Combining sediment source tracing techniques with traditional monitoring to assess the impact of improved land management on catchment sediment yields

This paper aims to demonstrate the potential value of combining sediment source tracing techniques with traditional monitoring approaches, when documenting the impact of improved land management on catchment sediment yields. It reports the results of an investigation undertaken in a small (1.19 km²) agricultural catchment in southern Brazil, which was monitored before and after the implementation of improved land management practices. Attention focussed on 50 storm events that occurred between May 2002 and March 2006 and which reflected the behaviour of the catchment during the pre-change, transition and post-change periods. Improved land management, involving minimum-till cultivation and the maintenance of good crop cover, was introduced in early 2003. The traditional monitoring provided a basis for evaluating the changes in storm runoff volume, storm hydrograph peak and storm-period sediment load and mean suspended sediment concentration. The results indicate that both storm runoff volumes and peak flows associated with a given amount of rainfall provided evidence of a significant decrease after the introduction of improved land management. Storm-period sediment loads showed a similar reduction, with a reduction by as much as 80% for low magnitude events and of ca. 40% for events of intermediate magnitude. However, there was no significant change in mean suspended sediment concentrations, indicating that the reductions in sediment load were primarily the result of the reduced storm runoff volume. Sediment source fingerprinting was used to explore the changes in the relative and absolute contributions to the storm sediment loads from the three key sources, namely the surface of the fields under crops, the unpaved roads and the stream channels. A comparison of the load-weighted mean contributions for the pre- and post-treatment periods indicated that the contribution from the field surfaces and unpaved roads decreased from 63% and 36% to 54% and 24%, respectively, whereas the contribution from the stream channels increased from ca. 2% to 22%. By relating the absolute amounts of sediment mobilised from each individual source group to variables representing the runoff and precipitation associated with the events, it was possible to identify changes in the response of the individual sediment sources to the changes in land management that occurred within the catchment. Sediment mobilisation from the stream channel during individual events increased substantially over the whole range of flows after the introduction of improved land management in the study catchment, whereas the amounts of sediment mobilised from the surfaces of the fields and the unpaved roads showed a significant decrease during events of low and intermediate magnitude. The short monitoring period associated with the study, coupled with inter-annual variations in rainfall, necessarily limit the scope and rigour of the study reported, but it is seen to provide a useful demonstration of how the coupling of sediment source tracing with more traditional monitoring techniques can provide an improved understanding of the impact of improved management practices on the sediment response of a catchment, as well as important information to inform the design and implementation of effective sediment management and control measures.