Hillslopes of a tropical montane watershed in northern Thailand are analyzed to understand erosion processes. Plots of slope versus drainage area are useful to delineate erosion regimes in a watershed. For example, regions susceptible to landsliding will concentrate in one field of a slope-area plot, while regions susceptible to channel initiation by overland flow erosion or spring sapping will plot in other fields. Boundaries between these regimes can be modeled given certain hydrologic assumptions, and slope-area plots can therefore be used as predictors of erosion mechanisms in a watershed. These ideas have been extensively tested by others in many regions in the northwestern United States. Here, we apply these concepts to a 2 square km tropical watershed with steep slopes underlain by granitic and metamorphic rock mantled by 2-to-3 meters of red saprolite and about 0.24 meters of loamy dark soil.

The Pang Khum Experimental Watershed has been the site of ongoing hydrologic studies since 1997. Current efforts are focused on quantifying the impacts of swidden agricultural practices versus unimproved forest roads on sediment delivery to streams. To support this effort, we conducted a field survey of channel initiation mechanisms in the watershed and modeled the boundaries of erosion regimes in slope-area space using a high-resolution digital elevation model.

Field results show that three channel initiation mechanisms operate in the watershed: 1) shallow landsliding, 2) erosion by overland flow, and 3) spring sapping. Preliminary modeling results show that locations susceptible to these processes plot in distinct fields on a graph of slope versus area. Further, channel heads formed by any of the three mechanisms fall in a region on a slope area plot that separates channeled and unchanneled convergent topography. These findings are similar to those reported in studies in other humid-climate forested mountain regions.