Advanced earth observation approach for multiscale forest ecosystem services modelling and mapping (MIMOSE)

Abstract

In the last decade ecosystem services (ES) have been proposed as a method for quantifying the multifunctional role of forest ecosystems. Their spatial distribution on large areas is frequently limited by the lack of information, because field data collection with traditional methods requires much effort in terms of time and cost. In this contribution we propose a methodology (namely, Multiscale Mapping of ecoSystem services - MiMoSe) based on the integration of remotely sensed images and field observation to produce a wall-to-wall geodatabase of forest parcels accompanied with several information useful as a basis for future trade-off analysis of different eS. Here, we present the application of the MiMoSe approach to a study area of 443,758 hectares coincident with the administrative Molise region in Central italy. The procedure is based on a local high resolution forest types map integrated with information on the main forest management approaches. through the non-parametric k-nearest neighbors techniques, we produced a growing stock volume map integrating a local forest inventory with a multispectral satellite IRS LISS III imagery. With the growing stock volume map we derived a forest age map for even-aged forest types. Later these information were used to automatically create a vector forest parcels map by multidimensional image segmentation that were finally populated with a number of information useful for ES spatial estimation. The contribution briefly introduces to the MIMOSE methodology presenting the preliminary results we achieved which constitute the basis for a future implementation of ES modeling