Tracing shifting cultivation in the Nam Ton watershed (Lao PDR) by multispectral image-to-image change detection techniques with statistical verification

When looking for critical areas in the context of erosion and degradation in Laos, the regions that are affected by shifting cultivation are prior and most distinct visual. Up to now the distribution and extent of these areas is still unknown. This makes it unclear whether the extent and distribution of this practice has a cause in the further depletion of Laos natural forest resources. In order to find more accurate figures, the following study compares different satellite remote sensing techniques that monitor shifting cultivation. An unsupervised classification method of a tasseled cap transformation has a best overall accuracy of 84.5% in comparison with unsupervised classification according to principal components analysis (81.9%) and vegetation indexing (71.4%). It is concluded that the applied techniques are very useful applications in the monitoring of shifting cultivation in the catchments in Laos. This is due to their fast delivery of timely and accurate relevant information