AtmosModel

AtmosProblem

The default problem definition (initial and boundary conditions) for AtmosModel.

AtmosPhysics

An AtmosPhysics for atmospheric physics

Usage

AtmosPhysics(
    param_set,
    ref_state,
    energy,
    moisture,
    compressibility,
    turbulence,
    turbconv,
    hyperdiffusion,
    precipitation,
    radiation,
    tracers,
    lsforcing,
)

Fields

• param_set

  Parameter Set (type to dispatch on, e.g., planet parameters. See CLIMAParameters.jl package)

• ref_state

  Reference State (For initial conditions, or for linearisation when using implicit solvers)

• energy

  Energy sub-model, can be energy-based or θliqice-based

• moisture

  Moisture Model (Equations for dynamics of moist variables)

• compressibility

  Compressibility switch

• turbulence

  Turbulence Closure (Equations for dynamics of under-resolved turbulent flows)

• turbconv

  Turbulence Convection Closure (e.g., EDMF)

• hyperdiffusion

  Hyperdiffusion Model (Equations for dynamics of high-order spatial wave attenuation)

• viscoussponge

  Viscous sponge layers

• precipitation

  Precipitation Model (Equations for dynamics of precipitating species)

• radiation

  Radiation Model (Equations for radiative fluxes)

• tracers

  Tracer Terms (Equations for dynamics of active and passive tracers)

• lsforcing

  Large-scale forcing (Forcing information from GCMs, reanalyses, or observations)

AtmosModel <: BalanceLaw

A BalanceLaw for atmosphere modeling. Users may over-ride prescribed default values for each field.

Usage

AtmosModel(
    physics,
    problem,
    orientation,
    source,
    data_config,
)

Fields

• physics

  Atmospheric physics

• problem

  Problem (initial and boundary conditions)

• orientation

  An orientation model

• source

  Source Terms (Problem specific source terms)

• data_config

  Data Configuration (Helper field for experiment configuration)

flux_first_order!(
    bl::BalanceLaw,
    flux::Grad,
    state::Vars,
    aux::Vars,
    t::Real
)

Computes (and assembles) flux terms F¹(Y) in:

∂Y
-- + ∇ • F¹(Y) + ∇ • F²(Y,G) = S(Y, G),     G = ∇Y
∂t

Computes and assembles non-diffusive fluxes in the model equations.

For this fallback to work, several methods must be defined:

• prognostic_vars
• eq_tends
• get_prog_state

optionally,

• precompute

and individual flux kernels that are defined for each type that eq_tends returns.

flux_second_order!(
    bl::BalanceLaw,
    flux::Grad,
    state::Vars,
    diffusive::Vars,
    hyperdiffusive::Vars,
    aux::Vars,
    t::Real
)

Computes (and assembles) flux terms F²(Y, G) in:

∂Y
-- + ∇ • F¹(Y) + ∇ • F²(Y,G) = S(Y, G),     G = ∇Y
∂t

Diffusive fluxes in BalanceLaw. Viscosity, diffusivity are calculated in the turbulence subcomponent and accessed within the diffusive flux function. Contributions from subcomponents are then assembled (pointwise).

For this fallback to work, several methods must be defined:

• prognostic_vars
• eq_tends
• get_prog_state

optionally,

• precompute

and individual flux kernels that are defined for each type that eq_tends returns.

init_state_auxiliary!(
    m::AtmosModel,
    aux::Vars,
    grid,
    direction
)

Initialise auxiliary variables for each AtmosModel subcomponent. Store Cartesian coordinate information in aux.coord.

source!(
    bl::BalanceLaw,
    source::Vars,
    state::Vars,
    diffusive::Vars,
    aux::Vars,
    t::Real,
    direction::Direction,
)

Computes (and assembles) source terms S(Y) in:

∂Y
-- + ∇ • F¹(Y) + ∇ • F²(Y,G) = S(Y, G),     G = ∇Y
∂t

For this fallback to work, several methods must be defined:

• prognostic_vars
• eq_tends
• get_prog_state

optionally,

• precompute

and individual source kernels that are defined for each type that eq_tends returns.

init_state_prognostic!(
    m::AtmosModel,
    state::Vars,
    aux::Vars,
    localgeo,
    t,
    args...,
)

Initialise state variables. args... provides an option to include configuration data (current use cases include problem constants, spline-interpolants).

Compressible <: Compressibilty

Dispatch on compressible model (default)

• Density is prognostic

Anelastic1D <: Compressibilty

Dispatch on Anelastic1D model

• The state density is taken constant in time and equal to the reference density. This constant density profile is used in all equations and conversions from conservative to specific variables per unit mass. The density can be accessed using the dispatch function density(atmos, state, aux).
• The thermodynamic state is constructed from the reference pressure (constant in time), and the internal energy (which evolves in time).
• The state density is not consistent with the thermodynamic state, since we neglect buoyancy perturbations on all equations except in the vertical buoyancy flux.
• The density obtained from the thermodynamic state, air_density(ts), recovers the full density, which should only be used to compute buoyancy and buoyancy fluxes, and in the FV reconstruction.
• Removes momentum z-component tendencies, assuming balance between the pressure gradient and buoyancy forces.

HydrostaticState{P,T} <: ReferenceState

A hydrostatic state specified by a virtual temperature profile and relative humidity.

By default, this is a dry hydrostatic reference state.

InitStateBC

Set the value at the boundary to match the init_state_prognostic! function. This is mainly useful for cases where the problem has an explicit solution.

TODO: This should be fixed later once BCs are figured out (likely want

different things here?)

ReferenceState

Hydrostatic reference state, for example, used as initial condition or for linearization.

NoReferenceState <: ReferenceState

No reference state used

recover_thermo_state(atmos::AtmosModel, state::Vars, aux::Vars)

An atmospheric thermodynamic state.

TODO: Define/call recover_thermo_state when it's safely implemented

(see https://github.com/CliMA/ClimateMachine.jl/issues/1648)

new_thermo_state(atmos::AtmosModel, state::Vars, aux::Vars)

Create a new thermodynamic state, based on the state, and not the aux state.

recover_thermo_state_anelastic(atmos::AtmosModel, state::Vars, aux::Vars)

An atmospheric thermodynamic state.

TODO: Define/call recover_thermo_state_anelastic when it's safely implemented

(see https://github.com/CliMA/ClimateMachine.jl/issues/1648)

new_thermo_state_anelastic(atmos::AtmosModel, state::Vars, aux::Vars)

Create a new thermodynamic state, based on the state, and not the aux state.

DryModel

Assumes the moisture components is in the dry limit.

EquilMoist

Assumes the moisture components are computed via thermodynamic equilibrium.

NonEquilMoist

Does not assume that the moisture components are in equilibrium.

NoPrecipitation <: PrecipitationModel

No precipitation.

RainModel <: PrecipitationModel

Precipitation model with rain.

RainSnowModel <: PrecipitationModel

Precipitation model with rain and snow.

RayleighSponge{FT} <: TendencyDef{Source}

Rayleigh Damping (Linear Relaxation) for top wall momentum components Assumes laterally periodic boundary conditions for LES flows. Momentum components are relaxed to reference values (zero velocities) at the top boundary.

AtmosBC(momentum = Impenetrable(FreeSlip())
        energy   = Insulating()
        moisture = Impermeable()
        precipitation = OutflowPrecipitation()
        tracer  = ImpermeableTracer())

The standard boundary condition for AtmosModel. The default options imply a "no flux" boundary condition.

DragLaw(fn) :: MomentumDragBC

Drag law for momentum parallel to the boundary. The drag coefficient is C = fn(state, aux, t, normu_int_tan), where normu_int_tan is the internal speed parallel to the boundary. _int refers to the first interior node.

Impermeable() :: MoistureBC

No moisture flux.

PrescribedMoistureFlux(fn) :: MoistureBC

Prescribe the net inward moisture flux across the boundary by fn, a function with signature fn(state, aux, t), returning the flux (in kg/m^2).

BulkFormulaMoisture(fn) :: MoistureBC

Calculate the net inward moisture flux across the boundary using the bulk formula. The drag coefficient is C_q = fn_C_q(state, aux, t, normu_int_tan). The surface qtot at the boundary is `qtot = fnqtot(state, aux, t)`.

Return the flux (in kg m^-2 s^-1).

FreeSlip() :: MomentumDragBC

No surface drag on momentum parallel to the boundary.

PrescribedTemperature(fn) :: EnergyBC

Prescribe the temperature at the boundary by fn, a function with signature fn(state, aux, t) returning the temperature (in K).

PrescribedEnergyFlux(fn) :: EnergyBC

Prescribe the net inward energy flux across the boundary by fn, a function with signature fn(state, aux, t), returning the flux (in W/m^2).

NishizawaEnergyFlux(fn) :: EnergyBC

Calculate the net inward energy flux across the boundary following Nishizawa and Kitamura (2018). Return the flux (in W m^-2).

Adiabaticθ(fn) :: EnergyBC

Prescribe the net inward potential temperature flux across the boundary by fn, a function with signature fn(state, aux, t), returning the flux (in kgK/m^2).

BulkFormulaEnergy(fn) :: EnergyBC

Calculate the net inward energy flux across the boundary. The drag coefficient is C_h = fn_C_h(atmos, state, aux, t, normu_int_tan). The surface temperature and qtot are `T, qtot = fnTandqtot(atmos, state, aux, t)`. Return the flux (in W m^-2).

OutflowPrecipitation() :: PrecipitationBC

Free flux out of the domain.

ImpermeableTracer() :: TracerBC

No tracer diffusion across boundary

Impenetrable(drag::MomentumDragBC) :: MomentumBC

Defines an impenetrable wall model for momentum. This implies:

• no flow in the direction normal to the boundary, and
• flow parallel to the boundary is subject to the drag condition.

Insulating() :: EnergyBC

No energy flux across the boundary.

NoSlip() :: MomentumDragBC

Zero momentum at the boundary.

average_density(ρ_sfc, ρ_int)

Average density between the surface and the interior point, given

• ρ_sfc density at the surface
• ρ_int density at the interior point

RemovePrecipitation <: TendencyDef{Source}

A sink to q_tot when cloud condensate is exceeding a threshold. The threshold is defined either in terms of condensate or supersaturation. The removal rate is implemented as a relaxation term in the CloudMicrophysics.jl Microphysics_0M module. The default thresholds and timescale are defined in CLIMAParameters.jl.

CreateClouds <: TendencyDef{Source}

A source/sink to q_liq and q_ice implemented as a relaxation towards equilibrium in the Microphysics module. The default relaxation timescales are defined in CLIMAParameters.jl.

WarmRain_1M <: TendencyDef{Source}

A collection of source/sink terms related to 1-moment warm rain microphysics. The microphysics process rates are implemented in the CloudMicrophysics.jl Microphysics_1M module.

RainSnow_1M <: TendencyDef{Source}

A collection of source/sink terms related to 1-moment rain and snow microphysics The microphysics process rates are implemented in the CloudMicrophysics.jl Microphysics_1M module.

NoLSForcing <: LSForcingModel

No large-scale forcing

Container for GCM variables from HadGEM2-A forcing,

used in the AMIP experiments