Modelling soil landscape genesis — A “time split” approach for hummocky agricultural landscapes

Abstract

Here we present a scheme for modelling soil landscape development in hummocky agricultural landscapes of formerly glaciated terrains. Soil landscape development is regarded as discontinuous over time. Long-term progressive soil development during the Holocene was replaced abruptly by an intense period of regressive pedogenesis during the last 50 years, which is deduced from published 137Cs data and catchment scale mass balances. Consequently, our primary working hypothesis is that the recent soil pattern is mainly a result of human influence during the last 50 years. The paradigm behind our approach is that we have to couple geological–historical information with recent dynamics of relevant processes to understand soil landscape development. Spatially-distributed erosion and sedimentation are the most important processes during that period. The intensity and extent of these processes is, however, also a function of the spatial pattern of soils (erodibility), which evolved in the progressive period. Therefore we developed a “time split” modelling approach. For the progressive period relevant pedogenic processes (de-/carbonatization, silicate weathering, clay translocation) have to be quantified by a spatially-distributed, coupled water flow and solute transport model based on reconstructed initial and upper boundary conditions (geomorphic units, climate history). Relevant outputs, e.g., topsoil textural properties, are used as a spatially-distributed input for the model of the regressive period. For the latter we use a modified version of the dynamic, physically based EROSION-3D model [Schmidt, J., von Werner, M., Michael, A., 1996. EROSION 2D/3D — a computer model to simulate water erosion (in German). Sächsische Landesanstalt für Landwirtschaft, Dresden.], which allows spatially-distributed modelling of water-related soil erosion and deposition. Relevant topsoil properties are iteratively adapted to the rates of spatially-distributed erosion and deposition. Preliminary results suggest that a coupling with tillage erosion plays an important role in our landscapes. Explicit analyses of the soil pattern using digital soil mapping techniques will be applied for model validation.