The probability distribution applicability of short-time strong rainfall for 18 cities in the Pearl River Delta were contrastively analyzed by Using GPD

GEV and Pearson-Ⅲ type probability distribution model

respectively. The main conclusions were reached in the following: ① The results of goodness of fit test using Anderson-Darling test

probability plot correlation coefficient (PPCC)

Root Mean Square Error (RMSE) and Q statistic reflected that the GPD method is more suitable for the design quantile calculation of the extreme hydrological and meteorological events for the shorter fixed number of year of the observation sites; ② The shape parameters of the GDP models for the cities of Huadu

Guangzhou

Xinhui

Enping

Shunde

Zhongshan

Zhuhai and Shenzhen show that the occurrence probabilities of short-time strong rainfall are high

and the calculated the design precipitation intensity larger than that with GEV and P-III type models; ③ Referring to “near shortterm rainfall intensity hierarchies”standard

in the Pearl River Delta

the rainfall intensity in a period of 2 years can reach the torrential rain level

which is the main influence factor for urban waterlogging events.