Methodology:

The data is 'raw' output of the Massachusetts Institute of Technology general circulation model (MITgcm; mitgcm.org). In each model run the diagnostics tool output yearly the temporal average of various diagnostics, including those provided, over the preceding year. The runs were done using version c66j, with the addition of the changes described in Snow K et al (2017, Geophys Research Letters) as well as Jordan J et al (2017, JGR-Oceans, doi:10.1002/2017JC013251). All relevant parameters given in the supplement of Snow K et al (2017, Geophys Research Letters).

The files are in Network Common Dataform (netcdf) format. A model output is provided for each experiment. There are 4 different types of experiment:

COUPLED: output from synchronously coupled runs

UNCOUPLED_SYM: ice model-only with a transversely symmetric depth-dependent parameterised melt as detailed in Snow K et al (2017, Geophys Research Letters)

UNCOUPLED_ASYM: ice model-only with a transversely asymmetric depth-dependent parameterised melt as detailed in Snow K et al (2017, Geophys Research Letters)

SLOW: ice model-only with a transversely symmetric depth-dependent parameterised melt as detailed in Snow K et al (2017, Geophys Research Letters), and elevated basal drag and in order to yield a slower-flowing ice sheet and shelf

A different file exists for each experiment with naming convention [EXPT TYPE]_[PERIOD_IN_YEARS]_YEAR.nc, with the addition of a single file for each expt type: [EXPT_TYPE]_CONST_FORCING.nc in which all external forcing is time-independent.

Dimensions of these files are:

X (orthogonal horizontal coordinate in meters)

Y (orthogonal horizontal coordinate in meters)

time (years)

NB: as detailed in Snow K et al (2017, Geophys Research Letters), the domain is idealised and an f-plane is used for Coriolis force, so the domain can be rotated without loss of generality.

Fields are:

double THICKNESS(time, Y, X) - the ice thickness (floating and grounded) in meters

double MELTRATE(time, Y, X) - the melt/freeze rate in meters ice per year (positive where melting)

double HEIGHT_ABOVE_FLOAT(time, Y, X) - the portion of the ice column greater than the maximum floating thickness (calculated as double MASK(time, Y, X)) - a file indicating the parts of the domain that actually represent the computational domain (ignores walls, etc). Equal to 1 where there is ice sheet or shelf, 0 where there is open ocean, and -1 outside of the physical domain.

One additional file is provided with dimensions X,Y only, detailing the bedrock elevation: BED.nc.

Data quality:

No "cleaning" was involved. The accuracy is subject to the ability of the model to approximate the differential equations solved, and the ability of the differential equations solved to represent reality, both of which issues have been discussed extensively in the literature (e.g. Durran et al (1999), Numerical Methods for Fluid Dynamics:with Applications in Geophysics, Springer).