Assessment of Soil Salinity Problems in Agricultural Areas Through Spatial and Temporal Remote Sensing

Abstract

This study is aimed at addressing the capability of using remote sensing data in detecting and tracking soil salinization variability using a series of experimental methodologies. In a controlled experiment, the spectral reflectance changes associated with salt crust formation on soil surfaces were tracked in order to detect the optimum moisture levels for salinity detection and recognize the influence of soil texture on salinity-induced spectral changes. In another experiment, lettuce plants were utilized to assess plant biophysical responses to moderate salinity levels with canopy-level reflectance data. An FR-ASD spectrometer was used to collect reflectance data in the 400-2500 nm spectral region. Finally, MODIS satellite data were employed to analyze the temporal profiles of selected high (8-11 dS/m), moderate (4-6 dS/m) and none (1-3 dS/m), salt affected sites in the Nile Delta, Egypt. The analyses of spectral data revealed that the use of remote sensing data to discriminate salinity levels in soils is highly affected by moisture content and texture. At low moisture contents, salts have high reflectance in the VIS-NIR spectral region but low reflectance in SWIR region. Spectral ambiguity with soil salinity was found across soil texture types.Significant relationships were found between plant chlorophyll content and the REP index (R2 =0.97), and dry biomass with SAVI values (R2 =0.94) under different salinity treatments. The spectral vegetation indices (VI's), SAVI and REP, and water indices (WI's) were found to be effective in discriminating between plants growing under moderate conditions of soil salinity and a non-saline condition. The combination between VI's and WI's was found to be useful in improving the ability to assess salinity stressed plants from non-stressed plants. Finally, MODIS results showed separability between canopy seasonal growth under high saline (HS) and non-saline (NS) conditions based on phenology. Canopies growing under HS conditions had lower VI and WI values in the green-up period. It may be concluded that using plant biophysical response to detect soil salinity could be useful in detecting early stages of salinity. Also using the combination between VI's and WI's using MODIS data is a useful to discern between high saline and none saline areas.