Monitoring Forest Cover Change Using Google Earth Engine in Kibale District from 2014- 2020.
Abstract
Kibaale district is one of the districts with 15 forest reserves in western Uganda. However, over the years these forests have received a lot of pressure from people. This is due to increased population especially by migrants, climate changes and increased demand of forest products. This study therefore aimed at monitoring forest cover to detect change in the land cover and predict forest cover in 2030 using Google Earth Engine. Using time series and change detection analysis, NDVI has been observed to be decreasing over time and some time increasing in some areas. NDVI values within the forest keep reducing indicating that these forests are no longer intact. Anomaly maps were further developed using a Z- score to analyze these changes in NDVI. Land cover classification was done using Random forest classifier with 1492 samples using classes of trees, water, built area and cultivated area. This was done at an overall training accuracy of 99.5%, training Kappa of 99.3%, validation accuracy of 81.4% and validation Kappa of 74.4%. later these classification maps were using for prediction with CA-MC models to obtain a land cover map of 2030.