DeepMelt-GL v1: a neural network emulator of sub-shelf melt rates for the unrepresented regions of ice-shelf cavities in ocean models

Ockenden, Helen; Burgard, Clara; Mathiot, Pierre; Kittel, Christoph; Gellens, Achille; Agosta, Cécile; Jourdain, Nicolas C.

doi:10.5194/gmd-19-7979-2026

Articles | Volume 19, issue 16

https://doi.org/10.5194/gmd-19-7979-2026

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

https://doi.org/10.5194/gmd-19-7979-2026

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

Articles | Volume 19, issue 16

Model description paper

|

27 Aug 2026

Model description paper |

| 27 Aug 2026

DeepMelt-GL v1: a neural network emulator of sub-shelf melt rates for the unrepresented regions of ice-shelf cavities in ocean models

Helen Ockenden, Clara Burgard, Pierre Mathiot, Christoph Kittel, Achille Gellens, Cécile Agosta, and Nicolas C. Jourdain

Data sets

Data which accompanies the manuscript "A neural network emulator of ice-shelf melt rates for use in ocean models which partially resolve ice-shelf cavities'' Helen Ockenden https://doi.org/10.5281/zenodo.17358228

Model code and software

Code which accompanies the manuscript "A neural network emulator of ice-shelf melt rates for use in ocean models which partially resolve ice-shelf cavities'' Helen Ockenden https://doi.org/10.5281/zenodo.17358195

Short summary

Since numerical computing is expensive, climate models must decide between having a high spatial resolution or running for long time periods. Here, we develop a simple neural network to emulate small-scale processes occurring beneath Antarctic ice shelves, which allows sub-shelf melt and ice–ocean interactions to be included in global ocean models which can run for multiple centuries. This neural network will help us to understand how ocean circulation may change in the future.