In this paper, we apply extreme value theory, hierarchical model, and MCMC methods to analyze the annual maximum of daily precipitation observed in the middle and lower Yangtze river basin. First, we make a exploratory study of this dataset. Secondly, we apply a hierarchical model to describe the relationship among the annual maximum of daily precipitation, time, and the geographic location. Thirdly, we design the MCMC algorithm for the hierarchical model, and use the software R to complete it. In the course of designing, we discover that updating the GEV’s shift parameter and scale parameter must be done simultaneously in our Metropolis-Hastings algorithm. Finally, we estimate the quantiles of the annual maximum of daily precipitation for each station. |