This repository contains codes and sample notebooks of a data challenge for downloading and processing the high frequency SSH mapping with artificial SWOT and artificial conventional nadirs data in the Californian SWOT X-over.

The quickstart demo_oi can be run online on Google Colab by clicking here:

The quickstart perform_your_eval can be run online on Google Colab by clicking here: .

You only need to compress your reconstruction maps in a zip file, drop it in your google drive account, create a link to access it and copy paste the link id in the quickstart in place of XXXXX.

The goal is to investigate how to best reconstruct sequences of Sea Surface Height (SSH) maps in the Californian SWOT cross-over from artificial SWOT satellite and conventional nadir altimetry observations. This data challenge follows an Observation System Simulation Experiment framework: "reference" full SSH are from a numerical simulation with a realistic, high-resolution ocean circulation model: the reference simulation. Satellite observations are simulated by sampling the reference simulation based on realistic orbits of past, existing or future altimetry satellites. A baseline reconstruction method is provided (see below) and the practical goal of the challenge is to beat this baseline according to scores also described below and in Jupyter notebooks.

The reference simulation is the MITgcm LLC4320 simulation. The simulation is run with tidal forcing. The SSH maps are available hourly. The barotropic tide has been removed from the reference run.

The SSH observations are SWOT and conventional nadirs altimeter data simulated on the reference run. Hence, the barotropic tide is also not present in the SWOT observations.

Three mapping experiments can be performed using SWOT:

• without observation errors: SWOT no noise ('ssh_model' variable),
• with KaRIn errors only: SWOT KaRIn noise ('ssh_karin' variable),
• with all observation errors: SWOT all noise ('ssh_obs' variable).

As an illustration, see the demo_perform_oi notebook that performs an OI reconstruction in these three experiments.

The SSH reconstructions are assessed over the period from 2012-02-01 to 2012-04-30: 89 days. For reconstruction methods that need a spin-up, the observations can be used from 2012-01-04 until the beginning of the evaluation period. This spin-up period is not included in the evaluation. For reconstruction methods that need learning from full fields, the reference data and the observations can be used from 2012-06-01 to 2012-10-29. The reference data between 2012-05-01 and 2012-05-31 should never be used so that any learning period or other method-related-training period can be considered uncorrelated to the evaluation period.

SWOT no noise

SWOT KaRIn noise

SWOT all noise

µ(RMSE): average RMSE score.
σ(RMSE): standard deviation of the RMSE score.
λx: minimum spatial scale resolved.
λt: minimum time scale resolved.

You can follow the quickstart guide in this notebook or launch it directly from CoLab.

The data are hosted on TBD . The data are also temporarily available here. They are presented with the following directory structure:

• dc_obs_swot/: SWOT data, observations from SWOTsimulator on MITgcm reference;

.
|-- dc_obs_swot
|   |-- 2022a_SSH_mapping_CalXover_swot.nc

• dc_obs_nadirs/: conventional nadirs data, observations from SWOTsimulator on MITgcm reference;

.
|-- dc_obs_nadirs/*/
|   |-- dt_global_XXXXX.nc

where * can be one of the available satellites: alg/, c2/, h2g/, j2g/, j2n/, j3/ and s3a/ ; and XXXXX dates specifications.

• dc_ref_eval: evaluation data, SSH reference from MITgcm during the evaluation period;

|-- dc_ref_eval
|   |-- 2022a_SSH_mapping_CalXover_eval_****-**-**.nc

where ****-**-** stands for year, month and day. 

• dc_mod: training/validation data, SSH MITgcm outside of the evaluation period.

|-- dc_mod
|   |-- 2022a_SSH_mapping_CalXover_model_****-**-**.nc

where ****-**-** stands for year, month and day. 

To start out download the observation dataset (dc_obs, 77M) from the temporary data server, use:

wget https://ige-meom-opendap.univ-grenoble-alpes.fr/thredds/fileServer/meomopendap/extract/ocean-data-challenges/dc/dc_calXover/dc_obs_swot.tar.gz

and

wget https://ige-meom-opendap.univ-grenoble-alpes.fr/thredds/fileServer/meomopendap/extract/ocean-data-challenges/dc/dc_calXover/dc_obs_nadirs.tar.gz

the reference dataset for the evaluation (dc_ref_eval, 660M) using (this step may take several minutes):

wget https://ige-meom-opendap.univ-grenoble-alpes.fr/thredds/fileServer/meomopendap/extract/ocean-data-challenges/dc/dc_calXover/dc_ref_eval.tar.gz

the model dataset for training/validation (dc_ref_eval, 1.5G) using (this step may take several minutes):

wget https://ige-meom-opendap.univ-grenoble-alpes.fr/thredds/fileServer/meomopendap/extract/ocean-data-challenges/dc/dc_calXover/dc_mod.tar.gz

and then uncompress the files using tar -xvf <file>.tar.gz. You may also use ftp, rsync or curlto donwload the data.

The baseline mapping method is optimal interpolation (OI), in the spirit of the present-day standard for DUACS products provided by AVISO. OI is implemented in the demo_perform_oi Jupyter notebook. The SSH reconstructions are saved as a NetCDF file in the results directory. The content of this directory is git-ignored.

The evaluation of the mapping methods is based on the comparison of the SSH reconstructions with the reference dataset. It includes two scores, one based on the Root-Mean-Square Error (RMSE), the other based on Fourier wavenumber spectra. The evaluation notebook demo_perform_evaluation implements the computation of these two scores as they could appear in the leaderboard. The notebook also provides additional, graphical diagnostics based on RMSE and spectra.

Cross-functional modules are gathered in the src directory. They include tools for regridding, plots, evaluation, writing and reading NetCDF files. The directory also contains a module that implements the baseline method.

The structure of this data challenge was to a large extent inspired by ocean data challenge (2020a).