台灣地區之區域乾旱頻率分析

Regional Drought Frequency Analysis in Taiwan

蕭政宗(Jenq-Tzong Shiau) ； 楊志傑(Chih-Chieh Yang)

《農業工程學報》 52卷2期 (2006/06) Pp. 83-101

https://doi.org/10.29974/JTAE.200606.0009

摘要

本文研究目的在於探討台灣地區北部、中部、南部及東部區域之區域乾旱頻率，本文主要應用指數洪水法(index flood method)之概念配合線性動差(L-moment)推估參數以建立各區域之乾旱機率分佈並推估不同迴歸期之乾旱量。本文以標準化降雨指數(standardized precipitation index, SPI)定義乾旱，首先選定能代表各雨量站年雨量的機率分佈，而後轉換為SPI值，本文定義SPI小於0為乾旱的開始，並以連續負值SPI的一段時間為一乾旱事件，乾旱事件之累積SPI值稱為乾旱量，本文以乾旱量作為分析區域乾旱頻率的基礎。其次利用以線性動差為基礎的不一致(discordancy)、異質性(heterogeneity)及適合度(goodness-of-fit)估量評估同一區域之資料是否一致及均勻，並選取一合適的區域乾旱機率分佈函數。本文以台灣地區共34個紀錄年限超過30年雨量站的年雨量紀錄，經以前述步驟分析得北、中、南、東各區域最適於代表區域乾旱量的分佈分別為皮爾遜第Ⅲ型分佈(Pearson type Ⅲ distribution)、通用極端值分佈(generalized extreme-value distribution)、通用帕雷托分佈(generalized Pareto distribution)及皮爾遜第Ⅲ型分佈。之後即可建立各區域乾旱量與迴歸期之關係以進行區域乾旱頻率分析並探討各區域歷史乾旱事件之發生頻率。

關鍵字

乾旱 ； 標準化降雨指數 ； 區域頻率分析 ； 指數洪水法 ； 線性動差

並列摘要

This study aims to investigate the regional drought frequency for the north, central, south, and east regions in Taiwan. Index flood method associated with L-moment estimation are employed to establish the regional drought frequency distribution for each region and to estimate the return periods for various drought magnitudes. In this study, standardized precipitation index (SPI) is used to define droughts. Probability distribution for annual rainfall data is determined first, then followed by transforming the rainfall data into SPI. A thought event is defined as the period during which the SPI is below zero. The cumulative SPIs in a drought event is considered as the drought magnitude, which is the thought data used for regional drought frequency analysis in this study. The L-moment based discordancy, heterogeneity, and goodness-of-fit measures are used to detect the unusual sites and select the regional probability distributions of droughts. Thirty-four annual rainfall data with record length exceeding 30 years from different regions in Taiwan are selected to illustrate the proposed methodology. The results show that the best data fitted regional probability distributions for north, central, south, and east regions are Pearson type Ⅲ, generalized extreme-value, generalized Pareto, and Pearson type Ⅲ distributions, respectively. In addition to estimate the drought magnitude of various return periods for each region, the frequencies of severe historical thoughts for each region are also explored.

並列關鍵字

Drought ； Standardized precipitation index ； Regional frequency analysis ； Index flood method ； L-moment