臺灣氣象乾旱特性未來趨勢推估

臺灣地區的年總降雨量雖然豐沛，但由於降雨時空分布的不均，仍時常會發生乾旱事件。本研究使用以統計降尺度或動力降尺度兩種方法產製之共26組模式資料，在代表濃度路徑（Representative Concentration Pathways, RCP）8.5的暖化情境下，討論於21世紀三個時期（初、中、末）的乾旱特性變化趨勢。並將模式資料依照各模式之海溫空間分布分成三類（C1～C3）及整體平均（C0），分別討論不同海溫群集對未來乾旱特性的影響。乾旱分析的方法採用標準化降雨指數（Standard Precipitation Index, SPI）進行乾旱事件的定義與乾旱特性的計算。未來台灣整體而言，短時間尺度的乾旱事件（SPI3）發生時的強度會增強，但其他乾旱特性易受海溫影響，C1的乾旱特性變化趨勢與C3相反；長時間尺度的乾旱事件（SPI12）在世紀初期仍可看到不同海溫群集間有較大的差異，但在世紀中期及末期則呈現較為一致的變化，其乾旱發生頻率傾向於減少，且乾旱發生時的持續時間會縮短。C2的乾旱特性則不論在長、短時間尺度的特性變化在模式間的變動都較大，顯示數值模式對C2海溫群集及其相關的天氣或氣候系統掌握能力較差。進一步將臺灣依行政區分為北、中、南、東四個區域分析，則顯示出臺灣乾旱特性的變化趨勢有區域上的差異，但整體來說，全台短時間尺度之乾旱發生頻率及乾旱強度都會增加，長時間尺度則不一定。台北及東部不管長短時間尺度的乾旱頻率、嚴重度及強度都增加；乾旱延時則在長時間尺度時受海溫分布影響較大。

關鍵字

乾旱特性；標準化降雨指標；未來推估

並列摘要

Droughts occur frequently in Taiwan due to the non-uniform distribution of precipitation in space and time, despite the abundance of total annual precipitation. In this paper, we analyze 26 sets of model projections, which are statistical downscaled or dynamical downscaled, for three future periods (early, middle, and late 21^(st) century) under the Representative Concentration Pathways (RCP) 8.5 scenario to understand the possible changes of drought characteristics. The model projections are divided into three categories (C1~C3) based on their sea surface temperature (SST) pattern and C0 represents the ensemble mean of three SST patterns. The Standard Precipitation Index (SPI) is employed to define drought events and to calculate drought characteristics. Results show that the drought intensity on the short-term timescale (SPI3) will increase in the future, but other drought characteristics are associated with SST. The drought characteristics of C1 tend to have opposite changes to those of C3. On the long-term timescale (SPI12), the drought characteristics show some differences between three SST categories in the early century, but become relatively more consistent in the middle and late century. For example, both drought frequency and drought duration will decrease in the future. Drought characteristics of C2 show less consistency between models, indicating the poor performance of numerical models on C2 SST and the associated weather and climate phenomena. Furthermore, we divide Taiwan into four regions, according to administrative counties, to conduct regional drought characteristics analysis. The results show distinct regional characteristics, representing different precipitation climatology in the four regions. Overall, the drought frequency and intensity increase on short-term timescale in all regions, except the long-term timescale. In the north and east regions, the drought frequency, severity and intensity all increase on both timescales. Drought duration is sensitive to SST patterns on the long-term timescale.

並列關鍵字

Drought Characteristics ； Standard Precipitation Index ； Future projection

國際替代計量

臺灣氣象乾旱特性未來趨勢推估

全文下載

主題瀏覽