Soil salinization is a major threat to agricultural sustainability in arid and semi-arid regions, leading to the progressive loss of fertile land. The irrigated El Habibia–Mansoura zone in the Mejerda low valley (NE Tunisia) is highly affected by salinity due to drought, dam construction, poor irrigation water quality, and insufficient drainage. This study employed Google Earth Engine (GEE) machine-learning algorithms to map and monitor soil salinity over a 24-year period (2000–2023). Landsat imagery was processed using positively correlated salinity indices (SI), Shuttle Radar Topography Mission (SRTM) data, negatively correlated indices such as NDVI and Iron Index (IIn), MODIS-derived climatic features, and ground control measurements. A Support Vector Machine (SVM) model achieved robust predictive performance (R2 = 0.82; RMSE = 1.34). Results reveal strong spatio-temporal variability in salinity, with the most pronounced peak between 2010 and 2015, when 46.72% of the area was moderately saline and 1.92% highly saline, reflecting high evapotranspiration and low precipitation. Extremely saline soils, although limited in extent, showed a gradual but consistent increase from 0.03% in 2000–2005 to 0.12% in 2020–2023. Iron nutrient availability sharply declined after 2013, correlating negatively with salinity, while alkaline pH (> 7) and high Na⁺ and Cl⁻ contents further constrained soil fertility. These findings confirm that salinity dynamics are driven by both climatic and anthropogenic factors, including irrigation practices, groundwater salinity, and land management. Remote sensing integrated with machine learning offers reliable monitoring tools, providing essential decision-support for sustainable land management and food security in Tunisia and worldwide.
扫码关注我们
求助内容:
应助结果提醒方式:
