Exploring Common Themes in Terrestrial and Space Weather Computational Modeling and Data Analysis

The Earth system sciences attempt to study our planet as an integrated system. This integrative approach is at the nexus of several critical scientific problems important to society, such as climate change and understanding the Sun-Earth connection. There are many commonalities underlying the study of these systems through computer modeling. Their evolution is shaped by nonlinear fluid dynamics, which in turn gives rise to emergent structures on multiple spatial and temporal scales; multiple components within the system generally interact to further modify its collective behavior; and finally, our interest in modeling the system is driven by a desire to understand and ultimately predict its future behavior for the benefit of human society.

The rapid advancement of high-performance cyberinfrastructure over the past half-century has dramatically advanced our understanding of complex interacting systems. In practice, modeling of these systems is a continual rebalancing act in which the resolution chosen, the number of physical processes represented, the number of ensemble members produced, and the spatial and temporal domain selected, are all adjusted to match the science objectives to the available computational and data storage resources.

Looking to the future, challenges abound for the integrated systems approach. The increasing parallelism and complexity of each successive generation of computer, combined with the growing complexity of models and data assimilation systems, will present continuing scalability and complexity challenges to application developers. Beyond the ever-expanding demand for computing cycles in this field, there is also a need for massive, persistent, and federated data storage systems; parallel data analysis capabilities; and new technologies that enable large and often heterogeneous data sources to be integrated in order to facilitate scientific discovery. This presentation will discuss experiences at NCAR in pushing the boundaries of what is possible in each of these areas.