This paper examines recent debate concerning the concepts of 'sustainable knowledge' and 'hybridity' in environmental research. Hybrid studies employ local, indigenous knowledge alongside global, scientific techniques to achieve guidelines for sustainable development. The paper discusses the value of indigenous knowledge in testing theories of environmental degradation, and the problems of overcoming socio-political constructions of environmental problems over wide time and space scales, as identified by Regional Political Ecology. The paper focuses on the so-called theory of Himalayan environmental degradation and illustrates the study with a case from northern Thailand. In Thailand, upland shifting cultivators are blamed for causing lowland sedimentation and water shortages, and have been considered by lowland communities to lack awareness of environmental degradation. The study used indigenous knowledge alongside GIS analysis and the Caesium-137 technique for measuring soil erosion to test the assumptions that land shortage has increased cultivation on steeper slopes, and that erosion is a problem for upland degradation. This is the first time these new techniques have been used in testing assumptions related to Himalayan degradation. Results indicated that upland farmers deliberately avoid erosion by increasing frequency of cultivation of flatter slopes rather than steeper slopes, and consequently the problem of erosion is overstated. However, this does not imply that their local knowledge is useful over larger areas, but instead shows the adaptability of local communities and their awareness of environmental risk. It is therefore argued that developing effective management techniques depends on differentiating more clearly between locally-based knowledge about environmental processes; politically-constructed statements about the environmental impacts of other groups; and falsifiable scientific assertions aiming to develop effective management techniques with reference to several communities. Copyright (C) 1996, Elsevier Science Ltd