Real SSH mapping data challenge in the Western Mediterranean Sea

A challenge on mapping real altimetric data in the Mediterranean Sea created by Datlas and MEOM-IGE.

This repository contains codes and sample notebooks for downloading and processing the SSH mapping data challenge.

1. Context and motivation

The goal is to investigate how to best reconstruct sequences of Sea Surface Height (SSH) maps from partial satellite altimetry observations. This data challenge follows an Observation System Experiment framework: Satellite observations are from real sea surface height data from altimeter. The region of interest in this DC is the Western Mediterrannean sea [1°E x 20°E , 30°N x 45°N]. The practical goal of the challenge is to investigate the best mapping method according to scores described below and in Jupyter notebooks.

2. Specific experimental set up

Observations

The SSH observations include SARAL/Altika (alg), Haiyang-2B (h2b), Jason 3 (j3), Sentinel 3A (s3a), Sentinel 3B (s3b) and Cryosat-2 new orbit (c2n) altimeter data. This nadir altimeters constellation was operating and will be investigated during the 20210101-20210331 period. Note that for the mapping the Cryosat-2 new orbit altimeter data will not be considered in order to perform an independent assessment of the various reconstructions.

The observations (for reconstruction and for evaluation) are hosted and available for download on the MEOM opendap server: see Download the data section below.

Data sequence and use

The SSH reconstructions are assessed over the period from 2021-01-15 to 2021-03-15.

For reconstruction methods that need a spin-up or training, the observations can be used during the 15 first and the 15 last days before and after the two-month evaluation period. Note also that any other Nadir satellite observations available outside of the evaluation period can be used and are found on the CMEMS website. The evaluation is performed within the geographical box: [2°E x 19°E , 31°N x 44°N] so to avoid potential boundary issues which is a topic left for an other challenge.
The altimeter data from Cryosat-2 should never be used so that any reconstruction can be considered uncorrelated to the evaluation period.

Baseline

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 baseline_oi Jupyter notebook. The SSH reconstructions are saved as a NetCDF file in the results directory. The content of this directory is git-ignored.

Evaluation

The evaluation of the mapping methods is based on the comparison of the SSH reconstructions with the independent Cryosat-2 along-track dataset. It includes two scores, one based on the Root-Mean-Square Error (RMSE), the other based on Fourier wavenumber spectra. The evaluation notebook example_data_eval 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.

3. To get started

Installation

💻 How to get started ?

Clone the data challenge repo:

git clone https://github.com/ocean-data-challenges/DC_MapMed_OSE.git

or

git clone git@github.com:SammyMetref/DC_MapMed_OSE.git

create the data challenge conda environment, named env-dc-mapmed-ose, by running the following command:

conda env create --file=environment.yml 

and activate it with:

conda activate env-dc-mapmed-ose

then add it to the available kernels for jupyter to see:

ipython kernel install --name "env-dc-mapmed-ose" --user

You're now good to go !

Check out the quickstart

Download the data

The data are hosted on the opendap: ocean-data-challenges/dc_MapMed_OSE/. They are presented with the following directory structure:

. 
|-- dc_obs
|   |-- dt_med_alg_phy_l3_20210101-20210331_1-2_40-2.nc
|   |-- dt_med_h2b_phy_l3_20210101-20210331_1-2_40-2.nc
|   |-- dt_med_j3_phy_l3_20210101-20210331_1-2_40-2.nc
|   |-- dt_med_s3a_phy_l3_20210101-20210331_1-2_40-2.nc
|   |-- dt_med_s3b_phy_l3_20210101-20210331_1-2_40-2.nc 
|   |-- dt_med_mdt_phy_l4_nrt_0.5-20.0_29-45.nc 
 

containing the data from 5 nadir altimeters (alg, h2b, j3, s3a and s3B) and the gridded MDT map (also available along track in the nadir netcdf files) if needed for reconstruction.

. 
|-- dc_eval
|   |-- dt_med_c2n_phy_l3_20210101-20210331_1-2_40-2.nc

containing the data from c2n nadir altimeter that is only available for the evaluation and must not be used in the reconstruction.

To start out you can download the observation dataset (dc_obs, 1.6M) from the temporary data server, with:

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

and the evaluation dataset (dc_eval, 320k) using:

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

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

Check the demo notebooks to see how to download the data in a jupyter notebook!

Quick start

You can follow the quickstart guide in this notebook: quickstart.

4. Leaderboard

Method	µ(RMSE)	σ(RMSE)	λx (km)	Notes	Reference
BASELINE OI	0.674136	0.111694	148	Covariances BASELINE OI	evaluate_baseline_oi.ipynb
BFN-QG	0.720857	0.0941839	125	BFN-QG without coasts (open QG pool)	evaluate_BFNQG.ipynb
BFNQG_coast	0.724554	0.0933299	125	BFN-QG with coasts (Dirichlet cond.)	evaluate_BFNQG_coast.ipynb
WaveVar	0.759956	0.0877991	118	Implemented from Ubelmann et al. (2021)	evaluate_WaveVar.ipynb

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

Acknowledgement

The structure of this data challenge was to a large extent inspired by the second ocean-data-challenge created for the BOOST-SWOT ANR project in collaboration with the MEOM team at IGE lab, CLS and IMT Atlantic: 2021a_SSH_mapping_OSE.