definitive response from every landowner or a method to predict which parcels
could be purchased.

For prediction purposes, a determinant that can be specified down to the
landowner of each land parcel is needed. This determinant would allow for
individual predictions to be made about the purchasability of each parcel. In this
particular case, the length of ownership, age of the owner, and type of land could
be used to project a probability of sale onto each parcel if the respondents are
representative of the non-respondents. Using general information that could be

easily collected about the owner or the property would aid in determining where

32

 

Q 1:: Refuge
Proposed Acquisition

- N o
Undecided

    

Figure 2-4. Map of Shiawassee National Wildlife refuge and the acquisition area
(from US. Fish and Wildlife Service, 1996) with the parcels owned by survey
respondents and responses highlighted.

33

the gaps would be inside the acquisition boundary, what the resulting refuge
would look like, and how much useful wildlife habitat would be added. It would
also allow for the assessment of the effectiveness of the acquisition towards the
initial objectives, in this case, wildlife production and habitat conservation.

This survey also allows the planners to evaluate the reason landowners
would not sell their land and to develop strategies for targeting these issues.
Since the most stated reason for not selling was that there was no reason to sell,
the planners could begin to develop persuasive arguments to counter this. I am
not suggesting a devaluation of the property or condemnation; instead, I
recommend educating the owner as to why the acquisition is important, why their
land is beneficial to wildlife, and how difficult management of habitat tends to be.
Also, services desired by the landowner could be traded as extra incentive.
Some of the comments written on the survey suggest that landowners in this
area are concerned with the decreasing access to public lands for recreational
purposes. Programs could be initiated to address this decreasing access, and
perhaps, the landowners would feel a little more sympathetic to the goals of the
refuge. This method could possibly work if the only obstacle is lack of motivation
to sell.

It was unfortunate that the landowners who would not sell their land were
also less disposed to letting management occur on their land. This does not
necessarily imply that the owners are not interested in managing their lands to
improve wildlife habitat. It may be that the owners are not interested in letting the

USFWS have any involvement with their land. This aversion to the USFWS

34

managing private property may stem from a feeling of distrust or imposition, or it
could come from a perceived lack of credibility. To overcome this obstacle, more
effort would have to be made on the part of the agency in public relations.

To facilitate the execution and planning of an acquisition project, some
methods have been suggested, such as using Geographic Information Systems
(GIS; Wright and Tanimoto 1998) and predefined contracts in organized mailings
(Ramsey and Addison 1996). However, both of these proposed methods are
used after the planning is initiated. With these previous methods, an agency has
to execute the acquisitions before they find out whether or not the expended
effort will result in a new reserve that will meet their expectations and objectives.
To determine whether an acquisition is worth pursuing, it is necessary to
understand the attitudes of the landowners. A survey similar to the one
conducted in this study would provide the information needed to assess the
feasibility of a potential acquisition before all the time and effort are spent. It
would also provide some information that could be used to plan strategies for
persuading landowners to sell or cooperate in the management of the area.

The next step in this direction of inquiry is to create land acquisition
simulation models in order to facilitate the entire planning process. Wright and
Tanimoto (1998) have already suggested that GIS could be used to prioritize the
land parcels based upon such characteristics as physical location and land use
to facilitate the planning process of an acquisition project. Ramsey and Addision
(1996) have proposed a method for facilitating the actual purchases through

generalized pricing and mass contacts. However, to truly facilitate a land

35

acquisition project, a dynamic simulation model needs to be created that includes
biological, geographic, economic, and sociological features. A model would
allow the planners to know when to proceed, how feasible and effective the
project would be, when to give up and cut their losses, and what the long-term
consequences of different alternatives would be (Liu et al. 1995). This attitude
survey is a step towards the goal of simulation models that can be used as
guidance for agencies as they contemplate expansions for their management

areas.

Acknowledgements
Funding for this project was provided by the Michigan Agricultural
Experimentation Station and the Wildlife Division of the Michigan Department of
Natural Resources. I would also like to thank Robert Holsman of Michigan State
University for assisting in the preparation of the questionnaire and Scott
Winterstein and Henry Campa of Michigan State University for helping to acquire

funding for this study.

Literature Cited
DellaSala, D.A., J.R. Strittholt, R.F. Noss, and BM. Olson. 1996. A critical role
for core reserves in managing Inland Northwest landscapes for natural
resources and biodiversity. Wild. Soc. Bull. 24: 209-221.
Fitter, R. and P. Scott. 1978. The Penitent Butchers. Fauna and Flora

Preservation Society, London.

36

 

Gill, RB. 1996. The wildlife professional subculture: The case of the crazy aunt.
Human Dim. of Wildl. 1:60-69.

Kershow, W.W. 1975. Land Acquisition. National Recreation and Park
Association, Arlington, VA. 26 pp.

Liu, J., J.B. Dunning, Jr., and HR. Pulliam. 1995. Potential effects ofa forest
management plan on Bachman’s Sparrows (Aimophila aestivalis): linking
a spatially explicit model with GIS. Conservation Biology 9:62-75.

Liu, J., Z. Ouyang, W. Taylor, R. Groop, Y. Tan, and H. Zhang. 1999. A
framework for evaluating effects of human factors on wildlife habitat: The
case of the giant pandas. Conservation Biology 13: 1360-1370.

Meeks, G., Jr. 1990. State Land Conservation and Growth Management Policy:
A Legislator’s Guide. National Conference of State Legislatures; Denver.
64 pp.

Michigan Department of Natural Resources. 1996. Shiawassee River State
Game Area Master Plan. Lansing, MI. 32 pp.

Press, D., D.F. Doak, and P. Steinberg. 1996. The role of local government in
the conservation of rare species. Conservation Biology 10:1538-1548.

Ramsey, OJ. and US. Addison. 1996. Facilitating a multiparcel land acquisition
project in the Western Big Cypress Region of Collier County, Florida,
USA. Natural Areas Journal 16:36-40.

Reid, F.A., J.R. Kelley, Jr., T.S. Taylor, and L.H. Frederickson. 1989. Upper

Mississippi Valley wetlands — Refuges and moist-soil impoundments. pp

37

181-202 i_n Habitat Management for Migrating and Wintering Waterfowl in
North America. Texas Tech University Press, Lubbock. 560 pp.

Reinecke, K.S., R.M. Kaminski, D.J. Moorhead, J.D. Hodges, and JR. Nassar.
1989. Mississippi Alluvial Valley. pp 203-247 i_n Habitat Management for
Migrating and Wintering Waterfowl in North America. Texas Tech
University Press, Lubbock. 560 pp.

Saunders, D.A., R.J. Hobbs, and CR. Margules. 1991. Biological
consequences of ecosystem fragmentation: a review. Conservation
Biology 5:18-32.

Terbough, J. and GP van Schaik. 1997. Minimizing species loss: the
imperative of protection. pp 15 — 35 in Kramer, R., C. van Schaik, and J.
Johnson (ed.). Last Stand: Protected Areas & the Defense of Tropical
Biodiversity. Oxford University Press, New York. 242 pp.

Turner, J. and J. Rylander. 1997. Land Use: the forgotten agenda. pp 60-75 i_n
Chertow,M.R. and DC. Esty (ed.). Thinking Ecologically: the Next
Generation of Environmental Policy. Yale University Press, New Haven.
271 pp.

United States Fish and Wildlife Service. 1996. Final Environmental Assessment
for Additions to Shiawassee National Wildlife Refuge. Government
Printing Office, Washington, DC. 103 pp.

Wright, GR. and PD. Tanimoto. 1998. Using GIS to prioritize land conservation
actions: integrating factors of habitat diversity, land ownership, and

development risk. Natural Areas Journal 18:38-4.

38

Chapter 3

A Socio-ecological Simulation Model of Land Acquisition to

Expand a National Wildlife Refuge

Abstract

Land acquisition is a common practice for expanding nature reserves such
as wildlife refuges. However, the feasibility of executing an acquisition project is
rarely assessed before a lot of time and energy has been spent in the project.
My goal was to create a simulation model that would allow land managers to
assess the feasibility of an acquisition project before the project is executed.
Shiawassee National Wildlife Refuge in Lower Michigan was used as a case
study. The model used information on the landowners’ attitudes to project the
land purchases and reported the types of land that were purchased (is.
agricultural or forested). It also used a soil map to determine how much of hydric
and non-hydric soil areas were purchased in the simulation. I found that the
information on the landowners’ attitudes is important because it can affect the
types of land purchased as well as the amount of land that is purchased in the
simulation. Using hypothetical incentive levels, I found that increasing the
incentive to sell increased the amount of land that was purchased, but the non-
hydric areas showed more increase than the hydric areas did. I also found that
increasing the number of landowners contacted with purchase offers shortens

the time required to complete the acquisition. However, the relationship is not

39

linear so there would be a limitation where the cost of increasing the number of
contacts made would outweigh the shortening of the project duration. If the
survey information is assumed to be true across time, the expansion of
Shiawassee National Wildlife Refuge will take an average of 249.7 years to
complete. I found that less than half of the acquisition area would be
purchasable without using incentives of some sort. To facilitate the acquisition
project, the United States Fish and Wildlife Service should increase the number
of purchase offers made per year and offer non—monetary incentives to the

landowners for selling their land.

Introduction

Purchasing land for biodiversity conservation and management is a
common practice (Ramsey and Addison 1996, Wright and Tanimoto 1998).
Land acquisition is a technique used both by private and public conservation
organizations to create reserves, such as wildlife refuges, which attempt to
counter habitat degradation and destruction (Reinecke et al. 1989, Terbough
and Schaik 1997, Liu et al. 1999). The Nature Conservancy, United States Fish
and Wildlife Service (USFWS), and state organizations such as the Michigan
Department of Natural Resources have all used land acquisition to establish
protected areas for wildlife (Press et al. 1996, United States Fish and Wildlife
Service 1996, Michigan Department of Natural Resources 1996). Once the
land is under the control of a management agency, the natural habitats can be

restored and protected from future development (Press et al. 1996). It seems

40

that, after a wildlife refuge is established, it is often desirable to acquire more
land to expand the area of protection. Larger reserves, generally, buffer the core
area from human disturbances more than smaller reserves (DellaSalla et al.
1996) and, therefore, contain better wildlife habitat, especially for species that
are very specialized in their habitat requirements (Saunders et al. 1991).
Unfortunately, the process of acquiring land has uncertain results. For
organizations with limited financial resources and a desire to funnel their
energies towards extracting maximum benefits, it is necessary to assess the
feasibility of an acquisition project fulfilling its objectives before money is invested
into purchasing the land parcels.

A preliminary feasibility assessment of a land acquisition project can be
executed through examining the attitudes of the landowners within the acquisition
area (Chapter 2). A survey of the landowners can reveal the percentage of
landowners that may sell and which landowners will definitely not sell their land.
However, rarely will all of the landowners respond to the survey, so it is
impossible to determine what the exact configuration of the final purchases will
be. To assess the feasibility of an acquisition project using more information
than just the number of parcels purchased (e.g., connectivity and restorable
habitats), a spatial simulation model needs to be constructed.

The goal of this paper was to present a simulation model that can be used
for evaluating the feasibility of a land acquisition project beyond the attitude
assessment of the landowners. To demonstrate the method, I have chosen to

perform a feasibility assessment on a land acquisition project that is currently

41

 

 

 

0.2;. a

underway. My specific objectives were to examine the distribution of resulting
purchases and to project the types of habitats that would be purchased and could
be restored. In addition, I was interested in examining potential methods of
facilitating the acquisition process (e.g., incentives for selling and mass

contacting).

Methods
Study Area

Shiawassee National Wildlife Refuge (hereafter refuge) in Saginaw
County, Michigan is attempting an expansion. The refuge was initially authorized
in 1953 as a migratory bird sanctuary and flood containment unit. It consists
mostly of managed wetlands located at the juncture of four rivers: the
Shiawassee, the Titabawassee, the Cass, and the Flint (Figure 1-1). Presently,
the goals of the refuge include maintaining the biodiversity of the area, providing
a migratory staging area for waterfowl and flood containment (United States Fish
and Wildlife Service 1996).

The refuge consists of 3680 ha, and the USFWS is interested in
purchasing land for an expansion of the refuge (United States Fish and Wildlife
Service 1996). The area of interest for the acquisition has been bounded and
consists of 3035 ha extending outwards on three of the four major rivers for the
area: the Shiawassee, the Titabawassee, and the Cass Rivers (Figure 1-2). The
majority of the land parcels in question encompass wetlands or are, at least,

within the flood plain of the rivers since these lands are the most useful for the

42

objectives of the refuge. Each parcel of land has been classified as high priority,
medium priority, or low priority based on its evaluated habitat potential for both
upland and wetland species (United States Fish and Wildlife Service 1996). The
landowners are being selectively contacted for the land acquisition project and
decisions about purchases are being made based upon the available monies for

the project.

Simulation Model

A script was created to simulate land purchases within the proposed
acquisition boundary using the programming language Avenue (Arcview 3.1,
Environmental Systems Research, Inc., 1996). The script combined spatial data
contained in geographic information systems (GIS) map layers with landowners’
attitudes towards selling to determine which parcels would be acquired. Since
expansions such as the one proposed for the refuge are usually initiated with
objectives related to habitat management, more information was needed than
just which parcels would be purchased. Four map layers were used: the
individual land parcels within the acquisition area, the boundary of the refuge, a
land-use layer, and a hydric soil layer. The first two map layers were acquired
from the USFWS. The hydric soil layer was derived from a digital soil map of
Saginaw County (Soil Survey Geographic Database) acquired from the United
States Department of Agriculture (USDA). Hydric soil areas represent places
that were historically “wet”, meaning that the refuge could restore some form of

wetlands in these places if the current habitat was not already wetlands. In

43

technical terms, a hydric soil is one that “formed under conditions of saturation,
flooding, or ponding long enough during the growing season to develop
anaerobic conditions in the upper part” (Soil Survey Staff 1999). This information
is important for land managers because they need to know what soils they have
to work with in restoration. The hydric/non-hydric layer identifies potential
wetland vs. upland restoration sites.

The land-use map was digitized from 1997 aerial slides acquired from the
USDA Farm Service Agency in Saginaw County and corrected against a digital
land-use map of 1978 Saginaw County created at the Center for Remote
Sensing, Michigan State University. Since the area had not changed overly
much, the 1978 land-use map was able to be used to determine classifications
that were difficult to ascertain from a visual inspection of aerial photos. The older
map also aided in the determination of wetland habitats. However, forested
wetlands were still difficult to identify so they were simply classified as forested
lands. I also drove through the area to compare the map layer with what was
actually present on-site to prevent errors such as classifying a housing
development as forested because of the neighboring woody vegetation. The
final land-use map had 5 categories of land use:

Developed/urban lands — patches of land that have been built upon by

humans (houses, factories, parking lots, etc.)

Agricultural lands — patches of land that appear to have been farmed with

obvious disturbance to the natural vegetation and top soil (tractoring

and plowing).

44

Grass/ands — patches of land that contain non-wet, non-forested, and non-
farmed lands that have no evident tall woody vegetation.

Forested lands — patches of land that contained clumped trees with a
dense canopy.

Non-forested wetlands — patches of land that are seasonally or
permanently inundated, but have no tall woody vegetation. This

category contains both emergent marshes and deep-water systems.

This land-use map was combined with the hydric soil layer to create a new
map layer that depicted each of the land-use categories broken down into hydric
and non-hydric subunits. Two of the categories did not have the split subunits.
These categories were the developed lands and the non-forested wetlands.
Wetlands have hydric soils and, therefore, lack a non-hydric counterpart. The
developed lands were consistently non-hydric. The combined layer allowed me to
determine how much of each land-use type would be purchased, how much of
each land-use type is potentially (if not currently) wet, and how much land-use
change could be initiated through the acquisition (e.g., agricultural lands changed

to wetlands).

Landowner Attitudes
The simulation requires that the land parcels have established
probabilities associated with whether or not they can be purchased. Without

information about the landowners’ attitudes toward selling, one could hypothesize

45

on the number of parcels that would be purchased, but all of the parcels would
have the same probability of being purchased resulting in a randomly scattered
accumulation of acquired land parcels. For a more realistic representation, a
predictive criterion specific to the individual parcels would be needed to weigh
the purchases so that the result would not be purely random. To address this
issue, I conducted a survey of the landowners within the acquisition boundary. In
the survey, I asked whether or not they would be willing to sell their land parcel
and when over the next 20 years they would be willing to sell (Chapter 2). What
the survey found was that a higher percentage (56.5%) of responding
landowners with parcels that included some wetlands were willing to sell their
land parcel than the landowners with non-wetland parcels (33.8%). Overall,
49.5% of the respondents replied that they would be willing to sell their land.
There was no way to conduct a non-respondent follow-up, so I did not
ascertain whether the non-respondent attitudes would match the respondent
attitudes. As there was no other information available, I was forced to project a
probability of sale to all the land parcels by assuming that the respondents were
representative of the non-respondents in terms of their willingness to sell.
Because this is not an ideal assumption for simulation modeling, | explored the
sensitivity of the non-respondent probabilities of sale to address the lack of a
non-respondent follow-up. I set the respondents’ probabilities to 1 or 0 in
accordance with their survey response (1 representing a sale and 0 representing
no sale) and varied the non-respondent probability of sale. I used three sets of

values. The first was a uniform non-respondent probability of sale of 0.500

46

where there was no difference between wetland-containing parcels and non-
wetland parcels. The second was a non-respondent probability that equaled the
respondent probabilities where wetland owners were more willing to sell. Finally,
the third value was a probability where the non-respondents’ attitudes were
opposite to those of the respondents (wetland-containing parcel and non-wetland
parcel probabilities are switched, i.e. non-wetland parcels = 0.565 and wetland
parcels = 0.338). An analysis of variance was used to determine significant
differences between the results. Tukey’s test was used to determine where the
differences lay.

The importance of the survey information was examined in two ways. The
first way was a comparison between hypothetical probabilities of sale and the
survey derived probabilities of sale. Specifically, I compared the total amount of
land purchased, the number of parcels purchased, the distribution of land types
purchased, and the amount of hydric and non-hydric soil areas purchased. I
used two hypothetical values, 0.500 and 0.333. The survey-derived probabilities
were 0.565 for wetland-containing parcels and 0.338 for non-wetland parcels
which yielded a total of 38.9% of the parcels being purchased.

The second examination of importance was to compare between two
simulation scenarios where the total number of parcels purchased were held
constant for both scenarios but the probabilities of sale were different. The
purpose of these simulations was to examine the amount of land purchased and
the distribution of land types purchased within an equal number of parcels. The

first scenario used the survey derived probabilities which differed between

47

wetland and non-wetland parcels, while the second had a uniform probability for
all land parcels (no difference between wetland and non-wetland containing
parcels). A simple un-paired T-test was used to determine significant differences
between the number of parcels purchased, the total area purchased, the existing

wetland area purchased, and the hydric-soil areas purchased.

Acquisition Scenarios

Three acquisition scenarios were simulated. The first scenario was to
introduce hypothetical incentive levels. To do this, I increased the survey-derived
probabilities of sale to include a portion of the landowners who responded
“undecided” to the question of whether or not they would sell. Since these
landowners have not definitely said no to selling, they could potentially be
persuaded to sell given sufficient encouragement. Two incentive levels (low and
high) were used. The low incentive level assumed that half of the undecided
landowners would sell. For example, 26.1% of the landowners with wetland-
containing parcels were undecided. Half of this number (0.130) added to the
original probability of 0.565 for the landowners that would sell yielded 0.695
(0.565 + 0.130) for the new probability of sale. Similarly, the new non-wetland
parcel probability of sale was 0.438 (0.338 + 0.100). The high incentive level
assumed that all of the undecided landowners would sell giving a probability of
sale of 0.826 for wetland-containing parcels and 0.538 for non-wetland parcels.
Of course, I expected that the total amount of land purchased would increase,

but I was interested in the amount of each land type purchased under the

48

 

incentive levels. Existing wetlands and hydric areas were of great interest since
the refuge is maintained primarily as a migratory staging area for waterfowl.

The next two scenarios operated at a no-incentive level and incorporated
a temporal aspect into the simulations. To do this, I had to first derive an annual
probability of sale. Since the original survey question was whether the
landowner would be willing to sell within the next 20 years, the survey-derived
probability of sale represents the probability that each parcel would be purchased
within 20 years. Therefore, I divided the original probabilities down to an annual
probability that would compound back to the original after 20 years. To do this, I

used the following formula,

P.=1—<1—p>"

where P” represents the probability of an event occurring within n years and 19
represents the probability that an event will occur in a single year. Using this
formula, I determined that the annual probabilities of sale were 0.041 for wetland—
containing parcels and 0.020 for the non-wetland parcels. Using these
probabilities, I could then determine the length of time required to complete the
acquisition.

The first temporal scenario mimicked the acquisition project conducted by
the USFWS where landowners are selectively and individually contacted with
purchase offers. I assumed that only a specified number of parcels would be
purchasable in the end. The number of parcels purchased was determined by
the percentage of landowners willing to sell found in the survey. Also, I assumed

that only a limited number of purchases would be attempted each year since

49

personnel limitations would prevent mass contacting of every landowner within
the acquisition area. Since it had been suggested that facilitating the contacting
of landowners with purchase offers would accelerate the acquisition process
(Ramsey and Addison 1996), I opted to test this theory with my model. I ran the
simulations with 5, 10, 15, 20, and 25 as the limits on the number of attempted
purchases and compared the length of time that was required to complete the
acquisition (only willing sellers have sold). It should be kept in mind that the
USFWS selects the landowners to contact based upon an impression that the
landowner would be willing to sell while the simulated contacts were randomly
selected. This random selection would lengthen the time required to purchase
the parcels from the willing sellers.

Whereas the first scenario assumed that only a number of landowners
were contacted each year and only the land owned by willing sellers were
purchased, the second scenario assumed that all of the parcels were contacted
each year and that all of the parcels were purchasable given sufficient time. In
other words, if the USFWS were persistent and patient, eventually, they would be
able to buy all of the land parcels within the acquisition boundary. The simulation
was run until all land parcels had been purchased to examine the length of time
that would be required to completely purchase the acquisition area.

Each scenario was simulated 20 times because of the stochastic nature of
the model. The results from each simulation were averaged for presentation in

this paper.

50

 

Results

In all simulations where only the willing sellers sold, the acquired parcels
were scattered within the acquisition area creating a matrix of purchased and
non-purchased parcels at the end of the simulation (Figure 3-1). This is
consistent with the results from the survey (Chapter 2) which suggested that at
least 612 ha would not be purchasable and these parcels would be interspersed
throughout the acquisition area. The intermixing of purchased and non-
purchased parcels would limit the total wetland area that could be restored
(hydric but not currently wet areas) because of, among other things, the high cost
and the inefficiency of managing small units (Harold Prince, pers. comm.)

Obviously, if none of the landowners sell, the refuge would remain the
same as it was before the attempted expansion. The refuge would consist of
3680 ha with approximately 850 ha of wet areas. If all of the landowners sell, the
refuge would have added the entire 3035 ha within the acquisition area, nearly
doubling its pre-expansion size. This 3035 ha includes 284 ha of developed
lands, 1287 ha of non-hydric agricultural lands, 546 ha hydric agricultural lands,
144 ha non-hydric grasslands, 36 ha hydric grasslands, 590 ha non-hydric
forested lands, 65 ha hydric forested areas, and 83 ha of non-forested wetlands.
This gives a maximum of 582 ha for restorable wet areas if one assumes that the

hydric forested areas are already wet.

51

.co:m_:E_w 822338 652 m E0: 9::st 28.8 88:23 u6 5:33:56 8:QO .6 anmxm Tm 2:9;

   
 

mam: 2 28-52 B
aim: 2 20m
. $2 coEmSB< 888$

  

magma

 

52

Sensitivity of Non-Respondent Probabilities of Sale

In examining the sensitivity of the non-respondent probability of sale, I
compared the total amount of acreage purchased, the amount of existing non-
forested wetlands purchased, and the amount of hydric areas purchased (Table
3-1). The results from the reversed non-respondent probabilities were different
from the other two (p<0.05). On the average, the total purchased hectares using
the reversed non-respondent probabilities were over 100 ha lower than either of
the other two even though the average number of parcels purchased fell between
the values for the other two scenarios. The total acreage purchased and the
existing wetland hectares purchased with the reversed non-respondent
probability were significantly lower (p<0.05) than in those other two scenarios.
The hydric-soil hectares purchased under the reversed probabilities were

significantly lower than under the uniform non-respondent probability but was not

Table 3-1. Examination of the sensitivity of the non-respondent probability of
sale. The numbers represent the average values (is.d.).

 

Number of Total Hectares of Hectares of
.. parcels hectares existing wetlands hydric-soil areas
PrObab'I'tY 0f sale purchased purchased purchased purchased
Non-respondents 111.1i 6.0 1058 i 265 34 i 7 248 i 30
equal to
respondents
Non-respondents 115.6 i 6.4 914 i 264 24 i 5 235 i 26
opposite from
respondents
Non-respondents 120.6 i 8.0 1087 i 258 31 i 6 285 i 35
uniform

 

53

significantly lower (p>0.10) than under the survey-derived probabilities. A
comparison between the survey-derived non-respondent probabilities and the
uniform non—respondent probability shows no significant differences (p>0.10) in
the amount of land purchased: total, wet, or hydric (Table 3-1). However, the
number of parcels purchased with the survey-derived probabilities was
significantly less (p<0.01) than the number purchased with the uniform

probability.

Hypothetical vs. Survey-Derived Probabilities of Sale

The total amount of land purchased using each of the probabilities was
consistent with the probabilities. Since the survey-derived probabilities of sale
fell between the hypothetical percentages, the total area purchased also fell
between those under the hypothetical simulations, with the low hypothetical
probability of sale yielding 1107 ha, the survey-derived probabilities yielding 1400
ha, and the high hypothetical probability yielding 1595 ha. However, examining
the land types purchased shows that the survey-derived probabilities result in a
different distribution of land types (Figure 3-2). In the non-hydric agricultural,
grass, and forested lands, I found a steadily increasing trend from the low
hypothetical probability to the survey-derived probabilities to the high hypothetical
probability where the survey-derived probabilities resulted in a value halfway
between the values for the hypothetical probabilities. In the developed and

hydric grassland categories, the survey-derived probabilities resulted in more

54

Sufism vmgozvcoo 290385 m E0: 8256 28 U6 mmEznmnoE mm :03 mm .28 36 $3205 $265093
Somdv :9: m vcm 8.8.8 32 m 9.6: umSomxo 295 823525 of. .uowmzoSQ on? 6E2 comm. ho EsoE< .N-m 93mm

8:353 3888 @886.“ wee—mmfiw waflmmfiw 35:38:? #83328 “332328
$828.52 ”Siam 26.3-52 USES 2:83-52 out»: 0533-82 0383-82

 

 

 

com

com

 

oov

 

3:53on 6388.093 swam com

 

paseuomd saJeioaH

5:53on 338284853 fl coo

 

.3253on Region»: 26A I 8“

 

 

 

com

55

Sufism voyosucoo >_m:o_>oa m E0: 8256 28 v6 mm:___nmnoa mm :95 mm 98 So E5305 3262693 88.8 :9;
m 98 88.8 32 m 9%: umSooxm 295 wcozmSEfi 9: .ummmcesq 89m :3 263-5: .82? 26>: .m-m 230E

25382

 

 

Seam

 

 

,
ll IL

. v
8:588 88582 85 m L
L

 

525303 Um>tofi$o>§m

 

bsEmnoa 3365093 30; I

 

 

 

com

com

Door

003

003.

paseqomd saiaioeH

56

land being purchased than the hypothetical probabilities. In the hydric forested
lands and the non-forested wetlands, the survey-derived probabilities gave the
same results as the high hypothetical probability. Grouping the categories into
hydric and non-hydric areas (Figure 3-3), I found that, overall, the non-hydric
areas steadily increased between the hypothetical and survey-derived
probabilities. However, the hydric areas show that the survey-derived
probabilities gave results more similar to the high hypothetical probability than
the low hypothetical probability. This shows that the survey information is
necessary if one wants to project the area purchased because the probability of
sale influences the number of parcels sold and the distribution of the acreage
across land type classifications.

In the second simulation scenario where I held the number of parcels

purchased constant (n=124), I found that the total acreage purchased was

Table 3-2. Simulation results to test survey significance. The numbers represent
the average values (is.d.).

Hectares of Hectares of
Number of existing hydric-soil
_ . parcels Total hectares wetlands areas
Type Of PrObab"'tY purchased Purchased purchased purchased

 

Uniform probability 124 1167 i 227 32 i 5 326 i 64
Survey-derived 124 1400 i 281 44 i 10 341 i 45
probability

 

57

significantly higher (p<0.05) with the survey-derived probabilities which favor
wetland-containing parcels (Table 3-2). The amount of existing wetlands that
was purchased was also significantly higher with the survey-derived probabilities
than with the uniform probability. Differing from this trend, though, was the
amount of hydric—soil areas purchased. Between the two probabilities, the hydric

areas purchased did not differ (p>0.5).

Incentives

As expected, the total amount of land purchased increased with the
increasing incentive levels, going from 1326 ha to 1695 ha to 2045 ha. However,
to evaluate whether or not it would be worthwhile to pursue incentives, I had to
examine the different land types, both existing and potential, purchased using the
different incentive levels. For the existing land types (Figure 3-4), the amount of
agricultural land purchased increased the most. This was to be expected since
the proposed acquisition area falls within a highly agricultural community. All of
the land types, except for the non—forested wetlands, showed an increase in the
amount of land purchased with the increasing incentive levels. The non-forested
wetlands, surprisingly, did not increase with the low incentive level. Grouping the
categories into hydric and non-hydric areas (Figure 3—5) showed that the areas
purchased increased with increasing incentives. However, the increase of hydric

areas was nearly linear while the increase in the non—hydric areas was not linear.

58

.wozEon cmE cam. 693505 32 .wm>_Eooc_ 0: 6mm: 9m; m_o>o_ o>=cmo:_ oQE. .8352: “6 flooto
or: mczmmzmgg co:m_:E_w 68:23 95 9: E0: 9539 womb Ucfl mczflxo “6 35.9% vwmmcosn. Tm 9:9“.

$5363 somehow 338$ 952$me wsfimmfiw 35:50ch #83323 tumour/ow
3388-82 035wa 253.202 0?ch 2:33.52 26am 353.52 0533.52

 

 

paseqomd sampeH

   

03:52: Boa

 

@652: 02 I

 

 

 

 

 

59

.wo>=cwo:_ :9; 5cm 69;:on >>o_ .moZEooE 0: 6mm: 9oz, m_o>w_ o>_Emo:_ 82:. .mo>_Emoc_ u6 flootm
or: mczmmzmmSE wcozmSEE 3283 9:2 m5 @555 68283 wmmzw __ow otgcéo: wzwco> 25>: .m-m 9:9”.

 

 

 

 

 

oeafioz 063m
o
08
84
H
.08 m
m
9
S
1. com d
n
l
O
u.
88 w
9
p
l 95:85 Bod “ 8S
96:88 02. .. 83

 

 

 

00$

60

Table 3-3. Simulation results using limited purchase attempts per year.

Years to completion of acquisition project
Number of attempts

 

 

per year (parcels) Minimum Average (s.d.) Maximum
5 759 830 (62.1) 925
10 402 437 (22.3) 470
15 234 297 (36.2) 345
20 198 222 (21.7) 260
25 154 177 (12.9) 196

 

Effects of Attempted Purchases on Time Required for Completing the Acquisition
Changing the number of attempted purchases per year significantly
changed the length of time required to complete the acquisition to the point when
all the land owned by willing sellers had been purchased (Table 3-3). As the

number of attempts increased, the length of time required to complete the
acquisition decreased. A polynomial regression (Figure 3-6) yielded a

relationship of

Y = 0.01x4 — 0.749638 + 21.3x2 — 285.71x + 1813.4
where x is the number of attempted purchases per year and Y is the years to
completion of the acquisition (R2 = 0.98). This finding supports the claim that
facilitating contacts would accelerate the acquisition process (Ramsey and

Addison 1996). However, a decision would need to be made on where the

61

om mm

.m:m__om m:____>> >: 8:30 gm. 9: 96 __m 39.65: 2 8:53:
m8: m5 vcm 59» 5Q 8938:: bmaEmrm 6 53:5: 9: :omznm: Q_:m:o:m_om .©-m 9:9“.

.83 E: 38:05: 689233 *0 33:52
om m: or m o

 

 

 

 

I. i| .!.:-ll.i . 00?

com

 

-gi .i-oom

iiiai: 004

com

- com

com

-. iii- ., cow

ii--, com

 

 

 

000?

(SM) uomsgnboe an; ;o uonaldwoo m1 amp, ;0 tnfiua-l

62

burden of the added effort and manpower required would discourage the desire
for a shortened time frame.

Assuming that all parcels would be purchasable given sufficient time, I
found that the acquisition project could be ongoing for many years, even if the
USFWS manages to contact every landowner every year. The time frame
ranged from 202 years to more than 300 years, and the average was 249.7 years
(s.d. = 37.9). The average length of time required (249.7 years) is slightly lower
than it should be because, in 2 of the 20 simulations, not all of the parcels had
been purchased in 300 years. Since 300 years was the duration of the
simulation, l have no data about how much longer than 300 years it would take to
complete the acquisition. However, since this only happened in 2 of the
simulations and because my purpose was only to demonstrate that the time

frame was lengthy, I decided to forego any further analysis.

Discussion

The survey results were important for the land purchase simulations. The
survey-derived probabilities of sale produced different distributions of the types of
land purchased than did the hypothetical probabilities of sale. Also, the non-
respondent probabilities were shown to be influential in determining the number
of parcels purchased, the total area purchased, and the distribution of existing
and potential land types purchased. Therefore, the lack of information on the
non-respondents could have significantly affected the results. A consideration for

other studies on the feasibility of acquisition projects would be to ensure that a

63

non-respondent follow—up can be accomplished. A non-respondent follow-up
would provide more information for the simulation and ensure an improvement in
accuracy of the simulation results.

The simulation results using the survey information show that less than
half of the acquisition area would be purchasable without using incentives of
some sort. Furthermore, the purchased parcels were interspersed with the not-
purchased parcels, affecting the continuity of the newly expanded refuge (Figure
3-1). Small isolated units may be too costly and difficult to manage for the
USFWS, so some of these small isolated units would not be managed until more
land could be purchased around them (Harold Prince, pers. comm). In addition,
the acquisition project could take as many as 250 years to complete.

If the USFWS should decide to proceed with the acquisitions, the results
also show that the USFWS should offer incentives to selling and increase the
number of purchase offers made to landowners per year. Financial incentives
that offer extra payment directly to the landowner would be impossible because
the USFWS’s policy is to buy land at the appraised value. Furthermore, offering
more money to one landowner would necessitate an equal remuneration for
other landowners to prevent disputes. Kershow (1975) states that incentives for
selling should come in the way of services and not money. For instance, the
relocation assistance offered by the USFWS would qualify as an incentive.
Whatever incentives the USFWS chooses to use, the incentives would have to

be tailored to each landowner for the greatest effectiveness.

64

Acknowledgements
Funding for this project was provided by the Michigan Agricultural

Experimentation Station and the Wildlife Division of the Michigan Department of
Natural Resources. Much appreciation goes to Doug Longpre for assisting in the
creation of the land—use coverages. David J. Peters was invaluable in providing
information about the status of the acquisition project. I am grateful to John J.
Connors for informing me of USFWS land acquisition procedures as applied to
the Shiawassee National Wildlife Refuge additions. Finally, I would like to thank
Drs. Scott \Mnterstein, Henry R. Campa, and R. Ben Peyton for their assistance

in acquiring funding for this project.

Literature Cited

DellaSala, D.A., J.R. Strittholt, R.F. Noss, and D.M. Olson. 1996. A critical role
for core reserves in managing Inland Northwest landscapes for natural
resources and biodiversity. Wild. Soc. Bull. 24: 209—221.

Kershow, W.W. 1975. Land Acquisition. National Recreation and Park
Association, Arlington, VA. 26 pp.

Liu, J., Z. Ouyang, W. Taylor, R. Groop, Y. Tan, and H. Zhang. 1999. A
framework for evaluating effects of human factors on wildlife habitat: The
case of the giant pandas. Conservation Biology 13: 1360-1370.

Michigan Department of Natural Resources. 1996. Shiawassee River State

Game Area Master Plan. Lansing, MI. 32 pp.

65

Press, D., D.F. Doak, and P. Steinberg. 1996. The role of local government in
the conservation of rare species. Conservation Biology 10:1538-1548.

Prince, Harold H. 2000. pers. comm. Professor, Department of Fisheries and
Wildlife, Michigan State University. East Lansing, MI.

Ramsey, OJ and D.S. Addison. 1996. Facilitating a multiparcel land acquisition
project in the Western Big Cypress Region of Collier County, Florida,
USA. Natural Areas Journal 16:36-40.

Reid, F.A., J.R. Kelley, Jr., T.S. Taylor, and L.H. Frederickson. 1989. Upper
Mississippi Valley wetlands — Refuges and moist-soil impoundments. pp
181-202 in Habitat Management for Migrating and Wintering Waterfowl in
North America. Texas Tech University Press, Lubbock. 560 pp.

Reinecke, K.S., R.M. Kaminski, D.J. Moorhead, J.D. Hodges, and JR. Nassar.
1989. Mississippi Alluvial Valley. pp 203-247 i_n_ Habitat Management for
Migrating and VVIntering Waterfowl in North America. Texas Tech
University Press, Lubbock. 560 pp.

Saunders, D.A., R.J. Hobbs, and CR. Margules. 1991. Biological
consequences of ecosystem fragmentation: a review. Conservation
Biology 5:18-32.

Soil Survey Staff. 1999. National Soil Survey Handbook, title 430—V1. United
States Department of Agriculture, Natural Resources Conservation
Service (Washington, DC, US. Government Printing Office, revised

September 1999).

66

Terbough, J. and GP. van Schaik. 1997. Minimizing species loss: the
imperative of protection. pp15 — 35 in Kramer, R., C. van Schaik, and J.
Johnson (ed.). Last Stand: Protected Areas & the Defense of Tropical
Biodiversity. Oxford University Press, New York. 242 pp.

United States Fish and Wildlife Service. 1996. Final Environmental Assessment
for Additions to Shiawassee National Wildlife Refuge. US. Government
Printing Office, Washington, DC. 103 pp.

Wright, GR. and PD. Tanimoto. 1998. Using GIS to prioritize land conservation
actions: integrating factors of habitat diversity, land ownership, and

development risk. Natural Areas Journal 18238-44.

67

Chapter 4

Conclusions and Recommendations

The landowner survey suggested that less than half of the landowners
within the acquisition area of Shiawassee National Wildlife Refuge would be
willing to sell their land without incentives, and landowners who owned portions
of wetlands would be more willing to sell than landowners who did not. Also, the
survey indicated that at least 612 ha of the land within the acquisition boundary
would not be purchasable for various reasons. The simulation model showed
that less than half of the land area would be purchased without some incentives,
and that the acquisition project would take over 200 years to fully complete. In
addition, over two—thirds of the area purchased would be non-hydric and only
manageable as upland habitat. While this may be good because it increases
habitat connectivity, upland habitat is not optimal for waterfowl management.
Furthermore, both the survey and the simulation model suggested that there
would be non-purchased parcels interrupting the overall continuity of the
expanded refuge. While continuity may not overly affect waterfowl that can fly to
other locations and are not dependent upon land corridors, it can be a
requirement for affordable management if the area is actively managed (meaning
that the units are constantly maintained at defined conditions). Small units are
extremely expensive to restore and manage (Harold Prince, pers. comm);
therefore, they probably would not be managed until other land around them

could be purchased.

68

The proposed acquisition for the expansion of Shiawassee National
Wildlife Refuge (refuge) will not fulfill the objectives of the United States Fish and
Wildlife Service (USFWS; objectives can be found in United States Fish and
Wildlife Service 1996). However, should the USFWS wish to continue with the
acquisition project and facilitate it, they can increase the amount of land
purchased by offering incentives for selling.

Also, the comments from landowners written on the survey questionnaires
indicated that some of the landowners who would not sell were not against
selling. They were just against selling their land to the USFWS because they did
not agree with the practices of the USFWS or felt that the USFWS abused its
power during the original establishment of the refuge. Communicating with these
landowners about the purpose of the expansion, the issues surrounding the
conservation movement, and the current methods employed by the USFWS
could sway these landowners into selling their land.

One way to overcome the bias against the USFWS is to acquire the land
through another agency. Currently, there exists a cooperation between the
Michigan Department of Natural Resources (MDNR) and the USFWS to acquire
and exchange land (James Dastyk, pers. comm. 1999; Brian Vogl, pers. comm.
1999). The MDNR operates a neighboring state game area and is interested in
acquiring a piece of the refuge that is excellent for goose hunting. The
agreement is that the MDNR will acquire an equal amount of land from the
acquisition area and then trade it to the USFWS. Using the MDNR to purchase

the land overcomes the barrier of landowners disliking the USFWS. In addition,

69

the MDNR can more conveniently pursue sales. The MDNR is housed within the
state with the main headquarters located approximately 90 miles away from the
refuge. This is much closer than the USFWS realty offices in Minnesota, which
are handling the refuge’s expansion. Further partnerships where another
organization is used to initiate the sales could expedite the acquisitions.

Another recommendation would be to alert the landowners within the
acquisition area that they are within the acquisition boundary. Since I received
comments back on the survey that indicated several landowners had been
unaware of the acquisition plan and several of these landowners solicited me for
USFWS contact information to discuss sales, I would assume that the number of
sales per year would be increased if all of the landowners knew that the refuge
was attempting to expand. This would at least provide more names to negotiate
with than just the people with whom the refuge manager has contact.

Finally, the sales should be approached using a spatial priority system to
address the issue of connectivity. One example of a spatial priority system would
be purchasing parcels immediately connected to the refuge and then moving
outwards. A purchase priority system of this type would ensure that any
purchases could be immediately incorporated into the management of the refuge
and would eliminate the small isolated units that would present difficulties for

managers.

70

Literature Cited

Dastyck, Jim. 1999. pers.comm. Refuge Biologist. Shiawassee National
Wildlife Refuge, United States Fish and Wildlife Service.

Prince, Harold H. 2000. pers. comm. Professor, Department of Fisheries and
Wildlife, Michigan State University. East Lansing, MI.

United States Fish and Wildlife Service. 1996. Final Environmental Assessment
for Additions to Shiawassee National Wildlife Refuge. US. Government
Printing Office, Washington, DC. 103 pp.

Vogl, Brian. 1999. pers.comm. Biological Technician. Shiawassee River State

Game Area, Michigan Department of Natural Resources.

71

 

 

 

5
. :5

U11: :

 

 

I .

III I
99262

LIBRARIE
I r‘

I

I”
2’07

I

III
30

I
I
I
I

II

I

I

I

III
II
3129

HIan STATE
I

My... 7 t

a. 3.11.62}.

 

 

 

 

 

 

r. 13.33..