Mapping soil salinity using a combination of vegetation index time series and single-temporal remote sensing images in the Yellow River Delta, China

Abstract

Accurate mapping of soil salinity is of great significance for agricultural safety. However, soil salinity mapping has proven difficult because of its complex causes. The vegetation index (VI) time series has great potential in soil salinity mapping on a larger scale, yet studies on the use of this index are limited. Therefore, two inversion models (Model A and Model B) are established to assess the effects of VI time series on soil salinity mapping in the Yellow River Delta (YRD). Model A is developed by combining VI time series and single-temporal remote sensing (RS) images, while Model B does not consider VI time series. The effects of VI time series are evaluated from several perspectives by comparing the performances of Model A and Model B. The results show that VI time series can improve the accuracy of soil salinity mapping. The accuracy of Model A (R2 = 0.80, root mean square error (RMSE) = 0.4 mS cm−1) is better than that of Model B (R2 = 0.64, RMSE = 0.68 mS cm−1). In addition, the involvement of VI time series alleviates the overestimation of low-salinity soils and the underestimation of high-salinity soils, eliminating the dividing lines caused by different image times in soil salinity mapping. This study provides a large-scale, high spatial resolution (10 m), accurate soil salinity mapping method. It may help to maintain agricultural and ecological security and promote regional sustainable development.