A Spatiotemporal Dataset of Soil Properties in Northeast China Based on Soil Sampling and Interpolation From 2009 to 2020

The Northeast region of China, serving as a crucial hub for grain production and an ecological security barrier, confronts significant challenges such as soil degradation and nutrient imbalance. Addressing the need for dynamic soil quality monitoring in the major grain-producing areas of Northeast China, this study innovatively develops a spatiotemporal sparse grid modelling framework and produces high-precision soil spatiotemporal datasets, based on soil testing and fertiliser recommendation data collected from various locations between 2009 and 2020. By integrating a spatiotemporal covariance function with the Kriging interpolation algorithm, the study systematically resolves the challenge of spatiotemporal collaborative modelling for multi-year discontinuous observational data. Consequently, continuous spatiotemporal datasets for soil pH, soil organic matter (SOM), total nitrogen (TN) and available potassium (AK) at a 500-m resolution in Yian County were successfully reconstructed. Various error metrics, including RMSE, MAE, MAXE, MINE and SE were employed to verify the high accuracy and reliability of the spatiotemporal Kriging interpolation method, with the relative error controlled at a minimum of 0.04. Geodetector analysis revealed significant spatial variability in soil properties (q > 0.8, p < 0.001). A spatiotemporal trend analysis framework, coupling Theil-Sen Median with Mann-Kendall, quantitatively demonstrated significant decreasing trends in pH, SOM and TN during the study period (with decreasing area proportions of 49.02%, 47.32% and 43.17%, respectively), while AK exhibited a significant increase of 41.96%. The spatial variability patterns were highly coupled with the spatial gradient characteristics of agricultural management measures. This dataset transcends the limitations of traditional static soil databases in spatiotemporal representation. Through a high-precision spatiotemporal continuous modelling technique system, it provides multi-scale spatiotemporal benchmark data support for precision agriculture, optimising conservation tillage of black soil, and simulation of agricultural carbon neutrality pathways. It holds significant scientific value for the sustainable management of farmland ecosystems in the context of global change. This dataset can be downloaded from https://doi.org/10.5281/zenodo.13978751.