Global Wind Atlas
dtu-wind
Global high-resolution wind resource maps: mean wind speed and power density at several heights, produced by microscale downscaling of a mesoscale simulation.
Before you use this
- Traces to the same reanalysis family as most European weather cutouts, through its mesoscale boundary conditions. Treating it as an independent check on a reanalysis-derived time series understates uncertainty.
- ProvenanceDerived & Traced
- DocumentationFully documented
- Currency & maintenanceNot yet assessed
How many layers of assumptions sit under this is not recorded. That is not the same as none.
The Global Wind Atlas provides wind climate statistics on a 250 m grid worldwide. It is produced in two stages: a WRF mesoscale simulation driven by reanalysis boundary conditions, followed by microscale downscaling that accounts for terrain and roughness. The two-stage derivation is why it appears independent of a reanalysis-derived time series and is not — the reanalysis is two hops upstream rather than one, which is exactly the depth an unbounded provenance walk is needed to find.
- Identifier
https://globalwindatlas.info- Domains
- Renewable resource & weather
- Keywords
- WRF, atlas, downscaling, resource, siting, wind
Fitness for purpose
- Supports
- Siting and Land Use
- Not suitable for
- Production Cost Modeling, Resource Adequacy
The exclusions are the half you cannot get anywhere else. A dataset that is wrong for your question is worth knowing about before you build on it.
Structure
- Network topology
- Not captured
- Impedances
- Not captured
- Voltage classes
- Not captured
- Provenance class
- modeled
- Upstream sources
- Not capturedWhich is a statement about what the catalog can show, not about how the data was produced.
- Licence
- CC-BY-4.0
- Redistribution
- Not captured
Access terms, per path
GeoTIFF, per country and per height
none
Geographic
- Bounding box
- Global
- Granularity
- gridded
- Native CRS
- EPSG:4326
- Geometry types
- raster
- Features
- Not captured
- As WKT
POLYGON((-180 -62, 180 -62, 180 84, -180 84, -180 -62))
Temporal
- From
- Jan 1, 2008
- To
- Dec 31, 2017
- Update cadence
- irregular
- Time resolution
- P10Y
| Field | Definition | Type | Unit | Concept | Value basis | Provenance | Report an issue with this field |
|---|---|---|---|---|---|---|---|
wind-speed_100mMean wind speed at 100 m | Long-term mean wind speed at 100 metres above ground, over the 2008-2017 simulation period. Caveat: A ten-year mean, not a time series. Two sites with the same mean can have very different generation profiles. | float32 | metre per second | Wind speed | modeled | Read from: gwa-wrf-mesoscale |
The catalog says why, rather than leaving the cell blank.
Three independent facets
Graded separately and never combined. A dataset can be perfectly current and completely unprovenanced, and one number would hide exactly that.
Provenance
BDerived & Traced
Values are modeled with full lineage to a recorded origin.
Documentation
AFully documented
All 1 attribute field(s) carry a definition and a unit where one applies. Allowed ranges are not assessed: the fields that carry them are deferred (PRD §4.3), so no record in this catalog states one and grading against them would mark every record down for the same absent field.
Currency & maintenance
–Not yet assessed
The stated cadence is 'irregular', which is not an interval, so 'past due' has no meaning for this dataset.
Not the same as a poor grade. Provenance and documentation are graded from field-level metadata, which this record does not carry yet.
Visible without signing in, deliberately: an evaluator comparing catalogs should not have to register first.
- 2Complementary
- 1Related by lineage
- 5Related
- 1 not independent
- ECMWF ERA5 reanalysisStrength 2 of 5Related by lineage
Directly related by lineage: one of these is built from the other, so agreement between them is not independent evidence.
- · Shares 2 concept(s), including Wind speed.
- · Both feed Siting and Land Use.
- · Coverage overlaps from 2008-01-01 to 2017-12-31.
- · Bounding boxes overlap over 81% of their combined extent.
- · Both are gridded.
- ESA WorldCoverStrength 2 of 5Complementary
Both feed: Siting and Land Use.
- · Both feed Siting and Land Use.
- · Bounding boxes overlap over 99% of their combined extent.
- · Both are gridded.
- NREL National Solar Radiation DatabaseStrength 2 of 5Complementary
Both feed: Siting and Land Use.
- · Shares 1 concept(s).
- · Both feed Siting and Land Use.
- · Coverage overlaps from 2008-01-01 to 2017-12-31.
- · Bounding boxes overlap over 24% of their combined extent.
- · Both are gridded.
- PyPSA-Eur weather cutoutsStrength 1 of 5Related
Different physics, complementary: shares Wind speed.
- · Shares 2 concept(s), including Wind speed.
- · Coverage overlaps from 2013-01-01 to 2017-12-31.
- · Bounding boxes overlap over 4% of their combined extent.
- · Both are gridded.
- Global Solar AtlasStrength 1 of 5Related
Related through shared coverage; no stronger connection is recorded.
- · Both are gridded.
- NASA MERRA-2Strength 1 of 5Related
Related through shared coverage; no stronger connection is recorded.
- · Both are gridded.
- NREL WIND ToolkitStrength 1 of 5Related
Related through shared coverage; no stronger connection is recorded.
- · Both are gridded.
- PyPSA-Earth weather cutoutsStrength 1 of 5Related
Related through shared coverage; no stronger connection is recorded.
- · Both are gridded.
OpenGrid redirects you to the source and never proxies the bytes. For a partial read of a large dataset, use the SDK — it fetches an access plan and reads only the slice you asked for.
GeoTIFF, per country and per height
● Checked and working- Size:
- 8 GiB
- No account needed:
- Yes