Remote sensing and automatic procedures: useful tools to monitor forest harvesting

Forests produce a wide range of ecosystem services, including the traditional wood production. Sustainable forest management approaches are used to locally calibrate wood harvesting on the basis of local conditions and should not adversely affect other ecosystem services. To assess forest harvesting sustainability and impacts it is essential to know their spatial distribution. At the present date accurate statistics on wood harvesting are not available in Italy. In this context, remote sensing and automatic mapping algorithms constitute an important tool for providing spatially explicit information to quantify wood harvesting and thus supporting more sustainable forest management approaches. In this work we tested an automatic mapping algorithm (3I3D) using multitemporal Sentinel-2 imagery to produce a map of forest disturbance produced by different types of wood harvesting in the province of Arezzo for the year 2018. Thanks to a photo-interpretation work of high resolution PlanetScope imagery and field data collected by the Carabinieri Forestali, we were able to calculate commission (2%) and omission (18%) errors of the automatic 3I3D map. The results of this work motivate the introduction of remote sensing tools as a support for monitoring and quantification of forest wood harvesting