An evaluation of landscape evolution models to simulate decadal and centennial scale soil erosion in grassland catchments

SummaryThere are a number of landscape evolution models now developed to the stage where they can be routinely used for both geomorphic evaluations and quantification of soil erosion rates and processes when subject to the action of rainfall and runoff. These models have considerable advantages over traditional modelling approaches as they remove the need to manually determine slope length and angle and because they can also determine both erosion and deposition. A further advantage of digital elevation based models is that they dynamically adjust the landscape in response to erosion and deposition, thus producing a better representation of slope length and angle over the duration of the simulation. A recent advance is that these models now have the ability to employ spatially variable hydrological and erosion parameters, the spatial distribution of soil particle size at user defined soil depths as well as several different flow direction algorithms. While these options are available in these models, minimal evaluation of these hydrological and geomorphological functions has taken place to assess whether they are correct. This study evaluates the well known SIBERIA and CAESAR models for their ability to predict landscape form and erosion for a grassland catchment in South-East Australia under similar rainfall conditions. The results demonstrate that both models predict similar hillslope form as well as erosion rates over a 1000-year modelled period. They also predict erosion rates within the range of independently determined field measured data using environmental tracers at decadal time scales for the site and region demonstrating the models are reliable in the setting examined here.

Research highlights
► Numerical landscape evolution models can be used for quantification of erosion rates.
► The SIBERIA and CAESAR models are evaluated here for a grassland catchment.
► Results demonstrate that both models predict similar hillslope form and erosion rates.
► Erosion rates are also within the range of independently measured data.
► The models are demonstrated to be reliable in the setting examined here.