Atmospheric Radar Research Seminar Series presents...
(in association with Kwo-Sen Kuo received his B.S degree in atmospheric science from National Taiwan University in 1983, M. S. degree in meteorology from South Dakota School of Mines and Technology in 1987, and Ph.D. in atmospheric science from Colorado State University in 1995. He became an IEEE Computer Society Certified Software Development Professional (CSDP) in 2011. Dr. Kuo is an expert in atmospheric radiative transfer, from which he has expanded to remote sensing, image processing, pattern recognition, and other information science and technologies. His recent research interests include 1) non-spherical particle scattering, primarily in support of algorithm development for NASA’s Global Precipitation Measurement mission, 2) tracking of mesoscale convective systems to quantify precipitation evolution through their life cycles, and 3) information technology (and system) to help scientists focus on science instead of getting distracted by data management issues)
Earth Science Investigations Enabled by High-Performance Computing and Data-Intensive Techniques
Dr. Kwo-Sen Kuo
NASA
05 October 2012, 12:00 PM
National Weather Center, Room 1350
120 David L. Boren Blvd.
University of Oklahoma
Norman, OK
In this presentation, I am going to introduce three on-going Earth science investigations that have benefited from high-performance computing (HPC) and data-intensive technologies/techniques (DIT). The first investigation involves the implementation of a now routine HPC technique, i.e. message passing interface (MPI) with domain decomposition, in an open-source discrete dipole approximation code, DDSCAT, to accommodate targets with much larger size parameters. The second involves the adoption of a tool, SciDB, that utilizes DIT to enable comprehensive and systematic identification of the occurrences of custom-defined Earth science phenomena in reanalysis data. And, the third is a system that combines observations from multiple sources to investigate a diverse set of properties associated with an event in its entirety, i.e. entire life cycle, rather than just the “snapshots” of the event. We hope to eventually integrate the capability of the first case and the system of the third case into the environment of the second case to allow researchers to investigate Earth science phenomena systematically, comprehensively, and, more importantly, interactively, all without burdening them with data and computation management issues.