Physics Switches

A "switch" here means a PhysicsConfig field that turns a physical process on/off or selects between parameterizations. They're grouped below by the same families used in src/config.jl, since a single physical process (e.g. clouds) is often controlled by two different switches with different mechanisms - one for the mean-climate state, one for the response to CO₂ forcing - so grouping by code family avoids duplicating that explanation. Build a config with create_experiment_config and override individual fields afterwards, e.g. cfg.log_rain = 1.

Two naming suffixes recur throughout: _dmc ("deconstruct mean climate") and _drsp ("deconstruct response"). Setting a _dmc switch to false removes that process from the model entirely and the model is normally re-run to a new equilibrium - this quantifies how much the process shapes the mean climate. Setting a _drsp switch to false keeps the control climate fixed (flux corrections are re-estimated so the baseline is unaffected) and only disables that process's contribution to the change under CO₂ forcing - this isolates the process's role in climate sensitivity rather than the mean state. This is the same deconstruction approach used in the MSCM experiments (see References below, MSCM §2.1–2.2).

Overview

Switch	Type	Default	Family	Effect
`log_clouds_dmc`	Bool	`true`	Mean climate	Zeroes cloud climatology at init when `false`
`log_vapor_dmc`^[1]	Bool	`true`	Mean climate	No effect - see Known Limitations
`log_crcl_dmc`	Bool	`true`	Mean climate	Disables circulation (advection+diffusion+convergence) when `false`
`log_hydro_dmc`	Bool	`true`	Mean climate	Disables hydrology (evap/rain/latent heat) when `false`
`log_atmos_dmc`	Bool	`true`	Mean climate	Master atmosphere switch - decouples atmosphere from surface when `false`
`log_co2_dmc`	Bool	`true`	Mean climate	Forces control-run CO₂ to 0 ppm when `false`
`log_ocean_dmc`	Bool	`true`	Mean climate	Master ocean-coupling switch - no mixed-layer/deep-ocean exchange when `false`
`log_qflux_dmc`	Bool	`true`	Mean climate	Governs whether flux corrections are computed, loaded, or zeroed
`log_clouds_drsp`	Bool	`true`	CO₂ response	Freezes cloud cover at a constant 0.7 when `false`
`log_crcl_drsp`	Bool	`true`	CO₂ response	Disables circulation's contribution to the CO₂ response when `false`
`log_hydro_drsp`	Bool	`true`	CO₂ response	Disables hydrology's contribution to the CO₂ response when `false`
`log_topo_drsp`	Bool	`true`	CO₂ response	Caps topography at 1.0 m ("constant topography") when `false`
`log_humid_drsp`	Bool	`true`	CO₂ response	Freezes humidity at a constant 0.0052 kg/kg when `false`
`log_ocean_drsp`	Bool	`true`	CO₂ response	Constant mixed-layer depth, no deep-ocean response, when `false`
`log_ice`	Bool	`true`	Circulation	Ice-albedo feedback - constant albedo/heat capacity when `false`
`log_hdif`	Bool	`true`	Circulation	Horizontal (heat) diffusion
`log_hadv`	Bool	`true`	Circulation	Horizontal (heat) advection
`log_vdif`	Bool	`true`	Circulation	Vertical/meridional diffusion (water vapor)
`log_vadv`	Bool	`true`	Circulation	Advection (water vapor)
`log_conv`^[2]	Bool	`true`	Circulation	Moisture convergence sub-step
`log_rain`	Int	`0`	Hydrology	Selects the precipitation parameterization
`log_eva`	Int	`-1`	Hydrology	Selects the evaporation/exchange-coefficient parameterization
`log_clim`	Int	`0`	Hydrology	NCEP-tuned hydrology coefficient override
`log_tsurf_ext`	Bool	`false`	External forcing	Loads surface-temperature anomaly forcing
`log_hwind_ext`	Bool	`false`	External forcing	Loads horizontal-wind anomaly forcing
`log_omega_ext`	Bool	`false`	External forcing	Loads vertical-velocity anomaly forcing

Mean-Climate Deconstruction Switches (`_dmc`)

These switches, forwarded as keywords by create_experiment_config(:decon_mean_climate; ...), remove a process from the model entirely so it can be re-equilibrated without it - quantifying that process's contribution to the mean climate (DF11 §3.8; MSCM §2.1).

Switch	Default	Effect when `false`
`log_clouds_dmc`	`true`	Cloud climatology zeroed at init - no cloud shortwave reflection or longwave-emissivity contribution (DF11 §3.1–3.2). Cloud cover is a prescribed seasonal field in GREB, not a simulated/responding one, so this removes the process outright rather than disabling a dynamic feedback.
`log_vapor_dmc`	`true`	No effect (dead switch - see Known Limitations).
`log_crcl_dmc`	`true`	`circulation!` returns zero tendencies - no advection, diffusion, or convergence. GREB's circulation is a fixed climatological wind field, not a responding dynamical core (DF11 §3.5), so this isolates how much of the spatial temperature/humidity pattern depends on that fixed transport.
`log_hydro_dmc`	`true`	Humidity climatology zeroed at init; `hydro!` (latent heat, evaporation, rain) disabled entirely - removes the hydrological cycle described in Hydro19 from the mean climate.
`log_atmos_dmc`	`true`	Master switch: disables sensible heat flux, freezes the downward-longwave feedback, and also gates circulation and hydrology - effectively decouples the atmosphere from the surface.
`log_co2_dmc`	`true`	Control-run CO₂ forced to 0 ppm - removes the greenhouse-gas baseline from the mean climate (DF11 §3.2, Eq. 4).
`log_ocean_dmc`	`true`	Master ocean-coupling switch: surface heat capacity uses the land value everywhere (no mixed layer), sea-ice heat-capacity blending is skipped, and the deep ocean contributes zero heat exchange (DF11 §3.7–3.8) - removes ocean heat storage/transport from the mean climate.
`log_qflux_dmc`	`true`	Governs whether the empirical flux-correction terms (the residual added to force the model's climatology to match observations, DF11 §3.8) are computed via spin-up, loaded from a precomputed file, or zeroed, depending on its combination with `log_topo_drsp`.

CO₂-Response Deconstruction Switches (`_drsp`)

These switches, forwarded as keywords by create_experiment_config(:decon_2xco2; ...), keep the control/mean climate fixed and instead disable a process's contribution to the change under a CO₂ doubling - isolating each process's role in climate sensitivity, since feedback strength is mean-state dependent and a fair "no-feedback" test must not also change the climate the feedback acts on (MSCM §2.1–2.2). :decon_2xco2 always applies the CO₂ doubling itself as the forcing (co2_concentration = 680.0); there is no log_co2_drsp because the response experiments' entire purpose is to measure feedbacks to that forcing, not to disable it.

Switch	Default	Effect when `false`
`log_clouds_drsp`	`true`	Cloud cover frozen at a constant 0.7 rather than the seasonal climatology - removes the cloud contribution to the CO₂ response while keeping some cloud cover present (MSCM p. 2161).
`log_crcl_drsp`	`true`	Removes circulation's contribution to the forced response (combined with `log_crcl_dmc` to fully gate `circulation!`).
`log_hydro_drsp`	`true`	Removes hydrology's contribution to the forced response (combined with `log_hydro_dmc` to fully gate `hydro!`).
`log_topo_drsp`	`true`	Topography capped at 1.0 m ("constant topography") - removes topographic modulation of the response; also interacts with `log_qflux_dmc` in the flux-correction branching.
`log_humid_drsp`	`true`	Humidity frozen at a constant 0.0052 kg/kg - removes the water-vapor feedback to CO₂ (DF11 §3.2, Eq. 5) specifically, while `log_hydro_drsp` covers the rest of the hydrological cycle's response.
`log_ocean_drsp`	`true`	Mixed-layer depth held constant (no seasonal variation) and the deep ocean contributes zero response tendency - removes the ocean heat-uptake response to CO₂ forcing (DF11 §3.7).

Circulation Components

GREB's atmospheric transport uses a fixed climatological wind field - it does not itself respond to forcing (DF11 §3.5). Diffusion approximates the smearing effect of weather disturbances with a roughly one-week lifetime; advection moves heat/moisture with the mean wind; both are scaled by topography. The vertical terms additionally let the model represent convergence-driven precipitation (e.g. the ITCZ) from the prescribed vertical-velocity field under continuity/hydrostatic assumptions (Hydro19 §3.1, §3.3, Eqs. 11, 17–18).

Switch	Default	Effect when `false`
`log_ice`	`true`	Disables the ice-albedo feedback: albedo becomes a constant value regardless of surface temperature, and heat capacity ignores ice fraction (DF11 §3.6, Fig. 3). GREB has no explicit sea-ice model - albedo and heat capacity near freezing are simple functions of surface temperature, which is what this switch removes.
`log_hdif`	`true`	Disables horizontal (heat) diffusion.
`log_hadv`	`true`	Disables horizontal (heat) advection.
`log_vdif`	`true`	Disables vertical/meridional diffusion of water vapor.
`log_vadv`	`true`	Disables advection of water vapor.
`log_conv`^[2]	`true`	Disables the moisture-convergence sub-step (water vapor only). Not exposed via `create_experiment_config` - set `cfg.log_conv` directly.

Hydrology Parameterization

Evaporation in GREB is a bulk formula on the saturation deficit, wind speed, and a prescribed (climatological, not dynamically simulated) soil-wetness fraction. Precipitation scales with relative humidity and vertical velocity (both mean and variability), which is what lets the model reproduce the ITCZ and midlatitude storm tracks rather than only a diffuse pattern (Hydro19, Abstract & §3.1–3.3, Eqs. 11–18).

log_rain - selects the precipitation regression coefficients:

Value	Meaning
`-1`	Original GREB
`0` (default)	Best GREB fit (ERA-Interim)
`1`	+ relative humidity
`2`	+ omega (vertical velocity) convergence
`3`	+ relative humidity & omega

When log_rain == 1, a rain-rate limiter is additionally applied.

log_eva - selects the evaporation/exchange-coefficient parameterization:

Value	Meaning
`-1` (default)	Original GREB (climatological wind + fixed gust term)
`0`	`Ts`-derived skin temperature + climatological wind speed + land/ocean exchange coefficients
`1`	Modified gust/coefficient variant of `-1`
`2`	Modified variant of `0` with different gust/coefficients

log_clim - NCEP-vs-ERA coefficient override:

Value	Meaning
`0` (default)	No override
`1`	If also `log_rain == 0`, overrides the precipitation coefficients to a hardcoded NCEP-tuned set - independent of which climatology files were loaded via `load_greb_jld2!`

Note: log_rain == 0 && log_clim == 1 is the only combination where log_clim has any effect - it's easy to set log_clim = 1 and see no change if log_rain isn't also 0.

External-Forcing Gates

These three switches only matter for :elnino, :lanina, and :rcp85. They gate whether the corresponding anomaly field is loaded from disk; the anomaly is then applied unconditionally in init_model! based on cfg.experiment alone, not on these switches. create_experiment_config already sets all three to true for those three experiments - you should not normally need to touch them directly.

Switch	Default	Effect when `true`
`log_tsurf_ext`	`false`	Loads the surface-temperature anomaly forcing field (RCP8.5/ENSO).
`log_hwind_ext`	`false`	Loads the horizontal wind (u/v) anomaly field.
`log_omega_ext`	`false`	Loads the vertical-velocity (omega) anomaly field.

Experiment-Level Fields

Scenario parameters rather than physics switches - set via create_experiment_config keywords or directly on the returned cfg:

Field	Type	Default	Purpose
`experiment`	`Symbol`	`:full_model`	Selects the experiment branch - see Experiment Presets below
`co2_concentration`	`Float64`	`340.0`	Static CO₂ (ppm) for experiments without a time-varying scenario table
`orbital_index`	`Int`	`0`	Row index into the `solar_scenarios` table for `:obliquity`/`:eccentricity`
`earth_sun_distance_pct`	`Float64`	`0.0`	Percent change in orbital radius for `:earth_sun_distance`
`co2_scenario`	`Dict{Int,Float64}`	`Dict()`	Year→ppm lookup, auto-populated for IPCC RCP/SSP/historical/custom-CO2 experiments
`custom_co2_path`	`String`	`""`	User-supplied "year CO2" text file path for `:custom_co2`

c_q, c_rq, c_omega, and c_omegastd also live on PhysicsConfig but are not meant to be set manually - set_hydrology_parameters! derives them from log_rain/log_clim at the start of every run.

Experiment Presets

What each create_experiment_config preset changes relative to :full_model:

Preset	vs. `:full_model`
`:full_model`	Nothing - the baseline
`:constant_topo`	`log_topo_drsp = false`
`:co2_double`	`co2_concentration = 680.0`
`:co2_quadruple`	`co2_concentration = 1360.0`
`:solar_plus27`	Solar constant +27 W/m²
`:elnino` / `:lanina`	The three `log_*_ext` switches set `true`; adds/subtracts ERA-Interim ENSO anomalies
`:paleo_231kyr`	`co2_concentration = 200.0`; paleo solar-forcing table
`:rcp85`	`log_*_ext` switches `true`; loads CMIP5 RCP8.5 anomaly fields
`:rcp26` / `:rcp45` / `:rcp60` / `:ssp119` / `:ssp126` / `:ssp245` / `:ssp460` / `:ssp585`	No switches change; loads a year→CO2 lookup table
`:historical_co2`	Year counter starts at 1850; loads the observed CO₂ record
`:custom_co2`	`custom_co2_path` set from the `co2_path` keyword
`:decon_mean_climate`	Exposes `log_clouds_dmc`, `log_ocean_dmc`, `log_atmos_dmc`, `log_co2_dmc`, `log_hydro_dmc`, `log_qflux_dmc`, `log_ice`, `log_hdif`, `log_hadv`, `log_vdif`, `log_vadv` as keywords (all default `true`, i.e. behaves like `:full_model` unless overridden)
`:decon_2xco2`	`co2_concentration = 680.0`; exposes `log_topo_drsp`, `log_clouds_drsp`, `log_humid_drsp`, `log_ocean_drsp`, `log_hydro_drsp`, `log_ice`, `log_hdif`, `log_hadv`, `log_vdif`, `log_vadv` as keywords

Further experiments

create_experiment_config covers all of these too. They change no switches relative to :full_model - the experiment symbol alone selects the branch in src/tendencies.jl's forcing, which sets the CO₂ or solar forcing per timestep. Read that branch for exact behavior.

Category	Symbols
CO₂ variants	`:a1b_scenario`, `:co2_10x`, `:co2_half`, `:co2_zero`, `:co2_sine_wave`, `:co2_step`
Orbital/paleo	`:solar_cycle_11yr`, `:paleo_solar_modern_co2`, `:modern_solar_paleo_co2`, `:obliquity`, `:eccentricity`, `:earth_sun_distance`
Other	`:sst_plus1`
Regional CO₂ (static mask)	`:regional_co2_nh`, `:regional_co2_sh`, `:regional_co2_tropics`, `:regional_co2_extratropics`
Regional CO₂ (dynamic/seasonal mask)	`:regional_co2_ocean`, `:regional_co2_land_ice`, `:regional_co2_winter`, `:regional_co2_summer`

Two of them take a parameter, passed as a create_experiment_config keyword:

Keyword	Applies to	Meaning
`orbital_index`	`:obliquity`, `:eccentricity`, `:paleo_231kyr`, `:paleo_solar_modern_co2`	Which row of the `solar_scenarios/*.jld2` table to load
`earth_sun_distance_pct`	`:earth_sun_distance`	Percent change in orbital radius; solar forcing scales as `(1 + 0.01·pct)^-2`

The 16 log_* keywords apply only to :decon_mean_climate and :decon_2xco2. Passing one for any other experiment emits a warning and is ignored - set the field on the returned PhysicsConfig instead.

:regional_co2_ocean and :regional_co2_land_ice derive their mask from the control run's annual-mean ice cover, so it is built once per run by apply_dynamic_co2_mask! between the control and scenario phases rather than by forcing. The control run itself always sees an all-ones mask.

Known Limitations

log_vapor_dmc is declared on PhysicsConfig but is never read anywhere in src/ - setting it currently has no effect on the model.
log_conv works like the other circulation-component switches (gates the moisture-convergence sub-step) but, unlike them, is not exposed as a keyword by create_experiment_config - it must be set with cfg.log_conv = ... after construction.

References

Abbreviations used above: DF11 = Dommenget & Flöter (2011); Hydro19 = Stassen, Dommenget & Loveday (2019); MSCM = Dommenget et al. (2019).

DF11 - Dommenget, D., and Flöter, J. (2011). Conceptual Understanding of Climate Change with a Globally Resolved Energy Balance Model. Journal of Climate Dynamics, 37: 2143. doi:10.1007/s00382-011-1026-0
Hydro19 - Stassen, C., Dommenget, D., and Loveday, N. (2019). A hydrological cycle model for the Globally Resolved Energy Balance (GREB) model v1.0. Geoscientific Model Development, 12, 425-440. doi:10.5194/gmd-12-425-2019
MSCM - Dommenget, D., Nice, K., Bayr, T., Kasang, D., Stassen, C., and Rezny, M. The Monash Simple Climate Model Experiments: An interactive database of the mean climate, climate change and scenarios simulations. Geoscientific Model Development, 12, 2155-2179. doi:10.5194/gmd-12-2155-2019

See the repository README's References section for the original GREB model homepage link.

1Declared in src/config.jl but never read anywhere in src/ - has no effect. See Known Limitations.
2Behaves like the other circulation switches but is not exposed as a keyword by create_experiment_config - set cfg.log_conv directly. See Known Limitations.