ClimaAtmos.jl

ClimaAtmos is the atmosphere model of the Climate Modeling Alliance (CliMA) Earth system model. CliMA is a consortium, led by Caltech and MIT, that has built a new Earth system model from scratch, in Julia, designed to learn from data. ClimaAtmos is its nonhydrostatic atmosphere component, spanning large-eddy simulation in Cartesian domains to global weather and climate simulation on the sphere [1].

The model can be run from the Julia REPL, with no namelists or batch scripts required; a first global simulation takes three lines of code (see Your First Simulation).

What the model does

• One model across scales. ClimaAtmos solves the fully compressible equations of motion for a deep atmosphere in a coordinate-independent formulation. The same equation set runs in Cartesian geometries, for cloud-resolving and large-eddy simulations, and on the sphere, for global weather and climate simulations.
• Conservative by construction. Energy, air mass, and water are conserved to floating-point precision, without ad hoc fixers. Using the specific total energy of moist air as a prognostic variable, together with an internally consistent thermodynamic formulation, guarantees closed budgets even in moist atmospheres and in the presence of subgrid-scale parameterizations.
• Performance portable. The model runs on CPUs and GPUs from the same code base and scales from a laptop to cloud supercomputers, with strong and weak scaling that make kilometer-scale global simulations achievable.
• Resolution-adaptive physics. The parameterization suite, built around the PROPHET scheme (an extended, prognostic eddy-diffusivity mass-flux scheme), avoids assumptions of scale separation that become inadequate as resolved scales approach the scales of parameterized processes such as atmospheric turbulence and convection. As resolution increases, the parameterized transport diminishes and hands over to the resolved flow.
• Built for calibration with data. All model parameters live in a central repository, ClimaParams.jl, and CliMA's calibration tools tune parameters against data, whether those are output from high-resolution simulations or Earth observations from space and from the ground.

To learn more, the dynamical core (concepts, numerics, and scaling) is described in [1], with companion papers to follow.

ClimaAtmos in the CliMA ecosystem

ClimaAtmos composes packages from the CliMA ecosystem into a full model; installing ClimaAtmos brings every dependency along.

• ClimaCore.jl provides the numerical methods for the dynamical core: the spectral-element/finite-difference discretization, fields, and discrete operators on which the governing equations are built.
• ClimaTimeSteppers.jl provides the implicit–explicit (IMEX) time steppers designed for Earth system models, used together with the implicit solver.
• Thermodynamics.jl and ClimaParams.jl are the shared foundation: one thermodynamic formulation for all CliMA components (the basis for energetic consistency) and a single source of truth for all model parameters.
• Physics libraries supply the parameterized processes: Insolation.jl, RRTMGP.jl, SurfaceFluxes.jl, and CloudMicrophysics.jl.
• ClimaCoupler.jl couples ClimaAtmos to the CliMA land, ocean, and sea-ice models, and ClimaCalibrate.jl drives the calibration-with-data workflows.

See The CliMA Ecosystem for the full architectural overview, including where each package enters the ClimaAtmos source code.

Finding your way around

Each section of these docs answers a different need; head for the one that matches yours:

• Getting Started: new to ClimaAtmos? Start here.

  • Installation – install the package, or run from a clone
  • Your First Simulation – build, run, and inspect one simulation
  • Script vs Config Interface – the two ways to configure a run

• How-to Guides: task recipes for running and configuring simulations.

  • Running simulations: single-column cases, global simulations, restarts, running on GPUs and MPI
  • Configuration: scripting simulations, custom configurations
  • Computing and saving diagnostics
  • Loading and visualizing output

• Explanation: the science and numerics behind the model.

  • The CliMA ecosystem – how ClimaAtmos composes the CliMA packages
  • Dynamics & numerics: governing equations, implicit solver, integer time (ITime)
  • Physics & parameterizations: PROPHET, microphysics, radiation, non-orographic gravity-wave drag, orographic gravity-wave drag, ocean surface albedo, topography

• Reference: look-up material.

  • API, Configuration options, Setups, Column Datasets, Grids, Surface conditions, Passive tracers, Trace gases, Available diagnostics, Glossary, Bibliography

• Developer Guide: contributing and extending the model (setups, diagnostics, tracers, column datasets, surface internals)

New here? Start with Installation and Your First Simulation, then pick the workflow that suits you in Script vs Config Interface. For column experiments, see Running Single-Column Cases.

ClimaAtmos is open source under the Apache 2.0 license. Questions and bug reports are welcome on the GitHub issue tracker.