Background Science Applications Physics Climate Groundwater Fusion Energy Life Sciences Materials & Chemistry SciDAC Institutes Enabling Technologies Applied Mathematics Computer Science Visualization & Data Mgt. SciDAC Outreach Participating Orgs Grant Solicitations FY2007 FY2006 FY2005 FY2004 FY2001 Collateral Materials SciDAC Review magazine '06 Progress Report (pdf) Publications 2001-5 (pdf) Logos Search Collateral (from SciDAC 1) Browse by Affiliation Browse by Author Browse by Title DOE Office of Advanced Scientific Computing Research (ASCR) Program Managers only

Variable-Resolution Models for Regional Climate Studies

Michael S. Fox-Rabinovitz, ESSIC (Earth System Sciences Interdisciplinary Center),
University of Maryland, College Park
Ferdinand Baer, Dale Allen, Ernesto Hugo Berbery, Kenneth Pickering,
Department of Meteorology, University of Maryland, College Park
Joseph J. Tribbia, NCAR, Boulder, CO
Georgiy Stenchikov, Rutgers University of NJ

Summary

The project is devoted to development of variable-resolution stretched-grid GCMs (General Circulation Models) using advanced numerical techniques and parallel-processing terra-scale SciDAC supercomputers. The stretched-grid GCMs produce accurate and cost-efficient regional climate simulations with enhanced regional resolution. The advantage of the stretched grid approach is that it allows us to preserve the high quality of global circulation and provide consistent interactions between global and regional scales and phenomena. Using the new stretched grid design with multiple areas of interest with enhanced resolution allows us to simulate simultaneously different regional climates and major monsoonal circulations over North America, South America, South and Southeastern Asia, and Indonesia-Australia. It makes possible to study their long-term connections including those of related to the El Niño-La Niña events, and other major oscillations. The international SGMIP (Stretched-Grid Model Intercomparison Project) has been initiated and is being conducted as a part of these research activities. Our SGMIP simulations have been produced using SciDAC terra-scale supercomputers. The results of the successful 12-year SGMIP multi-model ensemble simulations of the U.S. climate are available at the specially developed SGMIP web site ( http://essic.umd.edu/~foxrab/sgmip.html ) to the climate modeling and broader communities including the U.S. and international climate programs. We are collaborating with two other SciDAC groups led by F. Baer and J. Tribbia (UMD-NCAR) and by J. Cote (Canadian Met. Service/RPN).

This research is devoted to:

(a) developing the variable-resolution GCMs using advanced numerical techniques, and (b) conducting climate studies on regional-to-global scale anomalous climate events, in a broader context of climate variability and predictability. The diversified impacts of enhanced horizontal regional resolution (at least 0.5º or finer) over the U.S. and other multiple areas of interest located within all global quadrants, are thoroughly investigated. The potential of using the ensemble integration mode is being explored.

The following major issues are addressed in this study using variable resolution GCMs: (a) the impact of enhanced resolution on producing consistent global and regional anomalies at meso- and larger scales, with an emphasis on El Niño-La Niña related anomalies and their impact on floods, droughts, and monsoonal circulations, when resolving mesoscales; (b) the possibility of providing efficient downscaling capabilities when using the stretched-grid approach with consistent interactions of meso- and larger scales; (c) improved understanding and modeling of the processes that affect climate variability and predictability at broad-range temporal and spatial scales; (d) the possibility of reducing uncertainties of global and regional climate simulations using the multi-model ensemble integration mode; (e) atmospheric chemistry impacts.

The companion study with the members of another SciDAC group, F. Baer and J. Tribbia, is devoted to developing a stretched-grid GCM using the advanced spectral-element technique modified for variable resolution, and the NCAR CAM physics. Collaboration with J. Cote of CMS/RPN and his group is the integral part of our effort.

One of the focal points of our effort is the international SGMIP (Stretched-Grid Model Intercomparison Project) that has been designed, initiated and conducted under these research activities. SGMIP is aimed at studying the new global variable-resolution/stretched-grid approach to regional climate modeling.

The key SGMIP scientific and computational objectives are: stretching strategies; approximation of model dynamics; treatment of model physics; using the multi-model ensemble mode; optimal performance on parallel supercomputers; efficient downscaling to realistic mesoscales; and studying anomalous regional climate events.

Regional climate simulations obtained with the state-of-the-art variable-resolution GCMs developed in the U.S., Australia, Canada, and France, have been produced for SGMIP. At the current stage, the maximum enhanced variable resolution allowed by all the SGMIP participating model designs/structures and/or computational requirements, is 0.5º over the major part of North America including the U.S. The 12-year (1987-1998) simulated data/products obtained with SGMIP variable resolution GCMs are being processed and analyzed, with the emphasis on the U.S. climate anomalies.

SGMIP provides multi-model ensemble simulation results of the higher quality than that of individual ensemble members. The strong coordinated international SGMIP effort puts us in a favorable position for a comprehensive investigation of the diversified impacts on regional climate simulations due to enhanced regional resolution, including the multi-model ensemble results. We are establishing a connection with the AMIP group on introducing the SGMIP data and their analysis as a regional U.S. climate project. Our initiation of and participation in SGMIP was possible only through using the SciDAC terra-scale supercomputers at ORNL where all our simulations have been produced. (Note that foreign SGMIP participants used their own supercomputers.)

Maintaining the developed and evolving SGMIP web site allows us to: (a) disseminate of SGMIP data/products and analysis results to the climate modeling community; (b) provide consultations on demand to potential users of SGMIP data; (c) make SGMIP data/products and analysis results available on demand to national and international programs and groups such as AMIP, IPCC, CLIVAR, WMO/WGNE.

The SGMIP international effort reflects a trend in climate modeling and broader communities to move towards more detailed regional climate assessments important for the U.S. public, business and policy decision-makers, and for productive international collaborations on climate-related issues.

For further information on this subject contact:
Dr. Michael S. Fox-Rabinovitz, ESSIC,
University of Maryland,
College Park, MD
Phone: 301-405-0050
E-mail: foxrab@essic.umd.edu

back to project page