Recent advances in modeling depth distribution of soil carbon storage

Abstract:

Depth distribution of soil carbon storage determines the sensitivity of soil carbon to environmental change. We present different approaches that have been used to represent the vertical heterogeneity of soil carbon both in mapping and modeling studies. In digital soil mapping, many studies applied exponential decay functions in soils where carbon concentration has been observed to decline with depth. Recent studies used various forms of spline functions to better represent the vertical distribution of soil carbon along with soil horizons. These studies fitted mathematical functions that described the observations and then interpolated the model coefficients using soil-forming factors and used maps of model coefficients with depth to predict the SOC storage at desired depth intervals. In general, the prediction accuracy decreased with depth and the challenge remains to find appropriate soil-forming factors that determine/explain subsurface soil variation. Models such as Century, RothC, and Terrestrial Ecosystem Model use the exponential depth distribution functions of soil carbon in their model structures. In CLM 4.5 the soil profile is partitioned into 10 layers down to 3.8 m depth and the carbon input from plant roots is assumed to decrease following an exponential function. Not accounting for soil horizons in representing biogeochemistry and the assumption of globally uniform soil depth remain major sources of uncertainty in these models. In this presentation, we will discuss the merits and demerits of using various profile depth distribution functions to represent the vertical heterogeneity of soil carbon storage.