NESOSIM with MCMC calibration

NESOSIM version 1.1 with modifications to enable parameter calibration with a Markov Chain Monte Carlo method. Original model by Alek Petty; modifications made by Alex Cabaj. Refer to the main model record at https://zenodo.org/record/4448356 for additional information.

A few additional files were added to enable model parameter calibration:

mcmc_nesosim_buoy_dens_clim_io.py implements a 2-parameter MCMC calibration with observational constraints from Operation IceBridge snow depth, historical Soviet drifting station snow density, and additional snow depth measurements form CRREL-Dartmouth snow buoys. Outputs an HDF5 file containing accepted and rejected values of parameters, along with some statistics. (Will overwrite existing files unless file output paths are changed.)

mcmc_nesosim_1par_ic_io.py implements the single-parameter version of this calibration, for reference.

io_helpers.py is a script enabling the loading of multiple input files to memory to enable repeated runs of NESOSIM in the MCMC calibration without needing to re-read the same files from disk each time.

NESOSIM.py and utils.py have been slightly modified to work with the MCMC calibration (mainly file I/O).

Calibration input data is from Operation IceBridge, as aggregated by Petty et al (available at https://zenodo.org/record/6338558), CRREL-Dartmouth buoys (http://imb-crrel-dartmouth.org ) and Soviet drifting station data (Radionov et al, 1997).