Incorporating Spatially Explicit Landscape & Connectivity Metrics Improves the Predictive Ability of Species Distribution Models

Species distribution models are one of the most important GIScience research areas in biodiversity. However, models rarely incorporate spatially explicit landscape metrics. Here, we investigated the importance of these features in determining the fine-scale spatial distribution of the critically endangered Eurasian curlew (Numenius arquata). Tracking data for 18 birds were collected, a highresolution land cover map was generated from Sentinel 2 imagery, and the land cover was parameterised using proportions, as well as connectivity metrics using a moving window of 1km. Accounting for the spatial configuration of the landscape generally improved the predictions of habitat preferences beyond the widely employed proportion of suitable land cover.