Disable "excessive fall velocity" error trap

[yantosca]

Your name

Bob Yantosca

Your affiliation

Harvard + GCST

Provide a clear and concise overview of the new feature requested.

In the Calc_Fallvel routine (in module GeosCore/ucx_mod.F90, there is an error trap that halts a GEOS-Chem simulation when an excessive fall velocity is detected:

    !=================================================================
    ! CALC_FALLVEL begins here!
    !=================================================================

    DATA ALPHA/1.49e-5_fp,5.02e-6_fp,2.64e-5_fp/

    ! Sanity check
    IF ((RADIUS.le.0.e+0_fp).or.(DENSITY.le.0.e+0_fp)) THEN
       VEL=0.e+0_fp
    ELSE
       ! PCENTER (kPa -> Pa) = *1.d3
       ! RADIUS  (cm  -> m ) = *1.d-2
       ! Therefore multiply PR by 10
       PR = PCENTER * RADIUS * 10e+0_fp
       VAL_X = -1.0e+0_fp*PR/(ALPHA(3)*TCENTER)
       VAL_X = ALPHA(2)*TCENTER*EXP(VAL_X)/PR
       VAL_X = 1.0e+0_fp + VAL_X + (ALPHA(1)*TCENTER/PR)
       Vy = g0*DENSITY*RADIUS*RADIUS*(1.e-4_fp)/(4.5*ETA*TCENTER)
       VEL = 0.893e+0_fp * Vy * VAL_X
    ENDIF

    ! Velocities should be of the order of 0.1 m/s
    IF (VEL.gt.10.e+0_fp) THEN
       CALL ERROR_STOP(' Excessive fall velocity? ', ' CALC_FALLVEL, UCX_mod')
    ENDIF

However, this has tripped up many users in the past.

@ltmurray writes:

That is a conservative trap in the stratospheric chemistry mechanism that is often triggered when applying a restart file that is out of sync with the stratospheric dynamical conditions.

We propose removing the error trap, which can be confusing to new GEOS-Chem users and may no longer be relevant.

Tagging @ltmurray @lizziel @sdeastham @msulprizio

Reference(s)

No response

Will you be implementing this feature yourself?

Yes

Additional information

See:

• GEOS-CHEM ERROR: Excessive fall velocity? STOP at CALC_FALLVEL, UCX_mod #2653

[sdeastham]

I'm very reluctant to remove this trap permanently. This only happens when aerosols form in the stratosphere with a fall speed of more than 10 m/s, or 36 km/h - that corresponds to some truly massive aerosol particles! I'd also be worried about the numerical stability at this point (the vertical CFL will be >>> 1).

[yantosca]

Thanks for your feedback @sdeastham. Maybe we need a better error message so that users will know why the run failed. I think that's more of the issue now.

[yantosca]

@sdeastham: Could you suggest wording for an improved error message?

[ltmurray]

I'd recommend switching it to a warning that prints to the output, rather than forcing the model to stop. This generally only happens in my experience when the restart file is out of equilibrium with the current stratospheric dynamics, e.g., starting a future GCAP run from a present day restart file, and then it never appears again in the simulation. But it occasionally happens with MERRA-2 as well when using the default restart file applied to a different year.

[sdeastham]

A warning might do. But I think in that case we should probably clip the fall velocity to be less than or equal to dz/dt; I'd need to review the settling code to verify the importance of this, but I'm concerned you might get some mass continuity issues if you just allow the code to proceed. A more detailed message would be something like "Excessive stratospheric aerosol settling speeds detected (>10 m/s). This suggests extremely large, likely nonphysical concentrations of aerosol".