Introduce initial validation script

[alexolinhager]

This PR introduces a new tool to validate Greenland ocean thermal forcing parameterizations using historical CTD casts. The script takes a MALI output file and calculates a Willmott model skill score of TF profiles at every cell/month where CTD data exists. If multiple casts align with the same cell in a given month, then the casts are averaged together before calculating a skill score. Currently, the script calculates two separate skill scores, an average skill score for cells with originally defined ocean cells (orig3dOceanMask=1) and an average of all extrapolated ocean cells. (Note that the MALI thermal forcing parameterization alters ALL ocean cells, so this script must be run on a simulation with config_extrap_sill_blocking = false and config_TF_iceberg_melt = false to calculate a skill score for the unaltered ocean forcing dataset).

Example usage:
python validate_mali_extrap.py --maliDiag output.2002.nc --HadleyDir HadleyEN4.2.2 --geojson UummannaqDisko.geojson --maliForcing ecco_to_mali_remapped.1to10km_r01_20251030.nc

where maliDiag is the MALI output file, HadleyDir points to the location of the directory containing the Hadley EN4.2.2 dataset to validate against, geojson is a geojson file delineating the geographic boundary to use for validation, and maliForcing is the input MALI ocean forcing file.

As of now, Hadley EN4.2.2 is the only validation dataset supported, although others should eventually be incorporated (see https://e3sm.atlassian.net/wiki/spaces/FAN/pages/5908889649/Uummannaq+Disko+Bay+Regional+Ocean+Evaluation). All relevant CTD datasets are stored at: /global/cfs/cdirs/fanssie/standard_datasets/UummanaqDisko_datasets/CTD_validation

To do:

• Incorporate other CTD datasets into validation, starting with Bartholomaus et al. (2016)

• Filter out unreliable CTD casts/data

• Calculate correlation between MALI/ECCO and CTD grounding line and depth-averaged thermal forcing – Frontal ablation may not care about full profile accuracy.

• Time-lagged cross-correlations between Barthomaus et al. (2016) mooring CTD time series and extrapolated MALI/ECCO thermal forcing at similar depths/locations.

• Create plots for above analysis