Gully development, evolution and erosion using a landscape evolution model

Gullies are a recognised driver of landscape evolution. They are common features not just in natural landscape systems but also in agricultural and post-mining landforms. It is therefore necessary that gully initiation, development and evolution be thoroughly understood for both improved land surface process understandings but also for improved landscape management. Computer based landscape evolution models that use a mesh of grid cells to represent a catchment have the potential to model the evolution of transient features such as gullies. Landscape evolution models use a DEM and have the designed so that they adjust elevation to both erosion and deposition. Therefore they are able to better represent the effect of changing hillslope and catchment morphology on erosion. This process is dynamic. In this study the SIBERIA model is used to evaluate gully evolution. SIBERIA has been extensively used and has been shown to be able to model long-term whole-landscape evolution as well as predict hillslope and catchment scale erosion rates. This is the first evaluation of the model for its ability to predict erosion features such as gullies at the catchment scale. The results show that the model is able to largely predict gully formation as well as morphology. This is the first time that a landscape evolution model has been used to examine gully development in an undisturbed catchment. It is important to understand gully initiation, development and evolution as they are drivers of landscape change and erode and transport considerable volumes of sediment through the channel network (Wainwright et al., 2006). They are common features not just in natural landscape systems but also in agricultural and post-mining landforms. Post-mining landforms are of particular interest as they are often constructed at steeper slopes than the original landscape and have little vegetation compared to natural surrounds. They are also constructed of unconsolidated materials and can be prone to gullying. A further important issue is that they often contain fine material from mineral processing (i.e. tailings) that can be readily eroded if exposed while metalliferous mines often have acid-generating material that requires long term coverage. Also in the case of uranium mines these materials can be radioactive. It is therefore necessary that gully initiation, development and evolution be thoroughly understood for both improved land surface process understandings but also for improved landscape management. Many of the models developed have been developed with a single focus such as soil erosion quantification, headward growth rate prediction or gully risk assessment (Poesen et al., 2003). Computer based landscape evolution models (LEMs) offer many of the features of soil erosion and gully prediction models and can offer insights into the behaviour of landscape features such as gullies. LEMs that use a mesh of grid cells to represent a catchment were initially developed in the 1970s. With ever increasing computer processing power these models evolved to encompass larger domains and more complex processes. The models have been widely used to examine landscape development over geological time scales. They also have the potential to model the evolution of transient landforms and features such as gullies (Hancock et al., 2000; 2008). Here the SIBERIA model is used to evaluate gully evolution in the Tin Camp Creek catchment, Northern Territory, Australia. SIBERIA has been extensively used in the region and has been shown to be able to model long-term landscape evolution (Hancock et al., 2002) as well as predict hillslope and catchment scale erosion rates (Hancock et al., 2008; 2010). This paper evaluates the model for its ability to predict transient erosion features such as gullies.