Atacama rasters · preliminary release
Download the rasters as GeoTIFF
CHIRPS bias-corrected with the rain gauge network. Choose the dataset, the period and the format: the GeoTIFF is built in your browser from the Zarr, the compressed format stored on the server.
What to download
How the file is generated
- The server delivers the month's Zarr. Each month is a single compressed object, with values stored as integers (precipitation in thousandths of a mm).
- Your browser decompresses it and converts each integer to the dataset's unit using the scale factor declared in the Zarr.
- It writes the GeoTIFF in Float32, compressed (DEFLATE), with EPSG:4326 and nodata = −9999: one per month, or one per year with one band per month. If several files are produced, they are bundled into a ZIP together with a
metadata.json.
The file is identical, pixel by pixel, to the one produced by the Python reference script (export_geotiffs.py). No GeoTIFF is stored on the server.
From Python
For analysis, read the Zarr directly: xarray downloads only the months and the area you request.
# pip install xarray zarr fsspec aiohttp rioxarray
import xarray as xr, rioxarray
ds = xr.open_zarr("…")
pp = ds["precipitation"].sel(time="2018-02-01", x=slice(-69.6, -69.0), y=slice(-18.0, -18.6))
pp.rio.write_crs("EPSG:4326").rio.to_raster("putre_2018_02.tif")
Preliminary release. The redistribution license is not yet defined. Do not cite or redistribute these files yet.