Oct. 8, 2019 — On July 2, 2019, a total eclipse of the sun was visible across the South Pacific, Chile, and Argentina. While the final model predicting the details of the eclipse was run on the Pleiades supercomputer at NASA’s Advanced Supercomputing division at the Ames Research Center, Predictive Science Inc. researchers used the National Science Foundation’s Extreme Science and Engineering Discovery Environment (XSEDE)’s allocated Comet supercomputer at the San Diego Supercomputer Center and the Stampede2 supercomputer at the Texas Advanced Computing Center for preliminary test runs.
Why are these preliminary predictions important?
With the solar corona modeling work, the researchers are answering fundamental science questions about how the corona is heated and how the supersonic solar wind is accelerated. This understanding will ultimately enable forecasters to make better space weather predictions, which are relevant for high-altitude air travel where adverse space weather can affect conditions at the same elevations that airplanes travel.
“Comet is a great resource for our initial investigations because we can run and analyze test cases very quickly,” said Pete Riley, a computational physicist at Predictive Science. “Without the XSEDE allocation for Stampede2 and Comet, our work on this latest prediction of the solar eclipse would have been considerably more difficult.”
Not only have Predictive Science researchers used supercomputers for their work on solar eclipse predictions, but they have also run calculations ranging from simulations of coronal mass ejections to forecasts of influenza outbreaks, which impacts three to five million people around the world every year.
.Specifically, Riley and his colleagues are working with the Centers for Disease Control and Prevention (CDC) to better predict seasonal influenza virus patterns and spread. As with their space weather studies, the influenza forecasting project is important because of the impact that this disease has on the general population.
“Our predictions have certainly become more accurate through the use of SDSC and XSEDE resources,” said Riley. “Perhaps more importantly, the support we receive from XSEDE staff has been incredible. It’s allowed us to focus on model development and analysis of the results, and not on any difficulties that might arise from using these complex architectures.”
More about Predictive Science’s recent solar eclipse work can be viewed here. XSEDE is supported by NSF Grant Number ACI‐1548562.
Source: The Extreme Science and Engineering Discovery Environment (XSEDE)