Assessing top- and subsoil organic carbon stocks of Low-Input High-Diversity systems using soil and vegetation characteristics

Abstract

The soil organic carbon (SOC) stock is an important indicator in ecosystem service assessments. Even though a considerable fraction of the total stock is stored in the subsoil, current assessments often consider the topsoil only. Furthermore, mapping efforts are hampered by the limited spatial density of these topsoil measurements. The aim of this study was to assess the SOC stock in the upper 100cm of soil in 30,556ha of Low-Input High-Diversity systems, such as nature reserves, in Flanders (Belgium) and compare this estimate with the stock found in the topsoil (upper 15cm). To this end, we combined depth extrapolation of 139 measurements limited to the topsoil with four digital soil mapping techniques: multiple linear regression, boosted regression trees, artificial neural networks and least-squares support vector machines. Particular attention was given to vegetation characteristics as predictors. For both the stock in the upper 15cm and 100cm, a boosted regression trees approach was most informative as it resulted in the lowest cross-validation errors and provided insights in the relative importance of predictors. The predictors of the stock in the upper 100cm were soil type, groundwater level, clay fraction and community weighted mean (CWM) and variance (CWV) of plant height. These predictors, together with the CWM of specific leaf area, aboveground biomass production, CWV and CWM of rooting depth, terrain slope, CWM of mycorrhizal associations and species diversity also explained the topsoil stock. Our total stock estimates show that focusing on the topsoil (1.63Tg OC) only considers 36% of the stock in the upper 100cm (4.53Tg OC). Given the magnitude of subsoil OC and its dependency on typical ecosystem characteristics, it should not be neglected in regional ecosystem service assessments.