Pilot study on the use of National Forest Inventories to downscale European forest diversity spatial information in five test areas, covering different geo-physical and geo-botanical conditions
The project started the 21st of November 2006 and was 24 months long. This study aimed to address the feasibility to integrate National Forest Inventories data and remote sensing derived forest spatial pattern data to downscale large-scale aspects of forest biological diversity.
This study was designed in the framework of the research activities aimed to address the feasibility to integrate National Forest Inventories (NFIs) data and remote sensing derived maps to downscale large-scale aspects of forest biological diversity in order to understand if different forest landscape patterns mapped with remotely sensed may be related to different forest biodiversity conditions. The project is based on the integration of two sources of information: (i) forest biodiversity indicators assessed by NFIs; and (ii) forest spatial pattern maps derived from remotely sensed imagery. The study was conducted in seven test sites for just less of 100000 km2: one in Sweden for the Boreal area, one in Germany for the Atlantic, two in Czech Republic for the Continental, one in Switzerland for the Alpine and two in Italy for the Mediterranean.
In each test site raw NFI data were acquired as well as forest maps with different geometric resolutions (25 m and 100 m). NFI data were stored in a common database made of 3262 NFI plots, 44410 trees and 7884 pieces of deadwood. Raw data were harmonised in order to calculate forest biodiversity indicators for each of the available NFI plot. Indicators were grouped in six areas: forest types, forest structure, deadwood, naturalness, and stand age. Forest maps were harmonised in terms of geometry and resolution also adopting the European Forest Types system of nomenclature recently proposed by the European Environmental Agency. Forest maps were processed by Morphological Spatial Pattern Analysis (MSPA) deriving spatial pattern maps on the basis of different configuration of the GUIDOS software developed by the JRC. In order to understand if different forest spatial pattern classes could be related to different levels in the biodiversity indicators assessed in the field, each NFI plot was classified on the basis of the spatial pattern class of the forest patch it belongs to. Several approaches for the spatial pattern classification of the NFI plots were tested for each of the different spatial pattern maps created with different GUIDOS configurations.
