GEWEX Cloud System Study: Working Group 3
GCSS WG3: Column Models
GCM COLUMN MODELS
Revision: Jan, 10, 1996 - Main New Addition
1. Introduction
Atmospheric column models have been used to study physical processes occurring on a wide range of spatial and temporal scales. Due to their simplicity and efficiency, column models have been commonly used as development tools for building physical parameterizations for use in multi-dimensional models. The greatest limitation of column models is that the physical processes occurring inside the model air column cannot be coupled interactively with its environment. In addition, since the horizontal dynamic tendencies for the prognostic variables cannot be resolved explicitly in column models, they must be prescribed. As such, a physical parmeterization developed within a column model typically need substantial "tuning" when used in conjunction with other model physics in a multi-dimensional model. Some information on GCM column models will be given in this section. In particular, the activities within the working group on the utilization of GCM column models to develop layer cloud parameterizations will be summarized. Please send suggestions and comments to Kit Szeto.
2. The formulation of GCM Column Models
Depending on the intended application of the model, GCM column models of different degrees of complexity can be formulated. A hierarchy of GCM column models will be described in the following with brief descriptions given on their model physics and the forcing that is used to drive them. Prof. Dave Randall (ranall@redfish.atmos.colostate.edu) of Colorado State University has also developed a homepage devoted to the usage of single column models in the development and testing of GCM parameterization schemes. Click here to go to Randall's column model homepage.
a. Simple GCM process column models
main applications:
- parameterization development
model physics:
- part or whole parameterization scheme for isolated physical processes (e.g. stratiform clouds)
forcing:
- (i) driven by the dynamic tendency terms from archived GCM output
- (ii) driven by tendency terms computed from observations
- (iii) driven by the large-scale tendency terms computed from archived high resolution models that contains sophisticated model physics (e.g. CRMs) outputs
- (iv) idealized forcing
b. Simple coupled GCM column models
main applications:
- to study climate sensitivity and feedback mechanisms
- tuning of schemes developed in (a) when they are coupled with selected model physics
model physics:
- parameterization schemes of selected physical processes (e.g. cloud and radiation schemes for cloud-radiative feedback studies, cloud and surface layer schemes for local water cycle studies. etc.)
forcing:
- typically (i) and (iv) of a.
c. Full physic GCM column models
main applications:
- to study climate sensitivity and feedback mechanisms
- tuning of the scheme developed in (a) when coupled to the full model physics package
- long term local climate simulations
model physics:
- full package of GCM model physics
forcing:
- typically (i) of (a)
3. Column modelling activities within WG3
One of the critical components in the GCSS strategy for improving GCM cloud schemes is to utilize GCM column models in conjunction with cloud-resolving model (CRM) results to develop and to improve the parameterizations of different cloud processes. The research activities of WG3 members on this subject will be summarized in this subsection.
a) K. K. Szeto (AES, Canada)
While it is common to drive a column model with dynamic tendency terms calculated from GCMs, the coupling of CRM outputs to column models have been a relatively new approach to GCM parameterization development. Large-scale dynamic tendency terms and the bulk impacts from cloud processes can be readily calculated from the CRM outputs to drive and validate a GCM column model (see for example the LARGE-SCALE DIAGNOSTICS section on the computation of these effects). The CRM results also provide a wealth of useful information of cloud properties and effects on the sub-GCM-grid scales that can be employed to aid cloud scheme development. For example, some of the cloud effects can be included in the CRM forcing to reduce the degrees of freedom in the cloud scheme in the column model during the developmental stage of the scheme. In this preliminary effort, a simple GCM column model based on the Sundqvist (1988, 1993) stratiform cloud scheme is used in conjunction with the CRM simulation results of an Arctic frontal system (Szeto and Stewart, 1995) to test out the various possibilities in coupling CRM outputs to GCM column models. The purposes of this study are (i) to examine the various ways of forcing column models with CRM outputs and (ii) based on the outcomes of the study, suggest to the working group members the CRM output requirements for the purpose of parameterization development with GCM column models. Future works include the collaboration with the Canadian Climate Center GCM group to develop a column model which includes the CCCGCMIII model physics to be forced and validated by CRM outputs (early 1996) as well as the utilization of results from other CRMs in the investigation.
4. Examples of GCM column models
Information of some GCM column models will be given in this subsection. If you have information of a particular GCM column model, please send them to Kit Szeto and he will pose the information here. If possible, please include information on the general availability of the model as well as the contact person.
a)The NCAR CDG CCM2 Column model
Description: (Excerpt from the Annual Report (FY94) for the Climate and Global Dynamics Division (CGD) at NCAR)
With support from the DOE CHAMMP program, Hack, Kiehl, and Michael Hoswell, in collaboration with Bruce Albrecht and William Frank (both of Pennsylvania State University, or PSU), have continued research on a simplified workstation-based, numerical framework for parameterization development. A single-column version of the CCM2 physics package was the central component around which dynamical forcing and analysis components were developed. The target platforms were SUN workstations. Scientists adopted a netCDF data format, along with filters for transparently converting between netCDF and the CCM2 history-tape formats (for which a substantial ECMWF analysis data base currently exists). They developed a sophisticated X-Windows graphical user interface (GUI), which requires integration of the FORTRAN-based CCM2 physics with the C-based GUI. Model namelist parameters are specifiable as command-line data or via X-Defaults. The point-and-click graphical interface streamlines the control of code flow including data-set selection; column-location (latitude/longitude) selection; modification of control variables (such as termination conditions, update frequencies, specification of history data, etc.); modification of initial data and the associated large-scale forcing (e.g., modification of vertical structures, amplitudes, etc.); and the visualization of output data (vertical profiles, time series, etc.). The adoption of a semi-Lagrangian approach to handle vertical advection allows for the accurate solution of vertical-advection terms for an arbitrary number of model variables (including the standard-state variables).
Availability: Yes
Contact person:
b)The UQAM FIZ-C Column Model
J.P. Blanchet and D. Therrien of the University of Quebec at Montreal (UQAM) developed a one dimensional version (column) of the Canadian CCC GCMii. This new model generates data for only one point and can be run on a pc-type computer. The column model is used to study the parameterization of physical processes in GCMii but it could also be used for impact studies. Further more, this Local Climate Model (LCM) can be driven by observed upper air time series. This permits to validate the physical parameterization package of the GCM. Once validated, this model will be used to study available long climatic time series at stations. Applications ranges from climate sensitivity studies, retrieval of unavailable observations/data as deduced from the physics, to drive other systems (ecology, lakes, permafrost, forest,...) for very long integrations.
Availability:
Contact person:
