Updated land perturbation scheme

[ClaraDraper-NOAA]

1. Clean up of the pre-existing land perturbation scheme (was called with do_sfcperts) that was intended for use with GEFS forecasts. This scheme is not being used, and needed some clean-up to work robustly. This scheme has been retained as lndp_type=1 (output is zero diff to old scheme) – think about removing in the future?

2. Addition of new land surface perturbation option (lndp_type=2), intended for use in cycling DA, and designed to sit outside of the forecast model code.

Regressions tests have passed on hera, and I’ve added a fv3_ccpp_lndp regression test for the lndp_type=2 option.

Detailed list of changes:

In stochastic physics:

• logical do_sfcperts replaced with integer lndp_type (to enable selection of multiple land pert scheme options)
• deleted land_sppt namelist entry (land SPPT has not been implemented, and experiments indicate it has minimal impact)
• changed namelist names for consistent formatting, and to specify land (lnd) rather than more generic surface (sfc)
• hard-wired npsfc (now nplnd) dimension for old scheme, as was not coded correctly for npsfc >1 (this is the number of patterns used for the surface perturbations)
• replaced hardwired index in old scheme for the perturbed variables in sfc_wts with scheme to assign indexes according to which variables are specified for perturbations
• merged run_stochastic_physics_sfc routine into run_stochastic_physics
• simplified namelist inputs, by replacing individual variables in each namelist with an input vector of variable codes for which perturbations are requested (lndp_var_list), and a vector of perturbation magnitudes (lndp_prt_list)
  *For lndp_type=2, added capacity to apply land perturbations throughout forecast (do_advance_pattern), and to include the land perturbations in the stochini file.

Also, fixed a couple of minor bugs, most notably:
*added missing error trapping in stochastic physics (may not have done this everywhere that is necessary).

In FV3:
*merged run_stochastic_physics_sfc routine into run_stochastic_physics

• added calls to new land perturbation scheme
• added error trapping for abnormal exit from run_stochastic_physics
• added GFS_land_perts module with routines to apply the new perturbations
• replaced hardwired index in old scheme for the perturbed variables in sfc_wts with scheme to assign indexes according to which variables are specified for perturbations
• various changes to variables in namelists and type_defs

In ccpp/physics:

*variable name updates

• replaced hardwired index in old scheme for the perturbed variables in sfc_wts with scheme to assign indexes according to which variables are specified for perturbations

In UFS:
• Updated variable names in namelists
• Addition of lndp test

ASSOCIATED PULL REQUESTS:
stochastic_physics 28
ccpp-physics 480
fv3atm 148
ufs-weather-model 176

[pjpegion]

Changes here look good

[climbfuji]

Thanks for making these changes, this looks good to me!

[ClaraDraper-NOAA]

I just sent you an email - I pushed it, then realised it's not compiling.

…

On Wed, Aug 26, 2020 at 4:27 PM Dom Heinzeller ***@***.***> wrote: ***@***.**** approved this pull request. Thanks for making these changes, this looks good to me! — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#28 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AH7BYD544R6263OMZJMH243SCWD3RANCNFSM4PJISWCQ> .

-- -------------------------------------------------------------------- Clara Draper CIRES / NOAA Earth System Research Laboratories Physical Sciences Laboratory 325 Broadway, Boulder, CO DSRC, Rm 1D-105 +1 303 497 4467 --------------------------------------------------------------------