Associate Professor Matthias Vanmaercke from Liège University in Belgium will give a talk entitled 'Modelling Gully erosion rates at regional and continental scales: Challenges and Opportunities'.
Gully erosion is a dominant geomorphic process in many regions worldwide. Depending on the geographical setting, gullies can be held responsible for easily 30% to 90% of the total amount of sediments exported from river catchments. Furthermore, gullies can strongly increase sediment connectivity, leading to a further indirect increase in catchment sediment yields. As such, gully erosion often induces a plethora of problems, including the direct loss of soil and land, water reservoir capacity losses, increased flooding risks, but also decreased crop yields, damage to infrastructure and casualties. Several studies indicate that GE will strongly increase in most areas (and in particular Africa) as a result of increasing land use pressure and potentially climate change. Being able to quantify and predict gully erosion is therefore of essential importance, including for the development of suitable land use and catchment management strategies. Nonetheless, our ability to do so currently remains very limited. Especially at regional to continental scales. Various local models have already been proposed. However, their applicability to larger areas is generally limited due to data constraints and/or the limited range of environmental conditions for which the model is calibrated.
We therefore discuss some recent advancements in the development of tools that allow to predict average gully erosion rates (GE, [t km-2 y-1]) at regional to continental scales. Similar to the Revised Universal Soil Loss Equation, we aim to develop a model that provides realistic and meaningful estimates of gully erosion at annual to decadal time scales (allowing scenario analyses) while keeping the data requirements as low as possible. To achieve this goal, we propose a model structure where total GE is simulated as the product of the gully density of an area (expressed as the number of gully heads per km²) and the average retreat rate of each gullyhead. With respect to the later, a review and meta-analyses of measured gully headcut retreat rates (GHR) worldwide, shows that average GHR is mainly controlled by the area contributing to each gully head and, especially, rainfall intensity. The combination of both factors allows to explain nearly 70% of the observed seven orders of magnitude of variation in GHR worldwide. Given the very strong observed correlation between rainfall intensity and GHR, our analyses also show that gully erosion rates are likely very sensitive to climate change. To develop a model that can predict the spatial variation in gully density at continental scales, we constructed a database of >1000 study sites across Africa where all individual gully heads were mapped based on available aerial photos (resulting in a total of > 45000 individually mapped gullyheads). Overall, gullyhead densities in Africa vary between zero and 1500 gully heads per km². Statistical analyses show that this density is mainly explained by differences in vegetation cover, topography and soil characteristics. A regression model based on these variables is capable of explaining around 50% of the observed variation in gully density.
A potential avenue to further improve the accuracy of gully density predictions is likely to apply a method that combines the slope-area threshold concept of gully initiation with a simple runoff model. First analyses indicate that this is feasible at the continental scale, using currently available datasets. Such approach would not only permit to better account for the spatial patterns and interactions between the factors controlling gully initiation. It would also likely allow to improve the prediction of GHR and directly couple both model components into an integrated model that predicts total GE.
Overall, our results indicate that especially GHR exerts a strong control on total average GE. Given that GHR can already predicted to a reasonable extent, this suggests that the development of a model that can simulate regional and continental patterns of gully erosion is becoming feasible. Nonetheless some important challenges remain. These include: differentiating between active and none-active gully heads, the large temporal variability that typically characterizes gully erosion, the prediction of ephemeral gullies and the contribution of other gully erosion processes (e.g. piping, gully widening, gully deepening) to total GE.
Matthias Vanmaercke is a geographer focussing on modelling and understanding geomorphic processes at larger spatial scales, with a strong focus on Africa. He completed his MSc and PhD in Geography at the University of Leuven (Belgium). Since November 2016, he is professor at the University of Liège (Belgium). In April 2018, he received the Outstanding Early Career Scientist Award from the European Geosciences Union – Division of Soil System Sciences for ‘his outstanding research on understanding the rates and impacts of soil erosion and catchment sediment export’ at regional and continental scales.’