AIMS Geosciences (Jan 2024)
Applications of the scanning test of trend changes in regression coefficients to monthly temperature over China and Globe
Abstract
The algorithm of the scanning t-test of regression slope-coefficients in two phases is introduced to detect trend change-points, along with a coherency analysis of changes between two time series. This new algorithm is different from the previously published scanning Fmax test of trend changes. Meanwhile, the fuzzy weighted moving average (FWMA) was employed to intuitively verify the results of segment regressions. Then, these algorithms were applied to two series of monthly temperature over mainland China and the globe for January 1901–December 2020. The applied results show that significant changes in segment trends may be classified into two gradations on interdecadal and intradecadal scales. The coherency of trend changes between the two series were mostly positive, with a few differences in the change dates. The global warming "hiatus" was detected as two processes on the intradecadel scale: a sharp droop-down from July 1998 to February 2000 following a short warming up; the second weaker droop-down happened from November 2003 to July 2009. Thus, it was featured on the interdecadel scale as the warming rate slowed down to be nearly stable from October 2002 to June 2009 in globally but without turning into cooling. Mainland China seemed to slow down weaker, but lasted longer than the globe. A somewhat unexpected finding is that the warming rate over Mainland China was lower than that for the globe in the case of standardized differences. This contradicts the previous conclusion that resulted from annual anomalies of temperature. It is suggested that the anomalies in the distribution N(0, s) might be referred to the "perceptual" index to compare variations in the same series or between two series but with the same variance and distribution, while referring to the standardized differences in N(0, 1) as a "net" indicator to compare fluctuations between two series with different variances, even in different distributions.
Keywords
