Resampling of wild bees across fifteen years reveals variable species declines and recoveries after extreme weather

Despite concerns about the status of wild bees, there have been few studies with repeated sampling of these insects at the same locations using the same method. There is an urgent need for longitudinal studies to understand species- and population-level dynamics and the drivers of change, especially in pollinator-dependent crop systems. We sampled wild bees in highbush blueberry fields in three multi-year sampling periods between 2004 and 2018 to determine changes in the wild bee community active during bloom. We captured 35% (162 species) of the known Michigan bee fauna across the three sampling periods. Andrena, Lasioglossum and Ceratina were the dominant genera, though species composition varied significantly between sampling periods. Wild bee abundance (measured as captures in the pan traps) declined 61% and richness declined 33% between the first (2004–2006) and second (2013–2014) sampling periods. This was followed by limited recovery in species abundance and richness in the third sampling period (2017–2018). Some species showed dramatic abundance declines, particularly the blueberry specialist Andrena carolina. By contrast, other species showed stable or increasing prevalence, and population trajectories were not correlated with life history traits. Potential causes of the observed decline include increased use of insecticides and extreme spring weather in 2012. While farms varied in pest management intensity and overall insecticide program risk increased over time, there was no correlation between bee declines and insecticide use. Insecticide program risk increased between the second and third sampling periods, while most species increased in abundance. We therefore conclude the observed declines were most likely caused by extreme spring weather in 2012 resulting in widespread damage to flowering plants that provide resources for bees. Long-term monitoring to identifying species more vulnerable to environmental perturbations can help guide conservation recommendations as extreme weather events become more common.