A CLIMATE DEPARTURE INDEX FOR THE STUDY OF CLIMATIC VARIABILITY

Abstract

Three versions of a Climate Departure Index (CDI) are presented for studying how “normal” or “unusual” a particular year or event is compared to the long-term average for the region under consideration. Comparisons of a Simple CDI, Absolute Value CDI and Least-Squares CDI are made through the use of hypothetical examples and two case studies involving seasonal snowfall variations in northern New England and last spring-freeze date variations in the southeastern United States. Results clearly show that the Simple CDI is the inferior formulation owing to a compensation problem whereby above and below average sites within a region for a particular year cancel each other when computing the index value. Little difference in identifying extreme years was found between use of the Absolute Value CDI and Least-Squares CDI in the case studies examined. Nevertheless, a hypothetical example suggests that the least-squares approach for closeness of fit is the more appropriate method, thus making the Least-Squares CDI the preferred version.