From pixels to plants: linking remotely sensed habitat data to roe deer diet from DNA metabarcoding in Scottish forest landscapes

Résumé fourni par la source

Understanding how landscape characteristics influence herbivore diet selection is fundamental to predicting ecological impacts and developing effective wildlife management strategies. We linked remotely sensed habitat data and dietary DNA metabarcoding to test whether landcover classifications can explain variation in European roe deer (Capreolus capreolus) diet across Scottish forest landscapes. We characterised the diets of 198 roe deer from six sites, integrating remotely sensed habitat mapping of cull locations for 190 individuals with location data. Our results revealed that site explained nearly twice as much variation in diet composition (18%) as seasonal effects (10%), emphasising the primacy of spatial over temporal factors. Localities with similar habitat composition supported more similar diets, with these associations strongest in winter and summer. Habitat composition explained a consistent but partial share of dietary variation, the remainder likely reflecting individual selectivity, forage quality and fine-scale vegetation structure. Woodland habitats generally showed positive associations with particular plant genera, while open and wet habitats show predominantly negative relationships. These findings demonstrate that landscape configuration influences roe deer foraging, with clear management applications: for example, the extent of riverine and fen scrub within a locality was associated with reduced winter browsing on commercially valuable conifers, indicating where browsing pressure is likely to concentrate and where protection or deer management effort should be targeted. More broadly, linking freely available landcover data to molecular dietary analysis at the scale of individual foraging localities offers a transferable approach for anticipating herbivore impacts across forested landscapes.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.