Detection of the Spatial Variability of Soil Salinity East of the Bitter Lake of the Suez Canal, Egypt, as a Basis for Management Zones Using Sentinel-2 Multispectral Imagery

• Published in: Communications in Soil Science and Plant Analysis Vol. 54; no. 14; pp. 1989 - 2001
• Main Author: Ali, Ali M.
• Format: Journal Article
• Language: English
• Published: Philadelphia Taylor & Francis 06-08-2023; Taylor & Francis Ltd
• Subjects: Agricultural land; Arid regions; Arid zones; Decisions; Electrical conductivity; Electrical resistivity; Fertilizers; Imagery; Interocean canals; Lakes; Management decisions; Management zones; Plant growth; precision mapping; Prediction models; Remote sensing; Salinity; Salinity effects; Soil; Soil conductivity; Soil degradation; Soil salinity; Soils; Spatial variations; spectral indices
• ISSN: 0010-3624; 1532-2416; 1532-4133
• DOI: 10.1080/00103624.2023.2211111

Soil salinity is one of the more serious problems that cause soil degradation, particularly in semi-arid and arid regions. In agricultural lands, it threatens plant growth and productivity, which requires specific management decisions for various inputs, such as fertilizers. Accurate prediction and mapping of soil salinity have become possible through remote sensing techniques. The focus of this research is to derive soil salinity from spectral indices measured from different bands of sentinel-2 satellite in order to delineate management zones (MZs). Four spectral salinity indices were used for the purpose of this study, using blue, green, red, and NIR bands in different combinations. Exponential regression best describes the relationship between salinity indices and soil electrical conductivity (EC) through groundtruthing. The results indicate that salinity indices 2 (
) and 4 (
) (SI2 and SI4) outperformed the other indices and could describe 70% and 71% of the variability in soil EC with prediction models of:
and
, respectively. The generated models were validated by normalized root mean square error and the d coefficient of agreement. The results of the best-fit models were 16.5% and 0.86 for SI2 and 16.7% and 0.84 for SI4, respectively. The box-and-whiskers diagram proposed that the field can be divided into four soil EC-dependent groups, which are considered MZs to support a variety of management decisions. This study shows that through sentinel-2 multispectral imagery, the spatial variability of soil salinity can be mapped and used as a basis for MZs within the studied region.