Spatio-Temporal Mapping of Soil Salinization in the Nile Delta Using Sentinel-2 Multi-Spectral Imagery and Machine Learning

African Journal of Agriculture | Vol. 13, No. 4, April 2026 | pp. 009–016

DOI: 10.46882/2026/AJA/000735

Article Type: Original Research Article

Title: Spatio-Temporal Mapping of Soil Salinization in the Nile Delta Using Sentinel-2 Multi-Spectral Imagery and Machine Learning

Names of Authors: Amgad E. El-Mahdy¹, Fatma H. Al-Sherif¹

Authors’ Affiliations: ¹Department of Soil and Water Sciences, Faculty of Agriculture, Alexandria University, Alexandria, Egypt

Abstract: Soil salinization presents a profound environmental threat to agricultural sustainability in the coastal zone of the Nile Delta. Conventional laboratory testing is highly accurate but logistically slow and resource-intensive for large-scale monitoring. This study explored the efficacy of combining high-resolution Sentinel-2 multi-spectral satellite imagery with three machine learning algorithms—Random Forest (RF), Support Vector Machines (SVM), and Gradient Boosting Machines (GBM)—to predict and map topsoil electrical conductivity (ECe). Soil samples (n = 180) were collected at a depth of 0–15 cm from smallholder farms across Kafr El-Sheikh Governorate during the dry season. Spectral reflectance values, vegetative indices, and salinity indices extracted from cloud-free images were utilized as modeling predictors. Laboratory soil analysis revealed a wide variation in ECe, ranging from 1.25 to 18.64 dS/m. Among the models evaluated, the RF model achieved the highest predictive accuracy, yielding a coefficient of determination (R²) of 0.84, a Root Mean Square Error (RMSE) of 1.12 dS/m, and a Ratio of Performance to Deviation (RPD) of 2.34. The salinity index (SI-1) and the Normalized Difference Vegetation Index (NDVI) were identified as the most influential variables in predicting spatial salinity variations. The generated high-resolution salinity map revealed that 34.2% of the surveyed agricultural zone is severely affected by salinity, driven by sea-level rise and brackish water irrigation. This remote sensing framework enables rapid, cost-effective soil monitoring for targeted regional remediation.

Keywords: Remote sensing, Support vector machines, Random forest, Soil degradation, Nile delta, Electrical conductivity

Manuscript Timeline: Received: January 14, 2026; Revised: February 20, 2026; Accepted: March 12, 2026; Published: April 08, 2026

Citation: El-Mahdy, A. E., & Al-Sherif, F. H. (2026). Spatio-Temporal Mapping of Soil Salinization in the Nile Delta Using Sentinel-2 Multi-Spectral Imagery and Machine Learning. African Journal of Agriculture, 13(4), 009–016. DOI: 10.46882/2026/AJA/000735

Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read