Analyze GRIB
GRIB Analysis and pre-op meteo processing
Step 1: Inspect GRIB structure
When a GRIB file is imported, WindRose Studio inspects the dataset exactly as it exists in the original file.
Each GRIB message is read to identify the variable name, its timestamp, and the grid definition.
From these timestamps, the system determines the dataset’s temporal coverage.
The resulting metadata view is a transparent inventory of the dataset content, including variables that are not strictly hourly.
Step 2: Reconstruct domain
WindRose Studio reconstructs the effective spatial domain directly from the grid points physically contained in the GRIB file. The domain is displayed on the map and can be exported for external verification.
Exports are available as:
-
KML files, which can be opened in tools such as Google Earth and many GIS applications;
-
Shapefile (SHP), delivered as a zipped bundle, typically opened in GIS software such as QGIS or ArcGIS.
If the area-averaged projection is selected, the exported file contains the effective GRIB domain and all grid-point centres physically present in the dataset.
If a point-based projection policy is selected, the exported domain also includes:
-
the ERA5 grid-point centres found in the GRIB file, and
-
the selected projection grid point corresponding to the user’s station reference.
This allows immediate visual verification of “what the file contains” and “what grid point is actually used”.
Step 3: Spatial projection
After the domain is reconstructed, WindRose Studio asks how the dataset should be interpreted spatially.
Two policies are available:
-
Point-based projection: WindRose uses the ERA5 grid cell point nearest to the user’s station point.
-
Area-averaged projection: WindRose aggregates all ERA5 grid points contained in the GRIB domain.
No spatial interpolation is applied in either mode.
Step 4: Process hourly data
WindRose Studio builds a raw hourly table by reconstructing a continuous hourly timeline from the first to the last hourly timestamp present in the GRIB file.
All variables found in the dataset are included as columns. For each hour:
-
if a measurement exists, the value is reported as read from the GRIB message;
-
if no measurement exists for that hour/variable, the cell is marked as NA.
At this stage, WindRose does not apply meteorological transformations: the table is a structured hourly view of the GRIB content, designed to make completeness and alignment immediately visible.
Step 5: Generate outputs
Analytical outputs are generated from the validated wind series derived from the essential ERA5 components 10u and 10v.
WindRose reconstructs wind speed and wind direction deterministically from these components and then performs the statistical analyses (wind rose, frequency tables, histograms and box plots).
To ensure consistency, analyses are computed on the main analysis year (the complete annual period) and do not include any extra messages that may exist outside the year boundaries.
SAMSON export is a separate workflow: it requires the presence of additional variables in the GRIB file and applies variable-by-variable transformations to produce a regulatory-ready formatted output.