Human land use effects on markers of bumblebee colony health
Understanding the mechanisms behind global pollinator decline are a pressing area of ecological research. Bumblebees (Bombus spp.) are an important native pollinator of wild plants and agricultural crops. Using aerial photography from the National Agriculture Imagery Program (NAIP), and the National Land Cover Database (NLCD), combined with USDA-provided pesticide sampling and colony population data, we examine the interrelationship between human land use and Bombus spp. colony health. During summer 2013, the USDA Agricultural Research Service Bee Research Laboratory (BRL) placed 26 bumblebee research colonies in 3 clusters of sites in Central Maryland and Washington, D.C. The clusters were located in Howard County, MD (10 sites), Beltsville, MD (6 sites), and Washington, D.C. (10 sites). Sites were selected to sample a wide range of land-use land-cover types including urban, suburban, forested, agricultural, etc. Sites were also spaced to minimize overlap. The colonies were initially weighed and placed in the field for 8 weeks and were then recovered for laboratory analysis. All colonies were deployed in July 2013 and were recovered in September 2013. Colonies were frozen and weighed upon arrival at the laboratory and the following parameters for each colony were counted based on brood development: egg pouches, larvae pouches, larvae and pupae, adults, queen pupae, and adult queens. In addition to these parameters, samples of workers and wax from each colony were tested for presence and concentration of approximately 175 types of pesticides.
Analyses are focused on percentage impervious surface within a 1 km radius of deployed colonies. This radius is indicative of the effective foraging range of bumblebees. Weight change during time in the field was determined, and totals of egg pouches, larvae pouches, larvae/pupae, adults, queen pupae, and adult queens were summed to comprise a colony health summary statistic. Linear regression shows a negative correlation between percentage impervious surface and percentage change in weight, and also percentage impervious surface and the colony health summary statistic. In addition, linear regression shows a positive correlation between total pesticide load and the colony health summary statistic. These results echo similar findings in the literature that colony health is influenced by various combined factors.