Responses of soil carbon variation to climate variability in China using the LPJ model

Abstract

In this study, we explored the maximal response of soil carbon in a part of China to climate change, including variations in climatology and climate variability, under the condition of global warming. A conditional nonlinear optimal perturbation (CNOP) approach was employed to discuss the above issue using the Lund–Potsdam–Jena (LPJ) model. The variation in the soil carbon was compared with those caused by a linear temperature or precipitation perturbation. The key difference between the CNOP-type and the linear perturbations depended on whether the perturbations brought the variation in the temperature or the precipitation variability in comparison with the reference temperature or the precipitation variability. The model results demonstrated that the variations in the soil carbon resulted from the CNOP-type and linear temperature perturbations in south of the study region, which was corresponding to part of South China, had different variations. We examined three components of the soil carbon in the LPJ model: fast-decomposing soil carbon, slow-decomposing soil carbon, and litter below the ground. The variations of these components derived by the two types of temperature perturbations were different in the chosen region. The reduction in the litter below the ground may be the main cause of decreased soil carbon in arid and semi-arid regions as a result of the two types of temperature perturbations. The different impacts of the two types of temperature perturbations in the south of the study region may be mainly caused by the variations in the fast-decomposing soil carbon. The variations in the soil carbon caused by the two types of precipitation perturbations were similar. In the arid and semi-arid regions, the soil carbon increased due to the two types of precipitation perturbations. This research implies that the variation in temperature variability plays a crucial role in the variations of the soil carbon and its components in the study region.