eddy_scheme and srf_flux_avg

[mark@atmos_colostate_edu]

I have a question that relates to subroutine flux_avg_run.  It looks this does smoothing of the surface sensible and latent heat fluxes if the eddy_scheme is set to either “diag_tke” or “HBR” instead of “HB”.  The namelist documentation says that srf_flux_avg should get set to 1 to activate this automatically if one of those eddy scheme is picked.   But I just noticed that I have a couple of cam5 cases where the namelist shows that the eddy scheme is “diag_tke” but srf_flux_avg = 0.  Is that a bug? namelist doc:   eddy_schemeType of eddy scheme employed by the vertical diffusion package.  'HB' forHoltslag and Boville; 'diag_TKE' for diagnostic tke version of Grenier andBretherton; 'HBR' for Rasch modified version of 'HB'.Default: 'HB'   srf_flux_avgSwitch to turn on adjustment of the surface fluxes to reduce instabilitiesin the surface layer.  Set to 1 to turn on the adjustments.Default: 0 if eddy_scheme is 'HB', 1 otherwise. 

 

[eaton]

It's not a bug (except in the documentation which is wrong).  This functionality was implemented during the development of the cam5 physics package when we were encountering stability problems when coupling less frequently than every cam timestep.  If I recall correctly it was determined that for a stable solution cam needed to couple every timestep, i.e., dtime = atm_cpl_dt.  You'll see in the namelist_defaults_cam.xml file that srf_flux_avg is now always set to zero.