Automatic Assessment of Forest Landscapes

Forest Spatial pattern, forest fragmentation and connectivity are policy-relevant indicators of sustainable forest management and biodiversity conservation. The monitoring of forested landscapes in space and time requires precise mapping and statistical analysis. In this paper, we use a new method for the classification and analysis of forest spatial patterns derived from forest - nonforest masks based on satellite imagery. Forested pixels are classified as "perforated", "edge", "patch", "core", "corridor" and "branch". The method applies morphological image processing routines, which allow the automatic detection of fragmentation and connectivity related features such as core, patch and corridors over large areas in a holistic approach. Results are demonstrated at local scale for a Natura2000 site in northern Italy and at country scale for Slovakia. The temporal dynamics of forest spatial pattern is demonstrated and statistically evaluated. The impact of qualitatively different input maps on the classification results is discussed.JRC.H.7-Land management and natural hazard