The impact of soil layering in a mechanistic soil-landscape evolution model

Mankind has been actively reshaping and altering the soil system since the dawn of agriculture in the Neolithic. As our impact on the soil system has grown throughout the millennia, so has the need to understand its complex dynamics. This has led to the creation of mechanistic soil-landscape evolution models to simulate the long-term development of soils and landscapes. These models discretize the soil column in layers, which represent how the soil profile is handled in the model, rather than real geopedological horizons. As a result, the layer thickness directly impacts the resolution of the geomorphic and soil forming processes being modeled. It is still unclear how different layering options impact the outcome of the simulation and therefore which discretization would be preferable when developing or choosing a soil-landscape evolution model.

 

This work aims to bridge this gap by investigating how different soil layering options impact simulations of soil thickness, organic matter content, soil texture and computational efficiency.

We tested the impact of a) the number of layers and b) the thickness of these layers. The former was investigated by comparing simulations carried out with 2, 8 and 16 layers, while the latter was explored by simulating layers of uniform thickness (UT) and layers which increase in thickness with depth (IT), which provides a higher resolution closer to the surface. To ensure a general yet realistic setting, the model was applied to a silty soil profile from Canneto Pavese, in the Oltrepò Pavese region (Italy).

For this work, the mechanistic soil-landscape evolution model LORICA was chosen due to its ability to simulate both geomorphic and soil forming processes in the entire soil profile and for its ability to simulate soil layers with varying and dynamic thickness. We simulated the processes of bedrock weathering, physical and chemical weathering and carbon cycle dynamics and observed their impacts on soil properties 10, 100, 1000 and 10000 years of calculation.

 

We found that a greater vertical interaction between layers resulted in differences in the outputs, which occur both when a higher number of layers is adopted, and in the IT mode compared to the UT mode.

This work provides insight into the impact of layering options in soil-landscape evolution models, so that other researchers will be able to select the most apt and efficient set up for their simulations depending on their specific circumstances and needs. In the future, the action of water erosion will additionally be assessed on a landscape scale to consider spatial variations in soil development as well.