Zoning the soil salinization levels in the northern China’s coastal areas based on high-resolution soil mapping

Abstract

Zoning the soil salinization levels in large-scale coastal areas is vital for understanding the spatiotemporal mechanism of soil salinization and guiding the coastal integrated management. However, it is difficult due to the complicated influencing factors and the high demands for a fine mapping resolution and precise simulation results. In the present study, the northern China’s coastal areas (> 6 4000 km²) were selected as the study area, and extensive field investigation, multispectral remote sensing images, and open-source land cover data served as the data source. A simulation unit of 100 m × 100 m was employed to precisely map the soil salinity (SS) based on a predictor system that covered different aspects of influencing factors, and high-resolution soil salinization maps were generated for the entire study area and different cities. Results indicated that the mapping obtained a relative-root mean squared error of 0.31, which was in a low level and denoted a high accuracy compared with previous studies. The soil salinization levels presented the following spatial heterogeneities: (1) The levels showed distinct polarization, that is, extremely low and high levels covered the most of the study area. (2) The SS exhibited a distinct decrease from the coastline to the inner land, and wetlands and water areas exhibited much higher SS than the remaining land cover types. (3) Muddy coasts suffered more severe soil salinization than rocky and sandy coasts in the alongshore areas. (4) The inner land was generally free from the salinization but some small patches of bare lands and water areas were still at risk. (5) Dongying, Binzhou, Weifang, and Yancheng Cities suffered much more distinct soil salinization than the remaining cities. The coastline and ecological quality were the major factors that determine the spatial pattern of soil salinization in the alongshore and inland areas, respectively.