Predicting soil organic carbon movement and concentration using a soil erosion and Landscape Evolution Model

Abstract

Explicit relationships and numerical models that link landscape and topographic characteristics with soil organic carbon (SOC) processes are needed. Here we calibrate a digital elevation model based Landscape Evolution Model (LEM) (SIBERIA) and assess both its ability to predict erosion and deposition and the spatial patterns of SOC. The LEM is capable of predicting both erosion and deposition at the hillslope and catchment scale. The LEM is calibrated for the site using a field data approach as well as using a laboratory flume. The predicted soil erosion rates from the LEM (1.7–2.1 t ha−1 yr−1) compare well with independently determined erosion rates using 137Cs (2.1 to 3.4 t ha−1 yr−1). We also investigate field measured and modelled soil organic carbon movement using the LEM in relation to predicted erosion and deposition patterns and find that erosion and deposition patterns are related to the spatial patterns of SOC. This is the first time that a DEM based LEM has been shown to provide reliable prediction of not just soil erosion but also SOC. The results demonstrate that the majority of SOC is being transported in the near surface soil layer (top 2 cm) and that turnover at greater depths is slower and does not correspond with any modelled patterns. The modelled erosion and deposition suggests that on average 0.06 t ha−1 yr−1 of SOC is exported by erosion from the hillslope assuming a good vegetation cover. However if the hillslope is subjected to disturbance (i.e. tillage, overgrazing) then the site will export 0.46 t ha−1 yr-1of SOC. Laboratory results using flume suggest that there was no enrichment of SOC in the eroded sediment. The methods outlined here provide a new approach to quantify the dynamic movement of sediment and SOC at both the hillslope and catchment scale.