Python scripts to process and analyze Argo floats

Opening GitHub notebooks in Colab

You can open the notebook directly from GitHub in Google Colab without any local setup:

[https://colab.research.google.com/github/oceanmargins/omi_argo_floats/blob/main/argo_pandas.ipynb]

You need a GitHub account and a Google account.

Local setup with pixi

pixi is a fast, cross-platform package manager. It handles Python and conda dependencies in one step.

1. Install pixi

curl -fsSL https://pixi.sh/install.sh | bash

Restart your terminal after installation, or source your shell profile to make the pixi command available.

2. Clone this repository

git clone https://github.com/oceanmargins/omi_argo_floats.git
cd omi_argo_floats

3. Install the project environment

pixi install

This reads pyproject.toml and installs all dependencies (NumPy, pandas, xarray, matplotlib, argopy, JupyterLab, etc.) into an isolated environment — no conda activate or pip install needed.

4. Start JupyterLab

pixi run jupyter lab

JupyterLab will open in your browser. Open argo_pandas.ipynb from the file browser on the left.

Using the notebook

The notebook argo_pandas.ipynb fetches Argo float data from the IFREMER ERDDAP server and loads it into a pandas DataFrame for analysis.

Workflow:

1. Configure the query — The notebook builds an ERDDAP URL to select a specific float (fileNumber) and a geographic bounding box (latitude/longitude range). Edit these values in the second cell to target a different float or region.

2. Fetch data — pd.read_csv(url, skiprows=[1]) downloads the data directly from ERDDAP into a DataFrame. Columns include time, latitude, longitude, pres (pressure/depth), temp, psal (salinity), doxy, turbidity, chla, and nitrate.

3. Explore and plot — The notebook demonstrates a temperature–salinity scatter plot and a map of the float track. Adapt these cells for your own analysis.

To query a different float:

Go to [https://erddap.ifremer.fr/erddap/tabledap/ArgoFloats.html] to browse available floats and build a custom ERDDAP URL, then paste it into the notebook.

Downloading model and wind data

Several scripts in this repo download external reanalysis/forecast data used to compare against Argo observations. They all save into a local data/ directory (created automatically, and git-ignored).

myocean.py

Downloads salinity and temperature fields from the Copernicus Marine Service (CMEMS) global ocean physics analysis/forecast product (cmems_mod_glo_phy-so_anfc_0.083deg_P1D-m / ...-thetao_...).

• login(username, password) — authenticates with Copernicus Marine using copernicusmarine.login. Requires a free Copernicus Marine account.
• retrieve(data_var="salt") — subsets and downloads one variable ("salt" or "temp") as NetCDF into data/. The bounding box, date range, and depth range are currently hardcoded in the function.

Running the module directly (pixi run python myocean.py) downloads both salinity and temperature.

copernicus.ipynb

A scratch notebook covering the same Copernicus Marine workflow interactively: logging in, subsetting a small region (currents, salinity, temperature) as Zarr, loading it with xarray, and plotting it with matplotlib/cartopy (salinity color maps, current streamplots, coastlines). Useful as a starting point for exploring a new CMEMS subset before scripting it.

Note: cartopy is used for the map plots in this notebook but is not yet listed in pyproject.toml; install it into the pixi environment if you want to run those cells.

winds.py

Downloads ERA5 wind data and reads met-tower data used alongside the Argo/CMEMS comparisons.

• download_era5(variable="10m_u_component_of_wind") — downloads a full year (YEAR) of 6-hourly ERA5 reanalysis data over a fixed bounding box (AREA) via the CDS API, saving a .grib file into data/. Requires a CDS API key configured (typically in ~/.cdsapirc).
• read_mettower() — fetches the OMI Ghana meteorological tower JSON feed and returns it as a DataFrame with a parsed time column.
• retrieve() — downloads both u/v wind components and the met-tower data, saving the latter to data/met_tower_data.h5.

match_argo_model.py

Matches each Argo profile observation to the nearest CMEMS model grid cell, for direct comparison of observed vs. modeled temperature and salinity.

• Reads Argo data from data/argo_data.h5 and the most recent *phy-thetao_*.nc / *phy-so_*.nc files in data/ (as produced by myocean.py).
• match_argo_to_model(argo_file=ARGO_FILE, temp_file=None, salt_file=None) — for each Argo row, finds the nearest model grid point in latitude/longitude (via a scipy.spatial.cKDTree), and the nearest model time step and depth level (via a sorted nearest-value search), then adds model_temp and model_salt columns to the returned DataFrame.
• Running the module directly (pixi run python match_argo_model.py) prints a preview comparing Argo temp/psal to the matched model_temp/model_salt.

data/argo_data_with_model.csv

The CSV export of match_argo_to_model()'s output — one row per Argo profile observation, with columns fileNumber, time, latitude, longitude, pres, temp, psal, doxy, turbidity, chla, nitrate (from Argo) plus model_temp and model_salt (the matched CMEMS values) for direct comparison.