How far should adjacent pan traps be placed for bee sampling?

Abstract

Among several methods of active and passive sampling bees (Hymenoptera: Apoidea), pan traps, also known as bee bowls, are commonly utilized for their efficiency and as a means of avoiding collector bias. When comparing pan-trap samples using traditional statistical methods, the assumption of independence among samples needs to be met. To determine the necessary distance between pan traps to obtain independent samples, we investigated spatial dependence in bee catches using pan traps at four sites. At each site, a regular grid of pan traps was laid out, with each sample having two colors of fluorescent trap (i.e. blue and yellow). These trap pairs were separated by 10 m, with 100–109 points in the grid. In addition, five points within the grid were chosen randomly, and an additional 40 traps were placed at 2 m and 4 m from the five sample points to sample bees at a minimum 2 m trap distance. Traps collected bees for five consecutive days and bees were counted for each point and each trap color. Bee count data were analyzed using geostatistics to determine the spatial dependency in trap catches for yellow traps, blue traps, and both colors combined. The highest value of the range parameter of the semivariogram found across the four sites and pan trap colors was 17 m for blue traps. From this, we recommend a conservative minimum distance of 17 m between adjacent pan traps to obtain spatially independent samples for optimizing sampling plans when independent samples are necessary.