To determine the stewardship responsibilities and conservation opportunities for birds on public lands and waters, we overlaid the best available U.S. bird distribution information onto a map of public land ownership to determine the percentage of each species’ distribution on public land. For this report, we focus on those species restricted to a single primary habitat, or habitat obligates. We use the term distribution to describe the breeding and wintering occupancy of each bird species based on our analysis. When reporting the percentage distribution for a group of birds, we use the group average. The term species of conservation concern refers to listings designated by the USFWS Birds of Conservation Concern 2008 and the 2007 WatchList produced by the American Bird Conservancy and Audubon from information compiled by bird conservation partnerships.

With an understanding of the percentage of species’ distribution on public land, we can assess both the degree of protection for each species based on the biodiversity protection category and the responsibility of each public land agency for the future of each species.

Figure 1. This map shows 107,000 unique locations (orange dots) within the contiguous U.S. with eBird data from 2004-09, used in analyses for this report. The 622,000 stationary and traveling counts submitted from these locations constitute the eBird Reference Dataset 2.0, which is available at www.avianknowledge.net . Map courtesy of the Cornell Lab of Ornithology

	Thank You to eBird Volunteers Our understanding of bird distributions has greatly improved thanks to the thousands of bird watchers who have contributed observations to eBird (www.eBird.org). This effort is especially important for tracking seasonal and fine-scale changes in bird distributions, which is not possible with other bird-monitoring programs. However, even this massive observation network provides only imperfect information for assessing the year-round status of birds on many remote public lands across the U.S., including Alaska, Hawaii, and island territories. We urge birders to submit more observations to eBird from public refuges, parks, forests, and wilderness areas. We also urge agencies to support the submission of current and historical records to eBird and other data archives.
	Image: Birders at Montezuma National Wildlife Refuge, New York, by Jessie Barry.

Understanding Bird Distributions 

Most birds are not evenly distributed across their ranges as depicted in field guide maps, and these distributions change throughout the year as birds migrate. To represent the most accurate breeding and wintering distributions of birds in the contiguous 48 states, we analyzed bird observation data from eBird (www.ebird.org), a rapidly growing citizen-science program administered by the Cornell Lab of Ornithology and Audubon.

For this report, National Science Foundation initiatives provided access to resources typically used to analyze large-scale data sets in physics and astronomy research (e.g., 70,000 hours of computer time on TeraGrid). We analyzed more than 600,000 bird checklists collected by eBird participants during 2004–09 at 107,000 unique locations (Figure 1). 

For Alaska bird distributions, we used vegetation layers to modify bird range data from the Alaska Gap Analysis Project and NatureServe. State of Hawaii biologists compiled and analyzed distributions for Hawaiian bird species. Bird distributions for Puerto Rico, U.S. Virgin Islands, Guam, Commonwealth of the Northern Mariana Islands, and American Samoa were based on distribution of suitable habitat identified by local experts. For most ocean species, we used the best available colony-nesting data to evaluate the breeding distribution.

With support from the National Science Foundation and Leon Levy Foundation, collaborators at Oak Ridge National Laboratory, DataONE, TeraGrid, the Institute for Computational Sustainability, and the Cornell Lab of Ornithology used statistical models to account for gaps and biases in volunteer-collected data and to associate bird distributions with important environmental factors, including land cover, elevation, local climate, and human housing density for 139 species with sufficient eBird data.

These models indicated occupancy for approximately 130,000 predicted grid points in the contiguous United States. Cornell Lab experts evaluated the accuracy of predicted occupancy models for each species. See Figure 2 for examples of distribution maps. For the distributions of 156 additional species with very small ranges or associated with wetlands, we used the frequency of each species reported on eBird checklists. These distribution frequency maps provided coarser data and summarized occupancy within 20-square-km blocks. Winter and breeding distributions were analyzed separately for migratory species within the United States. We used best available eBird data to represent the distribution of resident species.

Figure 2. Examples of breeding distributions for obligate species in four habitats. Clockwise, from top left: Kentucky Warbler in eastern forests; Brewer's Sparrow in aridlands; Upland Sandpiper in grasslands; Williamson's Sapsucker in western forests. Maps are based on the predicted occupancy during peak breeding season at roughly 130,000 grid points, modeled using data from eBird and associations with land cover and other key environmental variables. Brighter areas indicate higher probability of occurrence. See additional distribution maps.

Mapping Our Public Lands and Waters

We used the Protected Areas Database of the United States (PAD-US version
1.1) to determine land ownership and biodiversity protection status of all public lands for the continental U.S., Alaska, Hawaii, Puerto Rico, and the U.S. Virgin Islands. PAD-US is a national spatial database created from authoritative data sources by the U.S. Geological Survey’s Gap Analysis Program (USGS-GAP; gapanalysis.usgs.gov).

Our analysis identified lands managed by BLM, DoD, USFWS, USFS, NPS, other federal agencies, and state agencies. PAD-US also classified public lands according to biodiversity protection status. For this report we categorized lands into (1) lands protected to maintain natural habitats; (2) lands managed for multiple uses including conservation; and (3) lands with no permanent protection from development or conversion but that may be managed for conservation. The first category includes lands where natural processes are allowed without interference or are mimicked through management. All lands in the first two categories are protected from permanent conversion to urban or agricultural development. Many public lands in the third category offer some degree of current protection, but are not permanently protected.

To estimate the extent of each primary habitat, we used the USGS-GAP National Land Cover. This dataset is the most detailed, consistent map of vegetation associations available for the United States. The 590 ecological systems and land-use classes were categorized into primary habitat designations for the analysis. These data were then overlaid with PAD-US to calculate the area of each primary habitat on public lands (not including coasts, islands, and oceans).

We considered coastal waters and oceans to be public water areas. Even though these public waters were not mapped, most states have ownership within 3 nautical miles of the coastline, with federal ownership beyond.

In all our analyses, we used the best data available for the United States. These data are valuable and relevant for evaluating broad landscape-level conservation questions, such as those posed here. However, differences may exist between data used for analyses and reported by agencies within the chapters of this report.

USGS-GAP’s Protected Areas Database of the U.S. (PAD-US version 1.1) was used to determine land ownership and biodiversity protection status of all public lands for the continental U.S., Alaska, and Hawai'i.

Determining Stewardship Responsibilities

To calculate the percentage of each species’ distribution on public lands and biodiversity protection categories for the continental U.S., we projected the distribution model or frequency map for each bird species onto PAD-US. For the distribution model results, we calculated percentages at the locations where the model predicted occupancy. 

Because the frequency maps provided coarser data and the occupancy data were summarized within 20-square-km blocks, we projected these data onto public lands and summed over the ownership categories within the blocks to calculate percentage of management responsibilities and biodiversity protection. 

In Alaska and Hawaii, the bird distributions were overlaid with PAD-US to determine the percentage of public land and protection status categories within each species’ distribution. 

For Puerto Rico and U.S. Virgin Islands, bird distributions were overlaid onto PAD-US, whereas for Guam, Commonwealth of the Northern Mariana Islands, and American Samoa, we used a qualitative assessment based on territorial and federal government data for public lands. 

For coastal and marsh species, we used a qualitative assessment rather than a quantitative analysis. For ocean birds, we focused on best available data from breeding colonies to calculate the percentage of the global population occurring on public lands. 

For each primary habitat, we reported the average distribution across multiple obligate species. These percentages measure both the degree of protection for each species on public lands based on biodiversity protection category and the responsibility each public land agency has for the future of each species. 

PAD-US version 1.1 includes significant contributions and large aggregated data sets from BLM, USFS, GreenInfo Network, and The Nature Conservancy. USGS-GAP relies on authoritative sources, such as federal, state, local, and nongovernmental organizations, and land trusts to provide valuable spatial and attribute data to improve and expand PAD-US. We encourage agencies and organizations with protected areas data to contact USGS-GAP (gapanalysis.usgs.gov). PAD-U.S. 1.2, the newest update, is available at gapanalysis.usgs.gov. 

 

  Previous page                                                                  Next page