What's New in CESM2

Infrastructure - CIME

In CESM2.0, the coupling infrastructure, scripting utilities and data models are now in CIME (Common Infrastructure for Modeling the Earth), a separate and open-source repository (http://github.com/ESMCI/cime). CIME provides a Case Control System for configuring, compiling and executing Earth system models, data and stub model components, a driver and associated tools and libraries. The CIME Case Control System is a new object-oriented set of python utilities that provide many new capabilities for easier portability, case generation and user customization, testing functionality, and greatly increased robustness and flexibility. CIME can also be downloaded and tested in a stand-alone mode, without any of the other CESM prognostic components, as a first step in porting CESM2.0. CIME also contains a default coupling model architecture that permits active and data components to be mixed in any combination, thereby enabling component feedbacks to be easily controlled. CIME also contains additional tools, such as a new statistical consistency test, that will allow new users to quickly verify their port of CESM2.0.

Science Enhancements

Atmosphere CAM

CAM contains substantial modifications of every atmospheric physics parameterization except for radiative transfer. The Cloud Layers Unified by Binormals (CLUBB) scheme has replaced earlier schemes for boundary layer turbulence, shallow convection, and cloud macrophysics. CLUBB is a prognostic moist turbulence scheme that calculates joint higher-order moments of subgrid vertical velocity, water content, and liquid water potential temperature. Equations for these moments are closed using assumed joint binormal probability density functions (PDFs) for these quantities. In addition to calculating subgrid vertical fluxes, CLUBB’s PDF closure is also used to calculate large-scale condensation and cloud fraction. An improved two-moment prognostic cloud microphysics (MG2) has also been introduced. The major innovation in MG2 is to carry prognostic precipitation species – rain and snow – in addition to cloud condensates. MG2 interacts with the Modal Aerosol Module (MAM4) aerosol microphysics scheme to calculate condensate mass fractions and number concentrations. Deep convection has been significantly retuned to increase the sensitivity to convective inhibition. Two schemes to calculate subgrid orographic drag have been substantially modified. Topographic orientation (ridges) and low-level flow blocking effects have been incorporated into the orographic gravity wave scheme. The previous parameterization of boundary layer form drag – turbulent mountain stress (TMS) – has been replaced with the scheme of Beljaars et al. currently used in the European Center forecast model. In addition to these physics updates, new infrastructure for traceable generation of topographic forcing files has been developed.


CAM-chem now includes the tropospheric chemistry scheme MOZART-T1, combined with comprehensive stratospheric chemistry, coupled to MAM4 (Modal Aerosol Model with 4 modes) which includes a primary carbon mode. The tropospheric chemistry is also coupled to a VBS (volatility basis set) scheme for Secondary Organic Aerosols (SOA). The default specified dynamics configurations use MERRA2 meteorological fields from NASA/GMAO.


WACCM6 now matches all of the CAM6 physical parameterizations, and adds significant new capabilities in the middle and upper atmosphere. WACCM6 extends the Modal Aerosol Module (MAM4) to provide a prognostic representation of stratospheric aerosols from volcanic and non-volcanic source gases. Combined with a database of SO2 emissions from volcanic eruptions, this provides a better representation of the chemistry and climate responses to volcanic eruptions. With a default horizontal resolution 4 times greater than CESM1(WACCM), WACCM6 provides improved stratospheric variability, including an internally generated quasi-biennial oscillation, and an improved climatology of sudden stratospheric warmings. WACCM6 includes updated and unified atmospheric chemistry, adding detailed tropospheric chemistry to the middle and upper atmospheric chemistry provided in CESM1(WACCM). The result of these improvements to chemistry, stratospheric variability, and volcanic aerosols is greater skill in hindcasts of the evolution of the Antarctic ozone hole, as well as ozone loss in the Arctic.


This release includes an extended version of WACCM, WACCM-X v.2.0. This model extends through the thermosphere and ionosphere to above 500 km, and now includes global electrodynamics and ion transport. Other ionosphere developments include time-dependent solution of electron and ion temperatures, metastable oxygen ion chemistry, and capability for high-cadence solar forcing. Additional developments of the thermospheric components are improvements to the momentum and energy equation solvers to account for variable mean molecular mass and specific heat, a new divergence damping scheme, and cooling by atomic oxygen fine structure.

Land CLM

Developments for CLM5 build on the progress made in CLM4.5. Notable changes are made to soil and plant hydrology, snow density, carbon and nitrogen cycling and coupling, and human land management. Hydrology updates include a dry surface layer-based soil evaporation resistance and a revised canopy interception parameterization. The number of ground layers is increased from 15 levels in CLM4.5 to 25 levels in CLM5 to resolve the permafrost active layer and to allow for spatially variable soil thickness, which ranges from 0.4 to 8.5 m deep based on an input dataset. Fresh snow density more realistically captures temperature effects and additionally accounts for wind effects. The maximum number of snow layers and snow amount is increased from 5 layers and 1 m snow water equivalent to 12 layers and 10 m to allow for the formation of firn over ice sheets. A new plant hydraulic stress routine explicitly models water transport through the vegetation with stomatal conductance a function of prognostic leaf water potential. An emergent feature of plant hydraulics is soil hydraulic redistribution. The Ball-Berry maximum stomatal conductance is replaced with the Medlyn conductance model, which is preferred for it’s more realistic behavior at low humidity levels. Plant nutrient dynamics are substantially updated to resolve several CLM4.5 deficiencies. The Fixation and Update of Nitrogen (FUN) model accounts for expenditure of carbon energy for nutrient uptake. Static plant carbon:nitrogen (C:N) ratios utilized in CLM4.5 are replaced with variable plant C:N ratios which allows plants to adjust their C:N ratio according to the cost of N uptake. The flexible C:N ratio eliminates the unrealistic instantaneous downregulation of potential photosynthesis rates based on soil mineral nitrogen availability as a means of representing nutrient limitation. Furthermore, stomatal conductance is now based on N-limited photosynthesis rather than on potential photosynthesis. The Leaf Use of Nitrogen for Assimilation model calculates photosynthetic capacity based on leaf N optimization under different environmental conditions.

Representation of human land management (agriculture, wood harvest) is extended with a global crop model (corn, wheat, soybean, cotton, rice, and sugarcane crops). Fertilization rates and irrigation equipped area are updated annually. Land unit weights can be adjusted during a simulation, which allows for vegetation-crop and vegetation-glacier transitions. More sophisticated and realistic building space heating and air conditioning prognoses interior building air temperature and waste heat. Included with CLM5 is a functionally supported version of the Functionally-Assembled Terrestrial Ecosystem Simulator (FATES). FATES is a cohort model of vegetation competition and co-existence, allowing a representation of the biosphere which accounts for the division of land into successional stages, and for competition for light between height structured cohorts of representative trees of various plant functional types. FATES is not active by default in CLM5.

Land Ice CISM

CESM2 includes many improvements to support realistic simulations of land ice. Version 2.1 of the Community Ice Sheet Model (CISM) has a parallel, higher-order dynamical core that can accurately simulate not just slow interior flow, but also fast flow along the margins of ice sheets. CISM's parameterizations of basal sliding, iceberg calving, and other physical processes have also improved. As a result, the match to observed Greenland ice flow is much closer than with previous model versions. The surface mass balance of the Greenland and Antarctic ice sheets is now computed by default in the Community Land Model (CLM), with or without dynamic ice sheets. The surface melt climate of both ice sheets has improved with the inclusion of a deep firn model that allows for meltwater infiltration and refreezing, as well as realistic densification rates. Surface winds over ice sheets are more accurate with a new drag parameterization, and a bias in high-latitude longwave cloud forcing is much reduced. CESM now supports interactive coupling of CISM with CLM and CAM, allowing the land topography and surface types to evolve as ice sheets advance and retreat.

Ocean POP

Since the CESM1.2 release, there have been improvements to both the physical parameterizations and numerical methods in the CESM ocean component, Parallel Ocean Program version 2 (POP2). The physics improvements include a new parameterization for mixing effects in estuaries to improve the representation of the exchange of freshwater between the terrestrial and marine branches of the hydrologic cycle; increased mesoscale eddy diffusivities at depth to improve the representation of passive tracers; use of prognostic chlorophyll for short-wave absorption; use of salinity dependent freezing-point together with the sea-ice model; and a new Langmuir mixing parameterization in conjunction with the new wave model component. The numerical improvements include a new iterative solver for the barotropic mode to reduce communication costs, particularly advantageous for high-resolution simulations on large processor counts; a new time filtering scheme based on an adaption of the Robert filter to enable sub-diurnal coupling of the ocean model; and subsequently, use of one-hour coupling frequency to explicitly resolve (sub-)diurnal and inertial periods. In addition, the K-Profile vertical mixing Parameterization (KPP) is incorporated via the Community ocean Vertical Mixing (CVMix) framework and the Caspian Sea is no longer included in the ocean model as a marginal sea. Finally, ocean biogeochemisty has been modularized under the Marine Biogeochemistry Library (MARBL) to enable portability to alternative physical frameworks.

Sea Ice CICE

CESM2 has moved to the Los Alamos National Laboratory Sea Ice Model (CICE) version 5.1.2. Although its structure is very similar to that of CICE4, CICE5 has a number of new physics options. The default CICE5 configuration in CESM2 includes the following new features: a mushy-layer thermodynamics scheme, adding in prognostic salinity to the thermodynamic calculations; a level melt pond scheme that takes into account the roughness on the surface of the ice for melt pond fraction; and salinity dependent freezing point in conjunction with the ocean model. In addition, the default number of ice layers has been increased to 8 (from 4) and snow layers to 3 (from 1) to primarily help better resolve the thermodynamic calculation.

River Model MOSART

The River Transport Model (RTM) is replaced with the Model for Scale Adaptive River Transport (MOSART) in which surface runoff is routed across hillslopes and then discharged along with subsurface runoff into a tributary subnetwork before entering the main channel. A primary difference is that RTM uses a linear reservoir method for river flow, whilst MOSART river flow is determined with the more physically-based kinematic wave method. RTM only produces streamflow, whereas MOSART additionally simulates time-varying channel velocities, channel water depth, and channel surface water variations.

Wave WW3

A new capability in CESM2 is the incorporation of a version of the NOAA WaveWatch-III ocean surface wave prediction model as a new CESM component. The waves are driven by winds through the coupler, and deliver three new benefits. First, the waves are used to provide the forcing necessary for including Langmuir, or wave-driven, turbulence in the ocean model, which results in bias reductions in mixed layer depths, pycnocline ventilation, and subsurface temperatures. Second, the wave statistics from the CMIP6 experiments will be contributed to the Coordinated Ocean Wave Climate Project (COWCLIP) experiment, which studies the changes in wave climate under climate change. The CESM contribution will be unique in the comparison as the waves are run in a fully coupled system. Third, the wave statistics are available now for development of other wave-related process studies and parameterizations, e.g., drag and gas transfer coefficients depending on sea state, waves in the marginal ice zone, or sea spray aerosol production. The implementation of the waves has benefitted from the wave model developers at the NOAA National Centers for Environmental Prediction (NCEP). If the wave model is not used, a theory wave feature in the Community ocean Vertical Mixing (CVMix) framework provides similar Langmuir mixing effects at lower cost.

CIME Data Atmosphere Model DATM

The new default CLM forcing dataset was developed for the third phase of the Global Soil Wetness Project (GSPW3). It is also the default dataset for LS3MIP and LUMIP land-only simulations. It is a 3-hourly 0.5° global forcing product (1901-2014) that is based on the NCEP 20th Century Reanalysis (20CR) version 2. 20CR output was then dynamically downscaled to T248 (0.5°) resolution using the Global Spectral Model with aspectral nudging technique. Bias correction for temperature, precipitation, and longwave radiation, and shortwave radiation were made using CRU TS v3.21 (Climate Research Unit), GPCCv6 (Global Precipitation Climatology Centre), and SRB (Surface Radiation Budget) datasets, respectively. A wind-induced snow undercatch correction was also applied. The CRUNCEP forcing dataset that is used in TRENDY is also included. Additional forcing datasets can be readily incorporated by the user.

To overcome various deficiencies of the existing atmospheric data sets used for forcing ocean – sea-ice coupled simulations – referred to as the Coordinated Ocean-ice Reference Experiments (CORE), a new forcing data set has been developed and made available in CESM2.0. The new inter-annually varying data set is based on the JRA-55 reanalysis product from the Japanese Meteorological Agency (JMA) and covers the 1958 – 2016 period. The development of the data set has been done in collaboration with our colleagues from the JMA Meteorological Research Institute, CLIVAR Ocean Model Development Panel (OMDP), and CESM Ocean Model Working Group (OMWG).