The Second Breeding Bird Atlas of Vermont (2003-2007)

Study Extent

The atlas uses a stratified random sampling design. Each USGS 7.5-minute topographic quadrangle that occurs (including partially) in Vermont was divided evenly into six 5 km X 5 km blocks. For each quadrangle, blocks are randomly selected to be surveyed.

Sampling

In the first atlas (1976-1981), one block in each quadrangle in Vermont was randomly selected to be surveyed—these were called “priority blocks.” A total of 179 blocks were selected to be surveyed in the first atlas. In the second atlas, these blocks were called “Priority 1” blocks, and for each quadrangle, a second block was also randomly selected, and these were surveyed in order to double the land area covered; these are called “Priority 2” blocks. Although the data collected from Priority 2 blocks cannot be used to assess changes since the first atlas, data from twice as many blocks will be available to assess bird distribution changes between the second and third atlases. All other blocks are called “non-priority.” Priority 2 blocks were selected in such a way as to maximize the number of blocks that could be resurveyed, while maintaining a stratified design. First, we randomly selected a block from the five non-priority blocks in each of the 179 quadrangles used in the first atlas. Two quadrangles (Juniper Island and Lyme) did not contain enough land mass within Vermont to add a Priority 2 block in addition to the already existing Priority 1 block. If the randomly selected block had less than 69 species recorded during the first Atlas, and a different non-priority block had at least 69 species recorded (some non-priority blocks were surveyed during the first Atlas), we instead used the latter block (with one exception, in the Springfield quadrangle). In the first atlas, some quadrangles that did not fall entirely within Vermont were not included in the randomized block selection process, resulting in some areas near state borders that were not surveyed. In the second atlas, blocks were added in these borderline areas as Priority 2 blocks. The blocks added were Charlotte 5 and 6, Gilman 1 and 5, Port Henry 4 and 5, Stratford 2 and 3, and Woodsville 1. A total of 186 Priority 2 blocks were added in the second atlas: one in each of 177 of the 179 quadrangles used in the first atlas, plus nine blocks added in quadrangles along state borders. The total number of Priority 1 and 2 blocks were 179 + 186 = 365. The remaining 872 non-priority blocks in the state were not selected to be surveyed, although some were opportunistically surveyed and those data are also included in the database. Field work for the first and second atlas was carried out during 1976 - 1981 and 2003 – 2007, respectively. Each block was surveyed for 1 to 5 years, depending on the amount of effort expended on a block in a given year. Blocks without a dedicated volunteer were blockbusted; that is, surveyed intensively during a short period, usually by at least two people per block. Blockbusting was carried out by both volunteers and paid field technicians. In addition to observations collected during formal atlasing efforts, incidental observations were recorded and were especially encouraged for rare species in non-priority and priority blocks. The protocol for surveying a block followed that of other atlases and was in accordance with the general principles outlined in the North American Atlas Committee (NORAC) guidelines. Atlasers conducted extensive area searches of each block in all of the habitats represented. They recorded all bird species detected within safe dates and any breeding evidence observed for each bird species. Evidence of breeding was categorized as “Confirmed”, “Probable”, or “Possible”, and within each of these categories were one or more specific codes that best described an observation. With each visit to a block, atlasers recorded date, start and end time, number of people atlasing, and total party hours for each date. This provided a total number of hours spent on each block in each year and for all years of the Atlas. Data from other bird survey programs, blockbusting, and incidental observations were all coded differently from regular atlasing. Atlasers also recorded the first date on which they obtained a valid observation for each species, and the first date on which they confirmed breeding. A minimum number of species and confirmations were required to declare a block “completed.” This approach ensures a standard, minimum amount of coverage and survey effort on blocks. Observers were required to survey in as many different habitat types as possible on the block. Using a minimum number of species, while imprecise, is preferable to using the amount of time spent on a block as a measure of effort. The latter is problematic because the amount of time needed to survey a block adequately can vary depending on factors such as topography, habitat complexity and diversity, accessibility of habitats, times periods when surveying occurs, and especially skill level and motivation of the observer. In the first atlas, surveying on a block was considered complete when at least 75 percent of the species expected to occur in the block were found, and evidence of nesting was confirmed for at least half of those species. Based on an assumption that the average block in Vermont harbors 100 breeding bird species, a block was considered complete when at least 75 species had been documented, and at least 35 of those species were confirmed breeding. The 75/35 rule could not be strictly applied to blocks with relatively homogenous habitat. Forested blocks with few openings or wetlands, for example, do not have a diversity of habitats adequate to support 100 breeding bird species. Most of these blocks were located in the Northeastern Highlands biophysical region and in blocks in the Green Mountains, where blocks were extensively forested and supported few other habitat types. Several research and monitoring projects in Vermont provided data to the second atlas. The Forest Bird Monitoring Project and Mountain Birdwatch (Vermont Center for Ecostudies) conduct point counts in unmanaged forests and at high elevations in Vermont, respectively. Long-term research projects on grassland birds in the Champlain Valley (N. Perlut, University of New England and A. Strong, University of Vermont) and on Gray Jays in the Northeastern Highlands (B. Barnard, Norwich University), provided data on grassland and boreal specialists, respectively. Audubon Vermont submitted sightings from Important Bird Areas, many of which were “Unique and Fragile Areas” at the time of the first atlas, as well as grassland bird data from monitored airports and targeted surveys for Upland Sandpipers. Records from Vermont Department of Fish and Wildlife monitoring projects were incorporated into the Atlas database, including records of waterbirds in marsh systems, colonial waterbirds on Lake Champlain, Common Loon, Black Tern, Bald Eagle, Osprey, Peregrine Falcon, and Spruce Grouse. Raptor chicks that were brought into the Vermont Institute of Natural Science during 2003 to 2007 from known locations were also used as a source of confirmations. Results from the Champlain Valley and central and eastern Vermont duck box surveys, conducted by Vermont Department of Fish and Wildlife, were also included.

Quality Control

Observers for the first atlas were generally experienced birders. For the second atlas, training was provided for any observers requiring practice by pairing them with highly experienced birders and/or participated in birding field trips. Support materials provided to atlasers and playbacks were provided for nocturnal surveys for owls, and for marshbird surveys. In order to avoid falsely recording a migrant as a potential breeder, we established “safe dates” for each species. Within the safe date period, a species is highly unlikely to be migratory, and therefore if it is present on a block, it is mostly likely to be breeding or attempting to do so. Adhering to safe dates removes observer variation and bias in deciding when an observation is valid. Safe dates were established using data from Vermont (including the first atlas for atlas 2), Records of Vermont Birds (Vermont Center for Ecostudies unpubl. data), and research and monitoring projects within the state. Local experts provided feedback on the species with which they were familiar. When Vermont data or expertise was not available for a species, data from neighboring states were used to estimate appropriate dates for Vermont. Though safe dates for different parts of the state may vary slightly, only one set of dates was used for each species to avoid confusion. Conservative dates that would be “safe” anywhere in the state were employed. Observers were encouraged to spend time in habitat that could harbor species of conservation concern. Materials provided to observers highlighted rare species, particularly those that could be easily confused with other species. Observers were asked to document the basis for identification of some rare species, particularly those that could be confused with other species or are difficult to identify, and to provide the exact location of the observation. County coordinators confirmed records of rare species by following up on reports in the field or, if that was not possible, by reviewing the submitted form and discussing it with the observer. All documentation of observations of rare species was reviewed by the atlas director.

Method steps

1. We used an online database that was developed specifically for atlases by USGS Patuxent Wildlife Research Center. Online data entry was usually carried out by volunteers, but in some cases the county coordinator or the director entered data on behalf of a volunteer. All data on field forms were entered in the database for each date in each block at the end of each field season. The database immediately flagged records that were entered outside of safe dates or did not match codes expected for a given breeding confirmation, so that the person entering the data could correct errors at the time of entry. The system included a data review section that coded records, delineating whether they were potentially acceptable, rare species that required extra review, or were outside of safe dates. Coordinators reviewed each entry and determined whether to accept or reject it. In addition to the automated coding, coordinators were asked to consider whether each record was consistent with what could be expected in their county given the location and date of the record, as well as the behavioral characteristics of the species. For example, a DD (distraction display) code entered for Killdeer would be a realistic confirmation code, but it would not be appropriate to use for a hummingbird; CF (carrying food) would be appropriate for many songbirds, but is problematic for many raptors that cache food. A Rusty Blackbird record would be closely scrutinized if it were reported from a block with little or no boreal habitat. After coordinators completed their review, all data were reviewed again by the atlas director and T. Murin (coauthor, Birdwatching in Vermont) based on the same criteria, and problematic date entries and incomplete field cards were also searched. Flagged records were reviewed carefully, often in consultation with the observer.

Birds documented as breeding in the State of Vermont, USA.

1. Aves
   common name: Birds rank: class

State of Vermont, USA