Abstract/Summary

Wild bees provide a critical ecosystem service by pollinating globally important crops. Documented bee declines, notably in agricultural landscapes, therefore threaten future food security. Yet, evaluations of methods to inventory bees are rarely carried out in different crops or focus specifically upon crop pollinating species. We utilise standardised field datasets to elucidate differences in the capacity of transect walks, observation plots and pan traps to sample wild bee pollinator communities in four contrasting crops. Our results indicate that individual survey methods detect different components of crop pollinator communities, with guild and body size potentially important causal factors behind these differences. Transects detected half or less of the total potential pollinator community in three of our four study crops. Whilst transects were the most efficient method for sampling bumblebees, they often missed smaller solitary species, which were most efficiently sampled by yellow pan traps. Crop type is likely an important determinant of the most suitable survey methods to sample bee pollinator communities. Whilst transects alone are sufficient in crops pollinated predominantly by bumblebees, pan traps, and potentially observation plots, may be an important addition in some crops where smaller solitary bee species are potentially important pollinators. Our results indicate that the most efficient methods to sample bee species in agricultural landscapes are dependent upon crop type and pollinator community composition. We use our findings to make a set of recommendations on the inventorying and monitoring of bee pollinator crop communities that can inform regional and national monitoring programmes.