Last update:4/12/05

PIs: Michael Stein, Dmitry Beletsky, Rao Kotamarthi, Barry Lesht, Noboru Nakamura,
Dave Schwab, Jonathan Stroud

This project is actively developing statistical approaches to problems in which both monitoring data and output from a physical model are available to assess the state of the physical environment. This work can be organized into a number of subprojects covering a broad range of environmental applications including air pollution monitoring, adjustment of emissions inventories, sediment transport in Lake Michigan, model evaluation and model comparison.

The development of statistical models and methods for spatial-temporal processes is central to much of this work and is perhaps the area of statistics most in need of advancement in applications of statistics to air and water pollution. To this end, we have been attacking this area from theoretical and practical perspectives, with each perspective challenging and supporting the other. One particularly challenging problem that arises in many of the subprojects is the development of statistical models for the errors made by deterministic numerical models, which is of great importance to describing and understanding the ways that models go wrong and is a critical component to developing effective data assimilation schemes for pollution models. Another major focus is to develop classes of statistical models and the corresponding inferential and diagnostic tools for processes with complex spatial-temporal dynamics, especially on large or global spatial scalses, for which it is necessary to take account of the spherical geometry of the Earth's surface.