Seal formation and interrill erosion on a smectite-rich Kastanozem from NE-Mexico

On-site and off-site effects of interrill erosion can be highly significant for river and reservoir sedimentation and can cause serious water contamination. Interrill erosion changes during storms, reflecting the effect of seal formation on soil resistance to interrill erosion. Swelling clay soils appear to be particularly complex in their response to rainfall. Assessment of interrill erosion hazards on such soils requires effective quantitative description of the effects of seal formation on interrill erosion. However, short-term changes of seal properties are not reflected well in current event-based erosion models, leading to poor prediction quality. The objective of this study was to identify the changes in seal condition relevant to temporal change in interrill erosion on a smectite-rich Kastanozem from NE-Mexico. Formation and removal of a non-coherent washed-out layer, consisting of stable aggregates, was the most apparent change in seal condition associated with interrill erosion dynamics. The progressive exposure of a more cohesive washed-in layer resulted in reduced interrill erosion, which is attributed to lower erodibility of the washed-in layer. The results demonstrate that seal formation and interrill erosion dynamics of a freshly tilled Kastanozem are strongly linked. The amount of rainfall required for the removal of the washed-out layer (180 mm) shows that seal formation lasts for a significant part of the rainy season. The changes of interrill erosion caused by seal formation on the Kastanozem are not represented adequately in current methods of erodibility assessment and improving the quality of event-based erosion prediction remains difficult for two reasons. First, relationships between soil properties, seal formation and erosion dynamics are not well understood. Second, the quantitative description of seal property dynamics relevant for interrill erosion is characterized by strong spatial and temporal variability. Therefore, it is likely that even with best modeling efforts, the quality of event-based erosion prediction will be limited by the high uncertainty about erodibility. (C) 2002 Elsevier Science B.V All rights reserved.