Experiment with ingesting NWP data from raw files, including relevant params at cloud heights

[JackKelly]

Try ingesting directly from the NetCDF files from the Met Office. If that doesn't work then create a new script for generating an intermediate Zarr file of NWP data.

All of "file 3" listed in this Met Office PDF look useful for helping to predict the evolution of clouds.

The "dream" would be that the user could request a list of NWP parameters when nowcasting_dataset runs, and the code will figure out which Met Office GRIB file to load from.

Related issues:

• Check for -1 values in sat data #29
• Re-create NWP Zarr with one chunk per init_time and step and small bit depth #30
• Re-chunk Satellite Zarr with one chunk for all channels and full spatial extent #44
• Download additional UK Met Office UKV NWPs #196

[JackKelly]

Loading data from raw NWPs will be way too slow. It takes 1.8 seconds to open each grib file!

Here's the notebook: https://github.com/openclimatefix/nowcasting_dataset/blob/jack/experiment-reading-nwps/notebooks/load_raw_nwps_benchmark.ipynb

I'll start a new issue about converting NWP grib files to Zarr...