The primary goal of the winter surveys was to compile a list of the bird species that occur in each block during early winter (defined as November-December) and late winter (January-February). As with the breeding surveys, perfect attainment of this goal is unrealistic and species will have been missed from all block lists – but we aimed to achieve sufficient coverage to obtain a good sense of how each species is distributed across the state during these time periods. Data collection was primarily by volunteer birders, many of whom signed up to lead the survey work for a given block, although anyone was able to submit checklists for any block.

Survey methods

Many bird species distribute themselves differently during the non-breeding season, compared to the breeding season. While nesting, many species are territorial, and all are – at least somewhat – tied to nest sites. At other times of the year, territoriality is less common and many species form flocks or concentrate in particular places. Nonbreeding birds are also more able to move in response to weather and other changes in environmental conditions when they do not have a nest to tend. With these differences in mind, we encouraged the use of several different types of survey to improve our ability to document how birds use space during the winter months.

Most data collection involved typical birding trips in which birders were asked to keep a checklist on which they recorded the number of individuals of each species seen. We asked volunteers to break up their winter survey effort into a series of separate one-hour observation periods to provide consistent survey effort across each site visited. For longer visits, we encouraged them to conduct consecutive observations, but to keep a separate checklists for each hour and specific area covered. We also asked observers to spread out their visits so that they surveyed different parts of the blocks and different habitats. If other interesting birds were seen outside of these timed observations, we asked that they be submitted as incidental observations. Instructions were detailed in the following protocol:

• Identify 5-10 sites within your block, chosen so that they span a range of habitat types. A site should be an area you can search on foot (or similar – a kayak or small boat would be fine) within an hour. Do not chose just the best or most well-known sites, but ensure your set of sites represents the range of habitats present in the block. Including lakes, ponds, and shorelines is fine, but consider doing specialized surveys on these habitats as well. A site can be anything from a public park, a walk through a neighborhood, a trail through the woods, a length of beach, etc.
• Timed surveys should not be conducted in a car, as you will miss more birds that way. Nor should they be stationary. Records from feeder-watching should be submitted as incidental observations, or as stationary eBird checklists.
• Visit each site for a timed one-hour survey sometime in November or December, and then again in January or February (i.e., a total of two visits per site). If you have a countdown alarm on your phone, use that to ensure you do not spend longer than an hour.
• Surveying multiple sites on the same day is fine. Similarly, if there is one large area that has a variety of habitats and would ordinarily take more than an hour to survey thoroughly, then it is ok to subdivide it into sections, each of which is considered a separate site. For example, one could divide the portion of Hammonasset State Park that lies in block 98F into three areas (e.g., Meigs Point area, central rotary area, west end area) and visit all three in the same morning, spending an hour in each. Each one-hour survey should be submitted as a separate checklist. Other interesting birds seen outside these one-hour blocks can be reported as incidental observations.
• During each hour-long period, survey the chosen site on foot, counting all birds seen or heard, while doing your best to avoid counting the same individuals more than once. We are looking for a count of the number of different individuals of each species that you detect – not the total number that you estimate were there.
• Include all birds that are using the habitat in the area you are surveying. This includes flying birds such as hunting raptors, gulls flying along the shoreline, sparrows flying between bushes, blackbirds flying from one field to another, etc., but not birds that are merely flying overhead (e.g., migrating blackbirds or finches, high-flying flocks of geese).
• In coastal areas or on large lakes, birds offshore should be counted as long as they are within the block boundary.
• Timed surveys can be conducted at any time of day, except the first hour after dawn and the hour before dusk. During these periods birds may be traveling from or to roost sites, and recording their location could provide misleading information about where they spend their time during the day. More information about how to document roosting species is given on the specialized surveys page.
• Avoid surveys during weather that is likely to make birds hard to find (heavy rain or snow, strong wind).

In addition to standardized one-hour surveys, we encouraged the use of specialized surveys that focused on (a) inland lakes and ponds, (b) coastal shorelines, major rivers, and large lakes, and (c) nocturnal and high tide roosts. The specialized surveys were designed to supplement other data collection, and observers were told that when block boundaries bisected the site being surveyed it was not necessary to assign each bird to a given block since the focus of most uses of the data would be linked to the location and habitat, rather than the block designations. Protocols for these surveys are provided here.

Block completion

Few winter atlases have been conducted previously, and we went into the study without a good sense of how much time it would take to develop relatively complete species lists. Consequently, we simply mirrored the breeding season goal of approximately 20 hours work in each atlas block over the course of the study, with 10 hours in each of the early and late winter periods.

Data submission

Volunteers were encouraged to submit data via eBird, and were warned to ensure that each checklist was located at a point that is representative of the area searched (e.g., the start point or a point in the center of the area covered), and that the point and entire surveyed area lay within the focal block. We also asked that they submit the data as a traveling checklist and recommended using the “Record track” function in eBird so that their phone’s GPS would determine the distance traveled automatically. Finally, we warned about defaulting to the use of eBird hotspot markers because they generally provided less precise location information, and often misplaced surveys into the wrong atlas block.

Although eBird checklists were the preferred method of data submission, we also provided a data sheet that volunteers could print and use, either submitting the form on paper via mail, or electronically via email. Paper datasheets were scanned and added to a digital archive that includes datasheets submitted via email. Paper originals are also stored at the University of Connecticut. Data that were not submitted by observers via eBird, were generally entered into the atlas eBird account by the atlas team to make the data available for other research uses. For rare cases, where privacy or the birds’ welfare were concerns, we added data directly to the final atlas database without putting it into eBird.

Other data sources

To generate final block lists and maps, we supplemented the volunteer-collected data with information from other sources. The primary source of additional data was the standardized transect surveys conducted by atlas field technicians to provide systematically-collected abundance data. Christmas Bird Count data were incorporated when participants kept track of block boundaries and submitted checklists from their counts to the atlas. As for the breeding atlas, general eBird submissions were not automatically added, but we did review the larger eBird database to ensure that we were not missing particularly notable records. During this review, records of species that are rare or geographically restricted in the state, and which enhanced our broad understanding of a species’ distribution in the state, were added as incidental observations to the atlas database and are reflected in the final species accounts.

Data verification

Individual checklists were generally screened soon after they were submitted to check for potentially erroneous identifications or abundance estimates that would not have been caught by standard eBird filters. Due to logistical constraints, this screening was not exhaustive, but it allowed us to follow-up with observers quickly when potentially problematic reports were found. In general, we did not exclude reports unless they involved extremely out-of-place species, or exceptional counts, and lacked supporting evidence. Surprising, but unverified, submissions are often identified in the species accounts, however, and are typically referred to as “reports”, rather than “records”. 

UPDATES TO COME:

ADD FIGURES SHOWING SPECIES RELATIVE TO SURVEY TIME

ADD MAPS SHOWING SURVEY TIME IN EACH BLOCK

ADD MAPS SHOWING SPECIES IN EACH BLOCK