Bhuiyan et al. (2026) Improving coastal water level estimation by merging nadir-only satellite altimetry data into a hydrodynamic model

Identification

• Journal: Environmental Monitoring and Assessment
• Year: 2026
• Date: 2026-03-14
• Authors: Soelem Aafnan Bhuiyan, André de Souza de Lima, Tyler Miesse, Martin Henke, Celso Ferreira, Viviana Maggioni
• DOI: 10.1007/s10661-026-15166-8

Research Groups

• Department of Civil, Environmental and Infrastructure Engineering, George Mason University, Fairfax, VA, USA
• Pacific Northwest National Laboratory, Richland, WA, USA

Short Summary

This study evaluates a novel method to improve coastal water level (CWL) predictions by assimilating nadir-only satellite altimetry data from four missions into the ADCIRC hydrodynamic model along the U.S. East Coast, finding that combined assimilation significantly enhances model performance at over 80% of gauge locations.

Objective

• To evaluate the effectiveness and extent of improvement in coastal water level (CWL) estimations by assimilating nadir-only satellite altimetry data into the ADvanced CIRCulation (ADCIRC) hydrodynamic model.

Study Configuration

• Spatial Scale: U.S. East Coast, including the Atlantic Ocean and the Gulf of Mexico. The ADCIRC model mesh has a resolution as fine as 100 m near the coastline and a coarser resolution of up to 60 km towards the ocean boundary.
• Temporal Scale: 107-day simulation period, from July 16, 2023, to October 31, 2023, with hourly model outputs.

Methodology and Data

• Models used:
  • ADvanced CIRCulation (ADCIRC) v55 (hydrodynamic model)
  • KDTree algorithm for spatial indexing
• Data sources:
  • Satellite Altimetry: Nadir-only observations from SARAL, Jason-3, Sentinel-6, and Surface Water and Ocean Topography (SWOT) missions.
  • Observation Data: 140 National Oceanic and Atmospheric Administration (NOAA) tide gauge stations along the U.S. East Coast.
  • Atmospheric Forcing: ECMWF Re-Analysis 5 (ERA5) product (hourly barometric pressure at mean sea level and wind velocity at 10 m above sea level).
  • Tidal Forcings: Oregon State University (OSU) TPXO9v2 global tidal database (nine tidal constituents).
  • Bathymetry/Topography: GEBCO-2024 Digital Elevation Model (DEM) and NOAA NCEI Continuously Updated DEM (CUDEM).
  • Assimilation Scheme: Direct Insertion (DI) using Pseudo-Atmospheric Pressure (PAP) forcing.
  • Evaluation Metrics: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Pearson Correlation Coefficient (CC).

Main Results

• Assimilating nadir-only observations into ADCIRC improved model performance (RMSE) at 76% of gauge locations for SWOT-nadir, 73% for Sentinel-6, 74% for Jason-3, and 21% for SARAL.
• Combining observations from SWOT-nadir, Jason-3, and Sentinel-6 improved ADCIRC performance (RMSE) at more than 80% of gauge locations for the 107-day simulation.
• Nadir-only satellite altimetry observations, even if flagged as “poor quality” near the coast, can be useful for improving model performance. When flagged data were disregarded, SWOT improved ADCIRC at 78%, Sentinel-6 at 73%, Jason-3 at 53%, and SARAL at 21%.
• Model skill significantly increases when satellite overpasses coincide with storm surge events, as demonstrated during Tropical Storm Ophelia and Hurricane Idalia.
• Excluding SARAL from the combined assimilation (considering only Jason-3, Sentinel-6, and SWOT-nadir) further increased the number of improved stations for RMSE from 106 to 120 and reduced worsening stations from 34 to 19.

Contributions

• Demonstrates a novel method for improving coastal water level (CWL) estimations by assimilating nadir-only satellite altimetry data into the ADCIRC hydrodynamic model.
• Quantifies the performance enhancement of individual and combined altimetry missions (SARAL, Jason-3, Sentinel-6, SWOT-nadir) for CWL simulations along the U.S. East Coast.
• Highlights the critical role of satellite overpass frequency, particularly during storm surge events, in improving model accuracy.
• Provides evidence that even altimetry data flagged as "poor quality" near the coast can contribute positively to model performance.
• Offers insights into the limitations of simple data assimilation schemes and the impact of ramp-up times on correlation metrics, suggesting areas for future research.

Funding

• Institute of Digital InnovAtion (IDIA), George Mason University.

Citation

@article{Bhuiyan2026Improving,
  author = {Bhuiyan, Soelem Aafnan and Lima, André de Souza de and Miesse, Tyler and Henke, Martin and Ferreira, Celso and Maggioni, Viviana},
  title = {Improving coastal water level estimation by merging nadir-only satellite altimetry data into a hydrodynamic model},
  journal = {Environmental Monitoring and Assessment},
  year = {2026},
  doi = {10.1007/s10661-026-15166-8},
  url = {https://doi.org/10.1007/s10661-026-15166-8}
}