DETECTING THE AUTO-CORRELATION BETWEEN DAILY TEMPERATURE AND RELATIVE HUMIDITY TIME SERIES

摘要

Multifractal detrended fluctuation analysis (MF-DFA) for bivariate series has been used to study the auto-correlation between temperature and relative humidity series in Wuhan city, China. The results show that long-range persistence auto-correlation exists between the temperature and relative humidity series and the auto-correlation has multifractal characteristics. For the two climate records, the contribution of single series to multifractality is analyzed by utilizing chi square (χ2) test. By comparing the chi square test statistics of original series with those of shuffled and surrogate series, we conclude that the relative humidity is more responsible for the multifractality due to its long-range correlation, and the temperature and relative humidity series almost have the same degree of contributions to the multifractality due to a fatness of probability density function (PDF) correlation. On the whole, the relative humidity series has dominant effect in the auto-correlation.