Tr/add get fields atmos #1205

imreddyTeja · 2025-02-24T23:25:56Z

Purpose

Add get_field methods that are needed to couple with the full land model.

Also adds update_field for emissivity.

To-do

~~LHF/SHF~~

Content

Add air_pressure, air_temperature, humidity, co2, diffuse_fraction, SW_d, LW_d, and cos_zenith_angle
to get_field on AtmosModelSimulation in Interfacer.jl.

Add get_field methods for the above values in
ClimaAtmosSimulation

Add emissivity to update_field on AtmosModelSimulation in Interfacer.jl,
and add a method for it in ClimaAtmosSimulation

Remove co2 from update_field on AtmosModelSimulation in Interfacer.jl and
method definition for ClimaAtmosSimulation

I have read and checked the items on the review checklist.

src/Utilities.jl

experiments/ClimaEarth/components/atmosphere/climaatmos.jl

src/Interfacer.jl

szy21 · 2025-02-25T22:26:02Z

experiments/ClimaEarth/components/atmosphere/climaatmos.jl

+    else
+        direct_flux_dn = radiation_model.face_sw_direct_flux_dn[1, :]
+        FT = eltype(total_flux_dn)
+        diffuse_fractions = (total_flux_dn .- direct_flux_dn) ./ (max.(total_flux_dn, eps(FT)))


I wonder if we want to limit it between 0 and 1 just to be exra safe. I don't think it should matter, maybe we can also wait and see if there is a problem.

diffuse_fraction is now clamped to [0,1]

szy21

Thanks!

szy21 · 2025-02-25T22:53:20Z

experiments/ClimaEarth/components/atmosphere/climaatmos.jl

+function Interfacer.update_field!(sim::ClimaAtmosSimulation, ::Val{:emissivity}, field)
+    # sets all 16 bands (rows) to the same values
+    sim.integrator.p.radiation.rrtmgp_model.surface_emissivity .=
+        reshape(CC.Fields.field2array(field), 1, length(parent(field)))
+end


In the long term, maybe we can make the RRTMGPInterface assume every band is the same if only one value is passed in (maybe this is already the behavior not? I'm not sure). This way we don't need to hard code the length(parent) here.

juliasloan25 · 2025-02-25T23:01:41Z

We recently changed CO2 from being read in within ClimaCoupler and passed to ClimaAtmos, to being read in within ClimaAtmos. We should have removed the update_field! method for co2 but we missed it. Could you remove that method for ClimaAtmos, as well as the mentions of :co2 in src/Interfacer.jl and the tests, in this PR please?

When we use the full land model, we'll need to read in mean global co2 from the atmosphere and pass it to the land model, so we need a get_field for this quantity from atmos. Could you add that in this PR too please?

juliasloan25 · 2025-02-25T23:06:29Z

@szy21 @Sbozzolo Will the field atmos_sim.integrator.p.radiation.rrtmgp_model.volume_mixing_ratio_co2 only be filled if we run atmos with radiation and the MaunaLoa option for CO2? That seems to be implied by the changes in the ClimaAtmos PR moving the file read to atmos https://github.com/CliMA/ClimaAtmos.jl/pull/3522/files, e.g. in src/solver/type_getters.jl. If so, we'll need to think about what CO2 value to supply to land when atmos doesn't provide CO2

Sbozzolo · 2025-02-25T23:34:58Z

@szy21 @Sbozzolo Will the field atmos_sim.integrator.p.radiation.rrtmgp_model.volume_mixing_ratio_co2 only be filled if we run atmos with radiation and the MaunaLoa option for CO2? That seems to be implied by the changes in the ClimaAtmos PR moving the file read to atmos https://github.com/CliMA/ClimaAtmos.jl/pull/3522/files, e.g. in src/solver/type_getters.jl. If so, we'll need to think about what CO2 value to supply to land when atmos doesn't provide CO2

It is always filled.

https://github.com/CliMA/ClimaAtmos.jl/blob/cedbf1b6e27ed084d04a66a559451841a44c31b8/src/parameterized_tendencies/radiation/radiation.jl#L93-L96

Here's how it's filled in the fixedco2 case

juliasloan25 · 2025-02-25T23:50:29Z

It is always filled.

https://github.com/CliMA/ClimaAtmos.jl/blob/cedbf1b6e27ed084d04a66a559451841a44c31b8/src/parameterized_tendencies/radiation/radiation.jl#L93-L96

Here's how it's filled in the fixedco2 case

What if co2 is set to nothing and we aren't using radiation? I.e. this case:

    (isnothing(co2) && !with_rrtgmp) &&
        @warn ("$(co2) does nothing if RRTGMP is not used")

Sbozzolo · 2025-02-26T00:07:43Z

If it's nothing and you arent' using radiation, co2 is not set because rrtgmp is also not defined

juliasloan25

This looks good! We should update the docs too - could you add a section to the Interfacer.md docs ### AtmosModelSimulation - required functions to run with ClimaLandSimulation that explains which get_field and update_field! methods are required to run with the full land (i.e. the ones you've added here). It would be good to note which ones won't be present if the atmosphere doesn't have radiation, and that because of this we need to enable radiation in the atmosphere when running with the full land model. Thank you!

Add diffuse_fraction, SW_d, LW_d, and cos of the zenith_angle to get_field on AtmosModelSimulation in Interfacer.jl. Add get_field methods for the above values in ClimaAtmosSimulation Remove safe_div and use anonymous func Only calculate on first level for diffuse_frac Add update_field for emissivity in AtmosModelSim Add emissivity to update_field in the Interfacer, and add a method definition for it in climaatmos.jl All 16 bands are set to the same values. Clamp diffuse_fraction to [0,1] Remove :co2 from atmos update_field :co2 is removed from the default AtmosModelSimulation update_field. The method definition is also removed for ClimaAtmosSimulation. Add atmos get_field for :co2 Add a default get_field for :co2 with AtmosModelSimulation, and a method definition for ClimaAtmosSimulation. Note that this returns a scalar. Add new fields to interfacer_tests Revert to returning zero when dividing by zero In get_field(::ClimaAtmosSimulation, ::Val{:diffuse_fraction}

imreddyTeja · 2025-02-28T17:09:48Z

docs/src/interfacer.md

+| Coupler name      | Description | Units |
+|-------------------|-------------|-------|
+| `air_pressure`     | boundary air pressure | Pa |
+| `air_temperature`  | boundary air temperature | K |
+| `cos_zenith_angle` | cosine of the zenith angle | |
+| `co2`              | global mean co2 | ppm |
+| `diffuse_fraction` | fraction of downwards shortwave flux that is direct | |
+| `humidity`         | boundary humidity | kg kg^-1 |
+| `LW_d`             | downwards longwave flux | W m^-2 |
+| `SW_d`             | downwards shortwave flux | W m^-2 |


Are these descriptions accurate? Do the pressure and temperature values need air_ prepended to them?

These look good! Prepending air_ is a good idea to distinguish from surface_. Instead of "boundary air pressure/temp/humidity", I would write "air pressure/temp/humidity at the bottom of the atmosphere", to avoid any confusion between the boundary layer of the atmosphere vs boundary surface

Add them to the the generic get_field for AtmosModelSimulations and specific methods for ClimaAtmosSimulations

It is refered to as `cos_zenith` by atmos and `cos\thetas` by land. This unifies it with one of the models. [skip ci]

szy21 · 2025-02-28T22:00:12Z

experiments/ClimaEarth/components/atmosphere/climaatmos.jl

+    CC.Fields.array2field(sim.integrator.p.radiation.rrtmgp_model.face_lw_flux_dn, axes(sim.integrator.u.f)),
+    CC.Utilities.half,
+)
+Interfacer.get_field(sim::ClimaAtmosSimulation, ::Val{:humidity}) =


I would call it specific_humidity to be more accurate, but up to you.

szy21 · 2025-02-28T22:01:50Z

docs/src/interfacer.md

+
+| Coupler name      | Description | Units |
+|-------------------|-------------|-------|
+| `air_pressure`     | air pressure at the bottom of the atmosphere | Pa |


juliasloan25

I just had one comment about throwing errors if the get_fields aren't extended, but this looks good other than that!

juliasloan25 · 2025-03-01T00:38:02Z

src/Interfacer.jl

@@ -155,12 +155,20 @@ an atmosphere component model.
 get_field(


I don't think we should error if these get_fields aren't implemented - e.g. if we want to couple a simpler/different atmosphere with the bucket, it doesn't need to extend all these methods required only for the full land model. I was going to suggest that we could give warnings if these aren't extended, but that will print a lot of warnings, so maybe the best thing to do is just remove these fields from here and expect users to read what's required to extend from the documentation

I thought this would only error if the get_field is actually called for that value. So if we used some simple atmosphere with the bucket, it would not need to extend get_field for all the methods listed here, because get_field(::SomeSimpleAtmos, ::Val{:cos_zenith}) would never be called. Is that correct?

I think it might still make sense to split the get_field(::AtmosModelSimulation) default error into what is required for coupling with all land models, and just what is required for coupling with ClimaLand.

remove from interfacer tests

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch 3 times, most recently from f2d41d5 to 0aed291 Compare February 25, 2025 00:38

juliasloan25 reviewed Feb 25, 2025

View reviewed changes

src/Utilities.jl Outdated Show resolved Hide resolved

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from 0aed291 to 328d9b0 Compare February 25, 2025 18:27

imreddyTeja marked this pull request as ready for review February 25, 2025 18:48

imreddyTeja requested a review from szy21 February 25, 2025 18:50

szy21 reviewed Feb 25, 2025

View reviewed changes

experiments/ClimaEarth/components/atmosphere/climaatmos.jl Outdated Show resolved Hide resolved

experiments/ClimaEarth/components/atmosphere/climaatmos.jl Outdated Show resolved Hide resolved

src/Interfacer.jl Outdated Show resolved Hide resolved

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch 5 times, most recently from 854a22d to f7c38c9 Compare February 25, 2025 22:16

szy21 reviewed Feb 25, 2025

View reviewed changes

szy21 approved these changes Feb 25, 2025

View reviewed changes

szy21 reviewed Feb 25, 2025

View reviewed changes

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from beb25cc to 67d8d6a Compare February 26, 2025 23:45

imreddyTeja requested a review from juliasloan25 February 27, 2025 00:08

juliasloan25 reviewed Feb 27, 2025

View reviewed changes

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from f32cf97 to 11e4d33 Compare February 28, 2025 02:28

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from 11e4d33 to dd322ff Compare February 28, 2025 17:04

imreddyTeja commented Feb 28, 2025

View reviewed changes

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from dd322ff to f20ed1c Compare February 28, 2025 17:43

Add air pressure, air temp, and humidity get_field

2a48532

Add them to the the generic get_field for AtmosModelSimulations and specific methods for ClimaAtmosSimulations

Update Interfacer.md with atmos get_field changes

97874c1

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from f20ed1c to 9caf01b Compare February 28, 2025 21:47

imreddyTeja requested review from juliasloan25 and szy21 February 28, 2025 21:48

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from da3ca99 to 9caf01b Compare February 28, 2025 21:54

Change cos_zenith_angle to cos_zenith

0077d3c

It is refered to as `cos_zenith` by atmos and `cos\thetas` by land. This unifies it with one of the models. [skip ci]

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from 9caf01b to 0077d3c Compare February 28, 2025 21:55

szy21 approved these changes Feb 28, 2025

View reviewed changes

Change humidity to specific_humidity for atmos

97044ba

juliasloan25 approved these changes Mar 1, 2025

View reviewed changes

Remove not always required fields default get_field

c8329ff

remove from interfacer tests

imreddyTeja force-pushed the tr/add-get_fields-Atmos branch from aee4e44 to c8329ff Compare March 5, 2025 17:39

imreddyTeja enabled auto-merge (squash) March 5, 2025 17:40

imreddyTeja merged commit a62d04e into main Mar 5, 2025
11 checks passed

imreddyTeja deleted the tr/add-get_fields-Atmos branch March 5, 2025 18:59

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Tr/add get fields atmos #1205

Tr/add get fields atmos #1205

imreddyTeja commented Feb 24, 2025 •

edited

Loading

szy21 Feb 25, 2025

imreddyTeja Feb 27, 2025

szy21 left a comment

szy21 Feb 25, 2025

juliasloan25 commented Feb 25, 2025

juliasloan25 commented Feb 25, 2025

Sbozzolo commented Feb 25, 2025

juliasloan25 commented Feb 25, 2025

Sbozzolo commented Feb 26, 2025

juliasloan25 left a comment

imreddyTeja Feb 28, 2025

juliasloan25 Feb 28, 2025

szy21 Feb 28, 2025

szy21 Feb 28, 2025

juliasloan25 left a comment

juliasloan25 Mar 1, 2025

imreddyTeja Mar 3, 2025

	\| `air_pressure` \| air pressure at the bottom of the atmosphere \| Pa \|
	\| `air_pressure` \| air pressure at the bottom cell center of the atmosphere \| Pa \|

		@@ -155,12 +155,20 @@ an atmosphere component model.
		get_field(

Tr/add get fields atmos #1205

Tr/add get fields atmos #1205

Conversation

imreddyTeja commented Feb 24, 2025 • edited Loading

Purpose

To-do

Content

Choose a reason for hiding this comment

Choose a reason for hiding this comment

szy21 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

juliasloan25 commented Feb 25, 2025

juliasloan25 commented Feb 25, 2025

Sbozzolo commented Feb 25, 2025

juliasloan25 commented Feb 25, 2025

Sbozzolo commented Feb 26, 2025

juliasloan25 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

juliasloan25 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

imreddyTeja commented Feb 24, 2025 •

edited

Loading