Generation of Rainfall and Temperature Data by Using HHT

碩士 === 國立臺灣海洋大學 === 河海工程學系 === 102 === In this study, four meteorological stations (Taipei, Taichung, Tainan and Taitung) are chosen, meanwhile, the monthly temperature and rainfall data over 50 and 100 years respectively are adopted. Through the Hilbert-Huang Transform, meteorological data would be...

Full description

Bibliographic Details
Main Authors: Chang, Kuang-Ching, 張光謦
Other Authors: Huang, Wen-Cheng
Format: Others
Language:zh-TW
Published: 2014
Online Access:http://ndltd.ncl.edu.tw/handle/zrb8p2
Description
Summary:碩士 === 國立臺灣海洋大學 === 河海工程學系 === 102 === In this study, four meteorological stations (Taipei, Taichung, Tainan and Taitung) are chosen, meanwhile, the monthly temperature and rainfall data over 50 and 100 years respectively are adopted. Through the Hilbert-Huang Transform, meteorological data would be decomposed by Empirical Mode Decomposition into Intrinsic Mode Functions (IMFs) and trend. For generating the meteorological projection during 2014-2033, we would just combine the IMFs with projection of trend. Here the Kolmogorov-Smirnov Test is adopted for examining the distribution equality of Marginal Spectrums of different groups in a specific duration. Next, the ability of data generation is evaluated under a given trend. Finally, compose the IMFs with projection of the trend to generate the meteorological projection, and test the difference of means of the present and future. As a result, the distribution equality is approved in frequency domain. Moreover, for temperature, the difference between the projection and present is insignificant in all stations. For rainfall, the difference between the projection and present is significant in Taipei and Taichung station; meanwhile, the temperature at all stations would be insignificant from the present situation.