Severe local storms and computational science: What's next?

This talk will look toward the future of computational science and its role in severe local storm research. During the past 20-25 years, one of the principle barriers toward increasing our understanding of severe local storms was insufficient computer resources to perform simulations either at sufficient resolution to capture a phenomena or to be able to generate experiments that sufficiently span the vast environmental parameter space associated with severe convective events. Advances in computational infrastructure during the next 10-20 years will remove that barrier while at the same time creating new challenges for researchers. Increases in computer power has already enabled storm-scale modeling research to move from simplified and idealized studies of convection toward highly complex and multi-scale simulations that are essentially storm-scale numerical weather prediction. The inherent uncertanties associated with convective modeling along with recent developments in data assimilation methods makes it almost certain that the number of individual simulations will increase by several orders of magnitude, while at the same time individual simulations will continue to increase in size as well. Therefore data sets generated from modeling research experiments are likely to be 10³-10⁴ larger in size than before. Therefore it seems likely that researchers will need to carefully consider their experimental methodologies as well as needing new tools to sift through these vast sets of numbers. The talk will discuss these issues in detail and attempt to present new ideas and technologies to help deal with the oncoming "storm of bytes".