Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/4920
Title: | 結合經驗模態分解法與k-NN移動視窗法之新型態氣象繁衍模式與應用於新竹供水系統之氣候變遷風險評估 Novel Weather Generator Using Empirical Mode Decomposition and K-NN Moving Window Method and Application to Climate Change Risk Assessment of Hsinchu Water Supply System |
Authors: | Yuan-Hung Li 李沅泓 |
Advisor: | 童慶斌(Ching-pin Tung) |
Keyword: | 經驗模態分解法,氣象繁衍模式,k-NN移動視窗法,氣候變遷,水資源, Empirical Mode Decomposition,Weather Generation,k-NN,Climate Change,Water Resource, |
Publication Year : | 2015 |
Degree: | 碩士 |
Abstract: | 氣候變遷受到許多關注,惟人為氣候變遷外,長期低頻率之自然變異也必須納入考量。氣候變遷風險評估所需要之未來氣象資料大都透過氣象繁衍模式產生,如何發展氣象繁衍模式,除了根據氣候情境產生氣象資料外,使其能重現低頻率之氣候循環特性就非常重要。本研究最主要目的在發展能重現低頻率氣候特性之新型氣象繁衍模式與應用於評估氣候變遷下新竹供水系統之風險。
本研究第一部份提出之新型氣象繁衍模式主要包括兩步驟,其一為利用經驗模態分解法EMD(Empirical Mode Decomposition) 分解出的不同的本質模態函數(Intrinsic Mode Functions),在繁衍資料過程,利用本質模態函數之包絡線與相位之關係,重新建構本質模態函數,以組成新的月雨量資料。其二為利用k-NN移動視窗法(k Nearest Neighbor - moving window),將已知的月雨量合理地降尺度至日雨量資料以及繁衍出需要的氣象參數,並進一步利用k-NN的特性建構區域內其他鄰近氣象站的氣象資料。研究方法應用於新竹地區,繁衍出的日、月氣象資料,必須能重現歷史資料的統計特性和時間序列上的特性,並且與Richardson-Type的氣象繁衍模式進行比較。確認具有表現歷史資料特性的能力後,再加入GCM氣候變遷情境,繁衍具有氣候變遷情境特性的氣象資料。 第二部分是評估氣候變遷之風險,以新型氣象繁衍模式產出的新竹地區氣象資料,套用GWLF中之水文模式,模擬新竹系統河川流量,並且進一步應用於新竹地區水資源的系統動力模式分析缺水之風險。水資源系統動力模式結合現有之水利設施,及可考量不同氣候、水文、社會經濟條件,本研究選擇不同評估指標,以及加入根據GCM模擬結果設計之氣候情境,分析未來水資源系統可能面臨的缺水風險。研究結果並且與Richardson-Type的氣象繁衍模式所呈現的新竹地區水資源系統動力模式結果進行比較,比較兩種繁衍模式在氣候風險評估結果之差異。 Climate change has drawn many attentions. However, not only human-induced climate change but also natural low frequency climate variability should be concerned. Weather generator is often used to produce weather data for climate change risk assessment. Thus, a novel weather generator needs to be developed to not only produce weather data based on climate scenarios but also reproduce the low frequency climate characteristics. The main purposes of this study is to develop a novel weather generation to preserve the low-frequency climate characteristics of the rainfall data and apply the generated data to evaluate the climate change risk of the Hsinchu water supply system. The first part of this study is to develop a novel weather generator. The proposed weather generator includes two steps. The first step is to apply Empirical Mode Decomposition (EMD) to decompose the time series of historical monthly rainfall amount into Intrinsic Mode Functions (IMF) and trends. Each IMF represents a generally simple component of the rainfall time series. During generating process, the envelope curve and every single phases of each month derived from IMFs is used to form new ones. Assemble the new IMFs and the trend to generate monthly rainfall series. The second step is to appropriately downscale the generated monthly rainfall series into daily rainfall series. After the monthly rainfall amounts has been produced, the daily rainfall amounts could be allocated from the monthly rainfall by k-NN moving window method. Thus the newly simulated daily rainfall series with long-term and low frequency climate characteristics is produced. Furthermore, weather parameters of nearby stations can also be produced by k-NN method. The statistical characteristics of the generated monthly and daily weather data, such as mean, standard deviation and autoregressive coefficient, must be comparable to the historical ones. If their similarity is conformed, further use of adopting GCM scenarios can be applied by adding statistical properties on the historical data and redo all the steps to produce future weather data. The second part of this study is to evaluate the climate change risk of a water supply system. The weather data generated by the novel weather generator are used as inputs for the GWLF model to simulate stream flows of the Hsinchu area. Then the simulated stream flows are further used in the water supply system dynamics model to analyze the risk of water deficits. The water supply system dynamics model considers current hydraulic facilities, different climate scenarios, hydrological conditions, and social and economic development. This study introduces several evaluation index and climate scenarios to assess future risks of the Hsinchu water supply system. Furthermore, the results are compared with the results using Richardson-Type generated weather data to identify the contribution of the proposed new weather generator. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/4920 |
Fulltext Rights: | 同意授權(全球公開) |
Appears in Collections: | 生物環境系統工程學系 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
ntu-104-1.pdf | 7.68 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.