Efficiently access NetCDF files in S3

[forman]

Is your feature request related to a problem? Please describe.

We need to efficiently access a huge archive of climate data files in NetCDF format stored in a S3-compatible object storage via the data store framework. The data files are indexed using kerchunk.

Describe the solution you'd like

Add a new filesystem data store that supports the reference protocol.

Additional context

https://climate.esa.int/