Uncertainty assessment of spatial patterns of soil organic carbon density using sequential indicator simulation, a case study of Hebei province, China.

The spatial patterns of soil organic carbon (SOC) are closely related to the global climate change. In quantifying the spatial patterns of SOC density, the concept of uncertainty of the SOC density values at unsampled locations is particularly important because such uncertainty can be propagated into the subsequent global climate change modelling and has fundamental impacts on the ultimate results of the model. A total of 361 SOC density data of topsoil (0-20 cm) in Hebei province and sequential indicator simulation (SIS) were applied to perform a conditional stochastic simulation in this study to quantitatively assess the uncertainty of mapping SOC density. The results showed that a great variation exists in the SOC density data. The conditional variance of 500 realizations generated by SIS was larger in mountainous areas of the study area where the SOC density fluctuated the most, and the uncertainty was smaller on the plain area where SOC density was consistently small. Realizations generated by SIS can represent the possible spatial patterns of SOC density without smoothing effect. A set of realizations can be used to explore all possible spatial patterns of SOC density and provide a visual and quantitative measure of the spatial uncertainty of mapping SOC density. With a given threshold of SOC density, SIS can quantitatively assess both local uncertainty and spatial uncertainty of SOC density that is greater the threshold.