Sea Surface Height Interpolation Edition

How can we fill in the gaps from Sea Surface Height Observations

CNRS
MEOM

This is the first edition of the OceanBench framework.

Data Challenges

We us the OSSE and OSE framework for SSH interpolation.

OSSE Experiments

Each subsequent experiment adds more data that is available to the user for usage in their algorithms. However, it is encouraged to use ablations to see which of the datasets had the most impact on the learning scheme.

I - OSSE NADIR

We use satellite-based NADIR altimetry tracks. The original simulation stems from a high-resolution ocean simulation stemming from the NEMO model [Ajayi et al., 2020].

II - OSSE SWOT

This experiment uses both the NADIR and SWOT altimetry data. The SWOT satellite offers a higher spatial resolution but a lower temporal resolution. In principal, this would allow us to see more detail in the spatial structures [Gaultier et al., 2016]. However, this offers more challenges because the amount of data captured from the SWOT simulator is much, much higher than that of the NADIR altimeters.

III - OSSE NADIR + SWOT + SST

This experiments adds the use available SST data to aid in training. From a scientific perspective, there are deep connections between SSH and SST as they are often jointly estimated in large Ocean GCMs [Ajayi et al., 2020]. From a practical perspective, SST is more abundant in operational settings, at higher resolutions and often with less gaps.

OSE Experiments

References
  1. Ajayi, A., Sommer, J. L., Chassignet, E., Molines, J.-M., Xu, X., Albert, A., & Cosme, E. (2020). Spatial and Temporal Variability of the North Atlantic Eddy Field From Two Kilometric-Resolution Ocean Models. Journal of Geophysical Research: Oceans, 125(5). 10.1029/2019jc015827
  2. Gaultier, L., Ubelmann, C., & Fu, L.-L. (2016). The Challenge of Using Future SWOT Data for Oceanic Field Reconstruction. Journal of Atmospheric and Oceanic Technology, 33(1), 119–126. 10.1175/jtech-d-15-0160.1