*************************** ** THE GEM MESSENGER ** *************************** Volume 7, Number 36 September 3, 1997 -------------------------------------------------------- THE GEOSPACE GENERAL CIRCULATION MODEL: A STATUS REPORT -------------------------------------------------------- From wolf at alfven.rice.edu Wed Sep 3 07:44:22 1997 R. A. Wolf and M. Hesse GGCM HISTORY AND PHILOSOPHY The idea for the Geospace General Circulation Model (GGCM) originated in the GEM (Geospace Environment Modeling) research program. The report that originally defined GEM (Roederer, 1988) called for the the development of a comprehensive computer model that would embody the most advanced scientific understanding of the dynamics of the Earth's magnetosphere. The GGCM name comes from the analogy to atmospheric General Circulation Models (GCM's). Since 1992, GEM has had either a GGCM Working Group or a GGCM Campaign. The GGCM effort accelerated in summer 1996, when we started working on a specific plan. The plan, as it existed in the summer of 1996, was described by Wolf et al. (1996). The purpose of this message is to provide an updated status report on GGCM planning. The primary aim of the GGCM code is basic research, and specifically on the quantitative testing of physical ideas. It is consistent with the goals of the U. S. National Space Weather Program: construction of a comprehensive, fully physics-based research model of the magnetosphere will obviously help the effort toward an operational space-weather model. The GGCM will have a modular structure. MHD might provide an overall computational structure for a global magnetospheric model, but the magnetosphere clearly deviates from ideal MHD in important ways, for example the following: * Magnetopause transfer processes * Substorms * Many-component inner magnetospheric plasma (Radiation belts, plasmasphere...) * Ionosphere and thermosphere * Exosphere (controls particle loss by charge exchange) * Pitch-angle scattering (controls particle loss by precipitation) * Acceleration of auroral electrons downward (affects ionospheric conductivity, winds...) * Acceleration of ionospheric ions upward (source of magnetospheric particles). The GGCM will have separate computational modules to treat different regions, aspects, and phenomena. It will, in fact, represent the first comprehensive attempt to embed microscopic plasma effects in global magnetospheric code. A computational spine will control the modules and organize them into a comprehensive global model. There are two obvious ways of organizing such a modular code: * Region-modular (different modules control different regions of space, coupled by boundary conditions). * Global MHD spine (modules provide transport coefficients and supplementary information, and, when necessary, override MHD-computed values). The GGCM will be a community computer model. It will be an open code, available for use by any qualified scientist. An observer who is interested in a particular time period or phenomenon will be able to design appropriate model runs. A scientist will be able to design runs to test a new idea, which might require substituting a new module for an old one; global magnetospheric consequences of the idea could thus be computed and tested against observations. Making the GGCM easily accessible to the community will some redirection of effort. Specifically, it will require the following: * A staffed "center"; * Capability for handling a variety of input and output data; * Code streamlining, robustness, and documentation. However, the effort will pay off: easy community accessibility to the GGCM will multiply its scientific output. The GGCM could also be routinely tested in a pseudo-operational setting, with real data used as input, predictions routinely compared with real data, and results made available to the scientific community. Routine and massive testing against data would help separate good and bad ideas. CHANGING IDEAS Although most aspects of the present GGCM philosophy sprang directly from its original conception at the beginning of the GEM program, there have been a few substantial changes over time. In the original plan, the GGCM was to be developed in the final stages of the GEM research program and was to encapsulate the scientific understanding obtained in that program. The plan now visualizes GGCM as a scientific tool that will help achieve the scientific objectives of GEM. GGCM was always visualized as a community model, developed by the magnetospheric community and widely used by that group. It quickly became apparent that easy community accessiblity would require a small "community modeling center," with facilities and personnel to coordinate model development and help people with model use. The present idea, which evolved at the June 1997 GEM Workshop, envisages immediate steps to widen the use of magnetospheric models, including creation of the GGCM Center, which would initially allow a wide range of scientists convenient access to existing magnetospheric models. The full, integrated GGCM would evolve from existing models. Although present magnetospheric models are far from being a full GGCM, which will properly integrate microscopic and macroscopic processes, some quite powerful models currently exist. As now envisaged, the first major steps in the GGCM program will aim to increase use of those models. IMPLEMENTATION PLAN To develop more detailed concepts of how the GGCM Center will operate and to plan possible structures for the GGCM computational spine, NSF is currently funding three concept studies. Principal investigators are Tom Hill (Rice), John Lyon (Dartmouth), and Al Ronn (TRW). The final reports due at the end of 1997 and will be presented to the scientific community at a mini-workshop that will be held December 7, 1997, before the Fall AGU meeting. The GEM Steering Committee has approved the appointment of a GGCM Steering Committee, which will review results of concept studies and make strategic decisions. Members of that committee are George Siscoe (chair), Joachim Birn, Bob Clauer, Tamas Gombosi, Mike Heinemann, Michael Hesse, Janet Luhmann, Gary Olsen, Ray Roble, Ted Rosenberg, Howard Singer, Dick Wolf, and Larry Zanetti. The plan that developed at the June 1997 workshop envisages three phases of implementation. NSF has funded the concept studies and will hopefully fund phase 1. The more expensive later phases may require participation of other agencies. Phase 1, the Preliminary Phase, has the primary goal of making results from pre-computed model runs readily available to the scientific community by means of CD roms and World-Wide-Web sites, with appropriate graphics tools. This step will not require operation of a full GGCM Center, and the required funding will be relatively modest. The time scale will be 1-2 years, starting in FY98. The GGCM and GEM Steering Committees have recommended the submission of NSF proposals pursuant to Phase 1, for the Fall 1997 GEM deadline. (See email message sent to GEM mailing list 8/19/97.) Phase 2, the Transition and Development Phase, has the primary goal of enabling easy access to tailored model outputs. In this phase, existing models will be used for various runs, including event simulations, in response to requests from scientists. The scientific community will have open access to simulation output. The GGCM Center will begin to play an important role in Phase 2, which should start in FY 99 or FY 00 and last about two years. It will allow extensive modeling of the large storms expected near solar maximum. Phase 3. the Final Phase, centers on the running, evaluation, and improvement of models and the development of the full GGCM, and magnetospheric models will be widely and routinely run by many scientists. Model output will be routinely compared to observations. GGCM modules will be developed and incorporated in the global code. The GGCM Center will, in this phase, be playing a central role. SUMMARY Magnetospheric models have traditionally been developed by small groups. Progress has been slow, but powerful capabilities have been developed. In the present plan, the first two phases of GGCM implementation are designed to dramatically increase model use - for data interpretation, planning, and to try out scientific ideas. In the third stage, the community will cooperate in the development of a GGCM that includes effects of small-scale processes on large-scale dynamics, and vice versa. The timing is right to launch this effort now, because it builds on the success of theory/modeling and open data sharing in the ISTP program, because adequate computing power is becoming available, because solar max is coming, and because the need for reliable space-weather forecasts is becoming more and more obvious. References: Roederer, J. G., Editor, GEM Geospace Environment Modeling: A Program of Solar-Terrestrial Research in Global Geosciences, University of Alaska, 1988. Wolf, R. A., J. F. Drake, M. A. Heineman, M. Hesse, J. G. Lyon, N. C. Maynard, G. L. Siscoe, and R. A. Wolf, Plan for a Geospace General Circulation Model. Available at http://jimme.gsfc.nasa.gov/~hesse/GGCM_plan.html or through the GEM home page. +-------------------------------------------------------------------------+ |To add name to the mailing list, send a message to: editor at igpp.ucla.edu | |For message to whole GEM mailing list, send to: gem at igpp.ucla.edu | | | |URL of GEM Home Page: http://igpp.ucla.edu/gem/Welcome.html | |Please update your e-mail address. | |CAUTION: Do not send messages to gem at igpp.ucla.edu unless you want | | your message to go to everyone in the GEM mailing list! | +-------------------------------------------------------------------------+