 The study shows that random forest algorithm for classification of remotely-sensed data is highly sensitive to input data characteristics, training data selection, specific input variables, and class proportions. High-dimension data sets should be reduced to uncorrelated important variables, independent error assessments should be used, iterative classifications are recommended, and the size of the training data set also impacts results. The study recommends using randomly distributed and minimally spatially auto-correlated training data sets for improved classification accuracy. This article was authored by Kareen Millard and Murray Richardson.