Wind resource layers — multi-variable windowed read (live)¶

One download() call can fetch several Global Wind Atlas layers for the same bounding box. Each is read windowed from its remote Cloud-Optimized GeoTIFF (only the AOI transfers), and download() returns one written GeoTIFF per layer. Here we pull the mean wind speed, the IEC-class-2 turbine capacity factor, and the air density at 100 m for a small box near DTU (Denmark) and compare them.

Runs live against public figshare (no credentials).

In [1]:

from pathlib import Path

import numpy as np
from pyramids.dataset import Dataset

from earthlens.earthlens import EarthLens

from pathlib import Path import numpy as np from pyramids.dataset import Dataset from earthlens.earthlens import EarthLens

2026-06-28 01:20:50 | INFO | pyramids.base.config | Logging is configured.

Fetch three wind layers at once¶

Pass several layer ids in variables=; the backend resolves each against the catalog and fetches it by its transport (all three here are vsicurl windowed reads). The returned list has one GeoTIFF path per requested layer, in order.

In [2]:

layers = ["wind_100m", "capacity_factor_iec2", "air_density_100m"]
paths = EarthLens(
    data_source="solar-wind-atlas",
    variables=layers,
    lat_lim=[55.0, 55.5],
    lon_lim=[12.0, 12.5],
    path="wind_layers_out",
).download(progress_bar=False)
for layer, path in zip(layers, paths):
    print(f"{layer:22s} -> {path.name}")

layers = ["wind_100m", "capacity_factor_iec2", "air_density_100m"] paths = EarthLens( data_source="solar-wind-atlas", variables=layers, lat_lim=[55.0, 55.5], lon_lim=[12.0, 12.5], path="wind_layers_out", ).download(progress_bar=False) for layer, path in zip(layers, paths): print(f"{layer:22s} -> {path.name}")

2026-06-28 01:20:52.276 | INFO     | earthlens.solar_wind_atlas.backend:_fetch_one_layer:256 - solar_wind_atlas wind_100m: windowed /vsicurl read https://ndownloader.figshare.com/files/17247017

2026-06-28 01:20:55.174 | INFO     | earthlens.solar_wind_atlas.backend:_fetch_one_layer:256 - solar_wind_atlas capacity_factor_iec2: windowed /vsicurl read https://ndownloader.figshare.com/files/17281778

2026-06-28 01:20:55.988 | INFO     | earthlens.solar_wind_atlas.backend:_fetch_one_layer:256 - solar_wind_atlas air_density_100m: windowed /vsicurl read https://ndownloader.figshare.com/files/17281898

2026-06-28 01:20:56.906 | INFO     | earthlens.solar_wind_atlas.backend:_log_attribution:313 - solar_wind_atlas attribution: Global Wind Atlas 3.0, © Technical University of Denmark (DTU) / World Bank (ESMAP). Licensed CC-BY-4.0.

wind_100m              -> wind_100m.tif
capacity_factor_iec2   -> capacity_factor_iec2.tif
air_density_100m       -> air_density_100m.tif

Inspect each layer¶

Read every written GeoTIFF back with pyramids and summarise its value range. All three share the same EPSG:4326 grid and AOI; only the quantity differs.

In [3]:

rows = []
for layer, path in zip(layers, paths):
    ds = Dataset.read_file(str(path))
    a = np.asarray(ds.read_array(), dtype='float64')
    finite = a[np.isfinite(a)]
    rows.append((layer, ds.columns, ds.rows, round(float(finite.min()), 3), round(float(finite.mean()), 3), round(float(finite.max()), 3)))

import pandas as pd
pd.DataFrame(rows, columns=['layer', 'cols', 'rows', 'min', 'mean', 'max'])

rows = [] for layer, path in zip(layers, paths): ds = Dataset.read_file(str(path)) a = np.asarray(ds.read_array(), dtype='float64') finite = a[np.isfinite(a)] rows.append((layer, ds.columns, ds.rows, round(float(finite.min()), 3), round(float(finite.mean()), 3), round(float(finite.max()), 3))) import pandas as pd pd.DataFrame(rows, columns=['layer', 'cols', 'rows', 'min', 'mean', 'max'])

Out[3]:

	layer	cols	rows	min	mean	max
0	wind_100m	201	201	7.452	8.648	9.656
1	capacity_factor_iec2	201	201	0.462	0.568	0.647
2	air_density_100m	201	201	1.218	1.232	1.234

What the numbers say¶

• wind_100m is the mean wind speed at 100 m in m/s.
• capacity_factor_iec2 is the fraction of rated output an IEC-class-2 turbine would average here (0-1) — it tracks the wind speed.
• air_density_100m (kg/m3) scales the power a turbine extracts at a given speed.

Because all three are read windowed from their global COGs, fetching three layers for a small box still transfers only a few hundred KB each — not the multi-GB global files. (Global Solar Atlas layers like ghi would instead download their full ~2.7 GB archive once, then crop locally.)