Skip to content

Parameters API ​

AtmosTransport.Parameters Module
julia
Parameters

Constants and small immutable parameter objects shared across the package: physical constants, numerical defaults, met-source constants and planet parameters. Numbers that several places use are defined here once.

source
AtmosTransport.Parameters.ADAPTIVE_SUBSTEP_MAX_REFINEMENTS Constant

Most times the adaptive write-time substep schedule raises a window's substep count (to the count the measured Courant ratio implies) before accepting it; one or two normally suffice.

source
AtmosTransport.Parameters.AVOGADRO Constant

Avogadro constant [mol⁻¹] (SI 2019, exact).

source
AtmosTransport.Parameters.CARBON_MOLAR_MASS Constant

Molar mass of carbon [kg mol⁻¹] (IUPAC 2007 standard atomic weight, the value behind SPECIES_MOLAR_MASS.co2 = 12.0107 + 2 × 15.9994). Carbon-mass fluxes (kgC) become CO₂-mass fluxes by SPECIES_MOLAR_MASS.co2 / CARBON_MOLAR_MASS ≈ 3.66419, which conserves carbon atoms; GEOS-Chem's HEMCO uses the scale factor CTOCO2 = 3.6667 (44/12), 0.068 % more.

source
AtmosTransport.Parameters.COARSEN_SUBSTEP_CFL_TARGET Constant

Adaptive substep target of the cubed-sphere binary coarsener.

source
AtmosTransport.Parameters.COLUMN_CLOSURE_TOL_KG Constant

Absolute floor of the per-cell closure tolerance of the cubed-sphere column balance [kg per flux application]: τ_c = max(COLUMN_CLOSURE_TOL_KG, 8 ε_FT Σ_{f∂c} Σ_k |F_{f,k}|) bounds the mean-free column residual of every cell after the balance (balance_cs_column_mass_fluxes!, recheck_cs_column_closure) and, plus the rounding bound of the recursion, the cm residual folded into the column (diagnose_cs_cm!). The floor binds for Float64 payloads (ERA5 C24 golden: largest projected residual 0.10 kg against τ_c = 10 kg); Float32 payloads are bound by the rounding term (smallest τ_c of window 1 of the MERRA-2 C90 golden: 6.1e5 kg).

source
AtmosTransport.Parameters.CP_DRY_AIR Constant

Heat capacity of dry air at constant pressure [J kg⁻¹ K⁻¹] (GEOS; the decimal product 3.5 × 287.04).

source
AtmosTransport.Parameters.CP_OVER_R_DIATOMIC Constant

Ratio of the heat capacity at constant pressure to the gas constant of an ideal diatomic gas, 7/2.

source
AtmosTransport.Parameters.DAYS_PER_JULIAN_YEAR Constant

Days in a Julian year, for annual totals given as rates.

source
AtmosTransport.Parameters.DEFAULT_BACKGROUND_VMR Constant

Dry volume mixing ratio of tracers whose initial condition sets none (400 ppm).

source
AtmosTransport.Parameters.DEFAULT_CFL_LIMIT Constant

Runtime palindrome outflow budget of split-sweep advection (lat-lon, reduced-Gaussian and cubed-sphere): a transport step is split into n subcycles so that one palindrome's worth of outflow, 2 (out_x + out_y + out_z) / n, is at most this fraction of min(m_start, m_end), the smaller of the cell's masses at the start and end of the step. Every pass then removes at most half this fraction of the mass the cell holds when the pass starts.

source
AtmosTransport.Parameters.DEFAULT_GAUSSIAN_BLOB_SIGMA_DEG Constant

Width (standard deviation) of the Gaussian-blob initial condition when the config sets none [degrees].

source
AtmosTransport.Parameters.DEFAULT_GEOS_SMOOTHING_SWEEPS Constant

Smoothing sweeps of the GEOS preprocessing moisture filter and column-drift correction.

source
AtmosTransport.Parameters.DEFAULT_GEOS_SMOOTHING_WEIGHT Constant

Weight of one GEOS moisture-filter / column-drift smoothing sweep (see DEFAULT_GEOS_SMOOTHING_SWEEPS).

source
AtmosTransport.Parameters.DEFAULT_LINROOD_PPM_ORDER Constant

Lin–Rood PPM edge-value family when the config sets none: 5 (Colella–Woodward fourth-order edges) or 7.

source
AtmosTransport.Parameters.DEFAULT_MAX_STEPS_PER_WINDOW Constant

Most write-time substeps per window the cubed-sphere binary coarsener accepts.

source
AtmosTransport.Parameters.DEFAULT_MAX_SUBCYCLES Constant

Most CFL subcycles one split-sweep advection step may need before the run stops: a guard against mass fluxes that would need absurdly many passes.

source
AtmosTransport.Parameters.DEFAULT_MET_INTERVAL_SECONDS Constant

Length of a met window when the source or config does not give one [s].

source
AtmosTransport.Parameters.DEFAULT_POISSON_BALANCE_MAX_ITER Constant

Iteration limit of the horizontal Poisson flux balance.

source
AtmosTransport.Parameters.DEFAULT_POISSON_BALANCE_TOL Constant

Residual tolerance of the horizontal Poisson flux balance (conjugate-gradient solve).

source
AtmosTransport.Parameters.DEFAULT_POISSON_PROJECT_EVERY Constant

Interval, in conjugate-gradient iterations, at which the cubed-sphere Poisson solve projects out the null-space (global-mean) component.

source
AtmosTransport.Parameters.DEFAULT_POSITIVITY_CFL_LIMIT Constant

Positivity limit checked when a transport binary is written: the palindrome outflow of a substep, 2 (out_x + out_y + out_z), must stay below this fraction of min(m, m_next).

source
AtmosTransport.Parameters.DEFAULT_SUBSTEP_CFL_TARGET Constant

Courant number that the adaptive write-time substep schedule aims for.

source
AtmosTransport.Parameters.DRY_AIR_MOLAR_MASS Constant

Molar mass of dry air [kg mol⁻¹] (U.S. Standard Atmosphere 1976; GEOS-Chem AIRMW).

source
AtmosTransport.Parameters.EARTH_RADIUS Constant

Earth radius of every transport target mesh [m] (mean radius, rounded).

source
AtmosTransport.Parameters.ECHLEVS_ML137_TROPO34 Constant

TM5-MP r1112 levels/ml137/tropo34 interface list, exactly as written.

source
AtmosTransport.Parameters.ECHLEVS_ML137_TROPO34A Constant

TM5 cy3-4dvar levels/ml137/tropo34a (CAMS73) interface list, exactly as written.

source
AtmosTransport.Parameters.ERA5_L137_ECHLEVS_PRESETS Constant

Named ERA5 L137 interface selections accepted by preprocessing configs.

source
AtmosTransport.Parameters.ERA5_MET_INTERVAL_SECONDS Constant

Interval of the ERA5 hourly analyses, and accumulation period of its hourly accumulated fields [s].

source
AtmosTransport.Parameters.ERA5_N320_SUBSTEP_CFL_TARGET Constant

Adaptive substep target of the ERA5 N320 → cubed-sphere preprocessor.

source
AtmosTransport.Parameters.ERA5_NATIVE_LEVEL_COUNT Constant

Model levels of the ECMWF IFS L137 grid on which ERA5 is archived.

source
AtmosTransport.Parameters.GEOSCHEM_CONSTANTS Constant

GEOS-Chem's constants, used by the GEOS-Chem non-local boundary-layer scheme: gravity [m s⁻²], dry-air and water-vapour gas constants [J kg⁻¹ K⁻¹] and the von Kármán constant from Headers/physconstants.F90; dry-air heat capacity (cpair) [J kg⁻¹ K⁻¹] and latent heat of vaporisation (latvap) [J kg⁻¹] from GeosCore/vdiff_mod.F90; the potential-temperature reference pressure [Pa] is this model's convention.

source
AtmosTransport.Parameters.GEOSCHEM_SPECIES_MOLAR_MASS Constant

Molar masses of GEOS-Chem's species [kg mol⁻¹]: the MW_g entries of the GEOS-Chem species database (species_database.yml), CO₂ 44.01, SF₆ 146.06 and ²²²Rn 222.0 g mol⁻¹. GEOS-Chem converts between emitted mass and dry mole fraction with these and AIRMW (DRY_AIR_MOLAR_MASS), so a diagnostic that turns GEOS-Chem emission fluxes into dry-air-equivalent storage mass uses them, not the model's SPECIES_MOLAR_MASS (smaller by 1.1e-5 for CO₂ and 3.4e-5 for SF₆). The CATRINE diagnostics use the CO₂ value for the fossil CO₂ tracer too.

source
AtmosTransport.Parameters.GEOS_3HOURLY_INTERVAL_SECONDS Constant

Interval of the GEOS and MERRA-2 three-hourly collections (A3, I3) [s].

source
AtmosTransport.Parameters.GEOS_MASS_FLUX_DT_SECONDS Constant

GEOS FV3 dynamics time step [s]. GEOS-FP and GEOS-IT accumulate the mass fluxes MFXC/MFYC over this step, not over the hourly met interval.

source
AtmosTransport.Parameters.GEOS_MET_INTERVAL_SECONDS Constant

Interval of the GEOS-FP / GEOS-IT hourly CTM collections [s].

source
AtmosTransport.Parameters.GLOBAL_MEAN_DRY_SURFACE_PRESSURE Constant

Global-mean dry surface pressure the preprocessing dry-mass pin holds, for every meteorological source [Pa]. The pin holds the dry column below the model top, <sum_k m_k> g / A; the at-most 1 Pa above the model top is negligible.

The value is the dry mean shared by modern reanalyses: MERRA-2 is 98,324.7 Pa (constant under its dry-mass constraint and equal to GEOS-Chem Met_PSC2DRY), while ERA5 is 98,329 Pa over 2014–2025 (annual means 98,320–98,339 Pa). Trenberth and Smith (2005, J. Climate 18, 864–875) give 98,305 Pa (985.50 hPa × 5.1352/5.1480), 0.02 % lower.

With constant g on a sphere, as in IFS, GEOS, TM5 and this model, the implied dry air mass is 5.114 × 10¹⁸ kg. That is 0.41 % below Trenberth and Smith's 5.1352 × 10¹⁸ kg because their mass-pressure conversion includes Earth's shape and the variation of g with latitude and height. A thin-shell constant-g model cannot match both values; this constant matches the observed pressure.

source
AtmosTransport.Parameters.GLOBAL_MEAN_SPECIFIC_HUMIDITY Constant

Global-mean specific humidity [kg kg⁻¹] that turns the dry surface-pressure target into a total one where only total pressure is available: ps_total = GLOBAL_MEAN_DRY_SURFACE_PRESSURE / (1 − q).

source
AtmosTransport.Parameters.GMAO_CS_LONGITUDE_OFFSET_DEG Constant

Longitude by which the GMAO (GEOS) cubed sphere is rotated [degrees].

source
AtmosTransport.Parameters.HORIZONTAL_BALANCE_TAG Constant

Value of the transport-binary header key horizontal_balance written by every preprocessing path that Poisson-balances its horizontal mass fluxes: "column".

Every path that applies a Poisson balance closes a window's horizontal fluxes against its endpoint mass tendency with one solve for the column-integrated divergence and spreads each face's correction over the layers, so the layer winds keep their vertical structure and the vertical mass flux diagnosed from the explicit tendency carries the vertical motion. GEOS pressure_fixer and pfix_corrected keep their native horizontal fluxes and record "none". A per-layer balance would leave almost nothing for the vertical flux because preprocessing has no independent vertical-flux input (issue #11).

source
AtmosTransport.Parameters.HYBRID_LAYER_MIN_DB Constant

Smallest increment ΔB_k = B_{k+1} − B_k of hybrid B at which a layer counts as hybrid in the column-balance weighting HybridMassWeightedColumn: a layer with ΔB_k < 1e-6 is treated as pure-pressure and receives no part of a column correction. ΔB_k is the share of a surface-pressure change that changes the layer's thickness, Δ(Δp_k) = ΔB_k Δp_s, so at the threshold a 1000 hPa (1e5 Pa) change of p_s changes the thickness by 0.1 Pa. GEOS/MERRA-2 L72 layer 41 (150–177 hPa) has ΔB = 8.2e-9 (8.2e-4 Pa per 1000 hPa, 3.1e-7 of its 2654 Pa thickness at p_s = 1000 hPa), a rounding residue of its B table. The smallest genuine hybrid increment of the shipped tables is ERA5 L137 layer 55 (73–77 hPa, the topmost layer with ΔB > 0): ΔB = 7.0e-6, 0.7 Pa per 1000 hPa, 0.17 % of its 416 Pa thickness. The ERA5 level presets keep it at ΔB = 7.0e-6 (ml137_full, ml137_cfl94, ml137_cfl85) or merge it into a layer with 2.4e-5 (ml137_66L) or 5.9e-5 (ml137_tropo34, ml137_tropo34a); the next GEOS layer has 6.96e-3 (696 Pa per 1000 hPa).

source
AtmosTransport.Parameters.HYDROSTATIC_REFERENCE_TEMPERATURE Constant

Temperature of an isothermal atmosphere used for hydrostatic layer thicknesses when no temperature is available, dz = R_dry T / g · Δp / p_mid (layer pressure thickness over its midpoint pressure) [K].

source
AtmosTransport.Parameters.IFS_EARTH_RADIUS Constant

Earth radius of the ECMWF IFS [m], used only where reproducing IFS/TM5 source geometry requires it.

source
AtmosTransport.Parameters.OMEGA_FIDELITY_CORRELATION_MIN Constant

Write-time OMEGA-fidelity gate: the correlation between ω_cm and the source OMEGA must exceed 0.5 (MERRA-2 wind-derived C90: 0.886). A zero cm has correlation 0, a sign-flipped one −1.

source
AtmosTransport.Parameters.OMEGA_FIDELITY_NRMSE_LIMIT Constant

Write-time OMEGA-fidelity gate: the normalized RMS difference between the vertical velocity diagnosed from a binary's cm and the source OMEGA, NRMSE = RMS(ω_cm − ω_src) / RMS(ω_src), must be below 1.0. NRMSE = 1 is the score of cm ≡ 0, so the limit accepts only a vertical transport that is closer to the source than none at all. A sign-flipped cm scores about 2, a level-flipped one about 1.4; MERRA-2 wind-derived C90 scores 0.47 (2.1× margin; 2022-01 golden).

source
AtmosTransport.Parameters.OMEGA_FIDELITY_SLOPE_MAX Constant

Write-time OMEGA-fidelity gate: the regression slope of ω_cm on the source OMEGA must stay below 1.5 (see OMEGA_FIDELITY_SLOPE_MIN: a cm twice too large has slope 1.73 at MERRA-2 C90).

source
AtmosTransport.Parameters.OMEGA_FIDELITY_SLOPE_MIN Constant

Write-time OMEGA-fidelity gate: the regression slope of ω_cm on the source OMEGA must exceed 0.5 (and stay below OMEGA_FIDELITY_SLOPE_MAX). An amount or time-convention error (2n against n substeps, the 8× mass_flux_dt error, a factor g) scales cm uniformly: it scales the slope by the same factor but leaves the correlation unchanged, and NRMSE stays below 1 for factors down to 1/20, so only a bound on the slope catches it. MERRA-2 wind-derived C90 measures 0.865 (2022-01 golden); 0.5 = 0.865/1.73 and OMEGA_FIDELITY_SLOPE_MAX = 1.5 = 0.865 × 1.73 are symmetric about it in log space, so a factor-2 error (slope 0.43 or 1.73) fails. Zero and sign-flipped cm fail too.

source
AtmosTransport.Parameters.REPLAY_TOLERANCE_FLOAT32 Constant

Replay-check tolerance for Float32 transport-binary data, in the per-cell metric of REPLAY_TOLERANCE_FLOAT64. Stored Float32 binaries measure at most 1.0e-6 (about 9 eps(Float32): Float32 rounding) in C90 and C180 cubed-sphere binaries from ERA5, MERRA-2 and GEOS-IT alike (2026-10 goldens and production samples).

source
AtmosTransport.Parameters.REPLAY_TOLERANCE_FLOAT64 Constant

Replay-check tolerance for Float64 transport-binary data. The check replays each window's air-mass budget cell by cell, r = m_cur − 2n (Σ F_out − Σ F_in) − m_next over the cell's faces, and accepts the window when every cell has |r| ≤ τ (|m_cur| + |m_next| + 2n Σ |F|): the residual relative to the summed magnitudes of the budget's own terms, the scale that the rounding error of the sum is bounded by. Stored Float64 binaries (2026-10 goldens) measure at most 6.5e-14, set by the residual of the Poisson flux balance.

source
AtmosTransport.Parameters.RN222_HALF_LIFE Constant

Half-life of ²²²Rn [s]: 3.8235 days (ENSDF evaluation, NNDC; the value of the CATRINE protocol). In Float64, 3.8235 * 86400.0 is exactly the 330350.4 s that run configs give in [chemistry.half_lives_seconds].

source
AtmosTransport.Parameters.R_DRY_AIR Constant

Gas constant of dry air [J kg⁻¹ K⁻¹] (GEOS, MAPL MAPL_RDRY).

source
AtmosTransport.Parameters.SPECIES_MOLAR_MASS Constant

Molar masses of the transported species [kg mol⁻¹], for converting fluxes and mixing ratios to and from mass: CO₂ (IUPAC atomic weights), SF₆ and ²²²Rn.

source
AtmosTransport.Parameters.STANDARD_GRAVITY Constant

Standard acceleration of gravity [m s⁻²] (3rd CGPM 1901; WMO).

source
AtmosTransport.Parameters.STANDARD_LAPSE_RATE Constant

Temperature lapse rate of the troposphere in the U.S. Standard Atmosphere 1976 [K m⁻¹].

source
AtmosTransport.Parameters.STANDARD_PRESSURE Constant

Standard sea-level pressure [Pa].

source
AtmosTransport.Parameters.STANDARD_SEA_LEVEL_AIR_DENSITY Constant

Air density at sea level in the U.S. Standard Atmosphere 1976 [kg m⁻³].

source
AtmosTransport.Parameters.STANDARD_TROPOPAUSE_TEMPERATURE Constant

Tropopause temperature of the U.S. Standard Atmosphere 1976 [K].

source
AtmosTransport.Parameters.SURFACE_AIR_TEMPERATURE_RANGE Constant

Plausible range of 2 m air temperature in met input [K], an open interval: readers reject values outside it (or on its ends) as a sign of a wrong variable, unit or fill value.

source
AtmosTransport.Parameters.SURFACE_HEAT_FLUX_MAX_ABS Constant

Bound on the magnitude of the surface sensible or latent heat flux in met input; readers reject values at or above it [W m⁻²].

source
AtmosTransport.Parameters.THETA_REFERENCE_PRESSURE Constant

Reference pressure of the potential temperature, θ = T (p₀/p)^κ [Pa].

source
AtmosTransport.Parameters.TM5_CONSTANTS Constant

TM5's physical constants, used by the port of TM5's boundary-layer diffusion (tm5_bldiff.jl): gravity [m s⁻²], dry-air heat capacity and gas constant (Rgas · 1000 / 28.94, TM5 binas.F90) and water-vapour gas constant [J kg⁻¹ K⁻¹], latent heat of vaporisation [J kg⁻¹], von Kármán constant, and the potential-temperature reference pressure [Pa].

source
AtmosTransport.Parameters.TM5_CONVECTION_MASS_BASIS_KEY Constant

Transport-binary header key recording the air-mass basis ("dry" or "moist") of the TM5 convection sections.

source
AtmosTransport.Parameters.TM5_CONVECTION_SOURCE_KEY Constant

Transport-header key naming the source of the TM5 convection sections. ERA5 N320 binaries record "ec2tm_from_rates(udmf,ddmf,udrf,ddrf)" (or "none" without convection); the lat-lon → cubed-sphere regrid copies its source's value ("unrecorded" when the source has none).

source
AtmosTransport.Parameters.TM5_EC2TM_CONVENTION Constant

Current ERA5 → TM5 convection convention, as in TM5 (tmm.F90, ECconv_to_TMconv): ERA5 UDMF/DDMF level k is the flux through the top of layer k, none crosses the ground, and the cloud (downdraft) top is the layer above the first interface that carries flux. ERA5-derived TM5 sections without this value were built with the convention of issue #54 and are rejected when a binary is opened.

source
AtmosTransport.Parameters.TM5_EC2TM_CONVENTION_KEY Constant

Transport-header key recording the index convention of TM5 convection sections derived from ERA5 by ec2tm_from_rates!; its current value is TM5_EC2TM_CONVENTION.

source
AtmosTransport.Parameters.VIRTUAL_TEMPERATURE_FACTOR Constant

Coefficient of the specific humidity in the virtual temperature, T_v = T (1 + 0.61 q): R_vap / R_dry − 1 ≈ 0.608, rounded as in most models.

source
AtmosTransport.Parameters.VON_KARMAN Constant

Von Kármán constant of the model's own boundary-layer schemes (0.41, the value of the legacy PBL diffusion they port). The TM5 and GEOS-Chem ports use their models' 0.4 (TM5_CONSTANTS.karman, GEOSCHEM_CONSTANTS.karman).

source
AtmosTransport.Parameters.PlanetParameters Type
julia
PlanetParameters{FT}

Physical constants associated with the modeled planet.

source