 |
|
 |
||||||
|
||||||
 |
Home
| Mission
|
about SciDAC
|
Contact Us |
|||||
 |
 |
 |
||||
|
Alumni ProjectDecadal Regional Climate Studies and Applications with Variable-Resolution GCMs Using Advanced Numerical TechniquesM. Fox-Rabinovitz, D. Allen, F. Baer, H. Berbery, K. Pickering, University of Maryland,J. Tribbia, NCAR, G. Stenchikov, Rutgers University SummaryThis study is devoted to research in advanced numerical methods and parallel computer algorithms for variable-resolution stretched-grid (SG) general circulation models (GCM). The models are used for efficient regional climate modeling with enhanced resolution, and for atmospheric chemistry regional experiments driven by the model products. This project represents a trend in the modeling and broader communities to move towards regional and sub-regional climate and climate change assessments and applications important for the U.S. and Canadian public, business and policy decision-makers, as well as for international collaborations on regional climate related issues. This study is done in collaboration with the Canadian co-investigators led by J. Cote and R. Laprise of University of Quebec at Montreal. The finite-difference SG-GCM developed and thoroughly tested over the last few years has been used for a successful simulation of the major anomalous regional climate events of 1997-1998, with 50-km regional resolution. The events are simulated simultaneously due to using the new SG-design with the following multiple (four) areas of interest, one at each global quadrant: the U.S./Northern Mexico, the El-Nino/Brazil area, India-China, and Indian Ocean/Australia. The simulated anomalous regional climate events include: the April-June flooding in the Midwest and Northeast and drought in South of the U.S.; the December-1997 - May-1998 Mexican drought; the Indian summer monsoon; the severe summer flooding in China; and anomalous precipitation over Australia. The high quality of global circulation and consistent interactions between global and regional scales are maintained for the SG-GCM integration when using the SG multiple areas of interest that provides a quite homogeneous global grid-point distribution. A 12-year (1987-1998) limited ensemble simulation of the U.S. climate using the SG-GCM with 60 to 100 km resolution, has been run for the period that includes the recent ENSO (El-Nino Southern Oscillation) cycles. The goal of the experiments is to analyze the long-term SG-GCM ensemble integrations in terms of their potential in reducing the uncertainties of regional climate simulation while producing realistic mesoscales. A special attention is devoted to analyzing the variability of anomalous precipitation (summer floods and droughts) over the U.S. The internal variability of the SG-GCM has been assessed. The ensemble means appear to be closer to the verifying analyses than the individual ensemble members. The ensemble means capture realistic mesoscale patterns, especially those of induced by orography. Two ENSO cycles have been analyzed in terms of their impact on the U.S. climate, especially on anomalous precipitation. The ability of the SG-GCM simulations to produce regional climate anomalies has been confirmed. However, the optimal size of the ensembles depending on fine regional resolution used, is to be determined. The regional climate simulation results obtained with the SG-GCMs show the maturity of the SG-approach. The international SGMIP (Stretched-Grid Model Intercomparison Project) has been initiated by P.I., M. Fox-Rabinovitz and is under way with participation of four major centers/groups from the U.S., Canada, France and Australia, employing the SG-approach for regional climate modeling. SGMIP will be connected in the future to the well-established AMIP (Atmospheric Model Intercomparison Project) as a regional project. SGMIP is also related to the IPCC (Intergovenmental Panel on Climate Change) activities. The SGMIP multi-model ensemble integrations are run with enhanced mesoscale resolution (0.5°) for the 1987-1998 (and beyond) period. The SGMIP products will be used for assessing the recent anomalous U.S. climate events including the ENSO cycles. SGMIP will also allow us to assess the optimal performance on different, the U.S. and Japanese, parallel supercomputers used by the U.S. and Canadian participants. The joint SGMIP effort is focused on a better understanding of the SG-approach. SGMIP is aimed at introduction of the efficient SG-approach to a broader regional and global climate modeling community. The availability of the advanced SciDAC terra-scale supercomputers at ORNL (Oak Ridge National Laboratory) is essential for conducting the investigation. The SGMIP runs could not be accomplished without the access to the facilities. In the next year or two, we plan to conduct research on a thorough analysis of the SGMIP results including those of multi-model ensembles. It will include studying anomalous regional climate events (floods, droughts, etc.) and major monsoonal circulations at mesoscale resolution related to the ENSO cycles, and to studying the U.S. water and energy cycles. The interactions with other teams of SciDAC will be also continuing through our participation in the DOE-sponsored Workshops on “PDE (Partial Differential Equations) on the Sphere”. We will continue our close collaboration with the NCAR as well as with the UQAM (University of Quebec at Montreal) groups. ReferencesFox-Rabinovitz, M.S. et al., 2001: A variable-resolution GCM : Regional climate simulation, MWR, v. 129, No. 3, 453-469. For further information on this subject contact:Dr. Michael Fox-RabinovitzESSIC (Earth System Sciences Interdisciplinary Center) University of Maryland, College Park foxrab@essic.umd.edu
|
||||
Home | ASCR | Contact Us | DOE disclaimer |
|
|
