Multi-decadal remote sensing observations evaluating the spatiotemporal patterns of seawater intrusion impact on a crop-rotation system

Seawater intrusion (SWI) threatens coastal agricultural systems worldwide, particularly under climate change and water scarcity. The Po River Delta (Italy), one of the most vulnerable coastal regions in the Mediterranean, has experienced recurrent SWI, with severe implications for crop production and livelihoods. However, despite its importance, no systematic, long-term assessment of the spatial and temporal impacts of SWI on agriculture in the region has been undertaken. This study addresses this gap by integrating multi-decadal Landsat surface reflectance data (2000–2024) with a machine learning framework (XGBoost) to classify major summer crops (soybean, rice, maize, and alfalfa). Three seasonal spectral indices linked to crop stress and salinization were computed from June to September to capture vegetation stress patterns. The analysis provided the first spatially explicit assessment of crop dynamics and salinity stress in the Po Delta over 25 years. Results reveal that 35,000–41,000 ha of cropland exceeded critical thresholds depending on index and month, with June consistently emerging as the most vulnerable period. Recurrent hotspots were concentrated along coastal zones, while rice was the crop reaching critical values of stress more frequently. In addition, a two-period comparison (2000–2012 vs. 2013–2024) highlighted worsening conditions near the coastline. Findings offer a novel framework to identify salinity-prone areas, guide crop planning, and resilience against SWI and climate change.