CR 66:91-101 (2015)  -  DOI: https://doi.org/10.3354/cr01336

ABSTRACT: Macau has continuous meteorological observations for more than 100 yr, data which are rarely available in South China. This study first homogenizes long-term observations in Macau using procedures for change point detection and adjustment, and then applies the statistical theory of extreme values to investigate changes in meteorological extreme events, including extreme rainfall and summer heat waves, at this station. A Poisson-generalized Pareto (GP) model is applied to extreme rainfall modeling, while a statistical model established by extending the Poisson-GP method is applied to heat wave modeling. In this heat wave model, the frequency of summer heat waves is modeled by a Poisson distribution, their intensity is modeled by a GP distribution, and their duration is modeled by a geometric distribution. Results show that these statistical models permit realistic modeling of extreme events in Macau. Return levels of extreme rainfall and heat waves in Macau are estimated by the Poisson-GP model fitted to the daily rainfall amount and summer daily maximum temperatures, respectively. Trends of extreme events are introduced into these statistical models through changes in parameters. It is found that the frequency of extreme rainfall increases significantly in the observational record of Macau, while the positive trend in the intensity of extreme rainfall is nonsignificant. For changes in parameters in summer heat waves, a significant increasing trend in the frequency is found after an adjustment is applied to a detected change point in the temperature record, but the positive trends in the intensity and duration are nonsignificant.

KEY WORDS: Extreme rainfall · Heat wave · Theory of extreme values