In a recent large PR #960 the grid mean thermodynamic variables were converted from (ρθ_lid_ice, ρq_tot) to (ρe_tot, ρq_tot). This PR resulted in undesired behavioral changes, most notably in GABLS. We expect some of these changes to be related to the fact the TC.jl not has ρθ_lid_ice as prognostic variable in the subdomains and ρe_tot while the two are not fully consistent under the anelastic assumption.
This Issue and connected PRs will change the prognostic variables in the subdomains to ρe_tot along with a reformulation of the environmental buoyancy flux in terms of moist static energy.
Similar to #960 , this is going to be a large code change. I will try to make sure to merge pill off PRs
Producers
@yairchn
@charleskawczynski
Components
Inputs
See docs for required inputs
Results and deliverables
Our key deliverable is an EDMF in terms of total energy. This is crucial to allow it connect with the global model of similar variables and ensure conservation properties.
QA: Reproduce the behavior in TRMM_LBA, maybe even some of the behavior in GABLS.
Task breakdown
[x] peel off PR: #1073 diagnose env mean using total enthalpy
[x] peel off PR: #1102 compute buoyancy fluxes using moist static energy gradients
[-] for now I am working on #1204 still a draft and needs debugging, I would appreciate reviews and comments
yairchn
commented
2 years ago
Purpose
In a recent large PR #960 the grid mean thermodynamic variables were converted from
(ρθ_lid_ice, ρq_tot)to(ρe_tot, ρq_tot). This PR resulted in undesired behavioral changes, most notably in GABLS. We expect some of these changes to be related to the fact the TC.jl not hasρθ_lid_iceas prognostic variable in the subdomains andρe_totwhile the two are not fully consistent under the anelastic assumption.This Issue and connected PRs will change the prognostic variables in the subdomains to
ρe_totalong with a reformulation of the environmental buoyancy flux in terms of moist static energy.Design of model equations: https://www.overleaf.com/7149629736ygjfhcfqxqsp
Cost/benefits/risks
Similar to #960 , this is going to be a large code change. I will try to make sure to merge pill off PRs
Producers
@yairchn @charleskawczynski
Components
Inputs
See docs for required inputs
Results and deliverables
Our key deliverable is an EDMF in terms of total energy. This is crucial to allow it connect with the global model of similar variables and ensure conservation properties.
QA: Reproduce the behavior in TRMM_LBA, maybe even some of the behavior in GABLS.
Task breakdown
...
Reviewers
@charleskawczynski @trontrytel @ilopezgp