A COMPOSITE SPACE TIME APPROACH TO STUDYING OZONE DISTRIBUTION OVER EASTERN UNITED-STATES/

This work is concerned with the composite space/time analysis of ozone concentrations over Eastern U.S. A novel method is used, which introd uces a mode of reasoning that is a fundamental combination of inductiv e and deductive processes. The method is based on a spatiotemporal ran dom field that organizes information concerning ozone distribution by reference to a space/time continuum. Randomness manifests itself as an ensemble of realizations (possibilities, potentialities) regarding th e ozone distribution. Random field representations can take several fo rms which can embody physical characteristics of the ozone distributio n or transform the data into a form that has certain desirable feature s. Numerical applications of the composite space/time method show that it generates ozone estimates that are more accurate than the estimate s obtained from purely spatial mapping methods; and it also leads to a technique of space/time ozone trend determination that performs bette r than mean filtering. Composite space/time ozone maps offer a means o f gaining understanding and insight into the variation of ozone concen trations over the Eastern U.S. and are important tools in ascertaining compliance with ambient air quality standards. Theses maps are also t he necessary inputs to ozone reliability studies and to ozone exposure -health damage analysis. (C) 1998 Elsevier Science Ltd. All rights res erved.