Validating reconstructions on instrumental data

[CommonClimate]

1. Goal: Paleoclimate Reconstructions need to be thoroughly validated on independent data before they can be trusted. One or more notebooks can show how to this for various variables (e.g. SAT, MSLP, Precip) using early instrumental data that was not used for calibration.
2. Ideas: Validate both LMRv2.1 and LMRonline fields against HCLIM and GloSAT. For offline DA like LMRv2.x, pens will be useful.
3. References/data sources:

• LMRv2.1 should (still) be "analysis-ready cloud-optimized" (ARCO). Re-use @jordanplanders 's notebook to load it easily.
• We should explore making LMRonline output ARCO as well. However, if fields can be lazy-loaded in xarray directly from Zenodo, that may not be necessary - the only thing that counts is to make it easy to load!
• Early instrumental data: we have HCLIM and GloSAT

6. Keywords (science): validation, uncertainties, field
7. Keywords (tech): xarray, pens

[CommonClimate]

Ideas for notebooks:

• validating timeseries of temperature with a) visual comparison (at a few interesting locations, plot the ensemble mean of each LMR 20 MC run as thin gray lines, and the HCLIM series as a thicker colored line) + b) quantitative validation, hacking cfr's code, as in compare().
• spatial validation, for 1816/17, the "year without a summer" associated with the Tambora eruption. Re-use the "VICS dashboard" for this purpose and overlay values of SAT, Precip or MSLP as present in HCLIM.