Detecting 3D Salinity Anomalies from Soil Sampling Points: A Case Study of the Yellow River Delta, China

Rapidly capturing the spatial distribution of soil salinity plays important roles in saline soils' management. Existing studies mostly focus on the macroscopic distribution of soil-salinity changes, lacking effective methods to detect the structure of micro-regional areas of soil-salinity anomalies. To overcome this problem, this study proposes a 3D Soil-Salinity Anomaly Structure Extraction (3D-SSAS) methodology to discover soil-salinity anomalies and step forward in revealing the irregular 3D structure of soil-anomaly salinity areas from limited sampling points. We first interpolate the sampling points to soil voxels using 3D EBK. A novel concept, the Local Anomaly Index (LAI), is developed to identify the candidate soil-salinity anomalies with the greatest amplitude of change. By performing differential calculations on the LAI sequence to determine the threshold, the anomaly candidates are selected. Finally, we adopt 3D DBSCAN to construct anomalous candidates as a 3D soil-salinity anomaly structure. The experimental results from the Yellow River Delta data set show that 3D-SSAS can effectively identify the 3D structure of salinity-anomaly areas, which are highly correlated with the geographical distribution mechanism of soil salinity. This study provides a novel method for soil science, which is conducive to further research on the complex variation process of soil salinity's spatial distribution.