Time Series Analysis of Annual Temperatures

Publisher Website
Google Scholar
Abstract
The time series of annual mean, maximum, and minimum temperatures at Boulder, Denver, Fort Collins and Pueblo, Colo., were analyzed to determine whether or not they are randomly distributed. Tests were performed by means of autocorrelation function and power spectrum. For series with significant first-order serial correlation coefficient, the adequacy of a first-order Markov process was investigated. The analysis revealed that the series of annual maximum temperatures at Fort Collins, and those of annual minimum temperatures at Fort Collins and Pueblo are purely random (white noise). For the remaining nine series, the values of autocorrelation function r1 are significant, and a first-order Markov model seems adequate.

This publication has 0 references indexed in Scilit: