The North American Breeding Bird Survey (BBS) was established in 1966 in response to a lack of quantitative data on changes in the populations of many bird species at a continental scale, especially songbirds. The BBS now provides the most reliable regional and continental trends and annual indices of abundance available for >500 bird species. This paper reviews some of the ways in which BBS data have contributed to bird conservation in North America over the past 50 yr, and highlights future program enhancement opportunities. BBS data have contributed to the listing of species under the Canadian Species at Risk Act and, in a few cases, have informed species assessments under the U.S. Endangered Species Act. By raising awareness of population changes, the BBS has helped to motivate bird conservation efforts through the creation of Partners in Flight. BBS data have been used to determine priority species and locations for conservation action at regional and national scales through Bird Conservation Region strategies and Joint Ventures. Data from the BBS have provided the quantitative foundation for North American State of the Birds reports, and have informed the public with regard to environmental health through multiple indicators, such as the Canadian Environmental Sustainability Indicators and the U.S. Environmental Protection Agency's Report on the Environment. BBS data have been analyzed with other data (e.g., environmental, land cover, and demographic) to evaluate potential drivers of population change, which have then informed conservation actions. In a few cases, BBS data have contributed to the evaluation of management actions, including informing the management of Mourning Doves (Zenaida macroura), Wood Ducks (Aix sponsa), and Golden Eagles (Aquila chrysaetos). Improving geographic coverage in northern Canada and in Mexico, improving the analytical approaches required to integrate data from other sources and to address variation in detectability, and completing the database, by adding historical bird data at each point count location and pinpointing the current point count locations would further enhance the survey's value.
Effective conservation and management of wildlife populations requires reliable information about their status, how their status is changing over time, and the factors driving those changes (Baillie 1990). Such information is necessary to understand which species are in need of conservation or management action, what actions might be effectively undertaken to achieve conservation, and, if actions are undertaken, whether these actions are effective (Figure 1). This is particularly important in an adaptive management framework, where information on population change contributes iteratively to planning in an effort to reduce uncertainty (Williams 2011). This approach allows practitioners to learn about the system as they manage it and to adjust their management actions or policies as required (Williams 2011).
Well-designed, large-scale monitoring programs are generally the most effective way to assess and detect changes in the status of populations (Stem et al. 2005). Even without a proper statistical design, large changes in populations, such as the rapid disappearance of Brown Pelicans (Pelecanus occidentalis) and Peregrine Falcons (Falco peregrinus) due to DDT (Cade et al. 1971, Blus 1982), or the loss of billions of Passenger Pigeons (Ectopistes migratorius; Blockstein 2002), can be detected by general observation. However, by the time such large changes have occurred, conservation intervention becomes expensive and difficult, and may even be too late, as in the case of the Passenger Pigeon (Blockstein 2002). Quantitative, precise information on population change derived from standardized monitoring allows for more proactive, informed, and defensible conservation measures (Stem et al. 2005). Quantitative data can also be used to evaluate potential drivers of population change, which, in turn, can help to guide conservation actions (e.g., Butler et al. 2007, Conroy et al. 2011; Figure 1).
The North American Breeding Bird Survey (BBS) was established in 1966 by Chandler S. Robbins in the wake of Rachel Carson's book Silent Spring (Carson 1962; see Sauer et al. 2013, 2017a for more information). The creation of the BBS was in response to a lack of reliable large-scale data on changes in the populations of many bird species, especially songbirds (Robbins and Van Velzen 1967). The BBS was one of the first “citizen science” programs; it relies on thousands of highly skilled birders, most of whom are volunteers, to undertake roadside surveys across the North American continent during the peak breeding season (between May and early July, depending on latitude). Each route comprises 50 3-min point count stations (i.e. BBS “stops”) spaced roughly 0.8 km apart, as safety conditions allow, along secondary roads. A single observer identifies and counts all birds seen within 400 m of their stop, or heard at any distance (Robbins et al. 1986). The survey is coordinated at the national level by staff at the U.S. Geological Survey (USGS) Patuxent Wildlife Research Center (Laurel, Maryland, USA) and the Canadian Wildlife Service (CWS) of Environment and Climate Change Canada (Ottawa, Ontario, Canada), and, since 2008, the Mexican National Commission for the Knowledge and Use of Biodiversity (CONABIO; Mexico City, Mexico), with the assistance of provincial, territorial, and state coordinators. In the 50 yr since its inception, the survey has expanded to include information from >5,400 BBS routes covering much of the U.S. and Canada, and portions of northern Mexico (Pardieck et al. 2016), and now provides regional and continental trends and annual indices of abundance for >500 species (Sauer et al. 2017a).
Data from the BBS have been analyzed regularly by CWS and USGS staff to estimate how populations have been changing over time (e.g., Erskine 1978, Robbins et al. 1986, Dunn et al. 2000, Downes and Collins 2003, Pardieck and Sauer 2007, Ziolkowski et al. 2010, Sauer et al. 2014, Environment and Climate Change Canada 2017a). Analytical methods have evolved considerably as new statistical approaches have been developed; BBS analyses have moved from graphed indices of the ratios of area-weighted average counts (Erskine 1978) through route regressions (Link and Sauer 1998) and estimating equations (Link and Sauer 1994) to, most recently, hierarchical Bayesian models (e.g., Link and Sauer 2002, 2016, Sauer and Link 2011, Smith et al. 2014). Published estimates of trends and annual indices of abundance have become the quantitative foundation for landbird conservation in North America (NABCI 2016, Rosenberg et al. 2016, 2017), contributing information to many of the steps required for species conservation (Figure 1). In addition to the status and trend estimates (along with their associated reliability [Environment and Climate Change Canada 2017a] and regional credibility [Sauer et al. 2017b] measures) published by the 2 federal agencies, raw BBS data have been analyzed by researchers who have incorporated these results into hundreds of scientific publications. The topics covered include, but are not limited to, range shifts, responses to climate change, population turnover, ecosystem services, migratory connectivity, habitat and land cover associations, effects of diseases and pesticides, invasive competitors, and impacts of land use change (see the USGS's BBS bibliography for a partial list: https://www.pwrc.usgs.gov/bbs/about/bbsbib.pdf).
In this paper, we review some of the ways in which data from the BBS have contributed to conservation of birds in North America. Specifically, we examine the extent to which the BBS has been used to: (1) identify conservation priorities, including assessing the status of individual species at national and regional levels, identifying species that may be at risk, and identifying priority locations and/or habitats for conservation; (2) inform conservation actions by identifying potential drivers of population change; (3) motivate conservation actions by reporting on ecosystem health through environmental indicators; and (4) evaluate the effectiveness of conservation management actions, including informing harvest management. We also consider some of the factors that have limited the application of the BBS to conservation and how these may be addressed in the future.
Identifying Conservation Priorities
Assessing species population status and conservation priorities. The BBS's rigorous survey design, consistent field methods, volunteer commitment, and continental coverage have made it the most valuable source of status information available for many species. Data from the BBS have been, and continue to be, used in a wide variety of conservation assessment databases at regional to continental scales. Here, we highlight 5 examples of broad-scale assessments, each developed for different purposes, all of which rely to a large degree on BBS data.
Avian Conservation Assessment Database. Formerly known as the Partners in Flight (PIF) Species Assessment Database, the Avian Conservation Assessment Database (ACAD) relies heavily on the BBS. Originally developed by PIF just for landbirds in the U.S. and Canada, the database has been expanded to include all other bird groups (waterfowl, seabirds, shorebirds, and other waterbirds), as well as Mexico and Central America (Rosenberg et al. 2017). The ACAD provides an assessment of the population status for all North American, Mexican, and Central American bird species based on several criteria including distribution, population size, population trend, and threats (Carter et al. 2000, Rosenberg et al. 2017). The BBS has been used to generate continental population trends for 62% (287 of 460) of the landbirds in the database. BBS data have also been used to generate population size scores for 89% (274 of 308) of the landbirds that have the majority of their breeding range within the U.S. and Canada (Rosenberg et al. 2017). The database now provides the quantitative basis for several national conservation plans (e.g., the PIF Landbird Conservation Plan), which contain lists such as the “Common Birds in Steep Decline List” and “PIF Watch List,” as well as various metrics such as an extinction half-life, and PIF's population objectives (Rosenberg et al. 2016, 2017).
NatureServe Network. The NatureServe Network (www.natureserve.org) collects, compiles, analyzes, and disseminates species and ecosystem status assessments for the Western Hemisphere, with the aim of providing a basis for sound and effective conservation action. NatureServe is a nonprofit organization that coordinates a public–private–academic network of programs operating in the U.S., Canada, and Latin America. One of the network's many products is conservation status assessments, which estimate the risk of extinction and extirpation of species and ecosystems, respectively, at global, national, and subnational levels. While the ACAD approach was developed specifically for birds, the NatureServe criteria are designed to work for all taxa, including fungi, plants, and animals. The “conservation status ranks” are calculated based on 10 ranking factors, which are grouped into the following 3 categories: rarity, threats, and trends (Faber-Langendoen et al. 2012). The first 2 factors are scaled and weighted relative to their effect on the risk of extinction, and an initial score is created. That score is then modified by adding or subtracting the trends factor, which results in a final rank on a 1–5 scale. These rankings are then translated into status descriptions ranging from “secure” to “critically imperiled” (Faber-Langendoen et al. 2012). As of 2017, BBS trends had been considered in the most recent review of the conservation status rank of ∼600 avian species or subspecies that occur in North America (B. Young, Director of Species Science, personal communication).
General Status of Species in Canada report. Canada's Species at Risk Act requires the preparation of a “general report on the status of wildlife species” (Minister of Justice 2015: section 128) every 5 yr. All governments in Canada made a commitment to prepare such a report under the 1996 Accord for the Protection of Species at Risk. For the Wild Species 2015 report (CESCC 2016), provincial, territorial, and federal governments adopted the NatureServe protocol for their assessments of plant and animal species at national and subnational levels in Canada. These ranks were then integrated into the NatureServe Network. Species assessments were based on range, abundance (i.e. rarity), environmental specificity, threats, and short- and long-term population trends. Nonmigratory species were assigned only one rank, whereas migratory species received separate ranks for breeding, nonbreeding, and migration periods. The use of BBS trends varied among regions, but contributed information to more than half of all breeding bird species assessments in the regions for which we were able to obtain information. For example, in the Canadian prairies, BBS trends informed the short-term trend metric for 66% (174 of 264) of species breeding in Alberta, 60% (150 of 252 species) in Saskatchewan, and 60% (159 of 267 species) in Manitoba (E. Beck, CWS Biologist, personal communication). In British Columbia, BBS trends were used for 61% of species (180 of 295 breeding bird species; A. Norris, CWS Biologist, personal communication), while in Quebec, they informed 50% (140 of 278 breeding bird species; S. Légaré, CWS Biologist, personal communication).
Status of birds in the United States. Under the Fish and Wildlife Conservation Act (Public Law 100-653, 102 Stat. 3825), the U.S. Fish and Wildlife Service (USFWS) is required to “identify species, subspecies, and populations of all migratory nongame birds that, without additional conservation actions, are likely to become candidates for listing under the Endangered Species Act” (16 USC 2912, Sec. 13 [a]). This “Birds of Conservation Concern” list (USFWS 2008) provides the primary motivation for conducting species assessments in the U.S. The USFWS uses a tapestry of partnerships and programs formed around 4 major bird groups (landbirds, shorebirds, waterbirds, and waterfowl; USFWS 2004) to assess avian populations at the national level. These partnerships, programs, and initiatives (e.g., PIF, U.S. Shorebird Conservation Partnership, Waterbird Conservation for the Americas, and North American Waterfowl Management Plan) provide population information for focal species at varying intervals and scales that then feed into national species assessments (e.g., the Birds of Conservation Concern list). Work is underway to develop a unified national assessment process based on a single standardized database that would follow the ACAD model (Rosenberg et al. 2017, R. Dettmers, USFWS Biologist, personal communication).
Currently, reports for the 4 bird groups are created separately, and all have made use of the BBS to varying degrees. For example, the primary data source for landbird assessments is the ACAD, which relies heavily on the BBS (PIFSC 2012). The most recent shorebird assessment (USSCPP 2016) relied on BBS data and trends for 14 of the 52 (27%) shorebird species that were evaluated. The most recent evaluation of waterbird populations for the USFWS Birds of Conservation Concern report (currently in review) used BBS data for 7 of the 106 (7%) waterbird species evaluated (B. Andres, USFWS Biologist, personal communication). However, with one exception (Wood Duck [Aix sponsa]; USFWS 2016a), waterfowl population assessments have not used BBS data, relying instead on data derived from aerial surveys, banding, and harvest surveys conducted by the USFWS and North American Waterfowl Management Plan partners.
Status of birds in Canada. The “Status of Birds in Canada” website (http://ec.gc.ca/soc-sbc/) was developed by Environment and Climate Change Canada to guide Canadian conservation planning. This web-based database informs management agencies by identifying and tracking changes in the national status of 452 bird species that regularly breed or occur in Canada and by providing assessments and detailed trend information that may be used to flag candidate species for listing (Environment Canada 2014). The website presents summary information on the status of each species, including an evaluation of the reliability of each assessment, along with the underlying data. For 217 species (48%), including 17 waterbirds and shorebirds, trends and annual indices from the BBS in Canada were the primary or sole sources of information used to determine population status. For another 24 species (6%), BBS results were used to supplement other sources of information, such as the Christmas Bird Count, Breeding Bird Atlases, or species-specific surveys. All told, the BBS was used to inform the population status assessment of more than half (53%) of all bird species in Canada, and 88% of all landbirds. In the assessment of the reliability of each species' population status (ranked as high, medium, or low reliability, or as data deficient), 66% of the 149 species with highly reliable status assessments were landbirds whose assessments relied on the BBS.
Informing legal protection for species at risk. In addition to species status assessments, BBS data have also informed the listing process for individual species under both the U.S. Endangered Species Act (ESA) and Canada's Species at Risk Act (SARA). In Canada, the listing process includes criteria specific to the magnitude of population change, and so some species have been listed as a direct result of their BBS trend estimates. As noted above, trend estimates derived from the BBS have also contributed to the listing process indirectly by highlighting species in decline that may be candidates for future assessment under the ESA and/or SARA.
The U.S. Endangered Species Act. The ESA was the first piece of legislation to identify, protect, and recover imperiled species from extinction (Waples et al. 2013). The listing of terrestrial and freshwater species is determined by USFWS managers, and is based solely on the best available scientific information (i.e. economic and social factors are not taken into consideration). A species may be listed as either “endangered” or “threatened” if the species is at risk due to one of the following 5 categories of issue: habitat destruction or damage, overuse, disease or predation, inadequate protection from existing regulatory mechanisms, or other natural or manmade factors that endanger the species' existence (Waples et al. 2013). We reviewed the bird species listed under the ESA and searched for documents containing the words “Breeding Bird Survey” in the U.S. Federal Register under 50 CFR Part 17 – Endangered and Threatened Wildlife and Plants. We also queried the listed bird species on the USFWS website (https://www.fws.gov).
Of the 16 species and 24 subspecies of bird listed as “endangered” or “threatened” (as of November 2016) that breed in the continental U.S. (USFWS's Environmental Conservation Online System, https://ecos.fws.gov/ecp/), we found that BBS trends informed the assessments of 5 species or subspecies. Most listed species were extremely rare, with small populations or small breeding ranges that were not adequately detected by the BBS. Thus, it is hardly surprising that we were unable to find a single assessment that indicated that data from the BBS were pivotal in the decision to list the species. However, we found that trend information at the species level was used to inform 2 species assessments, although neither species was found to meet the criteria for listing (Cerulean Warbler [Setophaga cerulea]: USFWS 2006, and Mountain Plover [Charadrius montanus]: USFWS 2011). BBS trend information at a species, guild, and/or national level was also used to inform evaluations when insufficient evidence was available for assessment at the subspecies or regional level (e.g., Streaked Horned Lark [Eremophila alpestris strigata]: USFWS 2013, and western population of the Yellow-billed Cuckoo [Coccyzus americanus]: USFWS 2014). Finally, “declining detections” of parasitic Brown-headed Cowbirds (Molothrus ater) on BBS routes in areas of overlap with their at-risk host, the Black-capped Vireo (Vireo atricapilla), was cited as one of several reasons behind a delisting proposal for the vireo (USFWS 2016b). Although our search was restricted to the Federal Register, and thus dependent on cited references in these published entries, data from the BBS may play a more important role than our results imply. USFWS biologists evaluate more data than are cited in the 12-month petition findings for “not warranted” species or in the proposed and final rules for listed species. These sources, which include the BBS, are still informative in the overall assessment process, depending on the species (K. Gifford, ESA Listing Coordinator, personal communication).
The Canadian Species at Risk Act. In Canada, candidate species at risk are assessed by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC), an independent scientific advisory body. COSEWIC reviews status reports for candidate species and determines whether species meet the criteria for listing as “endangered,” “threatened,” “special concern,” or “not at risk” under the SARA. COSEWIC's designation is then submitted to the Minister of the Environment, who considers the scientific assessment in conjunction with any political, social, or economic factors. We reviewed all COSEWIC status reports for birds (including those for species or populations assessed as “not at risk”) and identified those that used BBS data as a primary or supplementary source of information.
We found that data from the BBS were considered in 57% (65 reports) of the 114 status reports available from 1978 to April 2016 for birds. Of these, 40% (26 reports) used BBS trends or indices as the basis for the designation, including reports for 18 species or subspecies that were assessed as either “endangered” or “threatened,” 4 that were assessed as “special concern,” and 4 as “not at risk.” Of these 26 reports, all but 1 (Black Tern [Chlidonias niger]) are for landbirds. The remaining 39 reports used BBS trend information in some capacity, often supplementing BBS trends with other sources of information (e.g., Christmas Bird Count, species-specific surveys, Breeding Bird Atlases), or using BBS trends to assess the likelihood of a rescue effect from the United States (based in part on trends from throughout North America). The BBS was not used for 43% of reports; these included reports for waterbirds and raptors that are not well monitored by the BBS, as well as reports for a few passerines that are extremely rare in Canada and therefore not detected often enough by the BBS to estimate trends.
Informing regional conservation and land use planning. BBS data have often been used to develop species distribution models, which can inform conservation actions by identifying important habitats and areas for conservation (e.g., Lipsey et al. 2015), to calculate regional stewardship metrics (e.g., Carter et al. 2000), and to inform environmental assessments and land use planning, which most often occur at regional or local levels. Here, we highlight some examples of the ways in which such BBS products have been used at the biome scale (e.g., Bird Conservation Regions [BCRs], Migratory Bird Joint Ventures [MBJVs], and the Boreal Avian Modelling [BAM] project), as well as at the state and provincial or territorial scale (e.g., State Wildlife Action Plans [SWAPs], and Breeding Bird Atlases).
Bird Conservation Region strategies. BBS trend estimates have contributed to Bird Conservation Region (BCR) planning and conservation strategies. As part of the North American Bird Conservation Initiative (NABCI), Canada, the U.S., and Mexico have committed to developing conservation plans for all birds at the scale of BCRs (Schmidt et al. 1998). In Canada, conservation strategies have identified priority species within BCRs to focus management attention and resources where they are most needed, to determine conservation needs and establish measurable objectives, and to make recommendations for conservation actions to reach these objectives. Priority species have been identified based on their vulnerability and population status using a quantitative assessment of population size, distribution, trend, threats, and regional abundance (Kennedy et al. 2012). In these assessments, the population trends for 393 of 544 species (234 landbirds, 76 waterbirds, 44 waterfowl, and 39 shorebird species) were based on the BBS for at least 1 subregion (intersection of BCRs and provinces or territories; 32 in total; CWS 2014. In total, for all species and subregions that indicated a data source, 28% (1,138 of 4,133) used BBS data, and another 19% (805 of 4,133) listed the PIF Species Assessment Database (2005 data version) as the main data source (CWS 2014). Also, the late 1960s was chosen as the baseline (target) population level for landbirds, in part because this time period coincides with the start of the BBS and thus provides a means to track progress and evaluate the effectiveness of conservation actions implemented under the Canadian BCR strategies (Kennedy et al. 2012).
Migratory Bird Joint Ventures. In the U.S., conservation implementation and planning are done through Migratory Bird Joint Ventures (MBJVs). MBJVs are large-scale, regional partnerships that focus on habitat conservation for all bird species across North America. These programs have incorporated the few U.S.-based BCR strategies that were developed (these strategies used BBS trends within a PIF-style assessment framework to determine regional conservation priorities specific to each BCR; R. Dettmers, USFWS Biologist, personal communication). The idea behind MBJVs was first outlined in the North American Waterfowl Management Plan in 1986 (USFWS and Environment Canada 1986). There are now 22 habitat-based and 3 species-based MBJVs (see http://mbjv.org for more information). The extent to which BBS data have been used or incorporated into the various MBJVs has varied, but, generally, BBS trends have been used to identify priority species and develop conservation plans for a given joint venture, while BBS abundance indices have been used to inform relative abundance and distribution maps (e.g., Upper Mississippi River and Great Lakes Region Joint Venture; Potter et al. 2007). For example, by integrating BBS data and land cover characteristics, Lipsey et al. (2015) created species distribution models for the Sprague's Pipit (Anthus spragueii). The resulting maps, and others (e.g., Niemuth et al. 2005, 2007), have been used to inform and direct wetland and grassland acquisition within the Prairie Pothole Joint Venture (e.g., http://ppjv.org/science/spatial-planning-tools, http://ppjv.org/science/projects/breeding-bird-survey-data-to-develop-species-distribution-models).
Boreal Avian Modelling project. The Boreal Avian Modelling project (BAM; Cumming et al. 2010) has also made use of data at the level of individual BBS point count stations (BBS “stops”) to model the distribution and relative abundance of birds throughout the boreal forests of North America. BAM was established to bring together point count survey data from a wide range of sources to inform the management and conservation of boreal birds, as well as to forecast impacts of human land uses (Cumming et al. 2010, Barker et al. 2015, 2016). As of March 2016, BAM's BBS dataset included data from >65,000 locations, which, over time, represent >600,000 BBS point counts (including all Canadian and Alaskan BBS routes, as well as some routes from the contiguous U.S.; Barker et al. 2016). The database also contains information from >250,000 other point counts, spanning 135 different projects (Barker et al. 2016). Considerable effort has been undertaken to develop statistical models to integrate point count data collected using diverse protocols, including listening times ranging from 3 to 20 min, observations with or without distance information, and a mixture of roadside and off-road counts (Matsuoka et al. 2012, Sólymos et al. 2013, Barker et al. 2015). These models have been used to estimate the distribution and abundance of 98 landbird species in relation to current habitat and climatic variables (Cumming et al. 2014), as well as to simulate the responses of 80 species to potential changes in climate across North America (Stralberg et al. 2015, 2016). Mahon et al. (2014) also simulated cumulative effects of climate and land use changes on songbirds in boreal Alberta, Canada. Map outputs from these models are available on BAM's website (http://www.borealbirds.ca/). Additional applications of the data include ongoing work to identify priority areas for conservation of select landbirds in Canada's boreal forest (A. Camfield, CWS Biologist, personal communication).
State Wildlife Action Plans. Beginning in 2005, all U.S. states were required by congressional mandate to develop a State Wildlife Action Plan (SWAP) to be eligible for a portion of the US$50 million of federal funding available through the State and Tribal Wildlife Grants Program (AFWA 2012, Public Law 106-291: www.gpo.gov/fdsys/pkg/PLAW-106publ291). SWAPs identify the species and habitats in greatest conservation need. Many of these SWAPs cite survey-wide and state-wide BBS trends in their species assessments, and/or incorporate BBS data through a PIF assessment framework to evaluate and prioritize landbird conservation actions (e.g., SCDNR 2005, NMDGF 2006, KDFWR 2013).
Breeding Bird Atlases. Breeding Bird Atlases, which have been undertaken in many provinces and states in North America, are specifically designed to provide information on the distribution and relative abundance of bird species at a moderate spatial scale (typically 5 × 5 km or 10 × 10 km square grids; Beck et al. 2017). This information is used in a multitude of ways (reviewed by Gibbons et al. 2007), but, in terms of environmental assessment, atlas data are mostly used to determine species presence and to provide a regional context against which site-level data can be compared (Beck et al. 2017). Maps from some of the more recent atlases have helped to identify species-rich areas that have then been flagged as priorities for acquisition and restoration or enhancement by the Pacific Birds Habitat Joint Venture and the National Wetland Conservation Fund (K. Moore, CWS Conservation Planner, personal communication).
While breeding bird atlases primarily focus on collection of new data during the atlas period, numerous state and provincial atlases have incorporated data from the BBS. We scanned the methods summaries from a selection of published atlases, and asked subscribers to the North American Ornithological Atlas Committee list-serve ( http://www.bsc-eoc.org/norac/index.jsp?targetpg=listserv&lang=EN) to provide examples of how their atlases used BBS data. Many atlases have included state-wide and/or survey-wide BBS trend information in the species accounts, especially to provide context for changes in distributions between atlases (e.g., Kleen et al. 2004, Ellison 2010, Davidson et al. 2015). BBS data have commonly contributed to the creation of distribution maps; species observed on BBS routes are assigned a breeding evidence code (usually “possible”) within the atlas grid squares where they were detected (e.g., Cadman et al. 2007, Schneider et al. 2010, Rodewald et al. 2016). Presence data have also contributed to “probability of observation” maps, which provide an index of abundance with which comparisons between atlases can be made (e.g., Davidson et al. 2015, Stewart et al. 2015). Finally, some atlases have used BBS data to map relative abundance at the stratum level within the state (e.g., Busby and Zimmerman 2001), and at the grid square level by combining BBS data with miniroutes that followed BBS protocols but were shorter so as to remain within an atlas grid cell (e.g., Jacobs and Wilson 1997). To date, BBS point count data have not been integrated with atlas-specific point counts to estimate the relative abundance of species among grid cells, largely because BBS point counts are 3 min long and atlas point counts are 5 min long. However, plans for the upcoming atlas in Saskatchewan, Canada, include dividing atlas point counts into 3- and 2-min intervals to allow for the possibility of incorporating BBS data into relative abundance maps (K. Drake, Saskatchewan Atlas Coordinator, personal communication). Some atlases are also asking their observers to capture GPS coordinates for each of their BBS stops so that their checklists can be tied to specific locations on the landscape (A. Peele, Virginia Atlas Coordinator, and N. Anich, Wisconsin Atlas Coordinator, personal communications).
Evaluating Causes of Population Change
The factors that drive population change must be understood if conservation and management actions are to be effective, particularly when these factors are driving population declines. Population trends, such as those derived from the BBS, do not, by themselves, indicate why populations are changing. However, trends can be used to evaluate potential drivers of population change if combined with covariates or if incorporated into adaptive management models (Sauer et al. 2013) or integrated population models (Schaub and Abadi 2011).
Even for a topic as focused as applied avian conservation, an exhaustive review of the uses of BBS data is beyond the scope of this paper. The USGS's BBS bibliography database ( http://www.pwrc.usgs.gov/rwp/database_descriptions.htm#Breeding%20Bird%20Survey) includes 578 entries from 1965 to 2014. A Google Scholar search on July 20, 2016, for peer-reviewed papers with the wording “North American Breeding Bird Survey” between 2014 and 2016 provided 96 additional papers. We selected examples of different approaches from this extensive literature.
Some early analyses laid the groundwork by inferring causes of change based on differences in trends among geographic regions or habitats (e.g., Robbins et al. 1989, Flather and Sauer 1996) or between groups of species (e.g., Sauer and Droege 1992). More recently, analyses have generally used 1 of 3 approaches to combine BBS data with other data sources to evaluate drivers of population change at different scales: integration with demographic data, integration with environmental data on the breeding grounds, or integration with data throughout the annual life cycle.
Integration with demographic data. BBS data have been linked with demographic data from the Monitoring Avian Productivity and Survivorship (MAPS) program to evaluate which demographic rates influence population change on the breeding grounds. For example, DeSante et al. (2005) modeled spatial variation in MAPS data as a function of spatial variation in BBS trends to investigate the relative influence of productivity and survival rates on population change. They concluded that low adult survival was the proximate cause of decline at both continental and regional scales for the Gray Catbird (Dumetella carolinensis). Saracco et al. (2008) used MAPS data and reverse-time capture–recapture models to evaluate the effects of recruitment and adult immigration on adult survival of Yellow Warblers (Setophaga petechia). They found that apparent adult survival, not productivity, likely drove variation in the species' BBS population trends. More recently, Ahrestani et al. (2017) formalized this union of data by creating and testing an integrated population model that combined BBS and MAPS data for the Gray Catbird and Wood Thrush (Hylocichla mustelina). These analyses have not only provided insights into the demographic processes driving observed changes in abundance in different strata, which have helped to focus local conservation measures on the most appropriate period of the life cycle; they have also potentially improved the robustness of parameter estimation and provided the ability to estimate latent parameters that are not measured by