Evaluation of the accuracy of multispectral land cover classification algorithms: Comparison of AI-based Machine Learning, Maximum Likelihood, and Mahalanobis Distance Algorithms

These are field data of forest types, species and extent of land cover type, including grassland, water extent. The data were the basis for analysis of forest degradation as depicted by a set of multispectral band-ratio indices and their relationship to climate variables and climate change data derived from General Circulation Models (GCM). These data allowed for the mapping of forest condition as related to forest degradation as a function of past and future climate.