Integrating citizen science and multispectral satellite data for multiscale habitat management

Abstract

Habitat management is necessary for the conservation of threatened species, yet best practices in fragmented human-dominated landscapes have remained difficult to generalise. We show that multi-scale vegetation management decisions in heathlands can be supported by integrating opportunistic citizen science data and multispectral satellite data. Opportunistic observations were gathered from ten typical, mostly threatened animal species of dry heathlands in Flanders as point records with specified precision. We considered vegetation structure at the local scale, quantified by image texture within 0.25 ha derived from multispectral satellite data, and heathland heterogeneity at the habitat scale, quantified by the diversity in heathland vegetation communities within 50 ha. Additionally, locations inside heathlands were attributed to an open, closed or anthropogenic landscape context. Point process models were used to test the impact of heathland size, vegetation structure and heathland heterogeneity on the habitat suitability of the studied species. We found that (1) heathland vegetation management can benefit habitat suitability in fragmented heathlands, but with a different approach for local management of vegetation structure in small versus large heathlands (e.g. due to micro-fragmentation effects), (2) the landscape induces positive and negative edge effects (e.g. due to a high versus low resource availability), especially in small heathlands and (3) habitat suitability is driven by both vegetation structure and heathland heterogeneity but with different relative importance for birds, butterflies and grasshoppers (e.g. due to differences in mobility).