Repository navigation
Conversation
…odes sgs_ice_uniform_fraction_ramped(ξ_ice, T, T_lo, T_hi): ξ_ice at node temperatures above T_hi, 1 (uniform) below T_lo, linear in between, so the excess reconstruction can be used for mixed-phase ice while cirrus ice stays uniform. Provisional parameters sgs_ice_uniform_ramp_T_low/high (defaults 0, 0 = ramp off, bitwise neutral); Microphysics1MEvaluator carries the two bounds and keeps its 18-argument constructor. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
sgs_ice_incloud_fraction (provisional, default 0 = bitwise neutral): a fraction β of the subdomain-mean cloud ice drawn on at each quadrature node is confined to the ice-supersaturated nodes (colder than freezing, non-precipitating water above ice saturation; SGSIceSupersaturatedFlag), whose quadrature weight cf_ice microphysics_tendencies_1m computes with one extra saturation pass. The node factor φ = (1−β) + β·flag/cf_ice (sgs_ice_incloud_factor) has quadrature mean 1, so the cell ice is conserved; cf_ice = 0 falls back to uniform. Physical reading: cloud ice sits in the cloudy part of the cell where the Bergeron transfer from the coexisting liquid acts on it, instead of being spread into the clear part where it sublimates. Microphysics1MEvaluator carries (β, cf_ice) and keeps the 18- and 20-argument constructors. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…the nodes sgs_precip_incloud_fraction (provisional, default 0 = bitwise neutral): a fraction β_p of the cell-mean rain and snow seen at each quadrature node is confined to the moist half of the PDF (nodes whose centred saturation excess S′ ≥ 0; SGSMoistHalfFlag, weight cf_precip computed by microphysics_tendencies_1m with one extra pass), with the node total water shifted by the same amount so the node vapour and the condensate reconstruction are unchanged. ⟨φ_p⟩ = 1 conserves the cell precipitation. Physical reading: precipitation falls through the cloudy part of the cell and the air beneath it, so below-cloud rain evaporation and snow sublimation see that air's humidity rather than the driest tail of the PDF (the precipitation-fraction treatment of IFS/CAM). Microphysics1MEvaluator carries (β_precip, cf_precip); the 18-, 20- and 22-argument constructors are kept. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
snow_deposition_sublimation: "DepositionOnly" selects CloudMicrophysics' DepositionOnly (snow deposition below freezing, no sublimation; attribution option). Parameter test covers all three variants. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…cement sgs_snow_incloud_fraction (provisional, default -1 = same as sgs_precip_incloud_fraction, bitwise neutral for existing configs): the snow seen at a quadrature node can be placed with its own fraction, so rain alone can be confined to the moist half of the PDF while snow stays uniform (or the reverse). Microphysics1MEvaluator carries β_snow; the 24-argument constructor sets it equal to β_precip. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The sgs_ice_incloud_fraction option (confining the cell-mean cloud ice to the ice-supersaturated quadrature nodes) is removed: offline screens and the campaign showed it adds nothing once the Hallett-Mossop source and the ξ_ice temperature ramp are present, and its concentrated node ice is autoconverted by the ice-to-snow threshold. The generic node factor is kept for the precipitation-fraction placement as sgs_placement_factor; the Microphysics1MEvaluator constructors are now 18, 20, 22 and 23 arguments. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…cement Diagnose the precipitation fraction a_p of each cell by maximum-random overlap of the cloud cover swept from the top (`set_precip_fraction!`, `ᶜprecip_frac`), reset where the environment carries no precipitation, and confine the placed rain and snow to the moistest a_p of the sampled PDF: the shaft threshold on the centred saturation excess is the closed-form Gaussian quantile `S_star = σ_S Φ⁻¹(1 − a_p)` with a smooth node weight, normalized by the discrete weight of the selected nodes so the cell-mean precipitation is conserved exactly. The discrete cloudy mass CF_d of the quadrature is stored with the SGS moments. Enabled by a non-negative `sgs_precip_overlap_decay` (default -1 keeps the moist-half placement unchanged). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Under the overlap precipitation fraction, `sgs_precip_shaft_random` places the shaft on every cloudy node plus a random share p = (a_p − CF_d)/(1 − CF_d) of the clear nodes at the in-shaft concentration q/A, A = CF_d + (1 − CF_d) p, so rain evaporation and snow sublimation are evaluated at the clear-sky humidity of the cell rather than at its moist tail. Each node's tendency is the mixture of the in-shaft and precipitation-free states with the same node vapour; the cell-mean precipitation is conserved exactly with the cloudy weight that accumulates CF_d. Default off. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…ction `sgs_precip_fraction_floor` (default 0.1, the previous hard-coded value) sets the smallest precipitation fraction the node placement uses. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The tendency broadcast had grown past ~30 arguments, which drops Julia's broadcast off its specialized getindex path and allocates inside the GPU kernel (InvalidIRError: gpu_gc_pool_alloc). `sgs_microphysics_options` packs α, the uniform fractions, the ice ramp and the precipitation-placement keys into a NamedTuple consumed by a forwarding method of `microphysics_tendencies_1m`; the positional form is unchanged. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
A NamedTuple is not a valid broadcast argument; `SGSMicrophysicsOptions` carries the same fields and broadcasts as a scalar. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…ysics splintering branch; building on CM main) Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Contributor
|
📖 Docs preview for this PR: https://clima.github.io/ClimaAtmos.jl/previews/PR4884/ |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Purpose
WIP. do not review yet
To-do
Content