Resolution problems in calculating landscape metrics

The variation of landscape metrics caused by varying the map input resolution has been investigated. Landscape metrics are spatial indicators used to link spatial patterns with the ecological processes that generate them. An aerial photograph was semi‐automatically classified at different resolutions by superimposing a grid with a variable cell dimension (10, 20 and 40 metres) in a GIS environment. The variation of the mostly used landscape metrics was investigated. This approach allows the up‐scaling of spatial indices, thanks to the objective ('a priori') definition of the Minimum Mapping Unit. Some metrics showed linear trends over the range of examined scales, and other metrics had non‐linear response curves. For the first type of metrics, the translation of information over a wide range of spatial scales seems to be very simple, while for the second this translation appears to be impossible.