Degradation–environment–society spiral: A spatial auto‐logistic model in Thailand

In this contemporary interpretation of the widespread land degradation problem in Southeast Asia, it is hypothesized that spatial interplay of environmental and socioeconomic predictors determines the occurrence of land degradation. Village surveys, remote sensing and spatial auto‐logistic modelling of the relationship between degradation and land use dynamics in Lam Phra Phloeng watershed of Thailand enabled 80.2% of land to be classified correctly in terms of the presence or absence of erosion and explained 53.2% of the total variation. Cultivation and dependence on agriculture for livelihood positively and significantly affect degradation. Lack of access to institutional credit and land titles significantly increased the probability of occurrence of degradation. On the other hand, education and social cohesion are negatively associated with the occurrence of degradation. The Relative Operating Characteristic (ROC) curve was used to measures the performance of the model. The calculated area under the curve (0.879) suggests that significant predictor variables in the model can be confidently used to forecast the likelihood of occurrence of degradation and thus to identify priority areas for intervention. Policies to reduce land degradation should include measures to reduce pressure on the land, including alternative income sources. Policies could mobilize capital to invest in encouraging nature‐based tourism and other off‐farm income options.