*Updated lateral boundary conditions for nested grids so that features translate from parent grid to nested grid more accurately and seamlessly. There are some hard coded coefficients in bc/rbnd.f90 and/or nesting/nest_intrp.f90 that could be customized as needed, but default values are reasonable for most situations. (Work done by Leah Grant)
*In microphysics, put a cloud droplet nucleation limiter at 0.01microns (10nm) so that if the aerosol distribution median radius drops below 10nm, nucleation is stopped. Our nucleation lookup tables do not go below 10nm median radius. In the past, nucleation could continue as the median radius of the aerosol distribution was forced to the minimum size. In some situations it seemed that overnucleation was occurring, and a limiter was perhaps needed.
*Nucleate cloud droplets at the minimum droplet size instead of diagnosing a size based on nucleation number and available vapor. This will reduce the amount of mass in the microphysics budgets for nucleation, but this is compsentated for via the vapor condensation routines.
In microphysics, updates were done to permit better prediction of heterogeneous ice nucleation when values are realtively small. Ryan Patnaude, (Kreidenweis & DeMott group) used INP data and ice crystal concentration data from the SPICULE field project from congestus clouds for initializing and validating RAMS ice nucleation scheme. The following code changes greatly improved the prediction:
*In mic_adj.f90, I modified a condition statement from cx(k,lcat)<=0.0 to cx(k,lcat)<cxmin to check cx and rx and set things to zero if needed. Former method occasionally created NaNs due to attempted division by zero.
*Add call to “psxfer” in the mic_driv.f90 after the call to “colxfers”. This is necessary so that cloud ice and snow and be moved to the correct category after collision-coalescence. Otherwise, the number adjustment routine “enemb” can artificially add pristine ice number concentration in order to keep the mean pristine ice diameter in bounds from 15-125 microns.
*In mic_nuc.f90 and mic_nucpre.f90 added an IIFN=2 constraint so that the DeMott formulas are only applied between temperatures of -35C to 0.0C. In the past we permitted this to occur at colder temperatures, which perhaps should not be done since the formula is not meant to be extrapolated below -35C.
*In mic_nuc.f90 heterogeneous ice nucleation, I modified the nucleation limiter code from “diagni-cx(k,3)” to “diagni-cx(k,3)-cx(k,4)” to also consider the number concentration of snow in addition to pristine ice since snow is just vapor grown pristine ice. This seems to help limit the heterogeneous ice nucleation. This provides more reasonable concentration compared to SPICULE observations mentioned above in this section header.
*In micphys.f90 change rxmin to 1.e-16 and cxmin to 1.e-5 to permit lower pristine ice concentrations that occur from INP at warmer temperatures.
*In mic_nuc.f90 heterogeneous ice nucleation section, decided to permit nucleation at temperatures less than -35C, but only allow the DeMott formulas to provide an air temperature down to -35C. So, if actual air temperature is -45C, the DeMott formula will use a max cold temperature of -35C. This should permit nucleation without over extrapolating the DeMott fomulas.
*Throughout the RAMS code, I changed the CCN and GCCN naming conventions since GCCN do now actually have to be giant. They can be whatever size and solubility we choose. So, providing more generalized names avoids the issue of thinking we can only specify giant particles in the GCCN variable. The 2 dust modes, for example, are named MD1NP and MD2NP for the mode 1 and mode 2 number concentrations. They could be any size we choose. So, I’ve changed CCN and GCCN number to CN1NP and CN2NP. I’ve also changed their mass to CN1MP and CN2MP. This way we could have 2 aerosol modes that are Aitken mode and accumulation mode or say, accumulation mode and ultra-giant mode. There’s flexibility here as to what the modes of aerosols can be.
*Added optional IFN(ice forming nuclei) / INP(ice nucleating particles) initialization option in the subroutine “init_ifn” in file mic_init.f90 that allows the user to initialize simulations with the:
- SPICULE field campaign fitted vertical profile from aircraft obs assuming the user sets IIFN=2, IFN_FORMULA=2, and CIN_MAX = -1.0 (or)
- MC3E field campaign fitted vertical profiles from aircraft obs assuming the user sets IIFN=2, IFN_FORMULA=1, and CIN_MAX = -2.0.
These are reasonable continental profiles and likely a better default choice than the former default exponentially decreasing profile with height. When using the exponential profile (for CIN_MAX > 0.0, it’s probably most realistic to get CIN_MAX >= 0.5 /mg. Smaller values are probably too small even for a default scenario.
*Added RAMSIN namelist flags ISPONG_PTS and SPONGE_TAU related to the nested grid sponge layer which works to better transition scalar from a parent grid to nested grid.
What's Changed
Full Changelog: v6.3.03...v6.3.04
