Assessing forest type variability in the Yellapur forest division of Uttara Kannada district, using a remote sensing approach

The study LULC classification of different forest types was conducted in the Yellapur forest division of Uttara Kannada district, geographically positioned between 74° 25' to 75° 6' East longitudes and 14° 44' to 15° 7' North latitudes. A Land Use and Land Cover map of these forest types was developed using Sentinel-2 satellite data processed through Google Earth Engine. For dry deciduous forests, vegetation was the dominant land cover (67.18%) of the area, followed by agriculture (28.64%) and built-ups (3.98%). Moist deciduous forests had the largest vegetation cover (79.79%), with built-up areas (15.79%) and agriculture covering 3.57 percent. In semi-evergreen forests, vegetation dominated (91.71%) followed by built-up land (5.07%) and agriculture (1.99%). Semi-evergreen forests displayed the highest vegetation density compared to the other two forest types.Normalized Difference Vegetation Index (NDVI) values indicated varying levels of vegetation health, with the highest values observed in semi-evergreen forests. The classification for dry deciduous forests achieved an overall accuracy of 97.29% and a kappa coefficient of 0.96, while moist deciduous forests had an accuracy of 98.36% and a kappa coefficient of 0.97. Semi-evergreen forests demonstrated similar accuracy with 98.11% and a kappa coefficient of 0.97.