Skip to content

Solar & Wind Atlas — introduction#

earthlens ships a single solar-wind-atlas backend that fetches global solar- and wind-resource climatology layers — long-term-average grids of irradiation, PV potential, wind speed and related quantities — subset to a requested bounding box and written as GeoTIFF. Two open data families are covered:

  • Global Solar Atlas (DTU / Solargis / ESMAP · World Bank) — solar-resource layers GHI, DNI, DIF, GTI plus PV output (PVOUT) and optimum PV tilt (OPTA).
  • Global Wind Atlas v3 (DTU / ESMAP) — mean wind speed at 100 m, Weibull shape parameter k at five heights, turbine capacity factor (IEC classes 1–3) and air density.

Both atlases are free and keyless (CC-BY-4.0). This page orients the backend. For the hands-on walkthrough see Usage and the layer ids on the Available layers page; the rendered API is the Reference page.

How it works — two transports#

Which transport a layer uses is set by how its atlas is hosted, not by a knob:

  • Global Wind Atlas layers are read windowed. They are range-accessible Cloud-Optimized GeoTIFFs (COGs) on figshare, so the backend reads only the bbox window over GDAL /vsicurl/ through the pyramids GIS backend (Dataset.read_part). A small area transfers a few hundred KB, not the multi-GB global file.

  • Global Solar Atlas layers are downloaded once, then cropped locally. They ship only as DEFLATE-compressed GeoTIFF ZIP archives, which cannot be range-windowed (decompressing any pixel needs the whole member). So the per-variable archive (~2.7 GB at 1 km) is downloaded once into a cache and the bbox window is read from the local member. The first fetch logs a one-time multi-GB-download warning; subsequent requests reuse the cache.

Either way, download() writes one cropped GeoTIFF per requested layer and returns their paths (list[Path]). All raster I/O goes through pyramids — earthlens never imports a competing array stack.

from earthlens.earthlens import EarthLens

paths = EarthLens(
    data_source="solar-wind-atlas",
    variables=["ghi", "wind_100m"],
    lat_lim=[55.0, 55.5],
    lon_lim=[12.0, 12.5],
    path="atlas_out",
).download()
# -> [Path('atlas_out/ghi.tif'), Path('atlas_out/wind_100m.tif')]

Two things that shape the backend#

  • The layers are static — there is no time axis. Each layer is a single long-term-average climatology grid, not a time series, so download() has nothing to reduce over time. The EarthLens facade forwards aggregate= to raster backends, so this backend explicitly rejects a non-None aggregate= with NotImplementedError.

  • A request is a bbox subset. The full atlas grids are global; this backend returns only the bbox you ask for — read windowed for wind, cropped from the cached download for solar.

Attribution#

Both atlases are CC-BY-4.0; cite them when you use the data (the backend also logs the attribution once per atlas fetched):

  • Global Solar Atlas — Global Solar Atlas 2.0, © The World Bank, Solar resource data: Solargis.
  • Global Wind Atlas — Global Wind Atlas 3.0, © Technical University of Denmark (DTU) / World Bank (ESMAP).