zarr-fluxnet2015
FLUXNET2015 as Zarr store
Rehost of the FLUXNET 2015 dataset as a Zarr store for easier access.
It currently serves as a demo of what is possible. Only six variables (and some quality flags) are included at the moment. The total Zarr store size is only 200 MB, so more variables can still be included here on GitHub.
How to use
You need a Python environment with the following packages installed: requests aiohttp zarr xarray
Then you can do:
import xarray as xr
uri = 'https://excited-co2.github.io/zarr-fluxnet2015/fluxnet.zarr'
ds = xr.zarr(uri)
ds["gpp_nt"].sel(
site=["AT-Neu", "US-Sta", "ZM-Mon"] # select three sites
).groupby(
ds["time"].dt.dayofyear # groupby day-of-year
).mean().plot.line(hue="site")
Citations
Pastorello, G., Trotta, C., Canfora, E. et al. The FLUXNET2015 dataset and the ONEFlux processing pipeline for eddy covariance data. Sci Data 7, 225 (2020). https://doi.org/10.1038/s41597-020-0534-3
How to reproduce
Download FLUXNET fullset for all sites.
Get page source, paste into links.txt. Then:
cat links.txt | grep -oP 'https(\S*)bschilperoort' > stripped_links.txt
xargs -n 1 curl -O < stripped_links.txt
Create a python virtual environment:
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
Then run the processor:
python3 create_zarr.py