A new species distribution modeling approach for biased citizen science data

In the field of ecology, understanding and predicting species distri bution is crucial for effective conservation strategies. Citizen Science initiatives have revolutionized the collection of species occurrence data, providing a cost-effective method for monitoring wildlife across various spatiotemporal scales. However, the lack of standardized sampling protocols within CS programs presents analytical challenges, resulting in biased sampling efforts that favor regions that are easily accessible or extensively studied. This study conducts a case study on the national butterfly monitoring program in the UK, utilizing a marked point process framework that integrates spatial and temporal covariates to analyze the spatial and temporal patterns of butterfly species occurrences. Our approach combines two essential components: (i) the locations visited by volunteers participating in these schemes and (ii) the presence records of a species of interest, using the INLA/inlabru framework. Our investigation highlights the importance of addressing sampling bias to enhance the accuracy and reliability of species distribution modeling. By combining the strengths inherent in CS data with rigorous modeling techniques, our approach paves the way for a better understanding of species distribution dynamics, strengthening the foundation for more effective conservation efforts.