AUTOMATIC FOREST DEGRADATION MONITORING BY REMOTE SENSING METHODS AND COPERNICUS DATA

Nowadays, forests are the most widely spread land cover and therefore play a significant role in ecology and create processes' dynamics. Forests are threatened by various harmful effects due to biotic (insects, fungi, viruses, weeds, animals) and abiotic (floods, fires, storms, droughts, polluted atmospheres etc.) damages. Also, human damage (anthropogenic impact) is numerous and varied. They are caused by direct human action on the forest and indirect activities and processes (damage due to grazing, consequent devastation and erosion of habitats etc.). Forest devastation and illegal logging are one of the immediate negative human activities that have a detrimental impact on the forest. The research refers to the state forests of two management units, Javornik and Corkovae-Karlice in the border area of the Republic of Croatia. Part of the forests (about 3,000 hectares) of these two management units are located in a mine suspected area along the state border with Bosnia and Herzegovina. This research aims to develop an automatic algorithm for forest degradation monitoring by remote sensing methods and Copernicus data. The developed algorithm was based on the Sentinel-2 (S2) optical satellite imagery and Google Earth Engine. The proposed automatic forest degradation monitoring algorithm was based on the Delta NDVI change detection approach. Accuracy assessment was done by independent data in higher, 3-m resolution based PlanetScope imagery. Preliminary results show very similar forest degradation values per all tested forest compartment/subcompartment for automatically generated S2 10-m imagery forest degradation map and 3-m forest maps obtained manually from PlanetScope imagery.