MeteoSwiss - Open Data > Understanding MeteoSwiss' Open Data products > E. Forecast Data
Forecasting systems calculate future atmospheric conditions on the basis of measurement data and observations. MeteoSwiss uses these weather models to create weather forecasts and to enable it to issue weather warnings in the event of imminent hazards.
The following forecast data are available:
- Short-term forecast data 🟡 documentation upcoming
- Numerical weather forecasting model data 🟡 documentation upcoming
- Local forecast data 🟡 documentation upcoming
Nowcasting involves high spatial and temporal resolution forecasts of weather developments for the next few minutes and up to a maximum of six hours ahead. MeteoSwiss uses these short-term forecasts to, among other things, predict thunderstorms, hail and heavy rainfall.
As MeteoSwiss is planning to replace the current 'INCA' nowcasting software, the following datasets are available from the start of our open data provision:
- Precipitation (10min values): quantitative chain (based on CombiPrecip, RR)
- Wind, wind gust and wind direction (10min values)
- Relative sunshine duration (10min values)
- Total cloudiness (10min values)
The following datsets will be provided next:
- Snowfall (10min values): quantitative chain (based on CombiPrecip, RS)
- ...
- ...
Data granularity is every 10min. Update frequency for the period 0h- +6h is specified per dataset in the table below.
Data format is NetCDF.
| Dataset | Update frequency | Example data file | Productive version file name | Estimated volume per file (MB) |
|---|---|---|---|---|
| Precipitation (10min values): quantitative chain (based on CombiPrecip, RR) | every 10min | RR_INCA_202106280700.nc | ogd-nowcasting_RR-INCA_(date and time code).nc |
1.7 |
| Wind, wind gust and wind direction (10min values) | every 10min | ... | ogd-nowcasting_(product name)_(date and time code).nc |
... |
| Relative sunshine duration (10min values) | 10min | SU_INCA_202106280700.nc | ogd-nowcasting_SU-INCA_(date and time code).nc |
6.4 |
| Total cloudiness (10min values) | 10min | SU_INCA_202106280700.nc | ogd-nowcasting_SU-INCA_(date and time code).nc |
6.4 |
| Snowfall (10min values): quantitative chain (based on CombiPrecip, RS) | every 10min | RS_INCA_202106280700.nc | ogd-nowcasting_RS-INCA_(date and time code).nc |
0.4 |
Parameter metadata is part of each NetCDF-File. See example data files in the table above.
The coordinate system is Swiss LV95 EPSG:2056.
See e.g. MeteoSwiss' ....
MeteoSwiss uses two models, ICON-CH1-EPS and ICON-CH2-EPS, to forecast atmospheric changes in Switzerland and its surroundings over a longer period than nowcasting, providing predictions for up to five days. Both models include ensemble data assimilation.
The documentation covers the following topics:
- 2.1 Model Specifications
- 2.2 Available Parameters
- 2.3 Accessing Forecast Data
- 2.4 3D Grid Structure and Representation
- 2.5 Example Notebooks: From Retrieval to Visualization
- 2.6 Retrieving Forecasts via REST API
- 2.7 Accessing Static Grid Information: Height, Longitude, and Latitude
- 2.8 Reading Forecast Files Using ecCodes
| Attributes | ICON-CH1-EPS | ICON-CH2-EPS |
|---|---|---|
| Collection | ch.meteoschweiz.ogd-forecasting-icon-ch1 | ch.meteoschweiz.ogd-forecasting-icon-ch2 |
| Horizontal Grid Size | 1 km | 2.1 km |
| Ensemble Members | 11 | 21 |
| Forecast Period | 33 h | 120 h |
| Grid | Native icosahedral | Native icosahedral |
| Temporal Resolution | 1 h | 1 h |
| Model Run Interval | every 3 h | every 6 h |
| Format | GRIB edition 2 | GRIB edition 2 |
Users can find information about available parameters, including metadata about height, longitude and latitude, in the collection level assets of the above collections.
The parameter metadata is part of each GRIB file.
Users can access forecast model data from the last 24 hours. Data older than this is no longer available. The data in each collection is described in the Model Specification table.
⚠️ WARNING: Data located at the boundary of the spatial domain may be random.
The following tables summarize the volume of the different forecast files for ICON-CH1-EPS and ICON-CH2-EPS.
ICON-CH1-EPS Data Volume
| Single-Level Files | Multi-Level Files | |
|---|---|---|
| Deterministic | 199.0 Bytes - 2.2 MiB | 19.7 - 177.4 MiB |
| Perturbed | 1.9 KiB - 22.4 MiB | 197.1 MiB - 1.7 GiB |
ICON-CH2-EPS Data Volume
| Single-Level Files | Multi-Level Files | |
|---|---|---|
| Deterministic | 175.0 Bytes - 564.7 KiB | 4.9 MiB - 43.9 MiB |
| Perturbed | 3.4 KiB - 11 MiB | 97.5 MiB - 877.5 MiB |
The model data is structured on both a horizontal and vertical grid. While some parameters extend across the entire three-dimensional grid, others are only available at specific vertical levels. Parameters are classified as either single-level or multi-level:
- Single-level parameters contain data at a specific vertical level.
- Multi-level parameters extend across multiple vertical layers.
For example, vertical velocity is stored at multiple vertical levels, while the two-meter temperature is available only at a single vertical level.
The vertical grid above the surface is a height-based coordinate system that follows the terrain and is divided into multiple layers. The closer the layer is to the surface, the narrower the layers are, as shown in the image below. The so-called half levels align with vertical grid points, while the full levels represent an averaged value over a vertical interval. There are 81 discrete half levels and 80 full levels in our data.
Illustration of ICON's vertical levels, Working with the ICON Model 2024, Figure 3.2
Most parameters are stored on full vertical levels, while some — such as the vertical velocity W — are stored on half (staggered) levels.
To determine the vertical positioning of a parameter, inspect the GRIB2 key typeOfLevel:
-
generalVerticalindicates half levels -
generalVerticalLayerindicates full levels
For details on reading GRIB key values, see 2.8.2 Decoding GRIB Files with ecCodes.
For more detailed information on the vertical grid, read section 3.4 in Working with the ICON Model.
In addition to the vertical grid above the surface, there is also a grid below the land surface. In this case, the level numbers correpond directly to depths in meters below ground. For example, the parameter "soil temperature" (abbreviated as T_SO) is defined using this subsurface vertical structure.
The horizontal grid of ICON-CH1-EPS and ICON-CH2-EPS model is based on a native icosahedral grid inherited by the original ICON model grid (illustrated below).
Illustration of the grid construction, Working with the ICON Model, Figure 2.1
Since the provided data is given in the native grid, note that the grid points correspond to the center of the circumcircle of each triangle and not to the vertices. Therefore, the longitude and latitude are based in the middle of each triangle on the grid mentioned before. For more detailed information on the horizontal grid, read section 2.1 in Working with the ICON Model.
If users prefer not to use the provided library to load the data, they can retrieve datasets directly via the REST API by following the step-by-step instructions in this section to obtain forecast data for specific models, variables, and other customizable parameters.
Filtering and querying forecast data must be done using a POST request. To retrieve a forecast, use a tool like curl and send the request to the API endpoint:
curl -X POST "https://data.geo.admin.ch/api/stac/v1/search" \
-H "Content-Type: application/json" \
-d '{
"collections": [
"ch.meteoschweiz.ogd-forecasting-icon-ch2"
],
"forecast:reference_datetime": "2025-03-12T12:00:00Z",
"forecast:variable": "TOT_PREC",
"forecast:perturbed": false,
"forecast:horizon": "P0DT00H00M00S"
}'
Each parameter in the request body serves the following purpose:
collections: Defines the forecast model to use (ICON-CH1-EPSorICON-CH2-EPS).forecast:reference_datetime: Specifies the desired forecast initialization time (e.g.,2025-03-12T12:00:00Z).forecast:variable: Indicates the meteorological parameter of interest (TOT_PRECfor total precipitation, for example).forecast:perturbed: Boolean flag determining if the data is deterministic (false) or ensemble-based.forecast:horizon: Defines the lead time of the forecast in ISO 8601 duration format (P0DT00H00M00Sfor instant data).
Upon a successful request, the response will contain a dictionary of metadata, including forecast file links under the assets key. Locate the href field containing the pre-signed URL.
Download the GRIB file using the following command:
wget -O <desired_filename> “<pre-signed URL>”
After downloading your forecast data, it's good practice to verify its integrity before use.
To ensure the downloaded file is not corrupted, compute its SHA-256 hash and verify it against the checksum provided in the file's header field.
Steps:
- Open a terminal and generate the SHA-256 checksum of the downloaded file:
sha256sum <downloaded_filename>
- Retrieve the checksum from the file’s header field
x-amz-meta-sha256using the following command:
curl -s -i "<pre-signed URL>" | awk -F': ' '/x-amz-meta-sha256/ {print $0}'
- Compare the two hash values. If they match, your forecast data file is safe to use.
Once the file is verified, you can proceed with decoding the GRIB file using the instructions in 2.8.2 Decoding GRIB Files with ecCodes.
Besides the current forecast files, each catalog contains two static files. They store permanent information about the height of the half levels (HHL) in the vertical grid and the center point coordinates of each triangle on the horizontal grid.
❗ NOTE: The forecast GRIB files contain no information on height, longitude and latitude. They have to be determined via the static vertical and horizontal grid parameter files.
In the static vertical file, the heights of the half levels of the vertical grid are provided in meters above see level. In order to associate a value from a data file (for a given parameter) to a height in meters above sea level, follow the steps below:
- Submit a GET request specifying the collection you want to retrieve the static vertical files from (e.g.,
ch.meteoschweiz.ogd-forecasting-icon-ch1for ICON-CH1-EPS):
curl -X GET https://data.geo.admin.ch/api/stac/v1/collections/ch.meteoschweiz.ogd-forecasting-icon-ch1/assets
- Locate under
assetsinid: vertical_constants_icon-ch1-eps.grib2thehreffield and copy the pre-signed URL. - Download the file with:
wget -O <desired_filename> “<pre-signed URL>”
- Once the static GRIB file is downloaded, verify that the
uuidOfHGrid(Universally Unique Identifier) key in the data file matches the one in the HHL file. - Retrieve the value for the
levelkey and inspect thetypeOfLevelkey by listing the GRIB messages:- generalVertical: The value of
levelcorresponds to a half level in the HHL file. For each level (i.e., each GRIB message), the variablehprovides the height in meters above sea level for every grid point. - generalVerticalLayer: The
levelvalue corresponds to a full level. To obtain the height in meters above sea level, average the heights of the two surrounding half levels (above and below). - Other types of level: These are usually specified directly in meters and are self-explanatory.
- generalVertical: The value of
❗ NOTE: We recommend inexperienced GRIB file users to take a look at the provided Jupyter Notebooks. The data retrieval with the Python API includes fetching longitude and latitude.
The static horizontal file stores the longitude and latitude of the center points of each triangle in the horizontal grid. To retrieve this information, follow the steps below:
- Submit a GET request specifying the collection you want to download the static horizontal files from (eg.
ch.meteoschweiz.ogd-forecasting-icon-ch1for ICON-CH1-EPS).
curl -X GET https://data.geo.admin.ch/api/stac/v1/collections/ch.meteoschweiz.ogd-forecasting-icon-ch1/assets
- Locate under
assetsinid: horizontal_constants_icon-ch1-eps.grib2thehreffield and copy the pre-signed URL. - Download the file with:
wget -O <desired_filename> “<pre-signed URL>”
- Once the static GRIB file is downloaded, ensure that the
uuidOfHGrid(Universally Unique Identifier) key in the data file matches the one in the static horizontal file.
Once you have the desired GRIB files, you need a tool to read them. We recommend installing ecCodes from ECMWF.
By default, a GRIB file shows the short names defined by ECMWF. However, the ICON model has its own definitions. In order to install them, apply the steps below.
- Clone the GitHub repository eccodes-cosmo-resources into folder
<name_of_your_folder>. - Clone the GitHub repository ecmwf/eccodes into the same folder
<name_of_your_folder>.
⚠️ WARNING: Make sure both repositories are in the same folder and run on the same version.
Finally, execute the following command to set the GRIB definition path:
export GRIB_DEFINITION_PATH=<name_of_your_folder>/eccodes-cosmo-recources/definitions:<name_of_your_folder>r/eccodes/definitions
❗ NOTE: This command must be executed every time you start a new terminal session.
This section provides a brief introduction to decoding GRIB files using ecCodes. For more details, refer to the ECMWF ecCodes documentation.
Use the following commands to
- Check ecCodes installation details:
codes_info
- List all the GRIB messages in a file:
grib_ls filename.grib
- Filter GRIB messages based on key-value conditions:
grib_ls -w key1=value1,key2=value2 filename.grib
- Specify a list of keys to be printed:
grib_ls -p key1,key2 filename.grib
- Get a detailed view of the content of all GRIB messages:
grib_dump filename.grib
- Get a detailed view of GRIB messages with filters:
grib_dump -w key1=value1,key2=value2 filename.grib
⚠️ WARNING: Some variables in the ICON model are not included in the WMO standard definitions but are instead defined in ICON's local GRIB definitions. If a variable is missing, users should check the eccodes-cosmo-resources files.
...
...
There are files of data granularity ..., ..., ..., ... and update frequency hourly (now), daily (recent) or yearly (historical) for each station.
Data format is CSV with an estimated volume of ... MB per file.
See example data files: ....
See example parameter metadata files of data granularity: ... and ....
See example station metadata file.
See e.g. MeteoSwiss' ....