Magnetic storms result when the magnetized plasma that propagates away from the sun interacts with the near-earth space plasma environment. Such storms typically begin with a sudden worldwide increase in the magnetic field that lasts several minutes to several hours, followed by the main phase, which typically lasts about a day, and features large reductions in the magnetic field strength. The global influence of magnetic storms was recognized over eighty years ago, and the currents that they induce in earth and space-borne systems can have considerable impacts. For example, damage from the March 1989 storm was estimated in the order of billions of dollars. Given that the magnetic field can vary significantly within the continental scale, and that its rate of change causes damage, there is a need to model and predict local variability. This project will lay the mathematical foundation for a localized magnetic storm prediction system, which will allow assessment of geomagnetically induced currents and implementation of damage limitation strategies where necessary. Information on fluctuations in the magnetic field can be gleaned from ground and satellite based magnetometers. Although individual magnetometers provide a wealth of temporal information, the paucity of ground-based magnetometers limits our predictive capability of the local behavior of magnetic storms. New mathematical techniques are required. To address the above issue, we propose to solve the following geophysical problems: (i) modeling and prediction of magnetometer time series; (ii) spatiotemporal modeling of the magnetic field; (iii) spatial interpolation of the magnetic field at unobserved locations. We will concentrate our study on the available high acquisition-rate magnetometers in North America. Although these stations offer limited spatial coverage, they provide an abundance of temporal data. The methodologies employed in this project go beyond existing approaches and take into account the very high time resolution of magnetometer measurements. This will compensate for the sparsity of spatial measurements and yield information at unobserved locations. The spatiotemporal mathematical and statistical modeling techniques presented in this proposal are novel and innovative. As such, this proposal will advance the state of the art in both geosciences and mathematical sciences. Members of the project team, which is composed of geoscientists and mathematical scientists, possess the expertise to achieve these aims in a truly interdisciplinary fashion. By providing a physical framework for the spatiotemporal variability of magnetic fields, the proposed work will allow for deeper insight into the underlying physical mechanisms of magnetic storms. Many monitoring systems perform spatially sparse measurements. The proposed work, which addresses the spatial interpolation problem, will therefore find applications in diverse areas of the geosciences and broader scientific community. The most developed societies, which rely upon high technologies for daily essentials such as energy and communications, are most susceptible to the effects of magnetic storms. As society becomes increasingly dependent upon high-technology earth- and space-borne systems, the need to model and predict spatiotemporal fluctuations in the magnetosphere will become increasingly important.
