Available Simulation Types

The simulation type or "mode" controls which component models are used and how they are configured. It is selected via the mode_name configuration option (see Input).

`mode_name`	Description
`amip`	Atmosphere + prescribed ocean/sea ice + land
`cmip`	Atmosphere + dynamic ocean/sea ice + land
`slabplanet`	Atmosphere + slab ocean + land
`slabplanet_aqua`	Atmosphere + slab ocean only (aquaplanet)
`slabplanet_terra`	Atmosphere + land only
`subseasonal`	Short-range forecast from ERA5 initial conditions using AMIP setup

All simulation types use ClimaAtmos.jl as the atmosphere model. For AMIP, CMIP, and subseasonal runs the land model may be either the bucket or integrated land model from ClimaLand.jl; for slabplanet, aquaplanet, and terraplanet runs the bucket land model must be used. See Available component models for details on each component model.

CMIP

CMIP (Coupled Model Intercomparison Project) is the most complex simulation type supported by ClimaCoupler.jl. In addition to the prognostic atmosphere and land models, the ocean evolves prognostically in response to atmospheric forcing, and sea ice is thermodynamically active.

Component models:

Atmosphere: ClimaAtmosSimulation
Land: BucketSimulation or ClimaLandSimulation (controlled by land_model)
Ocean: OceananigansSimulation
Sea ice: ClimaSeaIceSimulation

GPU recommended

The CMIP configuration is computationally expensive due to the complexity of all component models. Running on a GPU is strongly recommended; see the device option in Input for how to select the compute device.

Behavior at the poles

The Oceananigans and ClimaSeaIce models currently run on a capped latitude-longitude grid spanning 80°S to 80°N. To avoid a gap at the poles, the selected land model is used to fill the polar regions. As a result, the land model cannot be started from saved initial conditions in this configuration. This will change in the future when the models switch to a tripolar grid.

CMIP mode is enabled by setting mode_name: "cmip".

AMIP

AMIP (Atmospheric Model Intercomparison Project) is a standard experimental protocol of the Program for Climate Model Diagnosis & Intercomparison (PCMDI). It is used to evaluate atmosphere and land model components while sea surface temperatures (SST) and sea ice concentration (SIC) are prescribed from observational data (e.g., HadISST).

Component models:

Atmosphere: ClimaAtmosSimulation
Land: BucketSimulation or ClimaLandSimulation (controlled by land_model)
Ocean: PrescribedOceanSimulation
Sea ice: PrescribedIceSimulation

AMIP mode is enabled by setting mode_name: "amip".

Slabplanet

The slabplanet configuration is a more idealized setup than AMIP, designed for studying conservation properties and individual model behavior. The ocean is a simple thermal slab with a prognostic surface temperature but no dynamics, and there is no sea ice.

Component models:

Atmosphere: ClimaAtmosSimulation
Land: BucketSimulation or ClimaLandSimulation
Ocean: SlabOceanSimulation
Sea ice: none (ocean fills ice-covered regions)

Slabplanet mode is enabled by setting mode_name: "slabplanet".

Slabplanet aqua

An aquaplanet setup: the slab ocean covers the entire surface with no land or sea ice.

Component models:

Atmosphere: ClimaAtmosSimulation
Ocean: SlabOceanSimulation (entire surface)
Land: none
Sea ice: none

Aquaplanet is enabled by setting mode_name: "slabplanet_aqua".

Slabplanet terra

A land-only analog to aquaplanet: the land model covers the entire surface with no ocean or sea ice.

Component models:

Atmosphere: ClimaAtmosSimulation
Land: BucketSimulation (entire surface)
Ocean: none
Sea ice: none

Terraplanet is enabled by setting mode_name: "slabplanet_terra".

Subseasonal

Generates 3–4 week forecasts initialized from ERA5 reanalysis data. The setup is otherwise similar to AMIP, but uses specific ERA5-derived initial conditions for the land model. The era5_initial_condition_dir option must point to a directory containing the initial condition files described below.

Initial condition files can be generated using the WeatherQuest package.

Subseasonal mode is enabled by setting mode_name: "subseasonal".

Expected input files

Given a start_date formatted as YYYYMMDD, the following files are expected in era5_initial_condition_dir:

File	Contents
`sst_processed_YYYYMMDD_0000.nc`	SST variable `SST` in °C; converted internally to K
`sic_processed_YYYYMMDD_0000.nc`	Sea ice concentration `SEAICE` in percent; converted to fraction
`era5_bucket_processed_YYYYMMDD_0000.nc`	Bucket land IC; auto-inferred if `bucket_initial_condition` not set
`era5_land_processed_YYYYMMDD_0000.nc`	Integrated land IC; required for `land_model: integrated`
`albedo_processed_YYYYMMDD_0000.nc`	Optional surface albedo; used when `bucket_albedo_type: era5`

Bucket land IC (`era5_bucket_processed_YYYYMMDD_0000.nc`)

Variable	Units	Dimensions	Description
`W`	m	`(lat, lon)`	Subsurface water content
`Ws`	m	`(lat, lon)`	Surface water content
`S`	m	`(lat, lon)`	Snow water equivalent
`T`	K	`(z, lat, lon)`	Soil temperature profile
`tsn`	K	`(lat, lon)`	Snow layer temperature
`skt`	K	`(lat, lon)`	Skin temperature

For variables that are resolved vertically in the ERA5 data, the data is present at the following land level midpoints: 0.035, 0.175, 0.64, 1.945 m.

Integrated land IC (`era5_land_processed_YYYYMMDD_0000.nc`)

Variable	Units	Dimensions	Description
`swe`	m	`(lat, lon)`	Snow water equivalent
`swvl`	m³/m³	`(z, lat, lon)`	Volumetric liquid water fraction
`si`	m³/m³	`(z, lat, lon)`	Volumetric ice fraction
`sie`	J/m³	`(z, lat, lon)`	Soil volumetric internal energy
`stl`	K	`(z, lat, lon)`	Soil temperature
`tsn`	K	`(lat, lon)`	Snow layer temperature
`skt`	K	`(lat, lon)`	Skin temperature

For variables that are resolved vertically in the ERA5 data, the data is present at the following land level midpoints: 0.035, 0.175, 0.64, 1.945 m.

Surface albedo (`albedo_processed_YYYYMMDD_0000.nc`)

Used when bucket_albedo_type: era5. Contains sw_alb_clr (clear-sky surface albedo, fraction), with dimensions (time, lat, lon) representing monthly data that is temporally interpolated.