Add options for NonEq relaxation timescales as functions of N and T by oalcabes · Pull Request #676 · CliMA/CloudMicrophysics.jl

oalcabes · 2026-02-11T00:05:08Z

Draft PR to make new versions of the NonEq scheme, namely tau(T) and tau(N) functionality. Still need to review docs and think carefully about testing, but I'd love comments and thoughts on the PR! This may break atmos as I've made changes to the bulk_microphysics_tendency function.

codecov · 2026-02-11T00:20:27Z

Codecov Report

❌ Patch coverage is 62.06897% with 11 lines in your changes missing coverage. Please review.
✅ Project coverage is 91.62%. Comparing base (14ef3e6) to head (019d9a7).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #676      +/-   ##
==========================================
- Coverage   92.02%   91.62%   -0.41%     
==========================================
  Files          54       54              
  Lines        2321     2341      +20     
==========================================
+ Hits         2136     2145       +9     
- Misses        185      196      +11

Components	Coverage Δ
src	`92.56% <62.06%> (-0.43%)`	⬇️
ext	`69.47% <ø> (ø)`

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

trontrytel · 2026-02-14T00:23:52Z

 MicrophysicsNonEq.τ_relax
 MicrophysicsNonEq.conv_q_vap_to_q_lcl_icl
+MicrophysicsNonEq.conv_q_vap_to_q_lcl_MM2015
+MicrophysicsNonEq.conv_q_vap_to_q_icl_MM2015


Do we still want to keep MicrophysicsNonEq.conv_q_vap_to_q_icl_MM2015?

Ideally we would keep "MicrophysicsNonEq.conv_q_vap_to_q_lcl_MM2015" and "MicrophysicsNonEq.conv_q_vap_to_q_icl_MM2015" and delete "MicrophysicsNonEq.conv_q_vap_to_q_lcl_icl_MM2015" - I was leaving it now for backwards compatibility but if we're okay with a breaking release I can just delete it

trontrytel · 2026-02-14T00:24:36Z

+
+        # Using same limiter as ClimaAtmos for now
+        # Not sure why, but without intermediate storing of the tendencies for the
+        # if/else branch this code segfaults on julia v1.11 (works fine on v1.10)


No idea. But I have seen this before too :/

oalcabes · 2026-05-08T21:23:07Z

Hi @trontrytel and @sajjadazimi -- I'm coming back to this PR. The PR is doing three main things, namely: (1) refactors the NonEquilibrium parameterization to accept tau as an argument rather than calculating it inside of the function (2) introduce a version of tau_i calculated from Frostenberg et. al. (2023) and (3) add parcel and documentation examples for the Frostenberg option. For now I would like to focus most on the refactoring change -- if you all like it we can keep it in this PR, but if you have major comments perhaps I'll do it in a separate PR.

oalcabes · 2026-05-08T21:26:05Z

I specifically desire comments on src/BulkMicrophysicsTendencies.jl

oalcabes · 2026-05-08T21:31:12Z

+    if noneq_scheme isa NonEq_Constant
+        τₗ = lcl.τ_relax
+        τᵢ = icl.τ_relax
+    elseif noneq_scheme isa NonEq_N # this assumes we're using the Frostenberg parameterization for τᵢ


To decide: do we want to have liquid calculated like this as well, or just keep the ice relaxation timescale changing instead?

oalcabes · 2026-05-08T21:34:11Z

+- `warm_rain`
+    - `sb`: SB2006 parameters
+    - `aps`: air property parameters % OLIVIA CLEAN THIS UP -- also missing ice???
+- `noneq_scheme` Type of NonEquilibrium scheme to use (NOT NECESSARILY NEEDED BUT KEEP IT FOR THE FUTURE)


currently passing this flag but not using it yet -- may not be necessary if we decide to keep tau_l always constant for now and only change tau_i

sajjadazimi

Thank you for refactoring and adding the options for relaxation time scale. It looks good overall. I left a few minor comments.

sajjadazimi · 2026-05-08T22:07:21Z

+\end{equation}
+```
+
+where ``q_c`` is the mass of condensate and ``ρ_c`` is the density of the condensate (either liquid or water). Then, we calculate the relaxation timescale:


liquid or water?

sajjadazimi · 2026-05-08T22:12:13Z

+N_i(T) = ln(-(b \cdot T)^9 \times 10^{-9})
+```
+
+ ``\sigma^2`` is the variance, ``a`` and ``b`` are coefficients. The parameters defined in [Frostenberg2023](@cite) for marine data sets are ``\sigma=1.37``, ``a=1`` m``^3``, ``b=1``/C.


I only see b in the eq

sajjadazimi · 2026-05-08T22:17:44Z

+"""
+    NonEq Scheme
+
+Abstract type for NonEquilibrium microphysics parameterizations. Namely,
+this scheme specifies how to set the relaxation timescales (τₗ, τᵢ) for
+liquid condensation/evaporation and ice deposition/sublimation.
+NonEq_Constant sets τₗ and τᵢ are determined by the τ_relax arguments in
+lcl and icl, whereas NonEq_N calculates τₗ and τᵢ from liquid and ice
+number concentration.
+"""
+abstract type NonEqScheme end
+struct NonEq_Constant <: NonEqScheme end
+struct NonEq_N <: NonEqScheme end
+


Maybe it's better to move this struct definition to NoneqMicrophysics

sajjadazimi · 2026-05-08T22:24:15Z

+        ::Microphysics1Moment,
+        noneq_scheme::NonEqScheme,


noneq_scheme could be an element of Microphysics1Moment. I see Microphysics1Moment as an umbrella that could define what submodels should be including the non equilibrium scheme.
This is not a strong opinion, so ignore if it's not possible.

I agree. Otherwise we will start to have to carry more parameters in Microphysics1Moment than what we actually need for a single simulation

sajjadazimi · 2026-05-08T22:26:20Z

+    if noneq_scheme isa NonEq_Constant
+        τₗ = lcl.τ_relax
+        τᵢ = icl.τ_relax
+    elseif noneq_scheme isa NonEq_N
+        τₗ = CMNonEq.τ_N(q_icl, N_lcl, lcl.ρw, aps.D_vapor)
+        τᵢ = CMNonEq.τ_Frostenberg(icl, aps, ip, q_icl, T)
+    end


For code readability, you could dispatch over noneq_scheme and return the tuple of tau_l and tau_i

sajjadazimi · 2026-05-08T22:27:40Z

 - `dn_rai_dt`: Rain number tendency (1/kg/s)
 """
-@inline function warm_rain_tendencies_2m(warm_rain, tps, T, q_tot, q_lcl, q_rai, q_ice, ρ, n_lcl, n_rai)
+@inline function warm_rain_tendencies_2m(warm_rain, noneq_scheme, tps, T, q_tot, q_lcl, q_rai, q_ice, ρ, n_lcl, n_rai)


same as 1m: noneq_scheme could live under warm_rain model

sajjadazimi · 2026-05-08T22:28:28Z

 """
 @inline function bulk_microphysics_tendencies(
-    ::Microphysics2Moment, mp::CMP.Microphysics2MParams{WR, Nothing}, tps,
+    ::Microphysics2Moment, noneq_scheme::NonEqScheme, mp::CMP.Microphysics2MParams{WR, Nothing}, tps,


same as above

sajjadazimi · 2026-05-08T22:30:04Z

+@inline τ_relax(p::CMP.CloudLiquid) = p.τ_relax
+@inline τ_relax(p::CMP.CloudIce) = p.τ_relax


Here, you could dispatch over non-eq method too

trontrytel · 2026-05-11T23:37:32Z


+### Liquid and ice relaxation timescales as functions of number concentration
+
+We also provide functionality to set the relaxation timescales of liquid and ice, ``\tau_l`` and ``\tau_i``, as functions of cloud droplet number concentrations ``N_l`` and ``N_i`` following [MorrisonGrabowski2008_supersat](@cite) and [MorrisonMilbrandt2015](@cite). First, we approximate average radius of droplets from mass and number:


Suggested change

We also provide functionality to set the relaxation timescales of liquid and ice, ``\tau_l`` and ``\tau_i``, as functions of cloud droplet number concentrations ``N_l`` and ``N_i`` following [MorrisonGrabowski2008_supersat](@cite) and [MorrisonMilbrandt2015](@cite). First, we approximate average radius of droplets from mass and number:

We also provide functionality to set the relaxation timescales of liquid and ice, ``\tau_l`` and ``\tau_i``, as functions of cloud droplet and ice crystal number concentrations ``N_l`` and ``N_i`` following [MorrisonGrabowski2008_supersat](@cite) and [MorrisonMilbrandt2015](@cite). First, we approximate average radius of droplets from mass and number:

oalcabes changed the title ~~Add temperature dependent relaxation timescale for NonEq~~ Add options for NonEq relaxation timescales as functions of N and T Feb 13, 2026

oalcabes force-pushed the oa/tau_i_temp_dependence branch 3 times, most recently from cf55a81 to 9346e0c Compare February 13, 2026 22:50

oalcabes requested a review from trontrytel February 13, 2026 22:52

oalcabes added the NonEq 💧🧊 Non equilibrium microphysics label Feb 13, 2026

trontrytel reviewed Feb 14, 2026

View reviewed changes

Comment thread docs/src/MicrophysicsNonEq.md Outdated

trontrytel reviewed Feb 14, 2026

View reviewed changes

oalcabes force-pushed the oa/tau_i_temp_dependence branch from 0e64b5e to ab706fb Compare April 1, 2026 22:50

oalcabes force-pushed the oa/tau_i_temp_dependence branch 2 times, most recently from 0d2eda2 to 9bffcec Compare April 30, 2026 18:47

oalcabes force-pushed the oa/tau_i_temp_dependence branch 3 times, most recently from caabc3d to 22cafcc Compare May 8, 2026 20:38

oalcabes commented May 8, 2026

View reviewed changes

Comment thread src/BulkMicrophysicsTendencies.jl

oalcabes marked this pull request as ready for review May 8, 2026 21:20

oalcabes requested a review from sajjadazimi May 8, 2026 21:20

oalcabes commented May 8, 2026

View reviewed changes

Add new NonEq parameterizations for tau(N) and tau(T)

019d9a7

oalcabes force-pushed the oa/tau_i_temp_dependence branch from 22cafcc to 019d9a7 Compare May 8, 2026 21:37

sajjadazimi reviewed May 8, 2026

View reviewed changes

trontrytel reviewed May 11, 2026

View reviewed changes

		@inline τ_relax(p::CMP.CloudLiquid) = p.τ_relax
		@inline τ_relax(p::CMP.CloudIce) = p.τ_relax


		### Liquid and ice relaxation timescales as functions of number concentration

		We also provide functionality to set the relaxation timescales of liquid and ice, ``\tau_l`` and ``\tau_i``, as functions of cloud droplet number concentrations ``N_l`` and ``N_i`` following [MorrisonGrabowski2008_supersat](@cite) and [MorrisonMilbrandt2015](@cite). First, we approximate average radius of droplets from mass and number:

Conversation

oalcabes commented Feb 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov Bot commented Feb 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oalcabes Feb 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

oalcabes commented May 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oalcabes commented May 8, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oalcabes May 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sajjadazimi left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

oalcabes commented Feb 11, 2026 •

edited

Loading

codecov Bot commented Feb 11, 2026 •

edited

Loading

oalcabes Feb 14, 2026 •

edited

Loading

oalcabes commented May 8, 2026 •

edited

Loading

oalcabes May 8, 2026 •

edited

Loading