Use of Sentinel-2 satellite imagery for forest site evaluation and forest harvesting detection

The access to Earth Observation data leaded researcher to a different point of view in the forest sector. Immediately tropical forest deforestation drawn the majority of interests (Perbet et al., 2019; Tang et al., 2019; Shimizu et al., 2017; Asner et al., 2009), heading to the development of many different tools for tropical forest monitoring. This study was focused on the application of satellite remote sensing data (derived from Sentinel-2) to two cardinal aspect for Italian forest. Since wood production plays a key role in developing a rural economy and stimulating the use of sustainable raw material, an increment of Douglas-fir plantation is desirable because of his great growth potential. Therefore, it was necessary to investigate good indices in order to assess the Douglas-fir land suitability and fertility indices. Empirical models were developed and validated using different sets of variables derived from remote sensing data and field survey. Models validation reached good results for Site Index ranging from 0.63 to 0.97 R2 and Current Annual Increment ranging from 0.50 to 0.98 R2. Furthermore, remote sensing data were applied to calibrate and validate different approaches for forest change detection. Knowing where and when forest harvests are done is crucial for correctly applying sustainable forest management and for controlling illegal logging. In this study was demonstrated that there are already tools developed in tropical forest that they could be applied to Italian forest. The best method was the basic one, which uses only summer images avoiding the seasonal noise problem in the time series but losing near-real time ability. If the temporal accuracy is essential the best method for removing time series seasonality resulted the harmonic model fitting, but further analyses are needed expanding the validation area in order to corroborate these results.