請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/71047
標題: | 氣象乾旱指數在台灣的比較及其應用 Application and comparison of meteorological drought indices in Taiwan |
作者: | Chin-Hui Lee 李晉輝 |
指導教授: | 陳維婷(Wei-Ting Chen) |
關鍵字: | sc-PDSI,SPI,SPEI,氣象乾旱指數,溫度, sc-PDSI,SPI,SPEI,meteorological drought index,temperature, |
出版年 : | 2018 |
學位: | 碩士 |
摘要: | 本研究介紹一個常用的氣象乾旱指數,可用於分析台灣的氣象乾旱:sc-PDSI (self-calibrating Palmer Drought Severity Index);並且透過比較三個氣象乾旱指數,探討台灣地區在做乾旱分析時,是否需要使用考慮溫度的乾旱指數:sc-PDSI、SPI (Standardized Precipitation Index)、SPEI (Standardized Precipitation Evapotranspiration Index)。為了計算三個乾旱指數,我們使用臺灣氣候變遷推估資訊與調適知識平台的台灣地區高解析度(5公里*5公里) 的降雨和溫度資料。我們使用了1960至2012年以測站為主的資料以及從2006至2100年為未來推估資料。未來推估資料選擇 CESM1-CAM5的RCP 8.5 情境,以統計降尺度方式得到的高解析度資料。另外,我們也使用1960至2009年間的歷史旱災紀錄,此組紀錄資料用於研究台灣地區的氣象乾旱和歷史旱災的相關性。
我們利用兩個命中率來評估氣象乾旱指數在台灣的表現 (hit ratio_dryness和hit ratio_dry)。第一個命中率 (hit ratio_dryness) 表示在旱災月中有多少為氣象乾旱。此命中率越高,表示在歷史旱災紀錄中,氣象乾旱指數可以呈現越多的氣象乾旱。研究結果顯示 sc-PDSI、 SPI-n、SPEI-n (n = 3, 6) 比SPI-1和SPEI-1 有較高的hit ratio_dryness。顯示台灣的旱災的發生主要是短期至中期的缺水情況導致。另外,sc-PDSI也有高的hit ratio_dryness ,顯示sc-PDSI所指出的台灣的氣象乾旱的記憶效應至少為一個中期以上的缺水情形引發的。第二個命中率 (hit ratio_dry) 表示在氣象乾旱中,有多少為旱災事件。這個命中率整體是偏低的情況。表示氣象乾旱的次數遠比旱災事件的次數要多。 我們使用兩個方法分析溫度的改變效應,在過去變遷或是未來推估的氣候中的影響:分析乾旱指數的長期趨勢、相關係數的改變。研究結果顯示,不論是在過去變遷或是未來推估的資料,降雨在台灣地區仍是決定是否有乾旱產生的一個因子。過去資料顯示,台灣降雨趨勢比潛在蒸發散量的趨勢有更顯著的增加,因此三個乾旱指數都呈現台灣地區會有長期變溼的趨勢,也表示在過去,潛在蒸發散量對乾旱的發生的影響小於降雨的影響。然而,未來推估的資料 (CESM1-CAM5, RCP 8.5)顯示,在台灣地區做乾旱分析時,由於大氣對水的需求會隨著溫度的上升而明顯的增加,因此使用有考慮溫度的乾旱指數比較合適。相比於過去氣候,未來氣候的潛在蒸發散量趨勢有顯著的增加。因此兩個有包含溫度的乾旱指數(SPEI, sc-PDSI)顯示台灣會有變乾的趨勢,但是因為SPI並未考慮到溫度,因此仍然呈現台灣會變溼的趨勢。 The purpose of this study is to introduce a widely used meteorological drought index for drought analyses in Taiwan: the self-calibrating Palmer Drought Severity Index (sc-PDSI) and to evaluate whether temperature should be considered for drought analyses in Taiwan via comparing three meteorological drought indices: the sc-PDSI, the Standardized Precipitation Index (SPI), and the Standardized Precipitation Evapotranspiration Index (SPEI). To calculate three meteorological drought indices, we use the precipitation and temperature from observation-based dataset of Taiwan Climate Change Projection and Information Platform Project (TCCIP), with a high spatial resolution of 5 km by 5 km from 1960 to 2012, and the statistically downscaled simulation from CESM1-CAM5 from 2006 to 2100 under RCP 8.5 scenario in Coupled Model Intercomparison Project Phase 5 (CMIP5). Besides, we use historical drought event survey data from 1960 to 2009 to compare the relationship between drought events and three meteorological drought indices. We use two hit ratios (hit ratio_dryness, hit ratio_dry) to evaluate the performance of a drought index against historical drought events. First, a higher ratio_dryness means that a drought index can reflect more meteorological dryness in drought events. The results show that hit ratio_dryness of drought indices (sc-PDSI, SPI-n. and SPEI-n, n = 3, 6) is higher than that of the SPI-1 and the SPEI-1. The higher hit ratio_dryness suggests that the occurrence of historical drought events results from the short- to middle-period water shortage. Besides, the higher hit ratio_dryness of the sc-PDSI indicates that the memory effect of meteorological droughts in Taiwan indicated by the sc-PDSI is at least a middle-term period. Second, the hit ratio_dry implies that how much meteorological dryness is drought events. The lower hit ratio_dry of drought indices implies that the number of meteorological droughts is much more than that of drought events. We evaluate the effect of the change of temperature on drought analyses in Taiwan in both the past climate and the future projection with two methods: analysis of index trend and the change of correlation coefficients. The results show that precipitation is an essential factor in determining the occurrences of drought events in Taiwan in both the past climate and the future projection. Because the increased precipitation trend is more significant than the PET trend in the past climate, the trends from three meteorological drought indices show that Taiwan would be in a wetter condition, suggesting that PET has minor effects on the occurrence of droughts in Taiwan in the past climate. However, the temperature should be considered for drought analyses in Taiwan in CESM1-CAM5 under RCP 8.5 scenario to reflect increasing atmospheric water demands. The trend of PET increases more significantly in the future projection than that in the past climate; thus, the SPEI and the sc-PDSI is reporting a long-term drying trend in the future projection, while the SPI shows a long-term wetting trend because of the increased precipitation trend. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/71047 |
DOI: | 10.6342/NTU201802224 |
全文授權: | 有償授權 |
顯示於系所單位: | 大氣科學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-107-1.pdf 目前未授權公開取用 | 21.68 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。