Find in Library

Abstract Soil carbon is becoming increasingly difficult to predict due to uncertainties in climate warming. The main objective of this research is to estimate the scope of uncertainties in soil carbon, which is the difference between maximal and minimal estimations of soil carbon, caused by uncertainties in climate change in China, by applying the Lund–Potsdam–Jena (LPJ) model and an improved optimization method. Based on an original optimization method (conditional nonlinear optimal parameter perturbation, CNOP‐P), a cost function is revised to evaluate the scope of uncertainties in predicting soil carbon sequestration, avoiding the overestimation of uncertainties. Two types of climate change projections are obtained from 10 global circulation models (GCMs) under the Representative Concentration Pathway (RCP) 4.5 scenario using two methods. For the first type of climate change projection, the upper and lower limits of estimated future soil carbon sequestration are 102.7 and 65.5 Gt, respectively, whereas the reference soil carbon sequestration driven by an ensemble of 10 GCMs is 79.6 Gt. The estimated future soil carbon sequestration ranges from 79.4 to 91.6 Gt for the second type of climate change. The numerical results imply that these two types of climate change, constrained by the projection of 10 GCMs, are the main factors affecting the uncertainty range of estimated soil carbon.