Statistical Modeling of the Temperatures in the Northern Hemisphere Using Fractional Integration Techniques

Abstract

The temperatures in the Northern Hemisphere from 1854 to 1999 have been analyzed in this article by means of a testing procedure that permits one to consider fractional degrees of integration. The tests are valid under general forms of serial correlation and deterministic trends and do not require estimation of the fractional differencing parameter. The results show that the series follows a fractionally integrated process with the order of integration higher than zero and thus implying long memory behavior. The series was decomposed into four different subsamples, and it was observed that the degree of dependence between the observations substantially increased during the twentieth century.