A new view on EU agricultural landscapes: Quantifying patchiness to assess farmland heterogeneity

Mapping and assessment of ecosystem services in agricultural landscapes as required by the EU biodiversity policy need a better characterization of the given landscape typology according to its ecological and cultural values. Such need should be accommodated by a better discrimination of the landscape characteristics linked to the capacity of providing ecosystem services and socio-cultural benefits. Often, these key variables depend on the degree of farmland heterogeneity and landscape patterns. We employed segmentation and landscape metrics (edge density and image texture respectively), derived from a pan-European multi-temporal and multi-spectral remote sensing dataset, to generate a consistent European indicator of farmland heterogeneity, the Farmland Heterogeneity Indicator (FHI). We mapped five degrees of FHI on a wall-to-wall basis (250 m spatial resolution) over European agricultural landscapes including natural grasslands. Image texture led to a clear improvement of the indicator compared to the pure application of Edge Density, in particular to a better detection of small patches. In addition to deriving a qualitative indicator we attributed an approximate patch size to each class, allowing an indicative assessment of European field sizes. Based on CORINE land cover, we identified pastures and heterogeneous land cover classes as classes with the highest degree of FHI, while agroforestry and olive groves appeared less heterogeneous on average. We performed a verification based on a continental and regional scale, which resulted in general good agreement with independently derived data.