A Statistics Program at the National Center for Atmospheric Research

University Corporation For Atmospheric Res, Boulder CO

Investigators

Abstract

ABSTRACT PI: Douglas Nychka proposal: 0355474 The Geophysical Statistics Project (GSP) at the National Center for Atmospheric Research (NCAR) provides a unique opportunity to address important research problems at the interface between the geosciences and statistics. The statistical research includes contributions to the analysis of nonstationary spatial and spatio-temporal data, the design and analysis of computer experiments, filtering, the statistics of extremes, data mining, and statistical computing. These methods are adapted to large data sets and publicly available software is available to implement new methods for spatial data and extremes. The intellectual contributions to applications in the geosciences is equally broad and includes the introduction of stochastic modeling for sub-grid scale processes, estimating the climatology for extreme events, the construction of spatially coherent weather generators, and statistical improvements to methods of data assimilation. A grand scientific challenge for this century is to understand the complex interrelationships among the atmosphere, ocean, land processes, biosphere and human activities that define the Earth System. Coupled with this effort is the need to predict changes to our environment and to translate such changes into more immediate economic and societal impacts. There is a corresponding challenge for statistical science to tackle the large and complex data sets that are now the norm in many areas of science and engineering and to leverage the rich structure and prior knowledge afforded by traditional numerical models. Through the training of young researchers, a vigorous visitor program and participation within several of the NCAR strategic initiatives, this work reaches a wide community of statisticians and researchers in the geosciences. In this context, some specific problems are the design and analysis of large computer experiments that simulate the Earth's climate, extrapolating historical meteorological data sets to locations where they are not observed, improving the algorithms used to make numerical weather predictions, and identifying models that include random processes to represent complicated geophysical phenomenon. One common element throughout this research is quantifying the uncertainty in a derived method or result. Uncertainty estimates are often difficult to obtain without a statistical treatment but are particularly important when the analysis is subsequently used for decision making or to assess risks.