Attachment 9 Northern Goshawk Inv and Monitory Tech Guide.pdf

Northern Goshawk Inventory and Monitoring Technical Guide

United States Department of Agriculture

Forest Service

Gen. Tech.

Report WO-71

July 2006

Medicine Lake

Mt. Shasta Intermountain Great Basin

Cascade Sierra

West Coast

Colorado Plateau and SW Mtns

Northern Goshawk Inventory and Monitoring Technical Guide

Brian Woodbridge and Christina D. Hargis

Northern Goshawk Inventory and Monitoring Design Team

Brian Woodbridge Christina D. Hargis Richard T. Reynolds

James A. Baldwin Gregory D. Hayward Kimberly Titus

Alan Franklin Sarah R. Dewey Christopher W. Schultz

Alan L. Williamson Douglas A. Boyce, Jr. John J. Keane

U.S. Department of Agriculture, Forest Service, Washington Office

Ecosystem Management Coordination Staff

Watershed, Fish, Wildlife, Air, and Rare Plants Staff

United States Department of Agriculture

Forest Service

Gen. Tech.

Report WO-71

July 2006

Proper citation for this document is as follows:

Woodbridge, B.; Hargis, C.D. 2006. Northern goshawk inventory and monitoring technical guide. Gen.

Tech. Rep. WO-71. Washington, DC: U.S. Department of Agriculture, Forest Service. 80 p.

Cover Photo: The concept of bioregional monitoring is conveyed through three photos superimposed on a digital elevation model of the Western United States, including portions of the Pacific Coast and Intermountain Great Basin bioregions. The overlaid images depict three levels of the bioregional monitoring design: a sample of contiguous PSUs in northern California (top), a PSU with call point transect lines (middle), and a northern goshawk nest (bottom). Photo credit: Brian Woodbridge. Composite image designed by Dave LaPlante.

The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis of race, color, national origin, age, disability, and where applicable, sex, marital status, familial status, parental status, religion, sexual orientation, genetic information, political beliefs, reprisal, or because all or part of an individual’s income is derived from any public assistance program. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means for communication of program information (Braille, large print, audiotape, etc.) should contact USDA’s

TARGET Center at (202) 720-2600 (voice and TDD). To file a complaint of discrimination, write USDA, Director, Office of Civil Rights, 1400 Independence Avenue, S.W., Washington, D.C. 20250-9410, or call

(800) 795-3272 (voice) or (202) 720-6382 (TDD). USDA is an equal opportunity provider and employer.

Northern Goshawk Inventory and Monitoring Technical Guide iii

Acknowledgments

We gratefully acknowledge the previous work of several individuals in the realm of goshawk monitoring; their ideas, field work, and publications are the basis of this technical guide. In particular, we acknowledge the contributions made by S.R. Dewey, S.M. Joy, J.J. Keane, P.L. Kennedy, V. Penteriani, R.T. Reynolds, and D.W. Stahlecker.

The bioregional monitoring design presented in chapter 2 was created by the North-ern Goshawk Inventory and Monitoring Design Team, whose members are listed on the title page of this technical guide. We give special recognition to J.A. Baldwin for contributing substantial time toward developing the bioregional design and preparing all the statistical text in chapter 2. We are grateful to D. LaPlante and B. Allison for spatial analyses of primary sampling unit (PSU) size and for preparing figures and to

J. Wilson and H. Wang for preparing the figure in Appendix C. We thank the follow-ing individuals who substantially improved the quality of this technical guide through their review of earlier versions: D.E. Andersen, P.H. Geissler, T.A. Max, A.R. Olson, M.G. Raphael, L.F. Ruggiero, H.T. Schreuder, and J.R. Squires.

Authors

Brian Woodbridge is supervisor of the Forest Resources Branch, U.S. Fish & Wildlife

Service, Yreka, CA.

Christina D. Hargis (name changed to Christina D. Vojta) is a wildlife ecologist, Watershed, Fish, Wildlife, Air, and Rare Plants Staff, USDA Forest Service, Washington Office, Washington, DC.

iv Northern Goshawk Inventory and Monitoring Technical Guide

Northern Goshawk Inventory and Monitoring Technical Guide v

Contents

Acknowledgments ......................................................................................iii

Contents .........................................................................................................v

Chapter 1. Overview .................................................................................1-1

1.1 Overview ..................................................................................................... 1-1

1.2 Background and Business Needs ................................................................ 1-1

1.3 Key Concepts ............................................................................................... 1-3

1.4 Roles and Responsibilities ........................................................................... 1-5

1.4.1 National Responsibilities ..........................................................1-5

1.4.2 Regional Responsibilities ..........................................................1-5

1.4.3 Forest Responsibilities ..............................................................1-5

1.5 Relationships to Other Federal Inventory and Monitoring Programs ......... 1-6

1.5.1 Forest Service Programs ...........................................................1-6

1.5.2 Programs in Other Federal Agencies .........................................1-6

1.6 Quality Control and Assurance .................................................................... 1-7

1.7 Change Management ................................................................................... 1-7

Chapter 2. Bioregional Monitoring Design .........................................2-1

2.1 Objective ...................................................................................................... 2-1

2.2 Planning and Design .................................................................................... 2-3

2.2.1 Goshawk Natural History Relevant to the Bioregional

Sampling Design ................................................................................ 2-3

2.2.2 Description and Rationale for Monitoring Design ............................ 2-5

2.3 Data Collection .......................................................................................... 2-14

2.3.1 Data Collection Methods and Rationale .......................................... 2-14

2.3.2 Quality Control/Quality Assurance .................................................. 2-17

2.3.3 Data Entry Forms ............................................................................. 2-19

2.3.4 Survey Logistics ............................................................................... 2-19

2.4 Data Storage and Management .................................................................. 2-20

2.5 Data Analysis ............................................................................................. 2-21

2.5.1 Estimating the Bioregional Frequency of Occurrence

of Goshawks .................................................................................... 2-21

2.5.2 Assessing Changes in Goshawk Frequency of Occurrence

Over Time ........................................................................................ 2-22

2.5.3 Evaluating Change in Occupancy Rate in Relation to Change in Habitat or Other Environmental Variables ................................... 2-23 vi Northern Goshawk Inventory and Monitoring Technical Guide

2.6 Reporting ................................................................................................... 2-24

2.6.1 Expected Reports ............................................................................. 2-24

2.6.2 Reporting Schedule .......................................................................... 2-24

Chapter 3. Goshawk Survey Techniques ............................................... 3-1

3.1 Objectives .................................................................................................... 3-1

3.2 Planning and Design .................................................................................... 3-1

3.2.1 Aspects of Goshawk Natural History Related to Survey

Methodology ..........................................................................................3-1

3.2.2 Sampling Designs .............................................................................. 3-5

3.3 Data Collection ............................................................................................ 3-6

3.3.1 Survey Methods ................................................................................. 3-6

3.3.2 Quality Control/Quality Assurance .................................................. 3-18

3.4. Data Storage ............................................................................................. 3-19

3.5. Data Analysis and Interpretation of Survey Results ................................. 3-19

3.5.1 Presence .......................................................................................... 3-20

3.5.2 Occupancy ....................................................................................... 3-20

3.5.3 Breeding .......................................................................................... 3-21

3.5.4 Successful Nest ............................................................................... 3-22

3.5.5 Fledging Rate .................................................................................. 3-22

3.6 Survey Applications .................................................................................. 3-23

3.6.1 Territory Monitoring Application..................................................... 3-23

3.6.2 Small Area Survey Application ........................................................ 3-25

3.6.3 Large Area Survey Application ........................................................ 3-26

3.7 Reporting ................................................................................................... 3-27

Appendix A. Literature Cited ...................................................................... A-1

Appendix B. Interactive Spreadsheet for Determining Bioregional Sample Size .....................................................B-1

Appendix C. Sample PSU Map ...................................................................C-1

Appendix D. Guidelines for Constructing Field Data Collection Forms ................................................................... D-1

Northern Goshawk Inventory and Monitoring Technical Guide 1-1

Chapter 1. Overview

1.1 Overview

This technical guide provides information on all aspects of inventory and monitoring related to the northern goshawk (Accipiter gentilis) and is to be used by the U.S. De-partment of Agriculture (USDA) Forest Service consistent with national direction, lo-cal priorities, and available funding, and also by interested partners and collaborators.

When the protocols described in this technical guide are implemented, the resulting data will meet standards of the Data Quality Act and, therefore, will be legally and scientifically defensible and consistent with data collected elsewhere using the same protocols.

The technical guide is divided into three chapters: an overview, a bioregional monitoring design, and a description of inventory and survey methodologies. The technical guide was written for bioregional monitoring coordinators and their survey teams, biologists at forest and district levels, and any other agencies and organizations interested in northern goshawk inventory and monitoring activities.

This introductory chapter provides an overview of the technical guide and describes the business needs that motivate the USDA Forest Service to inventory and monitor goshawks. This chapter also describes the roles and responsibilities of implementing this technical guide and provides the context of goshawk monitoring in relation to other Federal inventory and monitoring programs.

1.2 Background and Business Needs

The northern goshawk has attracted substantial interest over the past two decades because management activities in forest environments have the potential to affect nesting habitat and, hence, population levels of this species. Goshawks tend to nest in mature forests (conifer in the West, deciduous in the East), building large nests that are used by the original pair or successors for many years (Squires and Reynolds

1997). A variety of forest types and structural stages are used as foraging habitat, but the important role of mature forests as long-term nesting sites has placed consider-able attention on the goshawk.

The goshawk has been designated a sensitive species in six of the eight USDA

Forest Service administrative regions within its geographic range. Because of sensi-tive species status, 71 national forests are required by USDA Forest Service policy

(Forest Service Manual [FSM] 2670 and 2672) (USDA Forest Service 1995, USDA

Forest Service 1991) to evaluate the effects of proposed management actions on

1-2 Northern Goshawk Inventory and Monitoring Technical Guide goshawks and document the findings in a biological evaluation that is specific to each proposed action. Any decisions made by a line officer “must not result in loss of species viability or create significant trends toward Federal listing” (FSM 2670.32).

Forest supervisors are given the responsibility to “determine distribution, status, and trend of threatened, endangered, proposed, and sensitive species and their habitat on

Forest lands” (FSM 2670.45). Regional foresters are to identify sensitive species that qualify for conservation agreements (FSM 2672.12).

In addition to sensitive species status, 53 national forests (as of 2004) have designated the goshawk as a “management indicator species” (MIS) in their land and resource management plans developed under the National Forest Management Act.

The combined designation of the goshawk as both a sensitive species and an MIS has resulted in a need for information on the status and trend of goshawk populations and habitats throughout its range. The broad geographic distribution of the goshawk has resulted in a need for greater consistency in how this information is collected.

The goshawk has also received attention from members of the public. Environ-mental organizations submitted petitions in 1991 (Babbitt et al. 1991, Silver et al.

1991) and in 1997 (USFWS 1998a) to list the northern goshawk as threatened or endangered in the Western United States. The U.S. Fish & Wildlife Service (USFWS) concluded that listing was not warranted, based on the best available information

(USFWS 1996, 1998a).The status review team that assembled information for this finding, however, noted that information was not cohesive and they made several recommendations for acquiring more information on population and habitat trends.

One of the recommendations was that “land managers should improve inventory and monitoring of goshawk populations. Improvements should include a standardized protocol to conduct goshawk surveys.” (USFWS 1998b).

The Queen Charlotte goshawk, a recognized subspecies occurring in southeast

Alaska, was petitioned for listing in 1994. The USFWS concluded that listing was not warranted (USFWS 1997), but interest remains high regarding conservation of this subspecies.

The northern goshawk is also protected under the Migratory Bird Treaty

Act. Executive Order 13186 of 2001 clarified responsibilities of Federal agencies regarding migratory bird conservation, and these responsibilities include inventory and monitoring.

In summary, the USDA Forest Service needs information on status and trends of northern goshawk populations and habitats for the following reasons:

• The goshawk is a sensitive species. Forest Service Manual (FSM) 2670.45

(USDA Forest Service 1995) requires forest supervisors to collect information on sensitive species in order to determine when change in management is warranted.

• Habitat and population information is needed by national forests that have designated the goshawk as an MIS.

Northern Goshawk Inventory and Monitoring Technical Guide 1-3

• The USFWS may receive new petitions to list the goshawk and will call on the

USDA Forest Service again for information on status and trends of populations and habitats.

• Many public entities, including environmental groups and forest product industries, will continue to ask the USDA Forest Service for information on the status of goshawks on National Forest System lands, because this species, along with mature forests, remains a topic of interest.

Most national forests have partial information on goshawk territories and suitable habitat, and some national forests also have multiyear data on goshawk nest activ-ity. Standardized field protocol for goshawk nest area surveys have been published

(Dewey et al. 2003, Joy et al. 1994, Kennedy and Stahlecker 1993, Penteriani 1999) and USDA Forest Service biologists frequently use them. Inventory and monitoring data, however, usually are not comparable across forests because of different defini-tions for nest and territory occupancy, different levels of survey efforts, and different definitions of habitat. Furthermore, a lack of sampling design, either within a given forest or across administrative units, precludes the ability to evaluate trends in either populations or habitats. Consequently, most existing information is limited to the spa-tial occurrence of nests and a rough estimate of territory size and distribution.

To obtain consistent, reliable information on the status and trend of goshawk populations and habitats, USDA Forest Service biologists, research scientists, and members of the academic community identified a need for the following:

• Bioregional population monitoring in relation to habitat changes.

• Forest-level monitoring of the local effects of management actions.

• Inventory and survey standards that are based on published field protocol.

This technical guide was developed to fulfill these information needs.

1.3 Key Concepts

The term “protocol” is often used to refer to standards for collecting field data. The

Inventory and Monitoring Issue Team of the USDA Forest Service has recommended a broader interpretation of protocols to include all aspects of an inventory or monitoring plan: sampling design, data collection, data analysis, and reporting. This technical guide follows this recommendation and includes all of these topics in the chapters that follow. The term “procedures” is used for describing steps within a specific pro-tocol. For example, the procedures for establishing a sampling frame and setting up strata are specific steps in the sampling design protocol.

Two key concepts related to goshawk monitoring is the notion that goshawks maintain territories and that territoriality influences, in part, the spacing of goshawk

1-4 Northern Goshawk Inventory and Monitoring Technical Guide breeding activity and the use of resources. Ecologically, the term “territory” is usually defined as an area that is defended, but since defensive behavior is rarely observed, another definition is “any exclusive area” (Ricklefs 1979). The area used by goshawks for nesting and fledging of young is exclusive and is therefore a territory.

Average territory size is estimated after the breeding history in an area has been established for many years. This estimation is done by determining the location of alternate nests associated with each territory, finding the geometric centroid of each cluster, and calculating the distance between clusters. A territory is said to be occupied if adult goshawks are present, but additional criteria for determining territory occupancy can be found under subheading B of section 3.5. Within an occupied territory, any given nest site can be active or inactive, depending on whether a nesting effort is currently in progress. The terms “active” and “inactive” refer to nest site status, whereas the terms “occupied” and “unoccupied” refer to territory status.

The bioregional monitoring design in chapter 2 is not intended to track goshawk territories, nor does it depend on previous knowledge of territory location to be implemented. The sampling design, however, is based on the concept of territories in order to sample at a scale that is appropriate for this species.

An important distinction exists between goshawk presence, as defined in the bioregional monitoring design, and territory occupancy. Goshawk presence can be determined by the detection of an individual, whereas territory occupancy requires that two or more criteria are met (section 3.5). In most wildlife studies, however, the term “site occupancy” is used to indicate simple presence of one or more individuals in an area (Geissler and Fuller 1987). Because of the more rigorous definition of occupancy in the context of goshawk territory status, this technical guide uses

“presence” as the variable of interest in the bioregional monitoring design rather than

“site occupancy.” Specific criteria for determining goshawk presence are in chapter 2 under the bioregional monitoring design, whereas criteria for territory occupancy and nest activity are in chapter 3, section 3.5.

A third key concept of goshawk monitoring, particularly for the bioregional monitoring design, is presence/absence. In the recent past, biologists referred to presence/absence as present/not detected, because absence cannot be absolutely determined. This term, however, confuses the state of being present or not present with the activity of either detecting or not detecting an organism. This technical guide adopts the term presence/absence with the argument that although absence cannot be determined, it can be estimated statistically using a known or estimated detection probability. More details are presented in chapter 2, but it is introduced here as a key concept of the bioregional monitoring design.

Northern Goshawk Inventory and Monitoring Technical Guide 1-5

1.4 Roles and Responsibilities

1.4.1 National Responsibilities

• Develop the Northern Goshawk Inventory and Monitoring Technical Guide and update as needed.

• Ensure that data standards, data fields, and data analysis capabilities are built into the National Resource Information System (NRIS) Fauna application in order to carry out goshawk inventory and monitoring activities.

• Ensure flow of information with USFWS as directed under Executive Order

13186.

• Encourage cooperative efforts and partnerships with other agencies and organizations to monitor goshawks.

• Assist USDA Forest Service administrative regions with attaining adequate funding for bioregional monitoring.

1.4.2 Regional Responsibilities

In cooperation with other regional offices—

• Identify a bioregional monitoring coordinator to oversee the bioregional monitoring program. Ensure coordination of data collection, analysis, and reporting with adjacent administrative regions that share the same bioregion.

• Work across administrative lines to maintain, to the extent possible, the bioregional boundaries as identified in this technical guide.

• Work cooperatively with States and nongovernmental organizations to disseminate and share information regarding goshawk population and habitat trends.

• Identify current status and future funding needs of the bioregional goshawk monitoring program in the region’s Inventory and Monitoring Program Plan.

• Provide training to field personnel as needed for data collection, storage, analysis, and reporting.

• Ensure that monitoring results are distributed to participating national forests and other monitoring collaborators in a timely fashion.

1.4.3 Forest Responsibilities

• Contribute to the bioregional monitoring program either indirectly through funding, or with field personnel and equipment.

• Conduct area inventories and project surveys using the survey protocols described in chapter 3.

• Provide stewardship of the NRIS Fauna module for forest and district-level goshawk data.

1-6 Northern Goshawk Inventory and Monitoring Technical Guide

1.5 Relationships to Other Federal Inventory and Monitoring

Programs

1.5.1 Forest Service Programs

Before the development of this technical guide, the USDA Forest Service did not have a national protocol for northern goshawk inventories or monitoring. Monitoring has been the responsibility of individual national forests, with guidance from region-al offices. In addition, monitoring has been conducted as part of established research programs on several national forests, often funded by the national forests as adminis-trative studies. Long-term monitoring and/or in-depth research studies have occurred on the Klamath, Modoc, Beaverhead/Deerlodge, Inyo, Tahoe, Sawtooth, Targhee, Tongass, Kaibab, and Dixie National Forests. Recently, national forests in Utah, in collaboration with Brigham Young University, have undertaken several studies that address goshawk dispersal and movements (Rodriguez 2004).

The Rocky Mountain Research Station has an ongoing research study of north-ern goshawks that has been in place on the Kaibab National Forest since 1991 (Reyn-olds and Joy 1998). This study has provided knowledge of goshawk life history, reproductive patterns, and detection rates that were instrumental in formulating the bioregional monitoring design (Reynolds 2002). The study also contributed to the design of a standard survey protocol that has been adopted in this technical guide

(Joy et al. 1994).

The goshawk bioregional design described in this technical guide relies on broad scale habitat information from the Forest Inventory and Analysis (FIA) program in order to look for correlations between broad scale habitat characteristics and gos-hawk populations. Each bioregion can also use data derived from the Common Stand

Exam protocol to evaluate habitat changes.

The USDA Forest Service developed a technical guide for the Multiple Species

Inventory and Monitoring (MSIM) protocol. The MSIM provides a framework for collecting presence/absence data on a variety of terrestrial vertebrate species, includ-ing raptors, over broad spatial extents. The goshawk bioregional monitoring design is complementary to the MSIM because it has a similar monitoring objective and obtains data at a similar spatial scale.

1.5.2 Programs in Other Federal Agencies

The U.S. Geological Survey Patuxent Wildlife Research Center has spearheaded a continent-wide Breeding Bird Survey (BBS) since 1966 (Robbins et al. 1986, Sauer et al. 2001). Although trend data are available from the BBS for many bird species, the research center has concluded that data on northern goshawks are not sufficient for determining trends, either survey-wide or for any individual state or province, Northern Goshawk Inventory and Monitoring Technical Guide 1-7 due to low detections of goshawks per survey route and low numbers of survey routes with goshawk detections. Most States and provinces have fewer than five survey routes with goshawk detections (Sauer et al. 2001). There are no other Federal pro-grams for collecting data on northern goshawk populations or habitats.

1.6 Quality Control and Assurance

The inventory and monitoring protocols described in this technical guide are based on published field protocols (Joy et al. 1994, Kennedy and Stahlecker 1993). The bioregional monitoring design was designed by researchers and statisticians with substantial knowledge of goshawk ecology and principles of sampling design.

The technical guide was reviewed by six qualified professionals, including four statisticians both within and outside the USDA Forest Service. The bioregional monitoring design underwent a separate peer-review and has been published (Hargis and Woodbridge 2006).

Quality control and assurance for implementing the bioregional monitoring design is discussed under the heading of Data Collection in chapter 2.

1.7 Change Management

This technical guide is considered a draft until the bioregional monitoring design has been implemented for at least 1 year in at least one bioregion. After the first year, the design team anticipates several changes in the technical guide. The description of creating a sampling frame for bioregional monitoring is currently sketchy, and more details will be added based on the first bioregion’s experience. The Data Storage section will be expanded to describe in detail the structure of the bioregional database and the data fields that will be routinely migrated to NRIS Fauna. Guidelines for constructing a field data entry form might be revised for better efficiency and/or clarity. The Data Analysis section will be augmented with analytical tools available either on a CD or a Web site.

This technical guide will be reviewed 5 years after the first year’s revision to determine if additional changes are warranted. Population monitoring at the bioregional scale will likely remain unchanged, but more details on FIA data and landscape habitat variables might be added. An additional chapter on nesting-effort monitoring might also be added to provide a method for national forests to quantitatively evaluate changes in goshawk breeding efforts over time.

1-8 Northern Goshawk Inventory and Monitoring Technical Guide

Northern Goshawk Inventory and Monitoring Technical Guide 2-1

2.1 Objective

Information is needed on the status and trend of northern goshawk populations and their habitats to meet a variety of information needs described in chapter 1.

This chapter describes a monitoring design that will enable the Forest Service and collaborating partners to (1) estimate the frequency of occurrence of territorial adult goshawks over large geographic areas; (2) assess changes in frequency of occurrence over time; and (3) determine whether changes in frequency of occurrence, if any, are associated with changes in habitat. The goal is to monitor goshawks annually into the long-term future, with analyses of change every 5 years.

Goshawk populations experience some level of change in abundance from year to year due to changes in a combination of environmental factors, most notably climate and prey abundance. The USDA Forest Service is specifically interested in population changes that exceed normal fluctuations and that may be due to management-induced habitat changes. The range of normal fluctuations in goshawk abundance is currently not known, nor is the exact magnitude of change that can be detected with monitoring. The monitoring design described in this chapter, however, is intended to be used with a sample size sufficient to detect a 20-percent change in relative abundance over a 5-year monitoring period. If a 20-percent decline were observed within a bioregion, this percentage would represent a trigger point for assessing whether an immediate change in land management within that bioregion was warranted.

Given the mobility of goshawks and the wide range of forest types they use, it is difficult, if not impossible, to define discrete breeding populations. The USDA Forest

Service has chosen to monitor goshawks within fairly large geographic areas that are referred to here as bioregions. Generally speaking, a bioregion is a large spatial extent defined by coarse scale similarity in ecological conditions. Descriptions of bioregions and rationale for boundaries are described under section 2.2.2 of Planning and Design.

The indicator used to determine the frequency of occurrence of goshawks is P, the proportion of primary sampling units (PSUs) (Levy and Lemeshow 1999) with goshawk presence, or in other words, the frequency of presence. A PSU is a square sampling unit of 600 ha. The sampling frame for each bioregion consists of a grid of

PSUs laid over all potential goshawk nesting and forested foraging habitat, both on

USDA Forest Service lands and on lands of all collaborators in the bioregion. Each bioregion will estimate P from a stratified random sample of PSUs, using a sample

Chapter 2. Bioregional Monitoring Design

2-2 Northern Goshawk Inventory and Monitoring Technical Guide size that is sufficient for attaining an estimate that is within 10 percent of the actual frequency 90 percent of the time.

The ability to detect changes in frequency of presence is currently unknown, be-cause it will depend on the persistence of goshawks from one year to the next in each individual PSU that is sampled. An examination of the data after 2 years will allow for an estimation of the amount of change that can be detected with specified power.

To look for possible correlations between changes in population and habitat, two types of habitat data will be used: (1) Forest Inventory and Analysis (FIA) data summarized for all FIA points in the sampling frame, and (2) landscape pattern data for all PSUs, obtained from Geographic Information System (GIS) analysis.

Selected habitat variables from both sources will be used as covariates in a logistic model to evaluate relationships between goshawk presence and habitat. These habitat variables also will be included in models to evaluate change in P between 2 or more years. Habitat variables will likely differ between bioregions, reflecting geographic differences in goshawk habitat relations.

Each bioregion will identify a bioregional coordinator to oversee the goshawk monitoring program. The coordinator can be affiliated with any agency, research facility, or university, either under salary or under a contract. The bioregional coor-dinator will work with biologists and biometricians to establish the sampling frame, determine sampling intensity, select relevant habitat variables, provide training for field personnel, oversee data collection, analyze the data, and prepare annual reports.

This monitoring design was created for the USDA Forest Service, but collaboration with other agencies and land owners is strongly encouraged, because the larger sample size attained through collaboration will yield better estimates of status and change over time. Once a monitoring program is in place, however, adding

PSUs is not recommended, due to the difficulty in evaluating year-to-year differences when the sampling frame has changed. Later collaboration is possible, but the added

PSUs would need to be evaluated separately.

The following paragraphs describe recognized limitations of the bioregional monitoring program. First, although the boundaries of each bioregion include all land ownerships, the sampling frame is composed of only national forest lands and the lands of any monitoring collaborators. Therefore, inferences from goshawk population trends are applicable only to the lands within the sampling frame and not to the entire bioregion.

Second, this monitoring design does not provide a means of estimating total population size in a bioregion. A goshawk detection may or may not represent the presence of a breeding pair, and further efforts beyond this sampling design would be needed to establish the location of active nests. Furthermore, this monitoring program is not designed to provide information on nesting efforts or reproductive success.

Northern Goshawk Inventory and Monitoring Technical Guide 2-3

Finally, this design could be limited in the ability to detect population trends if either the precision of each annual estimate is low or if the specific PSUs with goshawk presence change every year. In either case, it might be difficult to detect small but potentially meaningful trends in P. The power to detect change will not be known until 2 or 3 years of data are available.

2.2 Planning and Design

2.2.1 Goshawk Natural History Relevant to the Bioregional Sampling Design

The northern goshawk is a wide-ranging forest raptor found in boreal and temperate forests of the Holarctic zone. In North America, two subspecies are recognized by the American Ornithologists’ Union (1957): Accipiter gentilis laingi, which occurs along the insular coast from Vancouver Island north to Icy Strait and Lynn Canal in

Alaska, and A. g. atricapillus, which occurs throughout the rest of the species range.

Goshawks in southern Arizona and the mountains of central Mexico have been proposed as a third subspecies, but this proposal is currently under debate (Whaley and White 1994). Differences among the subspecies represent subtle breaks in clinal variation (Squires and Reynolds 1997).

Goshawks use a variety of forest types for nesting and foraging. Across the entire breeding range, goshawks nest in a broad range of vegetative communities, from extensive mature coniferous forest in coastal regions to small patches of aspen and pine in Great Basin shrubsteppe communities (Squires and Reynolds 1997). Within their home ranges, goshawks use a diverse array of habitats for foraging, both in terms of vegetation type and the degree of openness (Squires and Reynolds 1997). At the scale of nest-site selection, goshawks nest in the densest stands available, given the capability of the forest type; high canopy closure also appears to be an important habitat characteristic for the species (Hayward and Escano 1989). The size of forest patches used for nest areas appears to be highly variable across the species’ range.

Goshawk habitat selection theoretically follows a modified model of ideal free distribution that is limited by territorial behavior (Fretwell 1972). Under the ideal free distribution model, individuals choose to occupy the best habitats first and will settle into secondary habitat only when competition for resources in primary habitat outweighs the lesser availability of resources in secondary habitat. When a species is territorial, the presence of dominant individuals forces greater use of secondary habitats even before resources become limiting. It appears that goshawks follow this model because high-quality habitats contain a fairly fixed number of territories

(Reynolds and Joy 1998), and no evidence supports the idea that increases in prey result in increased density of breeding pairs in these habitats.

2-4 Northern Goshawk Inventory and Monitoring Technical Guide

Where forest habitats are continuous, the spacing between the nests of breeding pairs is fairly regular. On the Kaibab Plateau of Arizona, mean nearest-neighbor distance for 103 nesting pairs was 3.9 km (SD = 0.32) (Reynolds and Joy 2006). On the Klamath National Forest, the distance was 3.3 km ± 0.3 SE for 59 nesting pairs

(Woodbridge and Detrich 1994), and similar spacing was observed in northeastern

California on the Modoc National Forest (Woodbridge 1998).

Within territories, goshawks typically shift their breeding sites among several alternate nests up to 1.8 km apart (Squires and Reynolds 1997, Woodbridge and Det-rich 1994). Although most alternate nests are grouped within a forest stand or cluster of adjacent forest stands, a search radius of 1 km is required to locate 95 percent of alternate nests used over a period of several years (Reynolds et al. 2005).

It is important to understand how territoriality and habitat quality can influence the ability to detect changes in goshawk abundance over time. During a population increase, goshawk density in high-quality habitats would remain fairly constant due to territorial spacing. More goshawks might be detected in these habitats, but this increase would be due to the presence of “floaters” rather than to an increased density of territories. In contrast, we predict that goshawk numbers would increase in habitats of secondary or marginal quality as surplus individuals who are unable to find vacant territories in high-quality habitats establish new territories. Nevertheless, overall density might be lower than that of high-quality habitats due to limitations in the availability of nesting stand structure or to prey resources.

During periods of population declines, it is likely that marginal habitats would be the first to show a drop in numbers, with prime territories remaining fairly constant.

A decline in goshawk abundance in high-quality habitats is likely to represent either a dramatic overall population decline or a decline in the quality of the primary habitat itself, either through changes in nesting site availability or food resources. Nesting site availability could decline as a result of succession, climate, or management actions. Food resources could decline from a combination of changes in habitat, food, predators, competitors, disease, or weather.

Because of the different population responses that are expected in habitats of different quality, it is important that all potential habitats are included in a monitoring design, not simply the high-quality habitat.

Northern Goshawk Inventory and Monitoring Technical Guide 2-5

2.2.2 Description and Rationale for Monitoring Design

This section describes the bioregions’ boundaries and provides guidance for establishing the sampling frame within a bioregion. Each subsection consists of the

Procedure, which is the protocol for carrying out this bioregional monitoring design, and the Rationale, which describes the scientific basis for the protocol.

Delineation of Bioregions

Procedure

The ecological basis for the bioregions is the Forest Service National Hierarchical

Framework of Ecological Units (Bailey 1980, McNab and Avers 1994), overlaid with the geographic range of the northern goshawk (Squires and Reynolds 1997).

By aggregating neighboring polygons of similar adjacent ecological provinces, the boundaries of 10 goshawk bioregions were delineated: 8 in the coterminous

United States and 2 in Alaska (table 2.1, figure 2.1). If a relatively small polygon of one ecological province was enclosed within a larger polygon of a different ecological province, it was included in the bioregion of the larger province (figure

2.2). Boundaries were also influenced by the configuration of national forests, so that no national forest would be split between two bioregions. Exceptions to this rule occurred with the Toiyabe and Inyo National Forests, both of which occur in the

Cascade-Sierra and Intermountain Great Basin bioregions (figure 2.3). These national forests will have separate data from each of the two bioregions.

Table 2.1. Goshawk bioregions.

Goshawk bioregion Area (km2)

Pacific 121,590

Cascade Sierra 1,181,072

Intermountain Great Basin 620,861

Northern Rocky Mountains/Blue Mountains 480,028

Central Rocky Mountains 317,891

Colorado Plateau and Southwestern Mountains 514,700

Great Lakes 490,500

Northeast and Central Appalachian Mountains 517,225

Coastal Alaskan Forests 173,700

Interior Alaskan Forests 697,545

2-6 Northern Goshawk Inventory and Monitoring Technical Guide

Figure 2.1. Goshawk bioregions.

Figure 2.2. A polygon of an ecological province (342) that is embedded in the Northern Rockies bioregion is included in that bioregion to smooth bioregional boundaries.

Northern Goshawk Inventory and Monitoring Technical Guide 2-7

Figure 2.3. The Toiyabe and Inyo National Forests both are split between two bioregions.

Rationale

The bioregional scale was selected for monitoring northern goshawk populations and habitats after considering two other possible scales: each individual national forest and the entire range of the northern goshawk. The national forest scale was considered too small for both ecological and sampling reasons. Goshawks using a specific national forest are not isolated from goshawks on adjacent forests and other neighboring lands, so “population” trends for a given forest might not be meaningful.

Also, because of the inherent variability in population estimates, the sample size required to detect a significant change in abundance at the national forest scale would be unaffordable for most national forests.

The entire range of the goshawk was considered too large for aggregating and interpreting population and habitat data due to potentially wide variation in goshawk habitat relations across the species’ range. Climatic, physiographic, and ecological factors influencing goshawk populations differ among bioregions, potentially resulting in dissimilar population abundance and trends. Bioregions represent the largest scale at which data should be aggregated, analyzed, and summarized. Range-wide trends will be estimated through a composite analysis of the bioregional trends.

2-8 Northern Goshawk Inventory and Monitoring Technical Guide

The 10 bioregions delineated for the purpose of broad-scale population monitoring are arbitrary, with movements between bioregions likely and, in some cases, documented. For example, goshawks from the Kaibab Plateau in the Colorado

Plateau bioregion have been recaptured on the Dixie National Forest, in the

Intermountain Great Basin bioregion (Reynolds 2002). The delineation of bioregional boundaries, however, is primarily based on differences in ecological conditions that could affect goshawk status and trend.

The bioregions are truncated at the Canadian border (with the possible exception of binational collaboration in the Great Lakes bioregion), and the artificial nature of these boundaries is acknowledged. Transnational movement of goshawks will be considered when population trends are reported for the four bioregions that border

Canada.

Design of Primary Sampling Units

Procedure

The PSUs are a grid of squares, each approximately 600 ha, and are spatially oriented to nest within the grid framework used by the FIA program. The FIA grid is a coast-to-coast coverage of 2,402.7 ha hexagons that was established to conduct forest inventories on Federal and private lands. Ideally, the grid of PSUs is created by starting with the FIA grid and nesting the goshawk grid within it. If the bioregional coordinator cannot obtain access to the FIA grid, the PSU grid can be established from a randomly selected x, y coordinate within the bioregion. Although the orientation of the FIA grid in relation to the PSU coverage will not be known, the relationship will be roughly four PSUs for each FIA hexagon.

Rationale

The size and shape of the PSU reflect ecological factors and sampling considerations.

Although each PSU is not intended to represent a goshawk territory, the size is approximately the same as a territory so that a detection in a PSU will roughly correspond to one breeding pair. If the PSU is too small, several adjacent PSUs could constitute one territory, and sampling in each would not be independent. If the PSU is too large, more than one nesting pair could be present, yielding an underestimation of goshawk occurrence. The PSU size will—

• Maximize the probability of a PSU containing one territory.

• Reduce the probability of a PSU containing more than one territory.

• Be logistically feasible and cost-effective to survey.

To determine optimal size, we compared the spacing of goshawk territories in three geographical areas. Mean nearest-neighbor distances among goshawk territories

Northern Goshawk Inventory and Monitoring Technical Guide 2-9 on the Kaibab Plateau (AZ), Southern Cascades Mountains (CA), and Modoc Plateau

(CA) are remarkably similar, ranging from 3 to 4 km. One-half of this distance, a radius of 1.5 to 2 km, yields an area of 761 to 1,017 ha. Using this information, we created a range of potential PSU sizes from 405 to 1,214 ha, at 202.3 ha increments, and overlaid them in a GIS with several maps of goshawk territories at known density and spacing. The results indicated that, while the proportion of PSUs with goshawk territories increased with increasing plot size, PSUs with multiple territories began to appear when PSU size was at 607 ha (table 2.2). Thus, PSU sizes above 607 ha might underestimate goshawk relative abundance, because two territories rather than just one could be present in a PSU.

Another consideration in PSU size was the desire to make the sampling design compatible with the framework developed by the FIA program. FIA grid cells are too large for adequate goshawk sampling, but a “densified” grid, in which the number of grid points is doubled, falls within the range of goshawk PSU sizes that was determined through interterritory distance. A densified grid is already used by the FIA program in the States within the Great Lakes goshawk bioregion and is also used by USDA Forest Service Region 1, which includes part of the Northern Rockies goshawk bioregion.

The Multiple Species Inventory and Monitoring Protocol is also designed around the FIA grid and was tested in California on a densified grid. This protocol uses several detection methods to survey for a variety of terrestrial species, many of which are prey of goshawks. There are benefits in designing the goshawk sampling frame to be compatible with that used for potential prey.

Plot shape may influence survey results through edge effects. Compared with linear or rectangular plots, circular or square PSUs provide the lowest perimeter to area ratio, thereby decreasing the potential for miscounting individuals occurring near the boundary of a PSU (Krebs 1989). A square plot also facilitates the use of line transects for individual call points.

Table 2.2. PSU size in relation to the number of goshawk territory core areas (cores) within it.

PSU size Number of Percentage of PSUs

(ha) PSUs 0 cores 1 core 2 cores

405 429 85.3 14.7 0.0

607 292 78.8 20.9 0.3

809 229 73.4 25.8 0.9

1,011 182 67.6 30.2 2.2

1,214 158 64.6 31.6 3.2

PSU = primary sampling unit.

2-10 Northern Goshawk Inventory and Monitoring Technical Guide

Establishing the Sampling Frame

Procedure

In order for a grid cell to become a PSU in the bioregional sampling frame, it must meet two criteria: (1) some portion of the grid cell must contain potential goshawk breeding habitat (for nest sites or forested foraging), and (2) most of the grid cell must consist of USDA Forest Service lands or lands of monitoring cooperators.

Each grid cell receives a numeric address and is classified as goshawk habitat or nonhabitat, based on a GIS classification from remotely sensed data. The rule set used to classify goshawk habitat versus nonhabitat will be developed separately by each bioregion.