Evaluating the Atmospheric Correction Impact on Landsat 8 and Sentinel-2 Data for Soil Salinity Determination

Abstract

Remote sensing technology effectively determines and evaluates salinity-affected areas' spatial and temporal distribution. Soil salinity maps for large areas can be obtained with low cost and low effort using remote sensing methods and techniques. Remote sensing data are delivered raw as Level-1 data, and they can be further atmospherically corrected to surface reflectance values, Level-2 data. This study evaluates the atmospheric correction impact on Landsat 8 and Sentinel-2 data for soil salinity determination. The study has been supported with in-situ measurements in Alpu, Eskisehir, Turkey, where samples were collected from various agricultural fields simultaneously with the overpass of the satellites. 71 to different analysis cases have been used to determine the effect of atmospheric correction. The first is to examine the relationship between the measurements taken front the areas with mixed product groups and the salinity indices for both data types. The other is to investigate the relationship between the measurement values taken only from the wheat and beet groups and the salinity index values. The results show that atmospheric correction has a high effect on the relationship between spectral indices and in situ salinity measurement values. Especially in all cases examined in Landsat, it was observed that atmospheric correction led to an improvement of over 140%, while nearly 50% was observed in Sentinel on a product basis.