Comparative analysis of spectral indexes for soil salinity mapping in irrigated areas in a semi-arid region, Brazil

Abstract

Salinization reduces agricultural yield and sustainable development in the long term. In this study, we identified the spectral index that best estimates soil salinity in the area of the Baixo-Açu irrigated perimeter in northeast Brazil. We also established how vegetation canopy and satellite image spatial resolution influence the results. The electrical conductivity of 42 sampling points was determined and correlated with 20 spectral salinity indexes. The spatial resolution influence on index response was evaluated using OLI/Landsat8 and MSI/Sentinel2 satellite images, with spatial resolutions of 30 and 10 m, respectively. They were applied to bare soil and vegetation canopy areas to assess the vegetation canopy influence on the indexes. Normalized Differential Vegetation Index greater than 0.33 indicated soils with vegetation canopy. In all analyses performed, the bare soil areas had the best results. From the 18 analyzed indexes, Sentinel2 images outperform Landsat8 images. The Soil Index-1 showed an 80.34%-determination coefficient that outperformed the other 19 indexes in soil salinity identification. These results indicate that the association of the Soil Index-1, Sentinel2 images, and the Normalized Differential Vegetation Index less than 0.33 improves the evaluation of the salinized areas.