search

MPAS-Dev / MPAS

Repository for private MPAS development prior to the MPAS v6.0 release.
Other
4 stars 0 forks source link

Updates to dome (Jigsaw var res) and EISMINT1 test cases #1503

Closed matthewhoffman closed 6 years ago

matthewhoffman commented 6 years ago

This merge updates two test cases.

  1. It adds a new variable resolution mesh generated by Jigsaw for the Halfar dome test case. The mesh is not automated into a COMPASS test case. The script has to be manually run through Jigsaw in Matlab, and then the result needs to be manually run through the steps to make a landice mesh. These steps could be automated in the future. The mesh is used in the MPASLI model description paper. It also modifies the Halfar analysis script to calculate an area-weighted RMS metric.
  2. The EISMINT1 test is updated to work correctly for temperature, as well as thickness. This required changes to the test case setup and analysis, as well as the addition of a few new model config options.

New model config options added to support temperature test for EISMINT1:

  1. support for surface air temperature coming from a lapse rate.
  2. force tracer advection to be activated if temperature evolution is turned on
  3. option to disable grounded BMB field from being calculated from the thermal evolution
pwolfram commented 6 years ago

@matthewhoffman, if you want this fully automated with JIGSAW you could use a similar workflow to the one I used in https://github.com/MPAS-Dev/MPAS/pull/1492 to call JIGSAW directly.

pwolfram commented 6 years ago

I'm thinking we want some type of standard to facilitate this in the long term.

matthewhoffman commented 6 years ago

Thanks @pwolfram - good to note that. For the purposes of this mesh, simply dumping it in the repo for future reference is sufficient, but in the long run I could imagine building an automated test around it, especially if we are able to remove the Matlab wrapper from around JIGSAW (e.g. replace with python). But we won't want that kind of usage for awhile (timescale of many months).

matthewhoffman commented 6 years ago

After changes to EISMINT1 setup, new stats are:

====================================
Max modeled thickness (m) = 2976.68217741
EISMINT models ice thickness at divide (m):
 3d models (10 of them): 2978.0 +/- 19.3
 2d models (3 of them):  2982.2 +/- 26.4
====================================

====================================
Basal homologous temperature at divide (deg C) = -13.0860572743
EISMINT models basal temperature at divide (m):
 3d models (6 of them): -13.34 +/- 0.56
====================================