Agriculture has been implicated as a major source of sediments in South Africa. The aim of the knowledge gap analysis was to understand the production and delivery components of agriculturally derived sediments under South African conditions and to assess the predictive ability to address the fate of these sediments from field to catchment scales. An overview is given of important erosion processes and erosion modelling applied in South Africa at the field and catchment scale. A limitation of the sediment models is that gully erosion is not simulated; therefore, the models should be complemented with gully erosion predictions if gullies are an important sediment source. Field-scale models inadequately predict sediment production localised at hydrologically sensitive areas as a result of saturation excess flow and/or throughflow. The discussion on erosion modelling reveals that more complex models have had limited application in South Africa because they require large and detailed data sets, and may have parameters that are difficult to measure or to estimate. A modelling framework is discussed which allows linking of sediment models requiring readily available data, gully erosion models/maps and the use of other techniques to assess the fate of agriculturally derived sediments from field to catchment scale.