Title: Coupling NASA Advanced Multi-Scale Modeling and Concurrent Visualization Systems for Improving Predictions of Tropical High-Impact Weather (CAMVis)
Primary Author: Shen, Bo-Wen
Organization: UMCP/ESSIC and NASA/GSFC
Co-Author(s): G. Bryan, W.-K. Tao, C. Henze, S. Cheung, J.-L. F. Li, and P. Mehrotra
The ESTO AIST project, Coupling NASA Advanced Multi-Scale Modeling and Concurrent Visualization Systems for Improving Predictions of Tropical High-Impact Weather (CAMVis), has completed its first phase this year. It seeks to improve Tropical weather prediction by seamlessly integrating NASA technologies (e.g., multiscale modeling, supercomputing and visualization) to create a real-time, high-resolution, weather prediction tool for high-impact tropical weather. CAMVis will be able to inter-compare model simulations with multi-satellite observations (e.g., QuikSCAT winds and TRMM precipitation) and to provide advanced 3D visualizations of atmospheric moist thermodynamic processes and cloud-radiation-aerosol interactions.
Among scenarios in Decadal Survey Missions report (NRC, 2007), ‘Extreme Event Warning’was identified as one of top priority scenarios by the advanced data progressing group (ADP) at the ESTO AIST PI Workshop 2010. This scenario focuses on ‘discovering predictive relationships between meteorological and climatologically events and less obvious precursor conditions from massive data set.’To achieve this goal, our approach is to improve and integrate existing NASA technologies into the CAMVis information system, which will (1) integrate NASA multi- scale model system, including Goddard Cloud Ensemble model (GCE) and the finite-volume General Circulation Model (fvGCM), and the concurrent visualization (CV) system; (2) improve parallel scalability of the multi-scale modeling system to take full advantage of the next-generation peta-scale supercomputers; (3) significantly streamline data flow for fast processing and 3D visualizations. During the first year of the AIST Project CAMVis, we have finished the following tasks: (i) initial coupling of the high-resolution global modeling (fvGCM) and current visualization (CV) systems on the new NASA Pleiades supercomputer; (ii) development of meta-global Goddard Ensemble Model (mg-GCE); (3) development of CV system version 2.0 and implementation of parallel ‘M-on-N’data transfer between ‘M’computing nodes and ‘N’visualization nodes.
In this paper, we will first discuss the current progress of the aforementioned components. Secondly, with the initial coupled system (fvGCM-CV), we will show that CAMVis has been able to improve the insightful understanding of multiscale processes and their multiscale interactions with 3D visualizations, which are key to realistically predicting the formation of tropical cyclones (TCs) such as twin TC and Nargis (2008) and large-scale tropical weather system such as Madden-Julian Oscillations. TC Nargis is one of the 10 deadliest tropical cyclones in recorded history, causing 133,000 fatalities and $10 billion dollars in damage in Burma (Myanmar) in early May 2008. Finally, we will discuss future tasks that would enhance and enable CAMVis to support the following NRC Decadal Survey Earth Science missions, including CLARREO, ACE, PATH, 3D-Winds.