Computer based landscape evolution models: practical applications and issues with functionality and evaluation

Abstract:

Computer based landscape evolution models (LEMs) were initially developed in the 1970s with the first model of Ahnert. Since then they have grown in usefulness and complexity. There are many models available all with their strengths and weaknesses. They provide a tool with which to examine both the natural and anthropogenic environment and have provided significant insights into earth surface processes and dynamics. In some areas of the world (Australia) there is a now a requirement that any new post-mining landform be assessed for its long-term geomorphological behaviour using a LEM before it can be approved to be built. The models are routinely used by consultants for assessment of hazardous waste sites and contaminated land. However, like all models if the right model is not selected for the task and the correct inputs used then an erroneous prediction will occur. The models now have features and functionality beyond the availability of field data for calibration and validation. The new Soilscape models now evolve both the surface and subsurface. There are also many questions about how reliable these models are both in space and time. Here in a series of case studies model use and evaluation is discussed. The need for field data, long term field sites and tools to better assess model prediction are needed given that these models are becoming commonly used.