to do 2: load MAR NCs in pangeo
Closed this issue · 2 comments
Clone this repo and make a new notebook in https://us-central1-b.gcp.pangeo.io/ which loads the NCs.
Use the first few block of this notebook as a guide (through to the block titled "Loop through list of URLs").
The workflow in that NB is:
-- import all packages needed
-- define a filename pattern (start simple and just write it out then build up to something you can loop over).
-- loop over the NCs and use xarray.concat to join them together along the time domain.,
An alternative. which might work is using xarray.open_mfdatasets. If this works it is easier.
xarray.open_mfdatasets does work.
This notebook implements the workflow described above.
The next step is to save the whole dataset as a zarr. The above NB does this just for the runoff variable, which @glugeorge is most interested in initially.