A continuous implicit neural representation framework  with gradient regularization for sea surface height  reconstruction from satellite altimetry

Li, Dongshuang; Pan, Liming; Yu, Zhaoyuan; Yuan, Linwang

doi:10.5194/gmd-19-7349-2026

Articles | Volume 19, issue 15

https://doi.org/10.5194/gmd-19-7349-2026

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

https://doi.org/10.5194/gmd-19-7349-2026

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

Articles | Volume 19, issue 15

Development and technical paper

|

07 Aug 2026

Development and technical paper |

| 07 Aug 2026

A continuous implicit neural representation framework with gradient regularization for sea surface height reconstruction from satellite altimetry

Dongshuang Li, Liming Pan, Zhaoyuan Yu, and Linwang Yuan

Data sets

Dataset for "A Continuous Implicit Neural Representation Framework with Gradient Regularization for Sea Surface Height Reconstruction From Satellite Altimetry" Dongshuang Li https://doi.org/10.5281/zenodo.18748410

Model code and software

Code for: Continuous Implicit Neural Representation Framework with Gradient Regularization for Sea Surface Height Reconstruction from Satellite Altimetry Dongshuang Li https://doi.org/10.5281/zenodo.21232682

Short summary

Satellites do not measure every part of the ocean at every moment, so maps of ocean surface height often have gaps. We developed a computer method that learns from scattered satellite measurements to fill in these gaps while keeping the reconstructed patterns smooth and physically plausible. Tests with real and simulated data show that the method improves regional ocean surface maps and better preserves important ocean features, supporting future monitoring of ocean change.