Temporal characteristics and variability of point rainfall: a statistical and wavelet analysis

Abstract

AbstractRainfall characteristics at different temporal resolutions play a significant role in sustainable urban water management. In this study we attempted to identify the temporal characteristics and variability of point rainfall measured in Melbourne, Australia. Statistical moments, lag1 autocorrelation, the Buishand's Q test for homogeneity, the Mann–Kendall (MK) test for trend and wavelet analyses for temporal variability were carried out for rainfall intensities at resolutions of 0.1, 0.5, 1, 3, 6, 12 h and for the monthly rainfall depths and proportion dry ratios. Series of rainfall intensities at different temporal scales and the monthly rainfall depths and proportion dry ratios were accumulated from the high resolution rainfall dataset for the period from 1925 to 2002. Homogeneity of rainfall intensity was found to increase as the temporal scale increases. Both rainfall intensities and monthly rainfall depths were found to be serially correlated. The 2nd, 3rd and 4th statistical moments of rainfall intensities increased as the resolution increased. While no statistically significant trends were found using the MK test, there were indications that trends are more likely as the temporal scale increases. Wavelet power spectra identified a dominant frequency scale (0.25–1 year) in the 3, 6 and 12 h rainfall intensities that were periodically observed in a 5 to 10–year cyclic order. This phenomenon could be influenced by the inter‐annual variability of the El Niño Southern Oscillation (ENSO). Copyright © 2009 Royal Meteorological Society