Methods
Study system : The Massachusetts Butterfly Club (MBC) is a group of citizen scientists which documents sightings of butterflies and maintains a database of observation records. The MBC compiles records of butterfly sightings from organized group trips and as reported opportunistically by individuals. The current club database of records from 1992-2018 includes 30,707 species lists from these reports (hereafter referred to as trips) encompassing 162,670 sightings of 99 different butterfly species. We thinned this to 10,884 trips encompassing 126,087 sightings of 89 species suitable for analysis using multiple criteria. First, although the database includes historic records used to build the Atlas of Massachusetts Butterflies, we excluded records prior to 1992, the year in which the club started to record large numbers of lists (i.e., > 1200 sightings/year). Next, following Breed et al. (2013), trips with < 5 species seen were excluded, to minimize the effect of anecdotal sightings (see Appendix S1 in Supporting Information ). Unlike Breed et al. (2013), we combined and retained records for two species groups. Canadian Tiger Swallowtail Paplio canadensis and Eastern Tiger Swallowtail Paplio glaucus records were combined (rather than excluded) because they are known to hybridize and can sometimes be difficult to visually distinguish. Spring Azure Celastrina ladon and Summer Azure Celastrina neglecta were also combined because of a change in species reporting within the MBC during the study period. With these two exceptions, species were excluded if during the study period they underwent a taxonomic shift or realignment, or if there was an outbreak of the population uncharacteristic of conventional population dynamics (following Breed et al., 2013; see Appendix S2 ). Species were also excluded if there were fewer than 15 observations before the year 2000, to prevent outliers or a lack of available data from driving trends (See Appendix S4 ). MBC records were processed into a database, generating 89 species-specific lists of all trips. Each species file included presence/absence of the species on each trip, the date of each trip, the number of individuals of that species seen on each trip, and the total number of species seen on each trip (see Appendix S1 ).
To interpret patterns of phenology and abundance change we compiled data on two life history traits: voltinism and range type. Following Breed et al. (2013), voltinism was categorized as either obligately univoltine (a single generation per year) or multivoltine (two or more generations per year). Categorization agreed with Breed et al. (2013), with a few exceptions in which scatter plots of observation day of year supported an update (see Appendix S4 ). Following Breed et al. (2013), we categorized species’ range type based on the center of their geographic range distribution, relative to Boston, Massachusetts (the approximate mid-latitude for the state of Massachusetts). Species were ordinally identified as having ranges centered north, central, or south of the latitude of Boston (species classified as northern would therefore be closer to the southern limit of their range).