OGD Model Data Access & Processing

This repository provides Jupyter notebook examples for accessing and processing numerical weather prediction (NWP) model data from MeteoSwiss, released through Switzerland’s Open Government Data (OGD) initiative.

You can run the notebooks locally or directly in Google Colab.

Notebook	Description	Open in Colab
01_retrieve_process_precip.ipynb	Retrieve and load precipitation forecasts as an Xarray object, then process, analyze, and visualize the data using Python tools.
02_download_soil_temp.ipynb	Download forecast files to disk for offline storage, external tools, or advanced manual processing.
03_calculate_wind_speed.ipynb	Retrieve wind component forecasts as Xarray objects and derive the horizontal wind speed using meteodata-lab.
04_calculate_rel_humidity.ipynb	Retrieve specific humidity, temperature and pressure as Xarray objects and compute relative humidity.
05_interpolate_vertically.ipynb	Retrieve temperature forecasts and perform vertical interpolation from model levels to pressure levels or target altitude.
06_calculate_global_rad_flux.ipynb	Retrieve radiation fluxes and compute the global radiation flux.
07_where_will_it_rain_next_24h.ipynb	Visualize the probability of precipitation over Switzerland for the next 24 hours.
08_where_will_the_sun_shine.ipynb	Map the probability of experiencing over 6 hours of sunshine on the day after tomorrow.
09_constant_parameters.ipynb	Retrieve constant model parameters and verify grid consistency with forecast parameters.
10_icon_ch2_pollen_forecast.ipynb	Retrieve, convert, and visualize ICON-CH2-EPS pollen forecasts.
11_analysis_data.ipynb	Retrieve and visualize KENDA CH1 analysis data.

You can run the notebooks in two ways.

Use Google Colab if you want to open one notebook directly.

Click the Open in Colab badge next to the notebook in the table above, then run the notebook cells from top to bottom.

The first cell installs the required Python dependencies.

Clone the repository and install all required packages. This project requires Python >=3.11,<3.13 and Poetry to manage dependencies and environments.

1. Install Python dependencies using Poetry:

   poetry install

2. Install the Jupyter kernel:

   poetry run python -m ipykernel install --user --name=notebooks-nwp-env --display-name "Python (notebooks-nwp-env)"

3. Open the notebook in VS Code or JupyterLab and select the kernel Python (notebooks-nwp-env).

For more context on the available numerical weather forecast data and how it’s structured, see:

🔗 MeteoSwiss Forecast Data Documentation

Feel free to open issues to suggest improvements or contribute new examples!

When making a change:

1. Work on the notebook in developer_notebooks/.
2. Developer notebooks should use the _clean suffix, for example 09_notebook_clean.ipynb.
3. Commit the developer notebook without outputs. This keeps PR review manageable and avoids noisy diffs from generated output.
4. Once the developer notebook has been reviewed, run it and save the corresponding notebook with outputs at the top level of the repository, without the _clean suffix. This is the version users will look at.