How can we better understand the impact of global change on bees using mechanistic simulation models and monitoring campaigns?

Talk presented at the Annual meeting of the GfÖ - The Ecological Society of Germany, Austria and Switzerland. Freising. Germany. 2024 9th-13th September

How can we better understand the impact of global change on bees using mechanistic simulation models and monitoring campaigns?

Jürgen Groeneveld, Anna Wendt, Volker Grimm

Insect pollinators face multiple stressors, including land-use and climate change. However, it is difficult to predict and disentangle the impact of these multiple stressors on bee vitality. To help to fill this knowledge gap, we have analysed how temporal and spatial floral resource availability together with temperature and sun shine hours impact honey yield for 185 colonies distributed all over Germany.  Floral resources have been estimated from a land cover classification map by Preidl et al. (2020, Remote Sensing of Environment) that is based on Sentinel-2A data. This map has a resolution of 20 m, representing 19 land cover classes (e.g. oilseed rape, maize, forest) and is covering the whole of Germany. Honey yield was estimated using daily changes in hive weights using data from the TrachtNet campaign. We were using random forests as machine learning based methods and generalized linear models to analyse the data. As expected, daily changes in hive weight were mainly driven by temperature and daily sun shine hours. We were also able to identify the importance of interactions between land cover types such as oil seed rape and grassland. However, we could not detect direct effects from single mass flowering crops such as oil seed rape, indicating that it is very likely that biotic factors such as colony size and colony health need to be considered in our study. Therefore, we develop workflows for mechanistic honey bee (BEEHAVE) and bumble bee (BEESTEWARD) models to use the same land cover and weather information to simulate honey harvest and population dynamics for a variety of biotic scenarios. We will present a prototype digital twin for honey bees where this workflow is implemented and that can be freely be used by the community for any point in Germany.