Coastal ocean monitoring stands to gain a sharper view of freshwater dynamics under a dual-pathway framework proposed by researchers from Ocean University of China, the National Satellite Ocean Application Service, and the Institute of Oceanography, Chinese Academy of Sciences. Published in the Journal of Remote Sensing on July 10, 2026, the Perspective addresses the persistent inability of spaceborne L-band radiometers to deliver stable, high-resolution sea surface salinity (SSS) estimates near coasts, where land contamination, imaging artifacts, and incomplete sea-state representations degrade data quality.
The importance of this work lies in its potential to overcome a long-standing observational gap. Current satellite products offer only 40–100 km effective resolution, with coastal uncertainty of 0.5–1.0 practical salinity units (psu) and substantial data loss within 50–100 km of land. Such limitations hamper operational monitoring of river discharge, estuarine mixing, extreme rainfall, ecosystem stress, and freshwater transport—processes that are central to coastal ecosystems and circulation. By organizing improvements into two coordinated pathways, the framework charts a course toward 10–20 km resolution and accuracy better than 0.3 psu within 100 km of shore.
Pathway I focuses on cleaning the measurement chain by combining visibility-domain corrections with brightness temperature (TB)-domain corrections, which is particularly critical for interferometric microwave radiometers (IMRs). Techniques such as visibility phase adjustment (VPA) suppress Gibbs oscillations, while antenna-pattern-based corrections estimate residual land leakage. Pathway II deepens the forward model by incorporating wave development, fetch, wave age, foam, shallow-water effects, and current-induced roughness changes. Together, these routes address the land–sea contamination and incomplete physics that have long frustrated coastal retrievals.
Rather than proposing a single universal algorithm, the authors classify methods by Technology Readiness Level and connect them to a roadmap. Near-term priorities include standardization and testbeds; mid-term efforts center on co-design of instruments and retrieval systems; and long-term goals involve integration with data assimilation and coastal freshwater observing networks. Physics-aware artificial intelligence is suggested for structured residual correction and hybrid modeling, but the authors stress that learned components must remain anchored in transparent physical constraints and robust error statistics.
Quantitative evidence reviewed in the paper underscores why a system-level solution is required. In coastal zones, SSS uncertainty typically reaches 0.5–1.0 psu, while biases in river plumes can exceed 0.5 psu. Land-induced TB contamination may extend hundreds of kilometers offshore, and masking or windowing often sacrifices coverage and resolution. The dual-pathway framework explicitly links measurement cleaning with physical model deepening to close the loop between a “clean” measurement chain and a “deep” physical forward model.
The implications extend beyond salinity mapping. Better coastal SSS products could guide future L-band satellite design, coastal product reprocessing, and operational assimilation systems. Over the next decade, mission teams could jointly optimize antennas, calibration, land-contamination control, sea-state modeling, and current-aware retrieval. Longer term, salinity, sea surface height, currents, and wave state could be estimated together through satellite, radar, model, and in situ observations, creating a reliable coastal freshwater observing system. The framework does not claim one universally optimal correction; instead, it coordinates instrument teams, retrieval developers, and coastal oceanographers around measurable coastal-performance goals and transferable physical principles. As a Perspective, the work synthesizes findings from satellite missions, instrument studies, radiative-transfer research, wave and current modeling, and recent correction methods, with no new field dataset or laboratory experiment reported. The article states that no data are associated with the research, and the full perspective is available at 10.34133/remotesensing.1058.
