KEYWORDS: subgrid clouds, mass flux, local diffusion
SOURCE: Joseph Olson and Laura Fowler
DESCRIPTION OF CHANGES:
Problems:
1) bug fix for momentum transport - minor impact
2) underdiffusive in very stable conditions, especially high latitudes. This reveals itself as a cold bias at high latitudes in RRFS and the TKE is too often hitting the lower limit.
3) cosmetic issue with cloud fractions in unsaturated grid cells with nonzero mixing ratios
4) uninitialized scalars in unused subroutine planned for deletion
Solutions (1-1 mapping with problems listed above):
1) added the missing rho to the mass flux components in the u and v tridiagonal vectors.
2) reverted the blending of the mixing lengths to same as HRRRv4 and tweaked some parameters to slightly increase the diffusion, mostly isolated to very stable conditions.
3) removed arbitrary limit of RH (was 101%) and combined the ice/snow cloud fraction check together.
4) replaced uninitialized variables with variables that are initialized (changed names). Still plan to remove this subroutine after the downdrafts are fully revised (very close).
LIST OF MODIFIED FILES: list of changed files (use git diff --name-status master to get formatted list)
phys/module_bl_mynn.F
TESTS CONDUCTED:
Do mods fix problem? (1) Yes, (2) alleviated, (3) Yes, (4) Yes. Single case studies were performed. All physics changes has a small impact.
Are the Jenkins tests all passing?
RELEASE NOTE: This commit is just a minor update, probably not worthy of mention in the release notes.
Syncing MYNN-EDMF + bug fixes
TYPE: bug fix, enhancement
KEYWORDS: subgrid clouds, mass flux, local diffusion
SOURCE: Joseph Olson and Laura Fowler
DESCRIPTION OF CHANGES: Problems: 1) bug fix for momentum transport - minor impact 2) underdiffusive in very stable conditions, especially high latitudes. This reveals itself as a cold bias at high latitudes in RRFS and the TKE is too often hitting the lower limit. 3) cosmetic issue with cloud fractions in unsaturated grid cells with nonzero mixing ratios 4) uninitialized scalars in unused subroutine planned for deletion
Solutions (1-1 mapping with problems listed above): 1) added the missing rho to the mass flux components in the u and v tridiagonal vectors. 2) reverted the blending of the mixing lengths to same as HRRRv4 and tweaked some parameters to slightly increase the diffusion, mostly isolated to very stable conditions. 3) removed arbitrary limit of RH (was 101%) and combined the ice/snow cloud fraction check together. 4) replaced uninitialized variables with variables that are initialized (changed names). Still plan to remove this subroutine after the downdrafts are fully revised (very close).
LIST OF MODIFIED FILES: list of changed files (use
git diff --name-status master
to get formatted list) phys/module_bl_mynn.FTESTS CONDUCTED:
RELEASE NOTE: This commit is just a minor update, probably not worthy of mention in the release notes.