PS 31-180 - Determining an Efficient Wild Bee Sampling Schedule

Abstract: Wild bees are vital pollinators in ecosystems; however, there are anthropogenic threats to their populations. While monitoring bees is clearly important, few studies monitor the entire bee flight season. Most studies allocate sampling efforts over week- to month-long periods, with more focus on maximizing geographic range. Given the varied phenology of bees, a limited sampling season could be missing temporal variation in abundance/richness. Our bee monitoring program instead encompasses the entire flight season (typically, early April to the first frost in November) at six sites in an urban to peri-urban setting. Our now six-year bee community database is comprised of 16,933 bees of 36 genera, including 40 species that were not previously recorded in Plymouth County, MA. Using this expansive database, we aim to identify a more efficient sampling schedule that can catch all previously detected genera at a similar probability as sampling throughout an entire bee season. Since our exact sampling dates vary from year-to-year, we divided the sampling season into half-month time bins. We analyzed our abundance and genera richness data to identify the half-month sampling windows that would maximize the probability of detecting every genus we have observed, including singletons and other rarely sampled genera.

We first ranked the time bins by unique genera richness, using abundance in cases of ties. With this initial ranking, we identified late-July as the half-month time bin encompassing the greatest bee richness/abundance, representing 80% of detected genera (28/35) and 13.5% of average annual total bee abundance. Adding the second highest ranked bin, early-August, 91% (32/35) of genera were represented. Adding early-May, late-August, and early-October accounted for detection of the remaining three rare genera (Melitta, Triepeolus, and Panurginus, one genus per month, respectively). Apis was not considered because of their domesticated status. According to the seasonal variation reflected in our six-year dataset, the sampling schedule in Plymouth County that has a likelihood of detecting all 35 wild bee genera should include five sample dates including early-May, late-July, early-August, late-August, and early-October. Determining a schedule with a reduced number of sample dates that provides a reasonable expectation of thoroughly representing the wild bee community could allow reallocation of resources toward sampling a broader geographic area or facilitate collaborations with labs of fewer personnel (e.g. other community colleges). An efficient sampling schedule that accurately reflects abundance/richness could facilitate better monitoring of bees across urban landscapes.