Sea surface salinity downscaling using deep generative diffusion models

Forestier, Enzo; Ollier, Luther; El Hourany, Roy; Boutin, Jacqueline; Mejia, Carlos; Thiria, Sylvie

doi:10.5194/os-22-2357-2026

Articles | Volume 22, issue 4

https://doi.org/10.5194/os-22-2357-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/os-22-2357-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 22, issue 4

Research article

|

05 Aug 2026

Research article |

| 05 Aug 2026

Sea surface salinity downscaling using deep generative diffusion models

Enzo Forestier, Luther Ollier, Roy El Hourany, Jacqueline Boutin, Carlos Mejia, and Sylvie Thiria

Data sets

Global Ocean Physics Reanalysis E.U. Copernicus Marine Service Information https://doi.org/10.48670/moi-00021

Short summary

This study evaluates deep generative diffusion models for downscaling sea surface salinity in the Gulf Stream. Using a reanalysis dataset as a controlled framework, it assesses the added value of high-resolution sea surface temperature and sea surface height as auxiliary constraints. The results show that diffusion-based reconstructions preserve plausible fine-scale variability, highlighting the method’s potential for future applications to satellite products.