Predicting spatial patterns and intensity of soil erosion and sediment redistribution in landscapes can be problematic in areas where few experimental data are available. Yet it is in those data-scarce regions that the ability to extrapolate local field evidence could be most useful in practical applications. Conventionally, soil conservationists and policy makers look at the field scale to judge the on-site impacts of soil erosion on the catchment. Water resource managers, in contrast, usually rely on sediment yield data of gauged catchments to evaluate the off-site effects of erosion intensity and sediment production. In fact, a scale gap can be observed in the approaches used by respectively land and water resources managers. The increasing availability of satellite data, with spatial resolutions ranging from fine to coarse and temporal resolutions from hourly to monthly, provide a significant information source for mapping, monitoring and predicting current land degradation, as well as for monitoring signals that describe spatio-temporal variations in soil surface characteristics. Moreover, the integrated use of experimental field data with remote sensing imagery allows field evidence to be extended to larger areas. On the other hand, owing to the increased capabilities of existing GIS platforms, distributed erosion models have evolved and are now commonly used as the georeferenced backbone for the analysis of hydrological and erosion processes. However, the accuracy of their predictions is seriously hampered by the natural complexity and spatial heterogeneity of the processes acting on the landscape itself, coupled with the limited availability of spatio-temporal datasets in ungauged catchments. Therefore, the aim of this thesis was to use and further develop existing spatially distributed approaches in combination with different geo-information techniques for analysing erosion and sedimentation processes, allowing linkages to be defined between field and basin scales in data-scarce environments. In this study, conceptual models were selected because they represent a compromise between complex data demanding (physically based) models and simple-structure regression (empirically based) models. The province of Cochabamba in the Central Andes of Bolivia was the focus of this research. The region of the middle and upper valleys of Cochabamba plays an important role in the agricultural sustainability of rural and urban livelihoods. Owing to the semiarid regime and topographical and geological conditions, however, the impacts of soil erosion are clearly visible and are leading to an increase in the land degradation processes. Five conceptual model approaches were selected and used, and the results compared regarding their ability to represent erosion and sedimentation phenomena at the regional scale. Hence, erosion and deposition maps for Cochabamba province, with an area of about 54,100 km 2 , were derived using the global climate, topography, soil datasets and low-resolution satellite images for land cover. A categorical comparison method was used to validate model predictions based on existing thematic mapping of erosional
