The statistical analysis of global climate change studies

Abstract

The focus of this work is to contribute to the enhancement of the relationship between climatologists and statisticians. The analysis of global change data has been underway for many years by atmospheric scientists. Much of this analysis includes a heavy reliance on statistics and statistical inference. Some specific climatological analyses are presented and the dependence on statistics is documented before the analysis is undertaken. The first problem presented involves the fluctuation-dissipation theorem and its application to global climate models. This problem has a sound theoretical niche in the literature of both climate modeling and physics, but a statistical analysis in which the data is obtained from the model to show graphically the relationship has not been undertaken. It is under this motivation that we present this problem. A second problem concerning the standard errors in estimating global temperatures is purely statistical in nature although very little material exists for sampling on such a frame. This problem not only has climatological and statistical ramifications, but political ones as well. It is planned to use these results in a further analysis of global warming using actual data collected on the earth. In order to simplify the analysis of these problems, the development of a computer program, MISHA, is presented. This interactive program contains many of the routines, functions, graphics, and map projections needed by the climatologist in order to effectively enter the arena of data visualization.