Argo in Python¶
Hands-on tutorial to work with Argo program data in Python.
Aimed at people with an oceanography background who want to learn the practical workflow: how to download the data, understand the file structure, plot profiles, map trajectories, and run basic analyses.
Contents¶
| # | Chapter | What it does |
|---|---|---|
| 01 | Accessing the data | argopy in five minutes: fetch_float, fetch_box, fetch_region, R/A/D modes and QC. |
| 02 | Anatomy of the dataset | How the xarray.Dataset is structured, what variables it has, how to filter by quality. |
| 03 | T/S profiles | Plotting individual profiles and T-S diagrams with isopycnals. Water mass identification. |
| 04 | Trajectories | Mapping with cartopy. Reading the track in a dynamical context. |
| 05 | Seasonal variability | Monthly and seasonal climatologies in a South Atlantic region. |
| 06 | Mixed layer depth | MLD with a density criterion, summer vs winter. |
Before you start¶
You’ll need Python ≥ 3.10 and the following libraries:
pip install argopy xarray netcdf4 cartopy gsw matplotlib jupyterIf you’ve never used argopy, don’t worry. It abstracts access to the Argo GDAC and returns a ready-to-use xarray.Dataset.
Why this tutorial¶
Argo floats are the operating system of the deep ocean. There are ~4000 active floats measuring temperature, salinity and pressure down to 2000 m, and the data is public and free. But the first time you open one of those NetCDF files you run into a peculiar structure: N_PROF / N_LEVELS dimensions, quality flags, calibration modes. There’s a learning curve before doing science.
This tutorial is the thing I wish I had when I started.
Related material¶
Narrative scroll. Scroll-driven visualization of an Argo float cycle (recommended for building intuition).
Argo program website. Official documentation.
argopy docs. Full reference.
Daniela Risaro · 2026