Assessing monthly habitat quality for bumblebees in grasslands using a habitat trait model: a case study in Worcestershire, UK

Abstract Bumblebee decline in the UK has been associated with land use change from diverse meadows to intensive agriculture. While grasslands can offer the resources that bumblebees need to survive and thrive, the precise availability of resources through the year and between sites can vary due to differences in species composition and management. In this study, we surveyed bumblebee abundance using transects and monitored their habitat using 1-square-meter plots in 2021 and 2022 across five semi-natural grasslands located in Worcestershire, UK. These data were used to build a monthly bumblebee abundance prediction with a Generalized Linear Effects Model (GLME), which was used to rank the sites based on the predicted number of bumblebees. The selected model indicated that bumblebee abundance was most strongly associated with the proportion of the preferred bumblebee forage species and floral richness. Vegetation height and the variation of floral resources coverage had small, non-significant coefficients in this dataset, but were retained as ecologically relevant descriptors and for their contribution to cross-validated performance. Our model successfully identified sites with the lowest bumblebee abundance, and the monthly assessments provided information to guide management actions to improve the habitat quality at a site scale. Our findings are foundational, suggesting directions for future research in broader geographical contexts to validate and refine habitat assessment methods. Implications for insect conservation This study highlights crucial implications for insect conservation, demonstrating that targeted management actions based on monthly habitat assessments can significantly enhance the quality of grassland habitats. This approach not only supports bumblebee populations but also other pollinator species within agricultural landscapes.