CCSM4: Notable Improvements
CCSM4 contains totally new infrastructure capabilities that permit new flexibility and extensibility to address the challenges involved in earth system modeling. An integral part of CCSM4 is the implementation of a coupling architecture that takes a completely new approach with respect to the high-level design of the system. The CCSM4 coupling infrastructure now provides the ability to use a single code base in a start-to-end development cycle; from model parameterization development that might only require a single processor, to performing ultra high resolution simulations on HPC platforms using tens-of-thousands of cores. The CCSM4 coupling architecture also provides "plug and play" capability of data and active components and includes a user-friendly scripting system and informative timing utilities. Together, these tools enable a user to create a wide variety of "out-of-the-box" experiments for different model configurations and resolutions and also to determine the optimal load balance for those experiments to ensure maximal throughput and efficiency. CCSM4 is also targeting much higher resolutions than any previous CCSM coupled model and efforts have been made to reduce the memory footprint and to improve scaling in all components.
CAM contains notable improvements to the deep convection, arctic cloud fraction, radiation interface and computational scalability. The calculation of Convective Available Potential Energy (CAPE) has been reformulated in the deep convection to include more realistic dilution effects through an explicit representation of entrainment. Sub-grid scale momentum transports have also been added to the deep convection parametrization. A freeze-drying process contributes to a greater consistency between polar cloud fraction and water condensate properties. The finite volume dynamical core has been made the default and has improved accuracy for transport processes. The option of the HOMME spectral element dynamical core on a cubed sphere grid vastly increases the computational scalability of CAM. A new radiation constituent interface expands flexibility in the specification of properties for individual gas and aerosol species and their radiative interaction. The TROP-MOZART chemistry and aerosol package is now available and is used to generate evolving aerosol burden changes from the most recent emission inventories. Additional physical process improvements from CAM development will be available in CAM5.
CLM has been modified substantially and includes several new capabilities, input datasets, and parameterization updates. The model is extended with a carbon-nitrogen (CN) cycle model that is prognostic in carbon and nitrogen as well as vegetation phenology. A transient landcover change capability, including wood harvest, is introduced and the dynamic vegetation model is merged with CN (CNDV). An urban model (CLMU) is added and the BVOC model is replaced with the MEGAN model. The hydrology scheme is updated with a TOPMODEL-based runoff model, a simple groundwater model, a new frozen soil scheme, a new soil evaporation parameterization, and a corrected numerical solution of the Richards equation. The snow model incorporates SNICAR - which includes aerosol deposition, grain-size dependent snow ageing, and vertically resolved snowpack heating - as well as new snow cover fraction and snow burial fraction parameterizations. CLM4 also includes a new canopy integration scheme, new canopy interception scaling, and a representation of organic soil thermal and hydraulic properties. The ground column is extended to ~50-m depth by adding 5 bedrock layers (15 total layers). New surface datasets based on MODIS products have been derived, providing a basis for the transient land cover datasets. To improve global energy conservation, runoff is split into separate liquid and ice water streams that are passed separately to the ocean model.
The CCSM sea ice component is now CICE, the Los Alamos Sea Ice Model, sometimes referred to as the Community Ice CodE. The main areas of enhancement fall into two categories: physics and computation. The scientific enhancements include new tracers, a new shortwave radiative transfer scheme, a melt pond scheme, and aerosol deposition, all applied to the snow and sea ice. The new computational enhancements include: more flexible computational decomposition strategies, high resolution support, parallel input / output, and OpenMP threading capability.
The ocean model has been updated to the Parallel Ocean Program version 2 (POP2) of the Los Alamos National Laboratory. Many physical and software developments have been incorporated. The physical improvements include a near-surface eddy flux parameterization; an abyssal tidally driven mixing parameterization; an overflow parameterization to represent the Denmark Strait, Faroe Bank Channel, Weddell Sea, and Ross Sea overflows; a submesoscale mixing scheme; vertically-varying thickness and isopycnal diffusivity coefficients; modified anisotropic horizontal viscosity coefficients with much lower magnitudes than in CCSM3; and modified K-Profile Parameterization that uses horizontally-varying background vertical diffusivity and viscosity coefficients. The software developments include capability for multiple time-averaged history files and space-filling curves. The number of vertical levels has been increased from 40 levels in CCSM3 to 60 levels in CCSM4. POP2 includes passive tracer infrastructures and ecosystem codes.
The CCSM4 data models have been completely rewritten. They are now parallelized and share significant amounts of source code. The new data models have created a natural hierarchy in the system and methods for reading and interpolating data have been established that can easily be reused.