f.cam6_3_132.FMTHIST_ne30.sci_opt_roughTopo.001

PeterHjortLauritzen commented 10 months ago

Description:

Investigate if rougher topography has an impact on stratospheric wind biases in FMT.

Same setup as "science optimized" code (https://github.com/NCAR/amwg_dev/issues/470) but using a smoothing radius of 50km instead of 100km when smoothing PHIS.

Command used to generate PHIS:

qcmd -l walltime=4:00:00 -l select=1:ncpus=1 -- ./cube_to_target --grid_descriptor_file ../regression-test-data/ne30pg3.nc --intermediate_cs_name /glade/campaign/cgd/amp/pel/topo/cubedata/gmted2010_modis_bedmachine-ncube3000-220518.nc --grid_descriptor_file_gll ../regression-test-data/ne30np4.nc --output_grid ne30pg3 --smoothing_scale 50.0 --name_email_of_creator 'Peter Hjort Lauritzen, pel@ucar.edu'

Case directory: /glade/campaign/cesm/cesmdata/cseg/runs/cesm2_0/f.cam6_3_132.FMTHIST_ne30.sci_opt_roughTopo.001

On github: https://github.com/NCAR/amwg_dev/tree/$CASE

Sandbox:
Locally (if still available): /glade/work/hannay/cesm_tags/cam6_3_132 with code mods from https://github.com/NCAR/amwg_dev/issues/470
On github: https://github.com/ESCOMP/CAM/tree/cam6_3_132

Diagnostics:
ADF diags (if available) https://webext.cgd.ucar.edu/FMTHIST/f.cam6_3_132.FMTHIST_ne30.sci_opt_roughTopo.001/atm/f.cam6_3_132.FMTHIST_ne30.sci_opt_roughTopo.001_1996_2006_vs_f.cam6_3_132.FMTHIST_ne30.sci_opt.001_1996_2006/

Output: /glade/derecho/scratch/pel/archive/f.cam6_3_132.FMTHIST_ne30.sci_opt_roughTopo.001/

Contacts: @cecilehannay @PeterHjortLauritzen @JulioTBacmeister @tilmes @islasimpson

Extra details:

PeterHjortLauritzen commented 10 months ago

PHIS difference between rougher topography and default:

(Colorbar is from -4000 to 6000)

Note ~600m difference in elevation near Antarctic peninsular, Southern Greenland, ...

In terms of numerical noise the rougher topography does not seem be an issue in terms of common fields I look at:

10 year average of OMEGA500 (rough topo left and default right):

10 year average of PSL:

10 year average of absolute value of surface pressure tendency:

JulioTBacmeister commented 10 months ago

Image below shows September-mean zonal-mean zonal wind from rough topo (middle) control(left) and ERA (right, only one year)

The rougher topo looks like it's driving things in the right direction in the southern hemisphere. It appears more drag is still needed though.

islasimpson commented 10 months ago

Here's some aspects of the zonal mean zonal wind. I think the effect on the SH is pretty minimal.

uzm ush

islasimpson commented 10 months ago

Sorry, forgot the legend in the above plot. blue is ctrl and red is rough in the above plot. ERA5 is black.

JulioTBacmeister commented 10 months ago

It seems the rougher topography is a small step in the right direction in the SH and doesn't do noticeable harm in the NH. @adamrher is the dual polar grid easy to use? I think it might be interesting to see what the SH looks like in that configuration.