Prediction and analysis of the soil organic matter distribution with the spatiotemporal kriging method

For the monitoring of soil fertility and health, soil samples are collected continuously every year in specific regions in China. The high-precision spatiotemporal (ST) continuous distribution of soil properties, e.g., soil organic matter (SOM), must be obtained based on ST soil samples. A total 11,716 soil samples were collected in Zigui County, Hubei Province, China in 2006–2018. First, the experimental variograms of SOM for various spatial and temporal lags were calculated. Second, a hybrid theoretical model integrating variations at the spatial and temporal scales was constructed to model the ST variations in SOM data. Third, the ST ordinary kriging (STOK) method was employed to determine the ST SOM distribution in the study area from 2006 to 2018. Then, the results obtained by STOK were compared with those generated by spatial ordinary kriging (OK) using the soil samples collected in one year. Finally, the temporal trend and change of SOM were analyzed based on the obtained ST distribution data. The results showed that (1) the STOK method performed better than the OK method because the STOK results attained a higher estimation accuracy and a more stable estimation variance in the years with a limited number of soil samples; (2) from 2006 to 2018, the SOM showed significant downward and significant upward trends in 15.05% and 32.05% of the total regional area, respectively. However, compared in 2006, the SOM in 2018 showed significant decreases and increases in 37.2% and 44.57% of the study area, respectively.