U.S. Fish and Wildlife Service: Additional Flexibility Needed to 
Deal with Farmlands Received from the Department of Agriculture  
(18-SEP-07, GAO-07-1092).					 
                                                                 
Over the past two decades, provisions of the Food Security Act of
1985, among others, have allowed the Department of Agriculture's 
Farm Service Agency in partnership with the Department of the	 
Interior's U.S. Fish and Wildlife Service (Service) to add	 
farmlands found to have important resources to the National	 
Wildlife Refuge System. The Farm Service Agency transferred such 
farmlands to the Service through outright ownership ("fee	 
simple") or through conservation easements. Individual farmlands 
are managed by the nearest refuge office. GAO was asked to	 
examine (1) the extent of farmland received by the Service, (2)  
the extent to which the Service is currently managing its	 
farmlands, and (3) alternatives for managing these lands. To	 
answer these objectives, GAO visited five refuges and surveyed	 
managers responsible for a random sample of 98 farmlands.	 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-07-1092					        
    ACCNO:   A76389						        
  TITLE:     U.S. Fish and Wildlife Service: Additional Flexibility   
Needed to Deal with Farmlands Received from the Department of	 
Agriculture							 
     DATE:   09/18/2007 
  SUBJECT:   Conservation					 
	     Easements						 
	     Federal agencies					 
	     Federal property management			 
	     Land management					 
	     Land use						 
	     Land use law					 
	     Program management 				 
	     Public lands					 
	     Wildlife						 
	     Wildlife conservation				 
	     Interagency relations				 
	     Records management 				 
	     Farmland						 
	     National Wildlife Refuge System			 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-07-1092

   

     * [1]Results in Brief
     * [2]Background
     * [3]The Service Has Received at Least 1,400 Farmlands Covering 1

          * [4]Since 1986, the Service Has Received Mostly Small Farmlands
          * [5]The Service's Farmland Records Are Incomplete

     * [6]The Service Is Generally Not Managing the Majority of Its Fa

          * [7]The Service Performs Limited to No Management Activities on
          * [8]Several Factors Affect the Extent of the Service's Managemen
          * [9]Farmlands That Align Closely with Refuges' Goals Receive Mor

     * [10]The Service Has Limited Alternatives for Managing Its Farmla
     * [11]Conclusions
     * [12]Recommendations for Executive Action
     * [13]Agency Comments
     * [14]Appendix I: Scope and Methodology
     * [15]Appendix II: Survey of the U.S. Fish and Wildlife Service's
     * [16]Appendix III: Comments from the Department of the Interior
     * [17]Appendix IV: GAO Contact and Staff Acknowledgments

          * [18]GAO Contact
          * [19]Staff Acknowledgments

               * [20]Order by Mail or Phone

Report to the Subcommittee on Interior, Environment, and Related Agencies,
Committee on Appropriations, House of Representatives

United States Government Accountability Office

GAO

September 2007

U.S. FISH AND WILDLIFE SERVICE

Additional Flexibility Needed to Deal with Farmlands Received from the
Department of Agriculture

GAO-07-1092

Contents

Letter 1

Results in Brief 6
Background 9
The Service Has Received at Least 1,400 Farmlands Covering 132,000 Acres,
but the Actual Number Is Unknown 15
The Service Is Generally Not Managing the Majority of Its Farmlands 20
The Service Has Limited Alternatives for Managing Its Farmlands 30
Conclusions 35
Recommendations for Executive Action 36
Agency Comments 37
Appendix I Scope and Methodology 38
Appendix II Survey of the U.S. Fish and Wildlife Service's Farmland
Management 40
Appendix III Comments from the Department of the Interior 49
Appendix IV GAO Contact and Staff Acknowledgments 51

Tables

Table 1: Executive Orders and Legislation Relevant to the Farm Service
Agency's Inventory Program 13
Table 2: Size of Farmlands Received by the Service 16
Table 3: Estimated Travel Times between the Managing Refuge Offices and
Their Farmlands 27
Table 4: Selected Survey Questions 42
Table 5: Numbers and Percentages of Properties Inspected and Not Inspected
in the Past 2 Years, by Size, Distance, and Number of Properties, and Odds
and Odds Ratios Derived from Them 44
Table 6: Confidence Intervals Associated with the Estimated Percentage of
Properties Inspected, by Size, Distance, and Number of Properties 46

Figures

Figure 1: U.S. Fish and Wildlife Service Regions 11
Figure 2: Example of an Easement Placed on Farm Property 15
Figure 3: Concentration of Service Farmlands by State 17
Figure 4: Number of Farmlands Received by the Service, by Calendar Year 19
Figure 5: Estimated Frequency of Inspections from Calendar Year 2002
through 2006 22
Figure 6: Farming Activities in Violation of a Wetland Easement, April
2007 23
Figure 7: Location of the Farmlands That Michigan's Shiawassee National
Wildlife Refuge Is Responsible for Managing 26
Figure 8: Old Farmhouse in New York, Renovated to Serve as a Refuge
Office, April 2007 30

This is a work of the U.S. government and is not subject to copyright
protection in the United States. The published product may be reproduced
and distributed in its entirety without further permission from GAO.
However, because this work may contain copyrighted images or other
material, permission from the copyright holder may be necessary if you
wish to reproduce this material separately.

United States Government Accountability Office
Washington, DC 20548

September 18, 2007

The Honorable Norman D. Dicks
Chairman
The Honorable Todd Tiahrt
Ranking Member
Subcommittee on Interior, Environment, and Related Agencies
Committee on Appropriations
House of Representatives

Since the early 1900s, public lands have been set aside to protect natural
ecosystems, provide habitat for wildlife, and offer Americans recreational
opportunities. In the mid-1960s, Congress consolidated many of these lands
into the National Wildlife Refuge System, a vast, ecologically diverse and
valuable network of lands administered by the Department of the Interior's
U.S. Fish and Wildlife Service (Service). Today, this system comprises 548
wildlife refuges covering 96 million acres, of which more than 17 million
acres are located in the continental United States.^1 The mission of the
refuge system is to "administer a national network of lands and waters for
the conservation, management, and where appropriate, restoration of the
fish, wildlife, and plant resources and their habitats within the United
States for the benefit of present and future generations of Americans."^2
The system includes land that has always been federally owned, as well as
land that has been acquired from other parties, including state
governments and private entities.

Through the Food Security Act of 1985 (Farm Bill) and subsequent
legislation, some land was added to the refuge system through a
partnership with the Department of Agriculture's Farm Service Agency.^3
This agency was established in part to provide loans to farmers who were
otherwise unable to secure enough funding to operate their farms. If a
farmer was unable to repay such a loan, however, the Farm Service Agency
sometimes foreclosed on the loan and acquired the farm property.
Properties thus acquired went into the Farm Service Agency's inventory,
with the goal that they would eventually be sold back into private
ownership. Because of a distressed farm economy at the time, the Farm
Service Agency had over a million acres of inventory properties when the
1985 Farm Bill became law. Provisions of the Farm Bill and subsequent
legislation authorized the Farm Service Agency to acquire land or
interests in land for conservation purposes. Under these authorities, the
Farm Service Agency placed conservation protections on its inventory
properties that contained important resources, which it defined as
wetlands, riparian zones,^4 floodplains, coastal barriers, and other areas
of high ecological quality or important fish and wildlife habitat. In
carrying out these authorities, the Farm Service Agency generally
consulted with the Service to determine, among other things, which
properties should be protected. A memorandum of understanding signed in
1987 implements this and other environmental management responsibilities
of the two agencies. Given the significant number of inventory properties
in the mid-1980s, the Service reviewed thousands of these properties over
the following decade and if important resources were identified on any
portion of a property, the Service recommended that the Farm Service
Agency protect that portion.

^1Refuge lands in Alaska account for almost 77 million acres, and more
than 2 million acres are spread across American Samoa, Guam, Hawaii,
Puerto Rico, the U.S. Virgin Islands, and minor outlying islands of the
United States.

^2The National Wildlife Refuge System Improvement Act of 1997, Pub. L. No.
105-57, 111 Stat. 1254 (1997), codified at 16 U.S.C. S 668dd(a)(2).

The Service recommended placement of conservation protections on certain
lands, but the Farm Service Agency made the final decision, often
negotiating with the Service on the specific number and location of acres
on a property that would be protected. Such protections usually consisted
of placing a perpetual easement on the property before selling it back
into private ownership or, in certain cases, transferring outright
ownership in fee simple to another federal or state agency, although
preference was given to the Service. An easement is a nonpossessory
interest in land. With an easement, the original landowner retains
possession of the land, but, depending on which resources are being
protected, the easement places restrictions on the land's use. All of the
easements prohibit landowners from further developing the land subject to
the easement, and many do not allow any agricultural activities or
alteration of the habitat. Not all of the properties that the Service
recommended for protection were of equivalent ecological value, however.
In some cases, for example, the Service recommended protecting a site
whose soil or other characteristics indicated that it once was or could
again become a wetland, even if the site was no longer a wetland at the
time because of alteration of the habitat from farming practices, such as
draining or leveling the land. In addition, the 1990 and 1996 Farm Bills
restricted the agency's authority to establish easements under the
program.^5

^3Food Security Act of 1985, Pub. L. No. 99-198, title XIII, S 1318(a), 99
Stat. 1354, 1530-1 (1985); Agricultural Credit Act of 1987, Pub. L. No.
100-233, title VI, SS 612 and 616, 101 Stat. 1568, 1674 and 1682 (1988);
Food, Agriculture, Conservation, and Trade Act of 1990, Pub. L. No.
101-624, title XVIII, subtitle A, SS 1813(h) and 1815; and title XXIII, S
2388(j), 104 Stat. 3359, 3823-4, 3825-6, 4053 (1990); and Federal
Agriculture Improvement and Reform Act of 1996, Pub. L. No. 104-127, title
VI, subtitle D, SS 639, 642, and 646, 110 Stat. 888, 1097, 1102-4 (1996).

^4A riparian zone is the bank or corridor adjacent to rivers, streams, or
other water bodies. A vegetated riparian zone can act as a protective
buffer between land and water.

In addition to helping the Farm Service Agency assess inventory properties
for important resources, in most instances the Service agreed to serve as
the manager of the easement properties. Other federal or state agencies
and private nonprofit organizations could also assume this duty,^6 but the
Farm Service Agency gave priority to the Service. Moreover, Service policy
held that it would become the easement manager of farmlands^7 it had
recommended for protection, unless another federal or state easement
manager had been identified. It was not uncommon for a federal or state
agency, including the U.S. Forest Service, Bureau of Land Management, or a
state land management agency, to receive fee-simple lands or easement
management responsibilities, although the majority of the farmlands were
accepted by the Service. All farmlands received by the Service from the
Farm Service Agency--whether through an easement or fee-simple
ownership--were to be administered as part of the National Wildlife Refuge
System, with management responsibility for each farmland generally falling
to the nearest refuge office. Service guidance stipulates that refuge
offices are responsible for management activities, including regularly
inspecting farmlands to check for violations, determining whether
landowners are complying with all the easement restrictions, and taking
enforcement actions on violations. Because the lands, or the interests in
the lands, are government property, the Service may take civil or criminal
action against violators for the "destruction" of government property.^8
Refuge offices are also responsible for maintaining contact with
landowners, particularly new landowners; conducting maintenance, including
marking easement boundaries and sometimes installing or maintaining
fencing; and in some cases doing ecological restoration, such as restoring
wetlands.

^5Food, Agriculture, Conservation, and Trade Act of 1990, Pub. L. No.
101-624, title XVIII, subtitle A, S 1813(h)(2)-(3), 104 Stat. 3359, 3823
(1990); Federal Agriculture Improvement and Reform Act of 1996, Pub. L.
No. 104-127, title VI, subtitle D, S 639, 110 Stat. 888, 1097 (1996).

^6See the Food Security Act of 1985, Pub. L. No. 99-198, title XIII, S
1314(a)(2)(B), 99 Stat. 1354, 1526-7 (1985), for the authority regarding
private nonprofit organizations.

^7For the purposes of this report, we use the term "farmland" to mean the
property interests and ownership in properties the Service received from
the Farm Service Agency. We recognize that some of these properties are no
longer used for farming.

After enactment of the 1985 Farm Bill, the Service accepted management
responsibilities for farmlands that were sometimes situated miles from the
nearest refuge and others that, after years of being farmed, were
ecologically degraded. From the outset, some refuge offices had concerns
about receiving certain farmlands. Refuge officials were often not
involved in the initial assessment of the inventory properties or in the
decision to accept management responsibility as part of the refuge system.
Particularly in cases where the farmlands were widely scattered or a long
way from the nearest refuge office, refuge officials were concerned
because of the potential costs of managing such far-flung parcels. Service
policy, however, held that the relative size of a farmland or its
proximity to an existing refuge should not be an overriding consideration
in recommending a property for an easement or fee-simple acquisition;
rather, according to a memorandum from the Director of the Service in
1988, the Service should make every effort to protect all lands identified
as having important resources, regardless of acreage. In addition, some
procedural and coordination problems between the two agencies were never
fully resolved. For example, the Service was not always notified of the
Farm Service Agency's final decision on its recommendations for
protection, and in some cases it was several years after a recommendation
was made that the property was removed from inventory and an easement
placed on the land. Also, in many cases the Farm Service Agency did not
legally survey the land to outline precise boundaries.

^816 U.S.C. S 668dd(c) states that "no person shall disturb, injure, cut,
burn, remove, destroy, or possess any real or personal property of the
United States, including natural growth, in any area of the System . . .
unless such activities are performed by persons authorized to manage such
area, or unless such activities are permitted . . . by express provision
of the law, proclamation, Executive order, or public land order
establishing the area, or amendment thereof." 18 U.S.C. S 41 states that
"whoever . . . willfully injures, molests, or destroys any property of the
United States on any [refuge] lands or waters shall be fined under this
title or imprisoned not more than six months, or both."

In contrast, in the Prairie Pothole Region of the United States, which
comprises portions of the states of Iowa, Minnesota, Montana, North
Dakota, and South Dakota, farmlands received from the Farm Service Agency
have been managed much like other easements the Service has acquired
there. The Service has had a long-standing easement program in this region
to protect wetlands and grasslands for the benefit of migratory waterfowl.
It has run a regular program to manage all of its easement acquisitions in
the region, including farmlands received from the Farm Service Agency,
largely through routine aerial surveillance and on-the-ground inspections
of suspected violations. Today the Service has more than 28,000 perpetual
easements in the Prairie Pothole Region, covering more than 2.3 million
acres, including more than 68,000 acres of easements from the Farm Service
Agency. Because of the unique nature of this program, we excluded the
farmlands found in the Prairie Pothole Region from our review. Concurrent
with this report, we are reporting separately on the Service's land
acquisitions in this region.

For this report, outside of the Prairie Pothole Region, we examined (1)
the extent of farmland received by the Service from the Farm Service
Agency, (2) the extent to which the Service is currently managing its
farmlands, and (3) alternatives for managing these lands.

To address the objectives of this report, we visited and collected
documentation from five refuges located in five different regions. We also
interviewed officials at and collected documentation from the Service's
headquarters and eight regional offices, as well as the Farm Service
Agency's headquarters and four of its state offices. Additionally, we drew
a random sample of 98 farmlands, obtained from a list provided by the
Service, and conducted a telephone survey with the 51 refuge office
managers responsible for these lands about the management activities they
have undertaken on the lands during the years 2002 through 2006. We
received responses from 100 percent of our sample, and the results can be
projected to the population of documented farmlands, with an overall
margin of error of +-10 percentage points at the 95 percent confidence
level unless otherwise noted. Specifically, to determine the extent of the
Service's acquisition of farmlands, we obtained and reviewed property
records from the Service's headquarters and regional offices and the
refuge offices that we visited. We assessed the reliability of the data
and found it to be sufficiently reliable for the purposes of this report.
We were informed by some regional realty officials, however, that the
database was probably incomplete. Consequently, we contacted each regional
office to determine if the offices maintained additional farmland records
outside of the centralized database. Three Service regions provided us
with additional farmland records that were not contained in the database.
To determine the extent to which the Service is currently managing its
farmlands, we reviewed relevant agency guidance for managing easement and
fee-simple lands. Through our site visits and survey, we collected
information on the Service's management activities, including information
on inspections, enforcement, contacts with landowners, maintenance, and
restoration, and on factors affecting their management of farmlands.
Finally, to determine alternatives for managing these lands, we examined
the Service's management challenges and discussed them with Department of
the Interior officials, including officials from the Service's Realty and
Refuge Offices, and the Department of the Interior's Office of the
Solicitor. We performed our work from September 2006 through July 2007 in
accordance with generally accepted government auditing standards.
Appendixes I and II present a more detailed discussion of our scope and
methodology.

Results in Brief

The Service's records indicate that it has received at least 1,400
easement and fee-simple farmlands from the Farm Service Agency since 1986,
but the actual number is unknown because the Service's farmland records
are incomplete. According to the Service's records, the farmlands are
scattered across 38 states and range in size from less than 1 acre to more
than 2,200 acres, although most are smaller than 50 acres. The majority of
the lands were acquired between 1986 and the mid-1990s; acquisitions over
the past 10 years have been minimal. The Service is responsible for
managing more than 132,000 farmland acres, of which about half are
included in the Service's centralized land records database. We learned of
the remaining half from three regional offices, because the lands had not
been entered into this centralized database. Roughly 90 percent of the
lands have easements on them, and about 10 percent are owned in fee
simple. Most of the lands were received to protect wetlands, but some were
meant to protect other resources, including floodplains or specific
wildlife habitats. In addition, we identified farmlands that were not in
the Service's database or reported to us by the regional offices, and we
therefore conclude that the figures we report represent a conservative
estimate of the total farmland acreage received from the Farm Service
Agency. Specifically, at several refuges we visited, we found that the
refuge office kept files on additional easement farmlands. At one of the
refuges we visited, for example, we identified 5 additional easements and
another 31 potential, or unconfirmed, easements. Moreover, in some
instances, because of incomplete records, the Service may be unaware of
certain farmlands that it is in fact responsible for managing. For
example, in one region, a realty official told us that a landowner
recently called the Service with concerns about an easement on his land,
but the Service was unaware of the easement because no record of it
existed with the Service. After reviewing records at the county courthouse
in which the land was located, the official was able to confirm that an
easement had indeed been placed on the land.

The Service is generally not managing the majority of the farmlands it
received from the Farm Service Agency. Specifically, it has inspected only
an estimated 13 percent of its farmlands annually, on average, over the
last 5 years and, as a result, may be unaware of violations and unable to
ensure compliance with the restrictions on its easement lands. At four of
the five refuges we visited, we observed violations on at least one of the
refuge's easement lands, such as farming encroachment, grazing, or removal
of vegetation. Few refuge offices have tracked changes in land ownership
to ensure that new landowners are aware of easements, and very little
maintenance or restoration work has been completed on the farmlands.
Often, not even basic maintenance, such as maintaining fencing, has been
completed; yet when easement restrictions prohibit activities such as
grazing, fencing may be critical to ensure cattle do not enter and graze
the area. We found that the Service's management activities on its
farmlands are hindered by the following several factors:

           o First, devoting significant management attention to many of the
           farmlands has not been emphasized by the refuge offices. Refuge
           officials do not believe that these lands contribute significantly
           to the refuges' goals or mission, largely because the farmlands
           are too small, isolated, or located great distances from the
           managing refuge office. For example, over 75 percent of the
           farmlands are smaller than 100 acres. Several refuge managers
           commented that they would consider 100 acres the minimum size
           needed to support ecological viability, especially given the
           lands' often isolated nature. Additionally, 83 percent of the
           farmlands are located an hour or more away from their respective
           refuges, making it difficult for managers to visit them. Because
           managing the farmlands has not been a high priority, the lands may
           not be achieving the conservation purpose for which they were
           acquired.
           o Second, management of easement lands may be difficult because of
           uncertainty surrounding the scope and extent of the easements. For
           example, boundaries are not clear to many refuge managers because
           detailed surveys were not done when the easements were established
           on these lands. When easement boundaries are not well defined,
           refuge managers may not be able to determine if a violation is
           occurring on the land.
           o Third, the Service is operating with constrained resources and
           declining refuge staff, and therefore the staff are not able to
           complete all the activities they believe are necessary to manage
           the farmlands. Over the past several years, refuge staff have been
           reduced significantly, and the Service plans to further reduce the
           refuges' workforce in the next 3 years under current
           workforce-planning efforts aimed at cutting expenses. For an
           estimated 88 percent of farmlands, the responsible refuge managers
           believe that they have not had enough resources over the past 5
           years to manage their farmlands.

While the Service is generally not managing the majority of its farmlands,
in some instances we identified farmlands to which the refuges are
devoting resources and paying a significant amount of management
attention, largely because the lands more closely align with the refuges'
goals. For example, we visited one easement land that provided an
important connection between the refuge boundary and a critical river
running through the area, and the refuge had completed extensive
restoration work.

The Service possesses limited alternatives for managing the farmlands it
has received from the Farm Service Agency because the lands became part of
the National Wildlife Refuge System. The alternatives the Service has had
at its disposal generally include (1) resetting refuge priorities to
ensure that farmlands are given management attention, (2) requesting
additional resources, and (3) paying little or no management attention to
the farmlands. Over the past 20 years, in the rare instances where
farmlands significantly contributed to the refuges' goals and mission,
individual refuges made them a priority and obtained the necessary
resources to manage them. These well-managed farmlands have been the
exception rather than the rule, however. The majority of the farmlands are
neglected because they do not contribute significantly to the goals of the
refuges or because they are costly to manage relative to their real or
perceived ecological value. On these lands, the Service is not meeting its
management responsibilities and, in some cases, may be permitting the
destruction of government property. Although some of these farmlands may
have little or no value to the Service, they are now subject to the
National Wildlife Refuge System's restrictions on land disposal, so the
Service generally cannot dispose of unwanted farmlands. Other federal land
management agencies have authorities that give them more flexibility in
managing their lands, which could serve as models for legislation
providing similar authority to the Service on a limited or short-term
basis to resolve the issue of unwanted farmlands. For example, the
Department of Agriculture's Forest Service has the authority to sell
certain small, isolated, low-market-value parcels of its landholdings
under specific conditions.

To ensure that the Service is able to provide an accurate accounting of
its farmlands and has an accurate assessment of its overall farmland
management responsibilities, we are recommending that the Secretary of the
Interior direct the Director of the U.S. Fish and Wildlife Service to take
the necessary steps to ensure that the Service's records of its farmlands
are accurate and complete. In addition, we are recommending that the
Service develop a proposal to Congress seeking authority for additional
flexibility to deal with farmlands that it deems not in the best interest
to continue to include as part of the National Wildlife Refuge System.
Factors affecting how significantly individual farmlands contribute to the
mission of the refuge system, along with current and long-term management
costs, will be important considerations for determining which farmlands
may not be in the best long-term interest of the system. Once such
considerations are weighed, the Service can determine an appropriate
proposal for congressional consideration. Since it could take years for
the Service to develop a proposal and have the proposal acted on by
Congress, as well as for the Service to exercise any new authority that
may be provided, we are not making a recommendation at this time
concerning the current level of management occurring on its farmlands.
Nevertheless, we believe that the Service should be mindful of ensuring a
minimum level of management attention to the farmlands it ultimately
retains. In commenting on a draft of this report, the Department of the
Interior concurred with our recommendations and provided several technical
clarifications, which we have made as appropriate. Appendix III presents
the Department of the Interior's comment letter. The Department of
Agriculture did not provide comments.

Background

The National Wildlife Refuge System provides habitat for more than 5,000
species of birds, mammals, and fish, and these refuges protect unique
wildlife habitats and species. Under the National Wildlife Refuge System
Administration Act of 1966, as amended, the Service is responsible for
managing each refuge to fulfill the mission of the system as a whole, as
well as to fulfill the specific purposes for which the individual refuge
was established.^9 Therefore, refuges often have additional conservation
goals tailored to a specific purpose, such as protecting the habitat of
particular endangered or threatened species. Individual refuges may
consist of contiguous tracts of land--ranging from less than 1 acre to
more than 19 million acres--or separate tracts of land scattered over one
or more states. Each refuge may encompass land that is (1) completely
federally owned; (2) primarily federally owned, with isolated tracts of
nonfederal land; or (3) in a few refuges, primarily nonfederally owned,
with isolated tracts of federal land. Each refuge is managed by a refuge
manager and other refuge staff, including biologists, law enforcement
officers, and other specialized staff.^10 One of the Service's eight
regional offices oversees the activities of each refuge;^11 each region
reports to Service headquarters in Washington, D.C. (see fig. 1).

^916 U.S.C. S 668dd(a).

^10Some refuges are managed as refuge complexes, where a group of refuges
in geographic proximity and with similar purposes are managed by one set
of refuge staff.

^11In this report, we refer to the California-Nevada Office as a regional
office (region 8).

Figure 1: U.S. Fish and Wildlife Service Regions

The Service has a process for adding lands to the refuge system: it may
acquire land by purchasing the land itself or purchasing interests in the
land, by accepting donations of land from nonfederal entities, by
accepting land transfers from other federal agencies, or by requesting a
transfer of land from the public domain into the refuge system.^12 The
Service can also exchange tracts of land with nonfederal entities, such as
state agencies or private landowners, although the tracts of land must be
comparably valued. With few exceptions, all lands that become part of the
refuge system continue to be included in the system until otherwise
specified by an act of Congress.^13 Typically, when evaluating lands for
inclusion in the refuge system, the Service conducts an assessment of the
land, which includes an in-depth biological evaluation and opportunities
for public participation and also requires approval by the Service
Director or one of the regional directors. The Service keeps an inventory
of acquired lands, or interests in lands, in a centralized database
maintained by realty offices in Service headquarters and regions.

^12Public domain means that the title to the land has always remained with
the federal government. Almost 90 percent of the lands in the refuge
system came from the public domain.

^1316 U.S.C. SS 668dd(a).

The Service has also added lands to the refuge system through the
partnership it developed with the U.S. Department of Agriculture's Farm
Service Agency to help it implement provisions of the 1985 Farm Bill and
subsequent legislation. Even before the 1985 Farm Bill was passed, the
Farm Service Agency had an affirmative responsibility to protect any of
its inventory properties that had wetland or floodplain values. Two
executive orders issued in 1977 called on all federal agencies to minimize
the destruction, loss, or degradation of wetlands and to restore and
preserve the natural and beneficial values supported by floodplains.^14 In
carrying out their responsibilities, including acquiring, managing, and
disposing of federal lands, federal agencies were to preserve and enhance
the natural and beneficial values of wetlands.^15 The 1985 Farm Bill and
subsequent legislation provided the Farm Service Agency the authority to
protect any of its land that contained wetlands and floodplains, in
addition to other important resources. Specifically, the Farm Service
Agency had the authority to place perpetual conservation easements on
inventory properties that contained important resources or to transfer
lands in fee simple to other entities, such as federal, state, or local
government agencies.^16 Since the 1985 Farm Bill was enacted, subsequent
legislation substantially changed the amount and type of lands available
for protection. Notably, the 1990 and 1996 Farm Bills restricted the
agency's authority to establish easements under the program. For example,
the 1990 Farm Bill limited authorized easements on certain parcels of
lands to no more than 20 percent of the parcel available for agricultural
production. The 1996 Farm Bill removed these restrictions but imposed
strict procedural requirements for establishing an easement. Table 1
summarizes the major executive orders and legislation affecting the
protection of farmlands.

^14Executive Order 11990, Protection of Wetlands, and Executive Order
11988, Floodplain Management, both issued on May 24, 1977.

^15Executive Order 11990 also directs the executive branch agencies "to
avoid to the extent possible the long- and short-term adverse impacts
associated with the destruction or modification of wetlands and to avoid
direct or indirect support of new construction in wetlands wherever there
is a practicable alternative."

^16The 1985 Farm Bill also authorized the Farm Service Agency to develop
the Debt for Nature Program. This program allows the Farm Service Agency
to cancel a portion of delinquent borrowers' debt in exchange for the
borrowers' voluntarily establishing conservation easements on their lands.

Table 1: Executive Orders and Legislation Relevant to the Farm Service
Agency's Inventory Program

Executive orders and                                                       
legislation           Date         Major provisions                        
Executive Order 11988 May 24, 1977 Ordered federal agencies, in carrying   
on Floodplain                      out their responsibilities, to "take    
Management^a                       action to reduce the risk of flood      
                                      loss; to minimize the impact of floods  
                                      on human safety, health, and welfare;   
                                      and to restore and preserve the natural 
                                      and beneficial values served by         
                                      floodplains."                           
Executive Order 11990 May 24, 1977 Ordered federal agencies, in carrying   
on Protection of                   out their responsibilities, to "take    
Wetlands^b                         action to minimize the destruction,     
                                      loss, or degradation of wetlands, and   
                                      to preserve and enhance the natural and 
                                      beneficial values of wetlands."         
Food Security Act of  Dec. 23,     Authorized the Secretary of Agriculture 
1985 (1985 Farm       1985         to acquire easements in real property   
Bill)^c                            for conservation, recreational, and     
                                      wildlife purposes as part of farm loan  
                                      debt restructuring. A portion of a      
                                      farmer's outstanding farm loan(s) was   
                                      canceled when an easement was acquired  
                                      by the United States. The Secretary, or 
                                      any person or federal, state, or local  
                                      governmental entity, could be           
                                      designated as the party responsible for 
                                      enforcing the easement. The act also    
                                      authorized the Secretary to grant or    
                                      sell an easement, restriction,          
                                      development rights, or equivalent       
                                      thereof to a unit of local or state     
                                      government or to a private nonprofit    
                                      organization for conservation purposes. 
Agricultural Credit   Jan. 6, 1988 Authorized the Secretary of             
Act of 1987^d                      Agriculture, under certain              
                                      circumstances, to transfer without      
                                      reimbursement to any federal or state   
                                      agency for conservation purposes any    
                                      real property or interest in real       
                                      property (i.e., an easement).           
Food, Agriculture,    Nov. 28,     Imposed a number of restrictions on the 
Conservation, and     1990         Secretary of Agriculture when           
Trade Act of 1990                  establishing perpetual wetland          
(1990 Farm Bill)^e                 conservation easements. In establishing 
                                      easements, the Secretary was directed   
                                      to "avoid, to the extent practicable,   
                                      an adverse impact on the productivity   
                                      of the croplands." For example, in      
                                      certain cases, easement acreage was     
                                      limited to 10 or 20 percent of the      
                                      existing cropland, and the buffer area  
                                      adjacent to a wetland was generally not 
                                      to exceed 100 feet in average width.    
Federal Agriculture   Apr. 4, 1996 Modified the restrictions added by the  
Improvement and                    1990 Farm Bill and imposed new          
Reform Act of 1996                 procedural requirements for the         
(1996 Farm Bill)^f                 transfer authority authorized by the    
                                      Agricultural Credit Act of 1987. The    
                                      limitations on percentage of acreage in 
                                      the 1990 Farm Bill were replaced with   
                                      the limitation that the Secretary of    
                                      Agriculture should not establish a      
                                      wetland conservation easement on an     
                                      inventoried property that was cropland  
                                      on the date the property entered        
                                      inventory or was used for farming at    
                                      any time within 5 years of entering     
                                      inventory. The new procedural           
                                      requirements on the transfer authority  
                                      included public notices; consultations  
                                      with the applicable governor and at     
                                      least one elected county official; and  
                                      the possibility of at least one public  
                                      meeting, if requested.                  

Source: GAO analysis of executive orders and legislation.

Note: The provisions for the Farm Service Agency's farmlands easement
programs are codified at 7 U.S.C. SS 1985(g), 1997, and 2002.

a42 Fed. Reg. 26951, May 25, 1977.

b42 Fed. Reg. 26961, May 25, 1977.

cPub. L. No. 99-198, title XIII, SS 1314(a)(2)(B) and 1318(a), 99 Stat.
1354, 1530-1 (1985).

dPub. L. No. 100-233, title VI, SS 612 and 616, 101 Stat. 1568, 1674 and
1682 (1988).

ePub. L. No. 101-624, title XVIII, subtitle A, SS 1813(h) and 1815, and
title XXIII, S 2388(j), 104 Stat. 3359, 3823-4, 3825-6, and 4053 (1990).

fPub. L. No. 104-127, title VI, subtitle D, SS 639, 642, and 646, 110
Stat. 888, 1097, 1102, and 1103-4, (1996).

In establishing easements, the Farm Service Agency generally used
boilerplate language that it developed in coordination with the Service.
Largely, it placed very restrictive easements on those portions of the
property that contained important resources. Easement restrictions
prevented the landowner from developing the land, conducting agricultural
practices, altering the vegetation or hydrology of the land in any way, or
otherwise disturbing the land. In some cases, if the purpose of the
easement was to protect a floodplain, then less-restrictive easement
language was used. On these easements, most agricultural practices were
allowed, and the landowner was restricted primarily from developing the
land.

In delineating easement boundaries on a property, the Farm Service Agency
generally established easement areas embedded within noneasement farmland
because the agency was protecting only those portions of the property
encompassing important resources. For example, on a 150-acre farm, an
easement might have protected only the 23 acres that constituted a buffer
to a stream running through the property. Or if several wetlands were
identified on a 320-acre field, only those wetlands, possibly along with a
buffer, might have received protection, while the remainder of the field
could still have been farmed (see fig. 2). As a result, easement areas
might consist of multiple, noncontiguous parcels on the larger farm
property.

Figure 2: Example of an Easement Placed on Farm Property

The Service Has Received at Least 1,400 Farmlands Covering 132,000 Acres, but
the Actual Number Is Unknown

The Service's records indicate that it has received at least 1,400
easement and fee-simple farmlands from the Farm Service Agency since 1986,
but the actual number is unknown. According to its records, the Service is
responsible for managing over 132,000 acres of mostly small farmlands
scattered across 38 states. In addition, we identified farmlands besides
those in the Service's centralized database or reported to us by the
regional offices. Consequently, we conclude that the figures we report
represent a conservative estimate for the total amount of the Service's
farmlands.

Since 1986, the Service Has Received Mostly Small Farmlands Scattered across 38
States

According to its records, the Service has received at least 1,400
farmlands from the Farm Service Agency.^17 About 90 percent of the lands
are privately owned, with easements on them; the Service owns the
remaining 10 percent in fee simple. The farmlands cover more than 132,000
acres and range in size from less than 1 acre to more than 2,200 acres.
Most of the farmlands, however, are small. Specifically, more than half
the farmlands are smaller than 50 acres, and only about 1 percent are
larger than 1,000 acres (see table 2).

Table 2: Size of Farmlands Received by the Service

Size in acres Number Percentage 
                    197        14% 
10-49            606         43 
50-99            301         21 
100-999          292         21 
1,000+            16          1 
Total          1,412       100% 

Source: GAO analysis of U.S. Fish and Wildlife Service data.

Farmlands owned by the Service in fee simple are generally larger than
farmlands with easements. While more than 60 percent of its fee-simple
lands are larger than 100 acres, less than 20 percent of the easement
lands exceed 100 acres. Accordingly, in terms of acreage, the Service owns
about 37,000 acres (28 percent) of its farmlands in fee simple, and nearly
95,000 acres (72 percent) are in the form of easements. The Service's
farmlands are scattered across more than 500 counties in 38 states within
the continental United States, with the highest concentration of farmlands
located in the Midwest (see fig. 3).^18

^17In most cases, the number of farmlands represents the original number of
properties the Service received from the Farm Service Agency, either
through fee simple or an easement. That is, each farmland corresponds to a
legal deed placed on all or a portion of a property that was under single
ownership at the time the realty transaction was legally finalized. In
some instances, however, easement restrictions that were placed on
separate portions of a single property may have been tracked separately,
and thus the lands would be counted separately here even though they were
under single ownership at the time of the realty transaction. In addition,
if an owner subsequently sold a portion of a property covered by the
easement, both ownerships would still be covered by the easement
restrictions. In other words, through land sales, one easement land with
one landowner could turn into two or more easement lands, with two or more
landowners. These types of changes would not be reflected in the numbers
reported here, however.

Figure 3: Concentration of Service Farmlands by State

The most common purpose for establishing easements or accepting fee-simple
ownership was to protect wetlands. Specifically, the results of our survey
show that about 80 percent of the farmlands were received to protect
wetlands. Farmlands were also received for other conservation purposes,
such as protection of riparian zones; floodplains; or the habitats of
endangered or threatened species, including the bald eagle and San Joaquin
kit fox.^19

^18The Service manages more than 650 additional farmlands--more than 68,000
acres in easements and 7,000 acres in fee simple--within the Prairie
Pothole Region of the United States. These farmlands are not included
within the scope of this review.

The majority of the farmlands were acquired between 1986 and the
mid-1990s; acquisitions over the past 10 years have been minimal (see fig.
4). According to Farm Service Agency officials, acquisitions have declined
because significantly fewer farm properties have entered the Farm Service
Agency's inventory since the mid-1990s. This is largely a result of the
improving farm economy and additional financing options farmers now have
besides loan foreclosure, according to the officials. Also, the
restrictions included in the 1990 and 1996 Farm Bills significantly
reduced the type of inventory farmlands available for protection.

^19Effective August 8, 2007, the bald eagle was removed from the federal
list of endangered and threatened wildlife.

Figure 4: Number of Farmlands Received by the Service, by Calendar Year

Note: We were unable to obtain an acquisition date for four farmland
records and therefore did not include those farmlands in this analysis.
Also, the years graphed here represent the year that an easement was
placed on a farmland or the land was transferred to the Service in fee
simple, which in many instances occurred several years after the Service
reviewed the property and made a recommendation for protection. For
example, one of the properties received in 2004 was initially placed in
the Farm Service Agency's inventory in 1994, but it took 10 years to
remove the land from inventory and formally place an easement it.

About half the farmlands reported here are included in the Service's
centralized database. We learned about the remaining half of the farmlands
from three regional realty offices. These regions, for various
administrative reasons, have not entered their records into the database,
according to Service officials. For example, in one region most easements
were not entered into the database because final legal reviews had not
been completed.

The Service's Farmland Records Are Incomplete

We identified farmlands that were not in the Service's centralized
database or reported to us by the regional realty offices and therefore
conclude that the Service has an incomplete account of its farmlands.
Specifically, through our site visits and contacts with refuge managers,
we found multiple instances in which individual refuge offices had files
for farmland easements that were not otherwise reported in the centralized
database or by the regional realty office. For example, one refuge manager
we spoke with said he was responsible for 29 additional farmlands not
included in the centralized database or reported to us by the regional
realty office. These farmlands equate to over 80 percent (more than 900
acres) of the farmlands that this refuge is responsible for managing. In
other instances, the Service had made a recommendation to the Farm Service
Agency to establish an easement, but the refuge office lacked a final deed
confirming the easement's establishment. As a result, some questions exist
as to whether the Service had received management responsibility for an
easement that was established on the land. At one of the refuge offices we
visited, for example, we identified 31 potential, or unconfirmed,
easements in addition to 5 confirmed easements, comprising over 1,500
acres, that had not been reported to us. The refuge office's files for the
31 unconfirmed easements contained documentation, such as maps and records
of site visits, but did not contain a final deed to confirm that the
easement was official. Farm Service Agency and Service officials we spoke
with in this state explained that the Farm Service Agency did not
generally send a copy of the final deed to the Service when an easement
was established. Thus, for these 31 unconfirmed farmlands, it is probable
that an easement was established and the Service received management
responsibility, but a copy of the final deed was simply not included in
the file.

Because of the Service's incomplete records, even the refuge offices may
be unaware of all of the farmlands that they are in fact responsible for
managing. For example, in one region, a realty official told us that a
landowner had recently called the Service with a concern about an easement
on his land. The Service, however, was unaware that it was responsible for
managing the easement because it had no record of the farmland. After the
call from the landowner, the Service confirmed that an easement had been
established on the property by reviewing records at the courthouse in the
county in which the farmland is located. In contrast, we identified a
couple of instances where the refuge office maintained a file for a
farmland, or assumed it was responsible for managing a farmland, when in
fact the Service never received management responsibility, or an easement
had not been established.

The Service Is Generally Not Managing the Majority of Its Farmlands

The Service is generally not managing the majority of its farmlands.
Management activities the refuge offices are responsible for include
regularly inspecting their farmlands to check for violations; taking
enforcement actions on violations; maintaining contact with the landowners
of easement lands; conducting limited maintenance activities including
marking easement boundaries; and, in some cases, doing ecological
restoration. Overall, however, we found that most refuge offices are
carrying out these management activities on a very limited basis. Several
factors affect the extent to which refuge offices devote management
attention to their farmlands. First, devoting significant management
attention to many of the farmlands has not been a priority because refuge
officials do not believe that the lands contribute significantly to the
refuges' goals or mission. Second, management of easement lands may be
difficult because of uncertainty surrounding the scope and extent of the
easements. Third, the Service is operating with constrained resources and
declining refuge staff and therefore the staff are not able to complete
all the activities they believe are necessary to manage the farmlands. In
some instances, however, we identified farmlands that the refuges are
actively managing, largely because the lands more closely align with the
refuges' goals.

The Service Performs Limited to No Management Activities on Most of Its
Farmlands

The Service's refuge offices have carried out very limited management
activities on the majority of their farmlands. Through site visits to five
refuge offices and our survey of the refuge offices' management of their
farmlands (which covered 2002 through 2006), we found that most management
activities, including inspections, contacts with landowners, and
maintenance and restoration activities, are carried out on only a very
limited basis. For an estimated 91 percent of farmlands, the refuge
managers believe that over the past 5 years they were able to accomplish
only half or less than half of the activities they believed were necessary
to manage these lands. In general, we also found that the Service's
fee-simple farmlands tend to receive more management attention,
particularly inspections and completed restoration work, than its easement
lands.

While Service guidance calls for each farmland to be inspected at least
once every year, very few farmlands are inspected annually. On-the-ground
inspections are important to determine whether the landowners are
complying with easement restrictions and to check for violations. From our
survey, however, we estimate that only 13 percent of the farmlands, on
average, have been annually inspected (see fig. 5). Additionally, on
almost one-third of the farmlands, refuge managers did not believe (and
were unable to provide documentation to show) that the farmlands had been
visited by refuge staff at any point over the 10-year period from 1997
through 2006.

Figure 5: Estimated Frequency of Inspections from Calendar Year 2002
through 2006

Note: Percentages do not add to 100 percent because of rounding.

Because the majority of farmlands are rarely inspected, refuge officials
may not know the extent to which violations may be taking place on their
farmlands. The farmlands, or the interests in them, are government
property, so the Service may take civil or criminal action against
violators for the destruction of government property. Through our survey,
we found that of the properties that were inspected at least once in the
last 5 years, the most recent inspection found violations on 26 percent of
the farmlands.^20 Common violations included some agricultural practices
such as grazing or haying, trespassing, dumping trash, hunting illegally,
or removing signs (see fig. 6). At four of the five refuges we visited, we
identified a violation on at least one easement land. For example, we
visited one easement where the landowner had recently cleared much of the
natural vegetation with a bulldozer and burned the land; at a different
site, we saw where the landowner had built a driveway and shed through the
middle of the easement area. At still another refuge, the refuge manager
stated that of the 34 easement lands he has personally visited in the last
3 years, 90 to 95 percent exhibit some kind of violation, most of which
degrade the ecological value of the land. At the fifth refuge, violations
appeared to be present on a few farmlands, but we could not confirm them
because of a lack of clear easement boundaries.

^20The 95 percent confidence interval for this percentage is from 15
percent to 40 percent.

Figure 6: Farming Activities in Violation of a Wetland Easement, April
2007

Note: Violations on this 10-acre easement include tilling the land and
clearing some of the wooded area for farming.

One way the Service can help prevent violations is by maintaining regular
contact with landowners and ensuring that landowners, particularly new
landowners, clearly understand the easement terms; yet despite Service
guidance to this effect, refuge managers rarely keep up with such
contacts. For instance, only 13 of the 51 refuge managers we surveyed
reported that they tracked land ownership changes on their easement
lands--for example, by checking land sales records at county
courthouses--and only 11 reported that they notified new landowners about
the easement on their property. Several refuge managers said that
farmlands changed ownership frequently, making it difficult for them to
keep up with the current landowners for all their easement lands. For
example, one refuge manager stated that at least 70 percent of his 80
easement lands had changed ownership in the past several years.
Compounding these difficulties, if a landowner divides and sells a portion
of the farmland covered by an easement, the Service must then track all
landowners to ensure that all portions of the easement area are in
compliance. Through our survey, we found that multiple ownerships under
one easement are fairly common; in one case, eight current landowners had
a portion of a single easement on their property.

Additionally, refuge offices have conducted very little maintenance on
their farmlands. On the basis of our survey, we estimate that
approximately 22 percent of the Service's farmlands have received
maintenance over the past 5 years. While limited maintenance activities
are required, particularly on easement lands, certain basic activities are
called for. Specifically, Service guidance calls for all farmland
boundaries to be marked, or posted, in a manner that is clear to
landowners and the refuge officials responsible for managing them. During
our site visits, however, we found that the farmlands were often not
posted. At two of the refuges we visited, none of the easements were
posted. At the other three refuges, while some farmlands had a few
postings, others did not. In some cases, we also found that important
fencing had not been installed or maintained. For example, we visited one
easement that was being improperly grazed because the fencing that had
been put up years ago was no longer able to keep cattle from entering and
grazing the easement land.

Similarly, refuge offices have completed only limited restoration
activities on their farmlands. When the farmlands were initially received
from the Farm Service Agency, the Service expected that restoring
wetlands, planting uplands, and other rehabilitation would be an important
component of the program, and in some cases a significant early investment
was made to restore some lands. From our survey, however, we estimate that
over the last 5 years, about 9 percent of the farmlands have had
restoration work done, such as planting trees or altering water flow.
Overall, since the farmlands were initially received by the Service,
refuge officials whom we surveyed were aware of restoration work on only
24 percent of their farmlands.^21 Service officials in one region told us
that although considerable funding was initially made available for
restoration projects, subsequent funding has not kept pace with
restoration or maintenance needs. For example, one of the refuge managers
in that region said that his office was given, on average, more than
$225,000 per year for restoration over the first 5 years of receiving
farmlands from the Farm Service Agency. This funding enabled completion of
all necessary restoration projects initially, he said, but later funding
to maintain the projects was not provided. In contrast, a number of refuge
managers said that restoration work is unnecessary on some farmlands
because the lands adequately achieve the conservation purpose for which
they were acquired if they are simply left in their natural state.

^21The 95 percent confidence interval for this percentage is from 15 to 36
percent.

Several Factors Affect the Extent of the Service's Management Activities

We found that several factors affect the extent to which the Service's
refuge offices conduct management activities on their farmlands. First,
managing the farmlands has not been a high priority because refuge
officials do not believe that the lands contribute significantly to the
refuges' goals or mission, because they are too small, isolated, and
distant. While the size of the farmlands varies considerably--of the more
than 1,400 farmlands nationwide, size ranges from less than 1 acre to more
than 2,200 acres--over 75 percent of the farmlands are less than 100
acres. Several refuge managers commented that they consider 100 acres a
minimum size to support ecological viability, especially given the lands'
fragmented nature. Often the farmlands are fragmented among active
agricultural lands, have little or no connectivity to other wildlife
habitat, or are widely scattered over a vast area. The farmlands are also
often located many miles from the managing refuge office. At one refuge we
visited, for example, the office is responsible for managing more than 100
scattered farmlands, many of which are located more than 60 miles away
(see fig. 7).

Figure 7: Location of the Farmlands That Michigan's Shiawassee National
Wildlife Refuge Is Responsible for Managing

Through our survey, we estimate that about 83 percent of the Service's
farmlands are located an hour or more away from their respective refuges,
making it difficult for managers to visit them (see table 3). From a
statistical analysis of our survey results, we found that both larger
farmlands and those that are closer to the managing refuge were
significantly more likely to have been inspected in the past 2 years than
smaller or more distant farmlands.^22

^22See appendix II for a more detailed discussion of our analysis.

Table 3: Estimated Travel Times between the Managing Refuge Offices and
Their Farmlands

Time (in hours)                    Percentage of farmlands 
Less than 1                                            18% 
One hour or more, but less than 2                       40 
Two hours or more, but less than 3                      27 
Three or more hours                                     16 

Source: GAO survey of the U.S. Fish and Wildlife Service's farmland
management.

Note: The data in this table represent the responses from refuge managers
to the following question, "How much time, in minutes, does it typically
take to drive from the refuge office to this property?" Percentages do not
add to 100 percent because of rounding.

Further, because managing the farmlands has not been a high priority, the
lands may not be achieving the conservation purpose for which they were
acquired. For example, we visited one farmland where an easement had been
acquired for wetland purposes, but over time, the area had been filled in
and become overgrown with dense vegetation and trees. The refuge manager
said that in that state, the land exhibited no wetland characteristics and
thus provided no value to any waterfowl species. He said it would take a
significant amount of restoration work and subsequent continued
maintenance to restore and sustain any wetland values on the land. We
found that refuge managers believe that about 23 percent of the farmlands
are currently meeting their conservation purpose to a great or very great
extent. We also noted instances where the Service's assessment determined
that the land did not contain important resources worthy of protecting,
and the Service therefore did not request an easement on the land; the
Farm Service Agency nevertheless established an easement on the land and
named the Service the easement manager. In still other cases, the specific
acres protected did not match the areas the Service had identified as
containing important resources. For example, we reviewed one case where
the Service received a farmland in fee simple and also received an
additional parcel of land containing a potentially contaminated concrete
building left over from former farming operations.

Second, managing easement lands may be difficult because of uncertainty
surrounding the easements' scope and extent. For instance, because legal
surveys often were not completed on the farmlands when the easements were
first established, precise boundaries may not be available. Few of the
property files we reviewed on our site visits contained evidence that the
land had been legally surveyed to establish precise boundaries. Through
our survey, we found that about 49 percent of the farmlands' boundaries
were clear to refuge managers. The lack of clear boundaries has made it
difficult for some refuge officials to identify the location of all their
easement lands and has also made it difficult to identify violations, such
as farming encroachment or removal of vegetation, and to enforce
restrictions. For example, we visited an easement where the refuge manager
said the landowner moved the easement posting a few feet every year so as
to expand his farming operations onto the easement land. Because the
easement had not been legally surveyed, however, the refuge manager said
it would be very difficult to insist that the farmer move the posting back
to its original boundaries. The refuge manager said he was not even sure
where the posting should officially be, given the lack of clear
boundaries. In addition to unclear boundaries, some easement restrictions
have been difficult for refuge managers to fully interpret and, in some
cases, enforce. At one of the refuge offices we visited, for example, the
refuge manager explained that while the easement allows "normal farming
practices," it also prohibits altering the land's hydrology. He said that
one of the easement landowners used farming equipment to level his land.
According to the refuge manager, this leveling did alter the hydrology but
could also be considered part of "normal farming practices." Thus, what
constitutes a violation may sometimes be unclear because of ambiguity in
the language of some easement restrictions.

Third, the Service is operating with constrained resources and declining
refuge staff, and therefore the refuge offices are not able to complete
all the activities they believe are necessary to manage the farmlands.
Over the past several years, refuge staff have been reduced significantly,
and the Service plans to further reduce the refuges' workforce in the next
3 years under current workforce-planning efforts aimed at cutting
expenses. Currently, nearly 200 refuges are unstaffed, with this number
likely to increase in the future. Senior headquarters and regional
officials explained that the refuges have done what they can to manage
their farmlands, given the limited resources they have at their disposal.
For an estimated 88 percent of farmlands, the refuge managers believe that
they have not had enough resources to manage their farmlands over the past
5 years. A number of refuge officials also specifically cited the lack of
law enforcement capabilities as a resource constraint. At two of the
refuges we visited, the offices did not have any kind of permanent law
enforcement staff; at two others, the law enforcement officer split his
time between refuge offices. Several managers pointed out that they simply
would not have the law enforcement capability to spend significant time
taking enforcement actions for violations on farmlands. Our survey results
suggest that the number of farmlands any one refuge office is responsible
for varies greatly (from 1 to over 100) and that the more farmlands a
refuge office is responsible for, the less likely the office is to have
inspected a property in the past 2 years.^23

Farmlands That Align Closely with Refuges' Goals Receive More Management
Attention

While the Service is generally not managing the majority of its farmlands,
in some instances we identified farmlands that refuge offices are devoting
significant management attention to, largely because these farmlands more
closely align with the refuges' goals. Service officials explained that
farmlands that are or have the potential to be more ecologically valuable
generally receive more attention than those with less potential. For
example, the Service's largest documented farmland is a fee-simple
acquisition 2,263 acres in size. When this land was acquired by the
Service, it was severely degraded by agricultural practices. At present,
however, because of significant restoration work done by refuge staff, the
land encompasses over 1,000 acres of native hardwood trees and abundant,
varied wildlife, including significant populations of marsh and water
birds, shorebirds, and other waterfowl. Although the land is located more
than an hour away from the refuge office responsible for managing it, the
refuge manager said that refuge staff visit the property multiple times
per year for restoration and maintenance activities. Funding was allocated
for restoration of this farmland before fiscal year 2002, but since then,
the refuge office has used routine refuge operations-and-maintenance
funding to manage it.

In other instances, farmlands are located in areas the refuge offices are
already focusing on, and these lands therefore receive more management
attention. For example, we visited one refuge office responsible for
managing a 606-acre easement adjacent to the refuge boundary that provided
a critical connection between the refuge and a river running through the
area. On this easement, the Service developed a management plan to
maintain and enhance wetland ecosystems, riparian zones, and wildlife
habitats while also allowing limited grazing on certain portions of the
land. In the last few years, the refuge office received grant funding to
conduct restoration work to increase the wetlands and address
channelization and erosion problems. Currently, two landowners each own a
parcel of the easement, and according to the refuge manager, they have
both supported the refuge's efforts to restore the land. Overall, he
suggested, the situation is a good example of how Farm Service Agency
easement lands can successfully contribute to the mission of the National
Wildlife Refuge System.

^23See appendix II for a more detailed discussion of our analysis.

In another instance, we visited a farmland of 361 acres owned in fee
simple by the Service that, because of its location, was turned into the
headquarters of a wetland management district. In this district, the
Service is working to acquire more wetland and upland habitat for
waterfowl, migratory birds, and other wildlife. The Service conducted
wetlands restoration work on the property and turned the old, dilapidated
farmhouse into the refuge office (see fig. 8). Fishing access, a nature
trail, and other environmental education activities now complement the
farmland's restored wetlands.

Figure 8: Old Farmhouse in New York, Renovated to Serve as a Refuge
Office, April 2007

The Service Has Limited Alternatives for Managing Its Farmlands

Since the Service first started receiving farmlands from the Farm Service
Agency more than two decades ago, it has had limited alternatives for
managing them because the lands became part of the National Wildlife
Refuge System. The alternatives include (1) resetting refuge priorities to
ensure that farmlands are given management attention, (2) requesting
additional resources, and (3) paying little or no management attention to
the farmlands. Generally, when a farmland has significantly contributed to
the goals and mission of the responsible refuge, the Service has either
adjusted its refuge priorities to ensure farmland management or found the
resources necessary to actively manage the property. These instances have
been rare, however. For most farmlands, the Service has chosen the third
alternative--paying little or no management attention to its
farmlands--and it is unlikely to increase its management attention. Refuge
officials we spoke with said that they were unlikely to raise the priority
of farmland management activities in the future unless additional refuge
funding was provided or regional or national officials directed them to do
so. They explained that because of the nature of the farmlands, it simply
does not make sense to reset refuge priorities to increase management
attention to them. Refuge managers said that they would also be highly
unlikely to ask for additional resources to manage farmlands, except in
rare instances, because of the continuing shortfall in resources to
complete even their core refuge activities, such as protecting critical
habitat. Service headquarters and regional officials agreed that as staff
and budgets continue to be downsized, further management attention to the
farmlands is unlikely. By paying little or no management attention to its
farmlands, however, the Service risks allowing habitats to be degraded;
violations to go unchecked; and, in some cases, government property to be
damaged.

Because these farmlands are now part of the National Wildlife Refuge
System, the Service has limited options--restricted by federal law and
executive orders--for relinquishing management responsibilities for
unwanted farmlands or farmlands it deems not in the best interest to
retain in the refuge system. Under provisions of the National Wildlife
Refuge Administration Act of 1966, as amended, lands that are part of the
National Wildlife Refuge System cannot be transferred or otherwise
disposed of (except by exchange) unless the Secretary of the Interior
determines, with the concurrence of the Migratory Bird Conservation
Commission,^24 that the land or interest in lands is no longer needed for
the purposes for which the refuge system was established. Refuge lands
have never been transferred or disposed of in this manner, however. In
addition, the Service is required to abide by the provisions of Executive
Order 11990, covering wetland protection, and Executive Order 11988,
covering floodplain management, which require protection and preservation
of these resources. It has also been the policy of the federal government
since 1989 to achieve "no net loss" of wetlands in carrying out the
government's land management activities.^25

^24The Migratory Bird Conservation Commission considers and acts on
recommendations of the Secretary of the Interior for the purchase or
rental of land, water, or both for the conservation of migratory bird
habitat. The commission consists of the Secretary of the Interior, the
Administrator of the Environmental Protection Agency, the Secretary of
Agriculture, and two members from each house of Congress.

In a few isolated instances, the Service has successfully exchanged
farmlands for private lands under existing authorities. According to
Service officials, fee-simple farmlands were generally exchanged for
private lands to acquire lands nearer existing refuges, and easement
exchanges were done mainly to acquire land that was more ecologically
valuable or to enable landowners to accomplish their land use objectives.
Service officials told us, however, that the possibilities for further
land exchanges are limited, especially for easement lands. On fee-simple
lands, the main difficulty is finding private entities that are interested
in obtaining the Service's isolated parcels and also own lands the Service
is interested in acquiring. For easement lands, the only parties
interested in acquiring easement rights are usually the underlying
landowners, and rarely do they own other lands that would make sense to
exchange. In addition, for both fee-simple and easement properties, the
administrative and procedural aspects of accomplishing land exchanges are
often extremely costly and burdensome. For instance, we reviewed one
exchange, involving an easement, that took more than 3 years to complete
and required costly survey and appraisal work in addition to significant
time and effort on the part of Service realty and refuge staff.

In addition to exchanges, the Service can look to third parties, such as
state agencies or nonprofit organizations, to establish cooperative
agreements to manage its farmlands. For example, one refuge has an
agreement with a state agency to manage one of its farmlands because the
land is located near a state-owned natural area. Service officials
indicated that such situations are rare, however, and it is unlikely that
many additional cooperative agreements will be established, given the
location, size, and ecological value of the lands and the management costs
to these groups--the very reasons the Service does not itself accord the
lands significant management attention.

In contrast, other federal land management agencies have authorities that
provide them flexibility to dispose of lands found not in their best
interest to retain. These authorities may provide models for addressing
farmlands that the Service determines no longer serve the purposes of the
refuge system. For example, under the Southern Nevada Public Land
Management Act of 1998, the Department of the Interior's Bureau of Land
Management has the authority to sell or exchange public land within a
specific area around Las Vegas, Nevada.^26 A portion of the proceeds from
the sale of such lands is put into a fund that is then used for a variety
of purposes, including conservation initiatives and certain land
acquisitions. The account is also available for the reimbursement of costs
incurred by the local offices of the Bureau of Land Management in
arranging sales or exchanges under this act.

^25Given the value of wetlands, the administration set a national goal in
1989 of balancing losses and gains to achieve no net loss of wetlands.
Each subsequent President has reaffirmed and expanded this goal of
achieving net gains of wetlands over the long term.

Similarly, the Small Tracts Act gives the Department of Agriculture's
Forest Service the authority to sell certain small, isolated parcels of
its landholdings under specific conditions.^27 Specifically, eligible
parcels may be sold if they have a market value of less than $150,000, are
smaller than 40 acres, are not efficiently administered because of their
location and size, and cannot be sold under any other authority. The Small
Tracts Act authorizes the Forest Service to exchange lands or interests in
lands of approximately equal value without need of formal appraisals. This
authority can greatly facilitate land transactions where the cost of the
appraisals is significant enough to diminish the likelihood of the
transaction. Pending legislation would, among other provisions, increase
the acreage from 40 to 100 acres.^28

Additionally, the Federal Property and Administrative Services Act of
1949, as amended, governs the disposal of most federal property that a
federal agency no longer needs to carry out its mission, programs, and
activities.^29 Such property is reported as excess to the General Services
Administration and offered first to other federal agencies for their use
and then to state or local governments or to certain tax-exempt nonprofit
organizations through such mechanisms as negotiated or public sales. If
these entities do not wish to acquire the property, the General Services
Administration can then dispose of it through a competitive sale to the
public, generally via sealed bid or auction.

^26Pub. L. No. 105-263, 112 Stat. 2343 (1998).

^27Pub. L. No. 97-465, 96 Stat. 2535 (1983), codified at 16 U.S.C. S
521c-i.

^28H.R. 485, 110th Congress (2007). A similar bill, H.R. 1905, was not
enacted in 109th Congress.

^2940 U.S.C. SS 521-9, 541-59. See also GAO, Federal Real Property: Most
Public Benefit Conveyances Used as Intended, but Opportunities Exist to
Enhance Federal Oversight, [21]GAO-06-511 (Washington, D.C.: June 21,
2006).

If the Service had authority for its farmlands similar to the above
examples, it could consolidate some of its unwanted farmlands through an
arrangement where it sold fee-simple farmlands or the easement rights and
then directed the proceeds to the purchase of lands that were more closely
aligned with the goals and mission of the refuge system. Considerations
such as the time frame for allowing such sales and purchases, and whether
such activities could occur on a nationwide scale or whether purchases
would be limited to the states or region in which the sale took place,
would be important. An advantage to such an arrangement would be that the
Service could, at a minimum, offset the loss of any wetlands and
floodplains. According to Service officials, however, a significant
obstacle to this option would be that although farmlands owned in fee
simple can have considerable market value, many of the easements have
little or no tangible market value, and easement sales may therefore be
difficult or impossible. Therefore, granting the Service authority
specific to its farmlands similar to that granted under the Southern
Nevada Public Land Management Act, the Small Tracts Act, or the Federal
Property and Administrative Services Act, while useful, would not likely
be sufficient to address most of the Service's unwanted farmlands.

Further, the Service recently developed a proposal to divest some of its
limited-interest refuges in North Dakota.^30 The divestiture proposal was
developed to help ensure that future resources are expended on lands that
support the mission and goals of the refuge system. Under the proposal,
refuge lands were selected for divestiture on the basis of specific
criteria, including how well the lands achieved one or more of the goals
of the National Wildlife Refuge System, whether the lands met their
intended purpose, whether the lands had biological integrity, and whether
it was possible to restore them. This divestiture proposal would require
congressional authorization for the properties recommended for divestment.
The Service could expand this model and adapt the criteria to address
management limitations specific to the farmlands, such as property size,
distance from the refuge office, number of refuge staff available, as well
as ecological considerations. Using such criteria, the Service could
propose to Congress a list of specific farmlands to be divested, along
with written justification outlining the Service's judgments.

^30Limited-interest refuges are lands with a defined refuge boundary but
where the Service does not own the land in fee simple and instead retains
limited easement rights, largely the right to flood or maintain an
artificial lake for promotion of water conservation or wildlife habitat.
The underlying property remains in private ownership.

We encountered a wide array of perspectives with regard to the value the
farmlands bring to the refuge system, the costs of their management, and
what should be done with those lands that are not significantly
contributing to the refuge systems' goals and mission. For instance, some
refuge managers and regional officials we spoke with believed that the
ecological value of the farmlands is not calculable--especially from a
long-term viewpoint, such as 50 years hence--and therefore all the
farmlands should be retained. Some refuge managers spoke of the low-risk
nature of many of the farmlands--lands where threats of violations are
low, where maintenance needs are minimal, and where the undisturbed
natural state maintains ecological value. These types of farmlands do not
require active restoration, and management, namely regular inspections and
contacts with landowners, would take only minimal effort. Other
headquarters, regional, and refuge officials, however, spoke of
operational efficiencies in the face of continually constrained budgets.
These officials were apt to suggest balancing resource constraints against
the ecological value of the farmlands in comparison with the value of
refuge lands; they argued that in this context, it does not make sense to
expend any resources on some of the farmlands. When asked, most refuge
officials were generally able to identify farmlands that they believed
would not have ecological value sufficient for continued inclusion in the
refuge system, even with significantly increased resources or refuge
staff. Some officials were reluctant to suggest that the Service should
attempt to relinquish its management responsibilities on those lands,
however, for fear of opening the door to potential divestiture of other,
more critical refuge lands, which could bring about a significant
ecological loss to the refuge system.

Conclusions

The 1985 Farm Bill and subsequent legislation provided the Service, in
partnership with the Farm Service Agency, a mechanism to protect, as part
of the National Wildlife Refuge System, farmlands that encompassed
wetlands or other important natural resources. The Service is now
responsible for managing a significant number of both conservation
easement and fee-simple farmlands across most of the continental United
States. The Service does not have a complete accounting of all of its
farmlands, however, so the actual scope of its responsibilities is not
certain. As a result, it may be unable to accurately assess its management
needs. Given the farmlands' wide variation in size, location, and other
characteristics, a complete accounting is needed to best determine how
these lands should be managed in the future.

Since its first farmland acquisition more than 20 years ago, the Service
has generally used the tools at its disposal to deal with these lands as
best it could. We agree with the Service that its main priority, and its
limited resources, should be directed at managing and sustaining its core
refuge lands. However, by not conducting at least a minimum level of
management activities on the farmlands it has acquired, including
inspecting the lands on a regular basis, posting the lands, and ensuring
that landowners are aware of restrictions, the Service could ultimately
see the loss or degradation of the resources it was given the
responsibility to protect. Under current law, the Service has very few
alternatives for reducing or removing management responsibilities for
lands received from the Farm Service Agency that are not significantly
contributing to the mission and goals of the National Wildlife Refuge
System. Therefore, we believe the Service needs additional flexibility to
better align the farmlands it is responsible for with the goals and
mission of the refuge system.

Other federal land management agencies have authorities to dispose of
lands found not in their best interest to retain, and these authorities
could serve as a starting point for the possible development of a
legislative solution for farmlands the Service received from the Farm
Service Agency. Existing authorities have largely been applied to
fee-simple lands, and their applicability to easements may be limited,
largely because of the relatively low market value of many of the
easements. Partly for this reason, a variety of approaches may be
warranted and tailored to the specific farmlands that the Service
determines should no longer be included in the refuge system. In any
approach the Service takes to alter its current farmland management
responsibilities, to the extent possible, it should strive to maintain a
conservation focus. For example, the Service could seek authority to
invest the proceeds from the sale of any farmlands into conserving other
lands or to offer farmlands to state agencies or nonprofits for
conservation purposes. Additionally, because any federal land divestment
proposal raises significant concerns, crucial variables to consider
include a time limit to identify potential farmlands for divestment and
narrowly limiting the scope of any divestment proposal to only those
farmlands the Service identifies as unwanted.

Recommendations for Executive Action

To improve the effectiveness and efficiency of the Service's management of
its farmlands, we recommend that the Secretary of the Interior direct the
Director of the Service to take the following two actions:

           o ensure that the Service's records for all of its farmlands are
           accurate and complete by reconciling regional and refuge office
           records and property records to determine which farmlands were
           transferred from the Farm Service Agency, and
           o develop a proposal to Congress seeking the authority for
           additional flexibility with regard to the farmlands the Service
           determines may not be in the best interest to continue to include
           as part of the National Wildlife Refuge System.

Agency Comments

We provided copies of our draft report to the Department of the Interior
and the Department of Agriculture. The Department of the Interior
concurred with our recommendations and agreed to take action on both
recommendations by the end of calendar year 2009. In addition, the
department provided several technical clarifications, which we
incorporated as appropriate. Appendix III contains the Department of the
Interior's comment letter. The Department of Agriculture provided no
comments.

We are sending copies of this report to interested congressional
committees, the Secretary of the Interior, the Secretary of Agriculture,
the Director of the Service, the Administrator of the Farm Service Agency,
and other interested parties. We also will make copies available to others
upon request. In addition, the report will be available at no charge on
the GAO Web site at http://www.gao.gov.

If you or your staff has any questions about this report, please contact
me at (202) 512-3841 or [email protected]. Contact points for our Offices
of Congressional Relations and Public Affairs may be found on the last
page of this report. GAO staff who made major contributions to this report
are listed in appendix IV.

Robin M. Nazzaro
Director, Natural Resources and Environment

Appendix I: Scope and Methodology

To examine the extent of farmland received by the Department of the
Interior's U.S. Fish and Wildlife Service (Service) from the Department of
Agriculture's Farm Service Agency, we interviewed officials at and
reviewed documentation from the Service's headquarters and regional
offices. Specifically, we requested all farmland acquisition data from the
Service's centralized land records system--its database for tracking
lands, including interests in lands, under its control. To assess the
reliability of the data we received from this database, we interviewed
officials most knowledgeable about the collection and management of these
data, and we reviewed relevant data controls. On the basis of our review,
we concluded that the database records were sufficiently reliable for us
to report on what information the database contained. We were informed by
some regional realty officials, however, that the database was likely
incomplete. Consequently, we contacted each regional office to determine
if the offices maintained additional farmland records outside the
centralized database. Three Service regions provided us with additional
farmland records that were not contained in the database. We also visited
and collected additional farmland records from five refuge offices in five
different regions.^1 At each site, we completed a detailed review of the
available files on farmlands and compared this information with the
information we received from the Service's centralized database, regional
offices, or both, as applicable. We selected the sites to visit using the
data we were given from the Service's headquarters and three regional
offices to identify the refuge office responsible for managing the largest
number of farmlands in each Service region. The five refuge offices we
visited were Tishomingo National Wildlife Refuge near Tishomingo, Oklahoma
(region 2); Shiawassee National Wildlife Refuge near Saginaw, Michigan
(region 3); North Mississippi Refuge Complex near Grenada, Mississippi
(region 4); Montezuma National Wildlife Refuge near Seneca Falls, New York
(region 5); and Modoc National Wildlife Refuge near Alturas, California
(region 8). In addition, we also interviewed the Farm Service Agency
officials at its headquarters and four of its state offices to determine
the availability of records it may have maintained on farmlands it
transferred, either in fee simple or through an easement, to the Service.
We found that the Farm Service Agency generally did not maintain organized
historical records of all its transfers to the Service.

^1Early in our review, we also visited a Service field office located in
the Prairie Pothole Region of the Service's region 6. Partly on the basis
of this visit, we decided to omit Prairie Pothole Region farmlands from
this review because these farmlands are managed very differently from
those outside this unique region.

To examine the extent to which the Service is currently managing its
farmlands, we identified and analyzed applicable laws, policies, and
procedures governing the Service's farmland management, and we interviewed
officials at the Service's headquarters, regional offices, and the five
refuge offices we visited. During our site visits, we reviewed the files
for each farmland in detail, including inspection records; records of
landowner contacts; records of completed restoration and maintenance work;
and general information about the farmland, including the easement
restrictions, boundary demarcations, and other legal documentation. During
these site visits, we also saw a sample of farmlands that the refuge
office was responsible for managing and learned what inspection activities
entailed. We also observed the condition of each farmland, whether
easement signs were clearly visible, and whether there were any signs of
easement violations or other problems. We also designed and conducted a
survey, through structured telephone interviews, of refuge managers from
across the United States about the extent to which they are managing their
farmlands. We received a 100 percent response rate from the simple random
sample of 98 farmlands we selected from the population of 1,423 farmland
records that were provided to us by the Service's headquarters and
regional offices. Appendix II discusses our survey methodology in greater
detail.

To examine the Service's alternatives for managing its farmlands, we
interviewed headquarters and regional officials from the Service's Realty
and Refuge Offices in addition to the refuge office managers contacted
through our site visits and telephone survey. We also interviewed the
Department of the Interior's Office of the Solicitor. Further, we analyzed
relevant federal laws and Service policies governing options for managing
lands within the National Wildlife Refuge System, and we reviewed federal
laws governing land management in other federal agencies, including the
Department of the Interior's Bureau of Land Management and the Department
of Agriculture's Forest Service.

We performed our work between September 2006 and July 2007 in accordance
with generally accepted government auditing standards.

Appendix II: Survey of the U.S. Fish and Wildlife Service's Farmland
Management

To further examine the extent to which the Department of the Interior's
U.S. Fish and Wildlife Service (Service) is currently managing the
farmlands it received from the Farm Service Agency, we designed and
conducted a survey, through structured telephone interviews, of refuge
managers from across the United States about the extent to which they are
managing their farmlands. The questions were designed to obtain
information on the number of farmlands each refuge is responsible for and
on the nature and quantity of inspections, violations, landowner contact,
maintenance, and restoration activity that occur on each farmland. In
addition, the structured interviews solicited information on factors that
may hinder the refuge offices' ability to manage their farmlands. The
interview questions focused primarily on management activities that took
place over the 5-year period from 2002 through 2006. We selected a simple
random sample of 106 farmlands from the population of 1,423 farmland
records that were provided to us by the Service's headquarters and
regional offices as of February 22, 2007. After we selected our sample, we
determined that 8 of the farmlands were not under the management
jurisdiction of the Service, so our final sample consisted of 98
farmlands. We conducted structured telephone interviews with the 51 refuge
offices^1 responsible for managing these 98 farmlands, which included both
fee-simple and easement lands located in 24 states across the nation.

We developed a structured interview guide with the assistance of a GAO
survey specialist and pretested it over the telephone with refuge managers
and staff from 17 refuge offices; we then revised the structured interview
guide, as appropriate, on the basis of the pretest results. A second GAO
survey specialist independently reviewed our interview guide to ensure
that the questions followed general principles of survey research. To
conduct our survey, we e-mailed each refuge manager responsible for one or
more of the 98 farmlands in our sample and requested a telephone
interview. We held our structured telephone interviews from March 22,
2007, through May 1, 2007, with the person whom the refuge identified as
most knowledgeable about managing the farmlands in question--usually the
refuge manager--along with other refuge staff familiar with sample
farmlands. At the end of each telephone interview, we requested that the
refuge manager send us all documentation available to support the
information provided in the interview, such as records indicating when
inspections or restoration activities occurred. On receiving the
documentation, we compared it with the information given in the interview.
Reviewing this documentation gave us additional confidence in the data we
collected during our structured telephone interviews. We did not encounter
any significant discrepancies between the information collected during the
interviews and the documentation supplied by refuge officials.

^1In a number of cases, one refuge office managed two or more of the
farmlands in our random sample. When a refuge office was responsible for
more than one farmland in our sample, we asked the same set of questions
for each farmland.

Our interview response rate was 100 percent,^2 and the results can be
projected to the population of documented farmlands.^3 The results cannot
be projected, however, to the additional undocumented farmlands we became
aware of during our audit work because data for these lands were not
available at the time our random sample was chosen and were therefore not
represented in our sample. Percentage estimates based on our sample have
margins of error of +-10 percentage points at the 95 percent confidence
level unless otherwise noted. We also asked a few questions that were not
specific to a sampled farmland but, rather, applied to the refuge office's
management of all its farmlands. Responses to these officewide questions
cannot be projected to a population of wildlife refuges and are not
presented as such. We also analyzed responses to open-ended interview
questions to identify common themes and to provide context for the
responses to our specific questions.

During the telephone survey, we took steps to ensure data quality. Because
of the practicalities of conducting structured telephone interviews,
interviewees' responses may reflect errors; interviewees may misinterpret
a question, for example. We took steps in developing and conducting the
survey to reduce such errors. First, we conducted the interviews in a
standardized manner to ensure that each respondent was asked the same
questions, each with precise wording, and in the same order. Second,
during the interviews, we annotated responses with notes about
respondent's hesitations, inconclusiveness, inconsistencies, or inability
to answer the question directly or succinctly. We reviewed these notes
before data analyses to ensure response accuracy. Third, for some
questions, we asked respondents directly about their level of certainty
regarding a response, for example, whether their response was "precise" or
"approximate." Finally, we validated responses obtained from the
structured interviews with documentation from the refuges. For example, we
verified such items as inspection dates, acreage amounts, and violations
against the interview responses. All survey data were entered into a
statistical analysis program and verified for accuracy. The results were
then summarized and tabulated.

^2Interviews were completed for all 98 of the randomly sampled farmlands.

^3This population consists of all farmlands documented in the Service's
centralized database, as well as additional farmland records provided by
three regions (regions 1, 2, and 3). Later in our audit work, we
identified additional (undocumented) farmlands that were not included in
the Service's centralized database or in records maintained by the regions
and provided to us.

Table 4 summarizes the key questions we are reporting on that we asked
during the structured interviews about the management of each sampled
farmland. We also asked other questions that we do not specifically report
on to supply further context for interviewees' responses. For example,
after asking whether a certain management activity had occurred, such as
an inspection, we asked the respondent during which season and year the
activity took place and to describe the activity.

Table 4: Selected Survey Questions

Refuge-level questions                                                     
How many fee-simple properties is your refuge office responsible for       
managing?                                                                  
How many properties with easements is your refuge office responsible for   
managing?                                                                  
Does your refuge office identify and track land ownership changes on any   
of your Farm Service Agency properties with easements? If so, what process 
does your office use?                                                      
Property-specific questions                                                
General                                                                    
How many sub-tracts does this property consist of?                         
Do you know when the last time you or someone from your refuge staff       
visited the property?                                                      
At the time of its acquisition, was the property acquired in order to      
protect wetlands; a riparian zone; a floodplain; a coastal barrier;        
threatened or endangered species habitat; fish and wildlife habitat of     
local, regional, state, or national importance; an aquifer recharge area;  
an area of high water quality; an area of high scenic value; or any other  
conservation purpose?                                                      
In its current condition, to what extent is this property meeting its      
conservation purpose? Would you say it is meeting this purpose to a very   
great extent, to a great extent, to a moderate extent, to some extent, or  
to little or no extent?                                                    
Maintenance and restoration                                                
At any time during calendar years 2002 through 2006, have you or has       
someone else from your refuge conducted any maintenance or restoration     
work on this property? If yes, what type of maintenance or restoration     
work was conducted on the property?                                        
To the best of your knowledge, has any other restoration work ever been    
completed on this property since it was acquired by the Service? If yes,   
what type of restoration work was conducted on the property?               

Inspecting for violations                                                  
At any time during calendar years 2002 through 2006, have you or has       
someone from your refuge inspected this property for violations?           
Approximately how many times have you or has someone from your refuge      
inspected this property for violations during calendar years 2002 through  
2006?                                                                      
On your most recent inspection, did you identify any violations? If yes,   
what types of violations did you identify?                                 
Factors that hinder management                                             
Would you say that all of the terms (restrictions) of the easement are     
clear, or are there terms that are unclear? If terms are unclear, have     
unclear terms hindered your ability to manage this property?               
Would you say that all of the boundary segments of this property are       
clear, or are there some boundary segments that are not clear? If some     
boundary segments are not clear, have unclear boundaries hindered your     
ability to manage this property?                                           
How many miles is it from the refuge office to this property?              
How much time, in minutes, does it typically take to drive from the refuge 
office to this property?                                                   
Has the time required to travel to this property hindered your ability to  
manage this property?                                                      
Has the level of resources that your refuge office received over the past  
5 years been more than enough, about enough, or less than enough to manage 
this property? If less than enough, has this lack of resources hindered    
your ability to manage this property?                                      
On balance, would you say that over the past 5 years, your office has been 
able to accomplish more than half of the activities necessary to manage    
the property, about half of these activities, or less than half of these   
activities?                                                                

Source: GAO.

In addition, to further analyze our survey responses, we applied certain
statistical transformations to the data, which we then analyzed for
statistically significant patterns. Specifically, using responses about
when the farmlands were inspected with specific farmland characteristics
as discussed below, we created new variables to assess correlations.
During our site visits, interviews, and telephone survey, refuge managers
mentioned several reasons that some farmlands are inspected while other
farmlands are not. These reasons included farmland size; distance between
a farmland and the refuge office; and refuge workload, including the
number of other properties managed by the refuge office. To corroborate
these explanations, we performed a statistical procedure known as logistic
regression analysis. Logistic regression can be used to determine whether
farmlands that are inspected have different characteristics, on average,
from farmlands that are not inspected. We used logistic regression to
assess the size and the significance of the effects of the following three
factors: property size, distance to refuge office, and the number of
properties managed by a refuge.

We estimated the effects of these three different factors on the
likelihood of properties' being inspected, using straightforward
contingency-table methods (i.e., bivariate cross-tabulations) and using
bivariate and multivariate logistic regression models. The three sections
of table 5 show the bivariate relationships between inspection status
(i.e., whether the property had been inspected in the previous 2 years)
and each of the three factors: (1) property size (top section), (2)
distance from the property to the managing refuge (middle section), and
(3) the number of properties managed by the refuge (bottom section). When
we considered each of these relationships independently (i.e., one at a
time), we found that whether a property had been inspected in the
preceding 2 years was significantly associated with each of the three
factors:

           o Larger properties were more likely to have been inspected in the
           preceding 2 years than smaller ones. Although 67 percent of the
           properties larger than 100 acres were inspected, only 45 percent
           of the properties between 11 and 100 acres in size were inspected,
           and only 7 percent of those properties 10 acres or smaller were
           inspected.
           o Properties that were closer to the managing refuge were more apt
           to have been inspected than those farther away. Eighty-four
           percent of the properties closer than 50 miles from the managing
           refuge were inspected, but only 37 percent of the properties
           between 50 and 100 miles from the refuge, and 27 percent of the
           properties farther than 100 miles from the refuge, were inspected.
           o Properties managed by refuges with more properties to manage
           were less likely to have been inspected than properties managed by
           refuges with fewer properties to manage. Sixty-two percent of the
           properties managed by refuges that managed fewer than 20
           properties were inspected, while only 41 percent and 13 percent of
           the properties managed by refuges that managed 20 to 49 and 50
           properties or more, respectively, were inspected.

Table 5: Numbers and Percentages of Properties Inspected and Not Inspected
in the Past 2 Years, by Size, Distance, and Number of Properties, and Odds
and Odds Ratios Derived from Them

                                Inspected in preceding 2 years?
                                                     Observed     Odds ratios 
                                             Odds on     odds         (linear 
                       No     Yes    Total inspected ratios^a        model)^a 
Size                                                                       
10 acres or         14   1(7%) 15(100%)      0.07                          
smaller          (93%)                                                     
11-100 acres   32 (55)  26(45)  58(100)      0.81    11.38            4.00 
Larger than     6 (33)  12(67)  18(100)      2.00     2.47            4.00 
100 acres                                                                  
Total               52 39(43%) 91(100%)                                    
                    (57%)                                                     
Distance                                                                   
Closer than    3 (16%) 16(84%) 19(100%)      5.33                          
50 miles                                                                   
50-100 miles   24 (63)  14(37)  38(100)      0.58     0.11            0.31 
Farther than   24 (73)   9(27)  33(100)      0.38     0.64            0.31 
100 miles                                                                  
Total               51 39(43%) 90(100%)                                    
                    (57%)                                                     
Number of                                                                  
properties                                                                 
Fewer than 20       14 23(62%) 37(100%)      1.64                          
                    (38%)                                                     
20-49          17 (59)  12(41)  29(100)      0.71     0.43            0.32 
50 or more     21 (88)   3(12)  24(100)      0.14     0.20            0.32 
Total               52 38(42%) 90(100%)                                    
                    (58%)                                                     

Source: GAO.

Note: The likelihood-ratio chi-square values testing the hypothesis of
independence in the three subtables for size, distance, and number of
properties are 14.24, 17.90, and 16.08, respectively, with two degrees of
freedom in each case. These values are large enough for us to reject the
hypothesis of independence with greater than 99 percent confidence. The
odds ratio for the linear model implies that there are linear differences
between the categories for acreage, miles, and properties, not that the
effects of actual acreage, miles, or properties are linear.

aThe observed odds ratios, and the odds ratios for the linear model, were
calculated by dividing the odds on being inspected for the second category
by the corresponding odds for the first, and then dividing the odds for
the third category by the odds for the second.

While sizable confidence intervals are associated with each of these
estimates of the percentage of inspected properties (see table 6), all of
these bivariate associations were statistically significant (at the p =
0.05 level or better). Moreover, when we used bivariate logistic
regression models to estimate these associations, we found that each
association was well described by simple models that constrained the
associations to be linear. To understand these models, it is helpful to
consider the three right-hand columns of table 5, which show for each
category of property size, distance from the refuge, and number of
properties managed by the refuge the observed odds on being inspected,
observed odds ratios, and odds ratios under a linear model.

Table 6: Confidence Intervals Associated with the Estimated Percentage of
Properties Inspected, by Size, Distance, and Number of Properties

                                                 95% confidence interval
                           Percentage inspected  Lower limit Upper limit 
Size                                                                  
10 acres or smaller                       7%           0%         32% 
11-100 acres                              45           32          57 
Larger than 100 acres                     67           41          87 
Total                                    43%          33%         53% 
Distance                                                              
Closer than 50 miles                     84%          60%         97% 
50-100 miles                              37           22          54 
Farther than 100 miles                    27           13          46 
Total                                    43%          33%         53% 
Number of properties                                                  
Fewer than 20                            62%          45%         78% 
20-49                                     41           24          61 
50 or more                                13            3          32 
Total                                    42%          32%         52% 

Source: GAO.

The observed odds are obtained by dividing, for each group of properties,
the number (or percentage) of properties inspected by the number (or
percentage) of properties not inspected. For properties 10 acres or
smaller, the odds on being inspected were 1/14 = 0.07, whereas for
properties 11-100 acres in size and larger than 100 acres, the odds on
being inspected were 26/32 = 0.81 and 12/6 = 2.00, respectively. The
observed odds ratios are obtained by dividing the odds for any given
category by the odds for the next lower category. In these data, we find
that properties in the middle acreage category were more likely to have
been inspected than those in the lowest acreage category (by a factor of
0.81/0.07 = 11.38) and that properties in the highest acreage category
were more likely to have been inspected than those in the middle category
(by a factor of 2.00/0.81 = 2.47). Like the percentages already mentioned,
these observed odds and odds ratios, derived directly from the sample
data, involve considerable error and sizable confidence intervals, and a
simple linear model estimating differences among size categories fits
acceptably. Under this model, the odds on being inspected were greater for
properties 11 to 100 acres in size than for properties 10 acres or
smaller, and greater for properties larger than 100 acres than for
properties 11 to 100 acres in size, in both cases by a factor of 4.0.
These ratios are shown in the right-hand column of table 5. This column
also shows that the effect of distance yields a linear odds ratio of 0.31
(properties farther away than 50 to 100 miles were less likely to have
been inspected than properties closer than 50 miles, by a factor of 0.31,
and properties farther away than 100 miles were less likely to have been
inspected than properties 50 to 100 miles away by that same factor). A
similar odds ratio of 0.32 estimates the linear effect of the number of
properties managed by the refuge responsible for each.

We also used multivariate logistic regression models to simultaneously
estimate the effects of these three factors on the likelihood of
properties' being inspected. These models allowed us to isolate the
independent effect of each factor by adjusting for the effects of the
other two factors. We also used robust procedures to account for the lack
of independence across the observations (i.e., properties), that is, the
clustering of properties within refuges. We obtained the following
results:

           o The effect of size remained pronounced and significant even
           after adjusting for the effects of distance and number of other
           properties managed by the responsible refuge. The adjusted odds on
           properties' being inspected were 3.5 times higher for properties
           11 to 100 acres in size than for properties 10 acres or smaller,
           and they were 3.5 times higher for properties larger than 100
           acres than for properties 11 to 100 acres in size. In other words,
           properties between 11 and 100 acres were 3.5 times more likely to
           be inspected than properties 10 acres or smaller. Similarly,
           properties larger than 100 acres were 3.5 times more likely to be
           inspected than properties between 11 and 100 acres.

           o The effect of distance remained pronounced and significant even
           after adjusting for the effects of size and number of properties
           managed. The adjusted odds on properties being inspected were
           lower for properties 50 to 100 miles from the responsible refuge
           than for properties closer than 50 miles, and lower for properties
           farther than 100 miles away than for properties 50 to 100 miles
           away, in both cases by a factor of 0.34. (Because odds ratios are
           symmetric, this result implies that the odds on being inspected
           were about 2.9 times higher for properties closer than 50 miles
           than for properties 50 to 100 miles away, and about 3.4 times
           higher for properties 50 to 100 miles away than for properties
           farther than 100 miles away.)

           o The effect of the number of properties managed was not
           statistically significant when size and distance were taken into
           account. Our sample of properties is small for detecting any but
           fairly sizable effects, however, especially in a multivariate
           context, and the net effect of number of properties, while
           statistically insignificant, is sizable enough to warrant further
           attention (the adjusted linear odds ratio is 0.54).

Appendix III: Comments from the Department of the Interior

Note: Interior's technical comments are not included.

Appendix IV: GAO Contact and Staff Acknowledgments

GAO Contact

Robin M. Nazzaro, (202) 512-3841, [email protected]

Staff Acknowledgments

In addition to the individual named above, Jeffery D. Malcolm, Assistant
Director; Mark A. Braza; Ellen W. Chu; Richard P. Johnson; Alyssa M.
Hundrup; Michael Krafve; Gretchen Snoey; and Arvin Wu make key
contributions to this report. Also contributing to the report were
Antoinette Capaccio, George H. Quinn Jr., Mark Ramage, Jena Y. Sinkfield,
and Douglas M. Sloane.

(360757)

GAO's Mission

The Government Accountability Office, the audit, evaluation, and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of GAO Reports and Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( [22]www.gao.gov ). Each weekday, GAO posts
newly released reports, testimony, and correspondence on its Web site. To
have GAO e-mail you a list of newly posted products every afternoon, go to
[23]www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: Voice: (202) 512-6000
TDD: (202) 512-2537
Fax: (202) 512-6061

To Report Fraud, Waste, and Abuse in Federal Programs

Contact:

Web site: [24]www.gao.gov/fraudnet/fraudnet.htm
E-mail: [25][email protected]
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [26][email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Susan Becker, Acting Manager, [27][email protected] (202) 512-4800 U.S.
Government Accountability Office, 441 G Street NW, Room 7149 Washington,
D.C. 20548

[28]www.gao.gov/cgi-bin/getrpt?GAO-07-1092 .

To view the full product, including the scope
and methodology, click on the link above.

For more information, contact Robin M. Nazzaro at (202) 512-3841 or
[email protected].

Highlights of [29]GAO-07-1092 , a report to the Subcommittee on Interior,
Environment, and Related Agencies, Committee on Appropriations, House of
Representatives

September 2007

U.S. FISH AND WILDLIFE SERVICE

Additional Flexibility Needed to Deal with Farmlands Received from the
Department of Agriculture

Over the past two decades, provisions of the Food Security Act of 1985,
among others, have allowed the Department of Agriculture's Farm Service
Agency in partnership with the Department of the Interior's U.S. Fish and
Wildlife Service (Service) to add farmlands found to have important
resources to the National Wildlife Refuge System. The Farm Service Agency
transferred such farmlands to the Service through outright ownership ("fee
simple") or through conservation easements. Individual farmlands are
managed by the nearest refuge office.

GAO was asked to examine (1) the extent of farmland received by the
Service, (2) the extent to which the Service is currently managing its
farmlands, and (3) alternatives for managing these lands. To answer these
objectives, GAO visited five refuges and surveyed managers responsible for
a random sample of 98 farmlands.

[30]What GAO Recommends

GAO recommends that the Service ensure that its records for all its
farmlands are accurate and complete and that it develop a proposal to
Congress seeking authority for additional flexibility in dealing with
farmlands the Service determines may not be in the best interest of the
National Wildlife Refuge System. In commenting on a draft of this report,
the Department of the Interior concurred with GAO's recommendations.

Since 1986, the Service reports that it has received at least 1,400
conservation easement and fee-simple farmlands covering 132,000 acres, but
the actual number is unknown because the Service's records are incomplete.
Scattered across 38 states, these farmlands range in size from less than 1
acre to more than 2,200 acres; most are smaller than 50 acres. In
addition, GAO identified farmlands that were not reported by the Service
headquarters or regional offices. Therefore, the numbers reported here
represent a conservative estimate of the total acreage received from the
Farm Service Agency.

The Service is generally not managing a majority of its farmlands. In the
past 5 years, only 13 percent have been inspected annually, on average.
The Service is thus not adequately ensuring landowners' compliance with
easement restrictions. GAO observed ongoing easement violations, including
farming encroachment (see photo below). Few refuge offices track changes
in land ownership, and basic maintenance has seldom been completed.
Several factors have hindered the Service's farmland management. First,
refuge officials do not emphasize managing most of the lands because they
do not believe they contribute to the refuges' mission. Second,
uncertainty about the extent or scope of some easements makes management
activity difficult. Third, constrained resources and declining staff
hinder completion of management activities. Nevertheless, GAO found that
farmlands most closely aligned with refuge goals receive considerably more
attention.

The Service possesses limited alternatives for managing its farmlands.
Alternatives include (1) resetting refuge priorities to ensure that
farmlands are given management attention, (2) requesting additional
resources, and (3) paying little or no management attention to the
farmlands. The Service has in most cases chosen the third alternative, and
refuge officials indicated that this approach is unlikely to change.
Because these lands are part of the National Wildlife Refuge System, under
current law the Service cannot dispose of unwanted farmlands, regardless
of their value to the refuge system's mission. Consequently, the Service
may need additional flexibility on a limited and short-term basis to
resolve the issue of unwanted farmlands.

Farming Encroachment in Violation of a Wetland Easement, April 2007

References

Visible links
  21. http://www.gao.gov/cgi-bin/getrpt?GAO-06-511
  22. http://www.gao.gov/
  23. http://www.gao.gov/
  24. http://www.gao.gov/fraudnet/fraudnet.htm
  25. mailto:[email protected]
  26. mailto:[email protected]
  27. mailto:[email protected]
  28. http://www.gao.gov/cgi-bin/getrpt?GAO-07-1092
  29. http://www.gao.gov/cgi-bin/getrpt?GAO-07-1092
*** End of document. ***