Add new implementation of downscaling to pipeline

[gilbertozp]

New implementation of downscaling retrieving data directly from ECMWF and running downscaling algorithm locally.

Options/steps for implementation:

• independent implementation
• incorporated into regular pipeline

[darpap]

I write here the summary of the discussion had with Gilberto:

• move to monthly netCDF with all vars (Global about 7-10 Gb?)
• available in different data processing centers
• netCDF open, extratction and processing directly implemented in ONEFlux. Better if available also as standalone tool
• ERA stored directly in Cluster, GDrive for backup may be

[gilbertozp]

Partially solved in PR #72 and portion not solved now in Issue #71