| Flood disaster caused by heavy rainstorm is the most serious natural disaster that threatens people's life and property.China's urbanization rate has gradually increased during the past years,especially in eastern China,which is highly urbanized.The urban system has exposures vulnerability in the event of heavy rainstorm and flood disaster,which leads to more serious flooding.Therefore,analyzing frequent occurrence rainstorm(usually referred to events with less than 20 years return period)in Xiamen and Hong Kong is necessary for further study to the whole East China area.This article mainly introduced the rainfall frequency analysis by means of the advanced Regional L-moments analysis in Xiamen and Hong Kong,respectively,focused on comparison of two sampling methods,AMS and AES,with respect to quantiles accuracy especially for the more frequently occurring events..Then to verify the applicability of the formula of Ven Te Chow,which explains the mathematical relation between the AMS and AES in return period.Furthermore,it is indicated that the 1-year event can be rationally estimated by utilizing the AMS data at an exceedance probability of 0.6321.The results of the Master's paper include the following:(1)Using the most advanced method of Regional L-moments via hydrometeorological approach for data available at four rainfall stations in Xiamen and 60 rainfall stations in Hong Kong for five durations(1h,3h,6h,12h and24h).By elaborately analyzing,the AMS data at the stations used in Xiamen and Hong Kong can be justified as a homogeneous region,respectively,for each of the five durations.The best fitted distribution for different durations in Xiamen are respectively GPA,PE3,GNO,GEV and GLO.The best fitted distribution for the five durations in Hong Kong are GNO,PE3,PE3,GNO and GEV.Estimated precipitation quantiles from each distribution based on AMS and AES samples in the two study areas have been checked through the temporal consistency adjustment test,respectively.(2)Analyzing the effects of AMS and AES on the precipitation quantiles has been undertaken from both theoretical analysis and practical data verification,respectively.Theoretically,the frequent occurrence rainstorm quantiles is smaller by using AMS sample than by using the AES sample.The calculation results of the actual data are consistent with the theoretical analysis results,except for some individual time periods.Compare the normalized result from AMS and AES sample,it finds the difference between the two was gradually decreased with the increase of the return period from 2-y to 100-y.Generally,the results by the real AMS and AES rainfall data verification are agree to the theoretical analysis.(3)The Chow V.T.equation is justified by comparison of the quantiles from the best fitted distribution based on both AMS and AES samples.It is pointed out that the nonexceedance probabilities used in calculation for return period of 2-y,5-y,10-y,25-y,50-y and 100-y events should be adjusted to 0.6065?0.8187?0.9048?0.9608?0.9802 and 0.9901,respectively,when the AMS data are used.(4)In this paper,a suggestion to rationally estimate the 1-year event is given that an nonexceedance probability of 0.3679 should be used in calculation when the AMS data are utilized.Hence,the precipitation quantiles for 1-y event for 5 durations in Xiamen and Hong Kong have been estimated based on the AMS sample data accordantly. |
PDF Full Text Request