Code, Input, and Outputs for Century-scale impacts of ice-sheet model initialization on Thwaites Glacier, West Antarctica

The following guide describes the code, input and data files contained in this repository. Code, Input and Scripts The file code_input_scripts.zip is mainly an archive of the github repo used to save modified MITgcm code; and matlab, python and shell scripts used to create binary inputs for and to run the coupled model and initialisation/spinup. code_ice/ All modified source files for the STREAMICE model used in the snapshot and transient calibrations as well as the coupled model, not in MITgcm version 69c. code_oce/ All modified source files for the MITgcm ocean model used in the spinup and coupled model, not in https://doi.org/10.5281/zenodo.5777459. coupling/ contains bash scripts and python utilities for building, running and managing coupled runs on ARCHER2 supercomputer. ice_init/ contains scripts for building, running and managing snapshot ice-sheet calibration on ARCHER2 supercomputer. ice_init_tc/ contains scripts for building, running and managing transient ice-sheet calibration on ARCHER2 supercomputer. spinup/ contains scripts for building, running and managing static ice-shelf ocean spinup runs on ARCHER2 supercomputer. python_scripts/ contains python utilities for running and managing workflow on ARCHER2 supercomputer. input/ contains matlab scripts for generating binary inputs for running the work flow input/start_2013_input/ice contains MITgcm and STREAMICE runtime parameter files: data (general MITgcm parameters) data.autodiff (Automatic differentiation runtime parameters) data.cost (Cost package runtime parameters) data.ctrl (Control package runtime parameters) data.diagnostics (Diagnostics package runtime parameters) data.optim (M1QN3 runtime parameters) data.streamice (STREAMICE runtime parameters) As well as all binary input files used in workflow for ice-sheet executable. Note that the run scripts modify the runtime parameter files, so the modified versions can be seen within the stdout streams of the ice- and ocean models (see below). input/start_2013_input/oce contains MITgcm, SHELFICE and STREAMICE runtime parameter files. data.obcs (open boundary condition package runtime parameters) data.cal (calendar package parameters) data.ext (external forcing package parameters) data.shelfice (runtime parameters of SHELFICE, ice-ocean interactions package) other parameter files described above As well as all binary input files used in workflow for ocean executable. In particular, the RCP85 boundary condition files are named with the convention: [NAME].OB[S/W].80_[YYYY] containing monthly gridded values for each year of the variable NAME, where NAME is 'salt', 'temp', 'uvel', or 'vvel'; S or W indicate the bottom y- or left x-boundary, respectively; and YYYY indicates the year. 80 is the code assigned to RCP85 experiments. The baseline climatogy files have the name pattern: [NAME].OB[S/W].40 Model Output used in Manuscript The file modeloutput.zip contains a number of outputs. It contains python bytestream (.pkl) files, containing model outputs that can be used to generate the figures in the manuscript. In addition, it contains the stdout streams from a single run of the ice and ocean executables. std1: the stdout stream from the running of the ice-sheet executable. The parameter values for each runtime parameter in running the model can be seen. std1555200: the stdout stream from the running of the ocean executable. The parameter values for each runtime parameter in running the model can be seen. Snapshot_Baseline_output.pkl: For the Snapshot-calibrated experiment with Baseline forcing, contains a python dictionary with gridded output for years: [2013,2014,2018,2100,2150,ridge_year,fin_year] where 'ridge_year' is a representative year for grounding on Upper Thwaites Ridge and 'fin_year' is the last year of the simulation. For each year, contains: Ice velocity (easterly and northerly) ice mask (where there is ice=1, ocean=0, or out of bounds=-1 in STREAMICE) grounded mask for ice (=1 where grounded) ice-shelf melt (m/a) bottom temperature (deg C) buttressing number In addition, ice and ocean grid coordinates and model bathymetry Snapshot_RCP85_output.pkl: similar to above for Snapshot-calibrated experiment with RCP85 forcing Transient_Baseline_output.pkl: similar to above for Transient-calibrated experiment with Baseline forcing Transient_RCP85_output.pkl: similar to above for Transient-calibrated experiment with RCP85 forcing Fig10_output.pkl: ice velocity and grounded mask for every panel in Fig. 10 (biannual plots of velocity change and grounding line over the final 12 years of Transient Baseline experiment) Time series outputs Time series outputs are provided in the files with convention [gl]_[type].pkl, where [gl] takes values ['Pig', 'Thw', 'Sm'] and [type] takes the values ['VAF', 'melt_arrays', 'Deepmelt_arrays', 'Meanmelt_arrays']. 'melt_arrays' refer area-integrated melt (Gt/a); 'Deepmelt_arrays' contain integrated melt deeper than 500m, and 'Meanmelt_arrays' contain area-weighted mean melt (m/a). The time series pkl files contain 10-member tuples of lists, not dictionaries. Members 1, 3, 5, and 7 contain time-level lists, and Members 2, 4, 6, and 8 contain the lists of quantities at the corresponding times, for the different experiments; member 9 contains a list of "forcing" codes (40 or 80), and member 10 contains a list of "initialisation codes", for the experiments. To access outputs in a file, use the python commands: >>> import pickle as pkl >>> f = open([filename],'rb') >>> Q = pkl.load(f)