WACCM bug in TH2d history variable dimensions

[mmills@ucar_edu]

In WACCM 3.1.9 the history variable TH2d is output as follows:

float TH2d ( time, lat, lon )
units : K
long_name : Zonal-Mean potential temp
cell_method : time: mean

The dimensions of this variable should be ( time, lat, lev ), not ( time, lat, lon ).

Fortunately, for standard 66-level runs at 4x5 resolution, the number of lons is greater than the number of levs. But potential temperature information may be truncated for 125-level runs.

 

[mmills@ucar_edu]

I've just been told this is a hack, not a bug. It is implemented thusly because for 2D fields, the history code only outputs time, lat, lon.

 

[marsh]

Yes, it is a work-around, since the history routines do not support zonal mean output required for the TEM diagnostics. Where nlevs > nlons I would avoid asking for TEM diagnostics 'MSKtem', 'VTH2d', 'UV2d', 'UW2d', 'U2d', 'V2d', 'TH2d', 'W2d' (usually saved the h4 history files).