isomip_plus

The isomip_plus test group (compass.ocean.tests.isomip_plus.IsomipPlus) implements variants of the ISOMIP+ experiments (see isomip_plus in the User’s Guide). Here, we describe the shared framework for this test group and 3 test case (Ocean0, Ocean1 and Ocean2) that have currently been implemented.

framework

The shared config options for the isomip_plus test group are described in isomip_plus in the User’s Guide.

Additionally, the test group has a shared namelist.forward file with a few common namelist options related to forcing and analysis members, as well as a shared streams.forward.template file that defines mesh, input, forcing_data, timeSeriesStatsMonthlyOutput, restart, output, globalStatsOutput, and land_ice_fluxes streams.

evap

The function compass.ocean.tests.isomip_plus.evap.update_evaporation_flux() is used to update the “evaporation” rate and other surface fluxes near the northern boundary that are used to mimic a spillway, preventing sea level from rising indefinitely due to the input of freshwater from ice-shelf melting.

geom

The function compass.ocean.tests.isomip_plus.geom.interpolate_ocean_mask() interpolates the ocean mask from the BISICLES grid of the input geometry to the MPAS-Ocean mesh. The mask can later be used to cull land cells from the MPAS mesh before interpolating other variables to the resulting culled mesh. Optionally, a thin film under the ice sheet can be used and grounded ice is not culled.

The function compass.ocean.tests.isomip_plus.geom.interpolate_geom() interpolates the remaining geometric variables:

• bottomDepthObserved – the bedrock elevation (positive up)

• ssh – the sea surface height

• oceanFracObserved – the fraction of the mesh cell that is ocean

• landIceFraction – the fraction of the mesh cell that is covered by an ice shelf

• smoothedDraftMask – a smoothed version of the floating mask that may be useful for determining where to alter the vertical coordinate to accommodate ice-shelf cavities

process_geom

The class compass.ocean.tests.isomip_plus.process_geom.ProcessGeom defines a step for processing ISOMIP+ geometry before interpolating it to the MPAS mesh. This includes applying a simple calving scheme based on a threshold in ice-shelf thickness, then apply smoothing to the topography data. Optionally, the ice draft can be scaled by a factor as a simple way to explore changing ice-shelf topography. Variables are renamed to those expected by MPAS-Ocean.

planar_mesh

The class compass.ocean.tests.isomip_plus.planar_mesh.PlanarMesh defines a step for generating a planar mesh.

cull_mesh

The class compass.ocean.tests.isomip_plus.cull_mesh.CullMesh defines a step for culling the mesh to include only ocean cells, with or without a thin film depending on the test.

initial_state

The class compass.ocean.tests.isomip_plus.initial_state.InitialState defines a step for setting up the initial state for each test case.

First, a mesh appropriate for the resolution is generated using mpas_tools.planar_hex.make_planar_hex_mesh(). Then, the mesh is culled to remove land cells. When a wetting-and-drying algorithm that relies on a thin film under land ice is used, a distance up to that specified by the config option nx_thin_film is retained. A vertical coordinate is generated, with 36 layers of 20-m thickness in the open ocean by default. By default, the Vertical coordinate is z-star, meaning the 1D grid is “squashed” down so the sea-surface height corresponds to the location of the ice-ocean interface (ice draft). The initial temperature and salinity profiles are computed along with zero initial velocity. Finally, forcing data fields are produced for restoring to temperature and salinity profiles at the northern boundary and for “evaporative” fluxes at the surface that are used to mimic a spillway, removing water at the northern boundary and preventing runaway sea-level rise from the the incoming ice-shelf meltwater.

For the time-varying version of a test case, initial_state also computes a set of time-varying landIcePressureForcing and landIceDraftForcing fields, based on the isomip_plus_forcing config options (see time_varying_Ocean0). The time evolution of the landIcePressure and landIceDraft fields is determined by linear interpolation in time between consecutive entries in the these forcing fields, which are stored in a file land_ice_forcing.nc.

Grounding line motion is allowed to occur for a subset of test cases with have the attribute thin_film_present set to true. For all other test cases, the grounding line and calving front are held fixed in time, so the field landIceFractionForcing is the same as landIceFraction in the initial condition for all time.

The initial_state step also generates horizontal sections through the domain of layer thicknesses and the mid-layer depth as well as horizontal sections of initial SSH, land ice presure, and total water column thickness.

ssh_adjustment

The class compass.ocean.tests.isomip_plus.ssh_adjustment.SshAdjustment performs sea-surface height adjustment described Ice-shelf cavities. Starting from the initial condition from initial_state, the test case performs a number of iterations (10 by default) of forward simulation followed by adjustment of the land-ice pressure field.

forward

The class compass.ocean.tests.isomip_plus.forward.Forward defines performance and simulation steps for running MPAS-Ocean from the initial condition produced in the initial_state step. A link to the MPAS-Ocean executable is created when the test case is set up and MPAS-Ocean is run (including updating PIO namelist options and generating a graph partition) in run().

The performance step is run for only 1 hour (appropriate for regression testing) except when tidal forcing is applied, in which case the run duration is 24 hours. Then, potential temperature and salinity are plotted at the top and bottom of the ocean and along a cross section of through the middle (y = 40 km) of the domain.

The simulation step runs for 1 month, then adjusts the “evaporative” forcing based on the average of the melt fluxes from that month. Then, namelist options are modified so the simulation is ready to run for another month.

See isomip_plus for a fuller description of how to use the performance and simulation steps.

streamfunction

The class compass.ocean.tests.isomip_plus.streamfunction.Streamfunction defines a step for computing the barotropic (vertically integrated) and overturning streamfunctions from the latest simulation results from the simulation step. This step is intended to be run repeatedly each time new simulation results come in, but can also be run once at the end of a longer simulation.

viz

The compass.ocean.tests.isomip_plus.viz.Viz class defines a step for performing visualization of ISOMIP+ results. This step should be run after running simulation any number of times and then streamfunction (unless you set plot_streamfunctions = False in the [isomip_plus_viz] section of the config file). Movie frames an time series plots will appear in the plots directory; The movies themselves in movies, and some time series averaged only over the deepest parts of the ice draft in timeSeriesBelow300m.

misomip

The compass.ocean.tests.isomip_plus.misomip.Misomip class defines a step for interpolating the results to the standard MISOMIP grid and writing out the results in the format expected by MISOMIP.

Note

There is currently an issue with fill values not being handled correctly that needs to be resolved before this step is fully useful.

isomip_plus_test

The same class, compass.ocean.tests.isomip_plus.isomip_plus_test.IsomipPlusTest, defines the Ocean0, Ocean1 and Ocean2 test cases at various resolutions and with various vertical coordinates. By default, these test cases only run 3 of the 7 available steps: initial_state to create and mesh and initial condition, ssh_adjustment to perform 10 1-hour simulations used to balance the land-ice pressure with the sea surface height, and performance to run a final 1-hour (15-time-step) forward simulation. If a baseline is provided when calling compass setup, a large number of variables (both prognostic and related to land-ice fluxes) are checked to make sure they match the baseline.

The optional simulation, streamfunction, viz and misomip steps, described above, are used to perform longer simulations and perform analysis and visualization.