next up previous contents
Next: 2 Obtaining the Source Up: 1 Introduction Previous: 1 Introduction   Contents

Subsections

1.1 What is CSIM4?

The Community Sea Ice Model (CSIM4) is a component of the Community Climate System Model (CCSM2), a fully-coupled global climate model that provides state-of-the-art computer simulations of the Earth's past, present and future climate states. The sea ice component includes the elastic-viscous-plastic (EVP) dynamics scheme, an ice thickness distribution, energy-conserving thermodynamics, a slab ocean mixed layer model, and the ability to run using prescribed ice concentrations. It is supported on high and low-resolution Greenland Pole grids, which are identical to those used by the Parallel Ocean Program (POP) ocean model. The high resolution version is best suited for simulating present-day and future climate scenarios while the low resolution option is used for paleoclimate simulations and debugging.

In addition to this User's Guide, there is a Scientific Document that describes the model physics and a Reference Guide with a detailed description of the ice model source code. This user's guide describes how to obtain and use CSIM4 and is intended for users who want to run the model with minor changes introduced via existing namelist options. Users interested in modifying the source code should see the Reference Document.

Although it is possible to run this model in a stand alone mode, this option is not discussed here. All configurations presented in this document exchange information with other model components via a flux coupler. An attempt has been made throughout this document to provide the following text convention. Namelist names and variables are typewritten. Directory and file names are given in italic.

1.1.1 Overview of CSIM4 within CCSM2

The directory structure of the fully coupled CCSM2 is shown in Figure 1. In general, CCSM2 consists of four models (atmosphere, ice, land, ocean) plus a coupler for exchanging information between these components. Each of the four models consists of an "active" version with state-of-the-art physics and a data, version that reads forcing input files. CCSM2 can be configured in a number of ways using various combinations of the components. These configurations are defined by letters. For example, B represents the fully-coupled run that consists of the Community Atmosphere Model (CAM2), the Community Land Model (CLM2), the POP ocean model and the CSIM ice model, all communicating through the coupler. This is the default setup for CCSM2. Definitions of all the model configurations should be in the CCSM2 User's Guide.

There are two supported configurations that include the ice model as the only active component along with the data ocean, a data atmosphere, and the data land model. These setups provide more rapid simulations than the fully coupled model and are also useful for testing the ice model. The directory structure of this configuration is shown in bold in Figure 1.

The first configuration uses a data ocean model called docn5 and a model output data atmosphere called datm5; this setup is called the D configuration. datm5 uses output from previous Community Climate Model (CCM) atmospheric simulations, and does not result in the best sea ice simulation. All forcing fields are read in from a single file, with a separate file for each monthly average. Running the ice model with docn5 will not allow ice growth in open-ocean grid cells, since the ocean mixed layer temperature is fixed at or above the freezing point of sea water. Congelation ice growth will occur with docn5.

The second configuration uses a climatological data atmosphere model called latm5 and the ocean mixed layer within the ice model; this is referred to as the M configuration. latm5 is more flexible than datm5 in that it allows for specification of multiple datasets (observed or reanalysis) for atmospheric forcing. It can also read in combinations of datasets with different temporal resolution, for example, 6 hourly air temperature and monthly averged precipitation. Unlike datm5, latm5 reads in each of the forcing fields from a separate file, which makes it easy to swap in different forcing datasets. One year of NCEP data has been provided with this release of CCSM2.

The recommended configuration for testing ice model physics and examining the ice cover response to external forcing is the M configuration, using latm5 on the fine resolution grid with the slab ocean mixed layer within the ice model.

The source code for the active components, the data models and the coupler are in the /models directory. CCSM2 and each component has a /doc directory containing model documentation. The Makefile and the Macros files reside in /models/bld. All configurations of CCSM2 use the same Makefile. The main scripts for running the coupled model and the setup scripts for the individual components are in the /scripts/test.a1 directory.

Figure 1: Directory Tree for CCSM2, with the active ice only (D) configuration shown in bold.
\includegraphics[width=5in]{ccsm2_tree}

1.1.2 Overview of Running CSIM4 within CCSM2

All configurations are run using two levels of C shell setup scripts. A model run is controlled by a main driver script that sets build and environment variables common to all components. The main script invokes the setup scripts for the coupler, atmosphere, ocean, land and ice components. These setup scripts collect the source code, build the executable, access necessary datasets, set individual component parameter values and generate a namelist file for each component. The main script then executes all of the components simultaneously.

1.1.2.1 What do I need to run CCSM2?

Two target architectures are supported: IBM and SGI. The current release will run on Compaq hardware, but will not restart exactly. This will be fixed shortly.

1.1.3 IMPORTANT: What's new in CCSM2.0.1?

Changes made to CCSM2.0 were limited so that the CCSM2.0.1 release would give bit-for-bit answers with CCSM2.0 when using all active models on the T42_gx1v3 grid. This is the configuration that produced the 1000 year control run. No changes have been made to the model physics in CCSM2.0.1.

New to CCSM2.0.1 is a paleoclimate configuration that runs on the T31_gx3v4 (coarse resolution) grid. Compared to the gx3 grid not officially released with CCSM2.0, the land mask in gx3v4 has been changed so that Hudson Bay is now open. It is important to note that the ice and snow albedos in this configuration have been reduced by 10% in the northern hemisphere from the values normally used in the sea ice model. These changes have been included as a seperate set of code modifications and setup scripts that is known by the undescriptive case name test.a2. Unfortunately, the CCSM2.0 control run has a very similar case name of test.a1, so be careful what scripts directory you choose to set up your model simulations.

CCSM2.0.1 also includes new forcing datasets for the data models. datm5 and dlnd5 use new forcing files produced from a coupled run using the latest version of CAM2. A monthly averaged climatology of ice concentration and sea surface temperature is available for use with dice5 and docn5. This ice concentration and sea surface temperature datasets were used to create the datm5 and dlnd5 datasets, so they should be consistent. The sea ice concentration dataset is used by the prescribed ice model.


next up previous contents
Next: 2 Obtaining the Source Up: 1 Introduction Previous: 1 Introduction   Contents
csm@ucar.edu