Vermont Butterfly Atlas 1.0 (2002-2007)

Study Extent

State of Vermont

Sampling

The Vermont Butterfly Atlas (VBA) was a six-year (2002-2007) census to document the relative abundance and distribution of butterflies across Vermont. VBA was closely modeled after the Massachusetts Butterfly Atlas (1986-1990), Connecticut Butterfly Atlas Project (1995-1999) and Breeding Bird Atlas of Vermont (Laughlin and Kibbe 1985). Data collection protocols generally followed those of other biological atlases. Because this was the first insect atlas ever conducted in Vermont, we planned for one trial year (2002), followed by five additional field seasons (2003-2007). Field biologists were hired in some years to augment survey coverage in areas lacking adequate volunteer coverage, including blocks in the Northeastern Highlands, Franklin County, and Windham County, as well as special habitats such as wetlands. VBA accepted butterfly records from anywhere in Vermont, but to ensure thorough coverage, we surveyed butterflies evenly and systematically across the state. We adopted a grid-based sampling scheme from the Atlas of Breeding Birds of Vermont (Laughlin and Kibbe 1985). This system relied on the 184 U.S. Geological Survey 1:24,000, 7½-minute quadrangle topographic maps (7.5 minutes = 1/8 of 1 degree of latitude or longitude) that cover Vermont. Quadrangles were divided into 6 blocks with each block covering about 25 km2. Because it would have been impossible to adequately survey all 1,177 blocks in Vermont, we randomly selected one priority survey block from each quadrangle for a total of 184 priority survey blocks scattered across the state. All data from both priority and non-priority blocks were accepted by VBA. Priority blocks constituted the minimum set of blocks requiring full surveys in order to obtain a valid sample of butterflies for the entire state. Butterflies were recorded as either voucher specimens (collection or photograph) or sight records (net-release, binoculars, visual-no aid). In the field each collected specimen was assigned a uniquely numbered voucher card and placed in a glassine envelope. Printed photographs were also handled in this manner after development. Digital photograph files were assigned a voucher card number and electronically transferred to VBA. An attempt was made to secure one voucher of each species in each priority block. Data collection fell under two categories: site surveys (timed counts) and casual observations (incidental records). Site survey forms were completed when a field worker visited a location specifically to conduct a timed survey and count of all butterflies observed, as well as to record habitat types in which they were found. Casual sightings did not require measures of time or counts of individuals seen, but simply noted that a given species was present at a particular location on a certain date. Field workers were asked to visit potential butterfly habitats in their adopted priority blocks at least once per month during the growing season (May-Sept). Many breeding bird atlases adopt minimum time and total species requirements; these define when a block was adequately surveyed (Smith 1982). Because we had no prior data to examine for VBS, we collected data for two field seasons before setting our standards. We ranked the detectability of each species on the checklist and calculated the minimum total number of species likely to be found in each block with a reasonable amount of fieldwork. The minimum standard was set at 30 species per block, with a preferred target of 40 species or more.

Quality Control

A data processing and retention protocol was outlined prior to the first field season. All data sheets were initially examined for completeness, matching codes and records, and other simple quality assurance and control (QA/QC) procedures. Site survey forms, casual observation forms and voucher card data were entered into the database and filed. Project staff and experienced volunteers completed most voucher determinations. For difficult species and specimens (e.g., Celastrina, Papilio glaucus and P. canadensis, Phyciodes, and Erynnis), experts were consulted for final identification. Because visual records yielded no physical material for determination, they were simply examined for potential errors in field identification by searching for any possible out of season dates, or odd locations and habitats. Any obvious mistakes resulted in deletions of records from the database or assignment to the genus level only. Vouchers with photographs that were not acceptable for determination were changed to visual records if the observer indicated a field identification.

Method steps

1. Final QA/QC was conducted after all data were entered. Although detailed priority block maps provided guidance to observers, most errors detected resulted from georeferencing mistakes made by observers, such as incorrect latitude and longitude or incorrect block names. Because we wanted records to be spatially referenced to a point and not just the block if possible, these corrections were very time consuming. Ambiguous records were only spatially referenced at the block level or higher (town, county or even state in rare cases). Exact reference points for each record will allow for more detailed analysis and modeling of species in the future.

Butterflies

1. Lepidoptera
   common name: Butterflies rank: order

State of Vermont, United States