Towards A Generalized Imaging Spectroscopy Framework for Surface Soil Salinity Retrieval in Coastal Wetlands: Evaluating Transferability and Hydroperiod–Salinity Relationships

Coastal wetlands are threatened by changing climatic and hydrologic pressures that may salinize fresh-brackish ecosystems, causing stress, transformation, or collapse. Monitoring soil salinity requires a generalizable remote sensing approach applicable across coastal environments. We collected in situ field spectrometer reflectance measurements and soil samples in California’s Carpinteria Salt Marsh Reserve (CSMR) to measure salt content. The first derivatives of these reflectances were used in a Partial Least Squares Regression (PLSR) model to predict soil salinity (n = 32, R2 = 0.90, Mean Absolute Error (MAE) = 5.00 ppt). We applied the model to NASA’s AVIRIS-3 imaging spectroscopy data to map soil salinity in the CSMR (Validation: n = 32, R2 = 0.53, MAE = 9.22 ppt) and to AVIRIS-NG data over South Africa’s Langebaan Lagoon to test model transferability (Validation: n = 11, R² = 0.67, MAE = 14.11 ppt). Primary mapped retrievals and associated analyses were restricted to non-submerged, soil-dominated pixels with MESMA-derived soil fractional cover ≥0.50. Reported retrieval performance therefore applies primarily to areas with substantial exposed soil. We used Sentinel-1 Synthetic Aperture Radar time series to map hydroperiod—the percentage of time a site was inundated over the data record—for comparison with the soil salinity maps. Soil salinity increased with hydroperiod at CSMR but decreased with hydroperiod at Langebaan Lagoon, consistent with differing hydrologic connectivity and inundation dynamics. These results demonstrate the potential for a transferable imaging spectroscopy framework for mapping surface soil salinity in coastal wetlands and for quantifying landscape-scale hydro-ecological controls on salinity patterns.