A total of 89 legacy soil organic matter (SOM) data (0–20 cm) sourced from the soil survey in 1980 in Hailun County of Northeastern China and 268 SOM data (0–20 cm) measured in the area in 2008 were collected in this study. The most straightforward mapping methods such as ordinary kriging (OK), inverse distance weighting interpolation (IDW), and the soil type-based method (ST) were used to derive the spatial patterns of changes in SOM; and the uncertainties associated with the mapping methods were quantitatively assessed through the uncertainty intervals of SOM realizations that were generated by sequential Gaussian simulation and/or Monte Carlo simulation. Results showed that the overall trend of changes in SOM over the last 28 years in the area decreased. The spatial patterns of changes in SOM derived by OK₂₀₀₈₋OK₁₉₈₀ (abbreviations before and after the dash represent the mapping method for SOM data in 2008 and 1980, respectively) and IDW₂₀₀₈₋OK₁₉₈₀, OK₂₀₀₈₋IDW₁₉₈₀ and IDW₂₀₀₈₋IDW₁₉₈₀, OK₂₀₀₈₋ST₁₉₈₀ and IDW₂₀₀₈₋ST₁₉₈₀ are similar. The change map obtained by the ST₂₀₀₈₋ST₁₉₈₀ method is also similar to those derived by OK₂₀₀₈₋ST₁₉₈₀ and IDW₂₀₀₈₋ST₁₉₈₀ methods, but local details of changes in SOM presented by the ST₂₀₀₈₋ST₁₉₈₀ method are coarser than those obtained by the OK₂₀₀₈₋ST₁₉₈₀ and IDW₂₀₀₈₋ST₁₉₈₀ methods. The OK₂₀₀₈₋ST₁₉₈₀ method, with an uncertainty tradeoff compared to the OK₂₀₀₈₋OK₁₉₈₀ and ST₂₀₀₈₋ST₁₉₈₀ methods, can be considered as the most suitable method for mapping changes in SOM in the area due to the limited legacy soil survey data and inaccurate sample location information records in the area. Copyright © 2014 Springer-Verlag Berlin Heidelberg.
CitationZhao, Y., Xu, X., Hai, N., Huang, B., Zheng, H., & Deng, W. (2014). Uncertainty assessment for mapping changes in soil organic matter using sparse legacy soil data and dense new-measured data in a typical black soil region of China. Environmental Earth Sciences, 73(1), 197-207.
- Soil organic matter (SOM)
- Spatio-temporal change
- Uncertainty assessment
- Sequential Gaussian simulation
- Monte Carlo simulation