Digital mapping of pre-European soil carbon stocks and decline since clearing over New South Wales, Australia

Abstract

Digital soil models and maps have been developed for pre-European (pre-clearing) levels of soil organic carbon (SOC) over New South Wales, Australia. These provide a useful first estimate of natural, unaltered soil conditions before agricultural development, which are potentially important for many carbon-accounting schemes such as those prescribed by the Intergovernmental Panel on Climate Change, carbon-turnover models such as RothC, and soil-condition monitoring programs. The modelling approach adopted included multiple linear regression and Cubist piecewise linear decision trees. It used 1690 soil profiles from undisturbed or only lightly disturbed native vegetation sites across all of eastern Australia, together with a range of covariates representing key soil-forming factors. The digital soil maps of pre-clearing SOC (% and mass) over New South Wales provide a more sophisticated alternative to currently available, equivalent maps. Independent validation of the SOC mass predictions over the top 30 cm revealed a concordance correlation coefficient of 0.76, which was 13% higher than the currently used map. Total pre-clearing SOC stocks amount to 4.21 Gt in the top 30 cm, which compared with a current stock estimate of 3.68 Gt, suggesting a total SOC loss of ~0.53 Gt over the entire state. The extent of SOC decline in both absolute and relative terms was found to be highly dependent on the climate, parent material and land use regime, reaching a maximum decline of 44.3 t/ha or 50.0% relative loss in cooler (moist) conditions over mafic parent materials under regular cropping use. The models also provide valuable pedological insights into the factors controlling SOC levels under natural conditions.