Predicting the impact of Farming Systems on Sediment Yield in the Context of Integrated Catchment Management
1059/1/03
December 2003

EXECUTIVE SUMMARY

1 MOTIVATION

Land disturbances associated with agriculture are one of the major sources of sediment. It is increasingly accepted worldwide that erosion control should be linked to both soil (on-site erosion) and water (off-site sediment) conservation initiatives. Predicting the on- and off-site impacts of agricultural practices and erosion control measures requires models in which we have some degree of confidence. Verification of these models, therefore, is critical. Models predicting the off-site effects of farmland erosion at the small catchment scale have found limited application in South Africa to date. The focus of this project was to verify selected models predicting the impact of farming practices on sediment yield at the hillslope to small catchment scale.

2 OBJECTIVES

The overall aim of the project was to verify the performance of selected models in predicting the impact of farming practices on sediment yield at the small catchment (<10 km2) scale in order to get an indication of the degree of confidence with which the off-site impacts of agricultural practices and erosion control measures can be predicted with the selected models. In order to achieve this aim, the following specific objectives were identified:

1. To improve international and South African methodologies for predicting the impact of selected land uses on sediment yield in local catchments.
2. To assess the impact of the following land uses on sediment yield: communal grazing, communal cropping, commercial grazing and pristine areas.
3. To assess the impact of quality, availability and the spatial distribution of input data on the accuracy of sediment yield prediction results.

The emphasis of the first objective was on verifying the performance of the selected models by comparing model predictions against measured data with those using the best available model input data (model verification). Verification was done by comparing model predictions against observed soil water contents at selected points in the catchments and runoff and sediment yield discharged at the catchment outlets.

The focus of the second objective was to get some indication of the ability of the selected models to predict sediment yield for typical South African agricultural practices by assessing the impacts through land management scenario predictions for the land uses occurring in the catchments.

Since the models will be applied in research catchments with detailed data, assessment of the impact of the availability and the spatial and temporal resolution of input data on the accuracy of sediment yield prediction results was verified (objective 3) in order to assess the performance of the models with input data with coarser spatial and temporal resolutions.