Factors affecting soil loss at plot scale and sediment yield at catchment scale in a tropical volcanic agroforestry landscape

Tropical deforestation and land use change is often perceived as the major cause of soil loss by water erosion and of sediment load in rivers that has a negative impact on the functioning of hydropower storage reservoirs. The Sumberjaya area in Sumatra, Indonesia is representative for conflicts and evictions arising fromthis perception. The purpose of this study as part of aNegotiation Support System approach was to assess sediment yield both at plot and catchment scale and to relate it to a variety of possible clarifying factors i.e. land use, geology, soil and topography. Sediment yield at catchment scale per unit area,was found to be 3-10 times higher than soil lossmeasured in erosion plots. A stepwise regression showed that the dominant factors explaining sediment yield differences at catchment scale in this volcanic landscape were a particular lithology (Old Andesites) and slope angle followed by the silt fraction of the top soil. In lithologically sensitive areas soil loss at the plot scale under monoculture coffee gardens decreases over time from on average 7-11 Mg ha−1yr−1 to 4-6Mg ha−1yr−1,mainly because of the development of surface litter layers as filters and top soil compaction in the areas without litter, but remains higher than under shade coffee systems or forest. The runoff coefficient under monoculture coffee remains on average significantly higher (10-15%) than under forest (4%) or under shade coffee systems (4-7%). In lithologically stable areas soil loss remained below1.8 Mg ha−1yr−1 and the runoff coefficient below2.5% under all land use types, even bare soil plots or monoculture coffee gardens. Less than 20% of the catchment area produces almost 60% of the sediment yield. The reduction of negative off-site effects on e.g. the life timeof a storage reservoirwould benefit greatly from an improved assessment of the lithologies in volcanic landscapes and the consideration of potential sediment source and sink areas. In lithologically sensitive areas, a shift fromsun to shade coffee systemsmay result in reducing surface runoff and soil loss, althoughwater erosion at the plot scale is not themain contributor to sediment yield at the catchment scale. The quantification of land use effects on dominant erosive processes such as river bank and river bed erosion, landslides and the concentrated flow erosion on footpaths and roads can contribute to more targeted efforts and relevant incentives to reduce (or live with) sediment load of the rivers