| In this study, Landsat 5 TM images obtained in 1995, 2000, and 2005 were used to categorize land use and land cover (LULC) and assess LULC changes and dynamics in the Nan watershed, Thailand using hybrid classification and post-classification change detection methods, respectively. Aerial photographs, digital LULC maps, and topographic maps were utilized to assess classification accuracy. Classified images together with ancillary data using Multi-Criteria Geographic Information System (GIS) Modeling were used to identify potential risk areas of soil erosion. The overall classification accuracy derived from the stratified random sampling pattern was 80.9% for 1995, 77.4% for 2000, and 86.0% for 2005. The results are satisfactory for LULC classification. Uncertainty of classification was mainly observed in green vegetation cover categories, which are forests, agricultural lands, and natural plants in wetlands. The study period revealed a significant increase in paddy fields and croplands at the expense of closed canopy forests. Paddy fields and croplands increased by 2.56% and 3.37%, respectively. Closed canopy forests, on the other hand, decreased by 3.64%. This change was caused by an increase in population within the watershed during the study period. Closed canopy forest, open canopy forest, and cropland categories were major drivers of LULCC in the Nan watershed. In 2010, the probability of change from forests to paddy fields and croplands are approximately 0.13 and 0.16, respectively. Expansions of high risk areas of erosion in northwestern and southeastern regions of the watershed in the two later years related to an increase of agricultural lands. Deforestation and environmental degradation in the Nan watershed are mainly influenced by population growth and increase demand for land uses to support human activities. |
