Researchers Proposed a New Method for Predicting Soil Organic Carbon Stocks Using High Accuracy Surface Modelling Combined with Soil Depth Information

The soil carbon pool, as the transfer medium for carbon migration and circulation, plays an important role in regulating the greenhouse effect and influencing global climate change. There are temporal heterogeneity and spatial heterogeneity in soil organic carbon (SOC) distributions, which have significant variations over short vertical distances. 

High accuracy surface modelling (HASM) is a surface method based on differential geometry surface theory developed in recent years for geographic information systems and ecological modelling. A combination of HASM and soil depth information will help solve the problem of the spatial prediction of soil properties in profile. 

Prof. SHI Wenjiao and her team members at the Institute of Geographic Sciences and Natural Resources Research of the Chinese Academy of Sciences provided a new method to predict SOC stocks based on the HASM and soil depth information, and compared the accuracy of 16 spatial prediction models (including single models, hybrid models, and HASM combined with single or hybrid models) on the spatial distribution of SOC stocks in Hebei Province (Figure 1). 

The results found that the combination of HASM and the generalized additive model (GAM) together with soil depth covariate (HASM_GAMD) achieved a better performance than other methods at soil depths of 0–30, 0–100, and 0–200 cm. 

The root-mean-square error and coefficient of determination values for predicting the spatial pattern of SOC stocks by the HASM_GAMD model demonstrated an improvement of 43 and 49%, respectively, compared with models without depth information. 

In addition, the prediction uncertainty of the HASM_GAMD model based on 90% prediction interval was lower than that of other models. The authors introduced an alternative method for modelling of SOC stocks and their findings are a valuable reference for assessing carbon stocks to support sustainable land management and climate change mitigation.


Figure 1 Prediction maps of the SOC stock in 0–200 cm using HASM_GAMD method in Hebei Province, China (Image by Prof. SHI Wenjiao’s group)

Download attachments:


SHI Wenjiao


Including soil depth as a predictor variable increases prediction accuracy of SOC stocks