Dr Ingrid Takken - Current Research

Current Research

For models to be suitable as tool to manage the on- and off-side effects of overland flow and erosion, they need to represent well where runoff will flow, where it may concentrate and where it may connect to the stream network. In line with the previous projects I have worked on, I continue to do research on model development and model improvement in this field.

In co-operation with researchers from different European countries I am working on the development of a model version of LISEM (i.e. LISEM-Breakthrough) that is capable of dynamically modeling flow directions. Although the TCRP model (described above) greatly improves our capability to predict erosion (and deposition) patterns, it still has limitations. The runoff pattern has to be created before LISEM can be run and acts as a static runoff pattern through the simulation of an event. In reality, the runoff pattern dynamically changes during an event, as the flow direction is a function of both oriented roughness and flow depth. LISEM-Breakthrough directly couples the TCRP model to LISEM to model flow directions in a dynamic way. The potential of the model to predict the effects of land management practices (roughness height, tillage direction) on soil erosion patterns will be evaluated.

To be able to model flow directions and locations of flow concentration accurately, high quality input data seem to be a prerequisite. The performance of present day models seem often more limited by the accuracy of the available input data, than by the model description itself. Results from the road-to-stream connectivity model (described above) show that e.g. the error on the calculated distance between a road drain and the stream network can be larger than + 100 meters. Although this could be partly due to the applied GIS algorithms, it is more likely to be related to the quality of the input data, in particular to the accuracy of the DEM (Takken and Croke, 2004). The generally available DEMs used by forest managers are DEMs of 20 or 25 m resolution developed from photogrammetric measurements. For one of the catchments included in my postdoctoral research, a 4 m resolution Lidar DEM is available. I am very interested in evaluating how much better results can be achieved with our present model based on this highly accurate DEM.