畢祿溪地區不同紀錄年限24小時降雨延時之設計雨型歷線

Design Hyetographs of 24-hour Duration from Different Rainfall Recording Periods at the Piluchi Area of Central Taiwan

盧惠生(Hui-Sheng Lu)

《中華水土保持學報》 30卷4期 (1999/12) Pp. 289-298

https://doi.org/10.29417/JCSWC.199912_30(4).0006

摘要

設計雨型歷線爲上游集水區治理時，推估設計流量歷線的重要輸入資料之一。本研究乃以台灣中部高山地區畢祿溪氣象站降雨紀錄，分析降雨型態特性，比較不同年限降雨紀錄及方法推求之設計雨型歷線，以建立適宜的設計雨型歷線推估式。獲悉畢祿溪地區24小時降雨延時歷線爲中峰型雨型，最大雨峰時間比率在20.8-95.8％間，最大雨峰降雨比率在6.3-20.5％間。採用Huff法、分配法、位序法及修正位序法等四種方法推求24小時降雨延時之設計雨型歷線，經誤差分析，顯示修正位序法最佳，其設計雨型歷線雨峰時間爲11小時，雨峰降雨比率爲11.51％。同時以修正位序法之設計雨型歷線，建立之指數型經驗推估式，可靠性甚佳。比較不同降雨紀錄年限推求之設計雨型歷線，修正位序法亦比其他三方法的穩定性爲優，至於推求設計雨型歷線所需最少的降雨紀錄年限，修正位序法約需5年以上，而位序法、分配法及Huff法則約需10年以上。

關鍵字

集水區 ； 降雨型態 ； 設計雨型歷線 ； 誤差分析

並列摘要

Design hyetograph was one of the important input data to estimate the design stormflow hydrograph with hydrologic simulation models for upstream watershed planning and management. In this study, analyses of storm pattern characteristics, comparisons of design hyetographs from different methods and rainfall recording periods, and establishment of empirical equations of de sign hyetograph were carried out with rainfall records of Piluchi weather station at the high mountain area of central Taiwan. The results showed that hyetographs of 24-h duration at Piluchi were mostly middle peak type of storm pattern with the ratio of time to rainfall peak as the percentage of duration ranging from 20.8% to 95.8% and rainfall peak ratio ranging from 6.3% to 20.5%. Based on error analyses, the modified ranking method showed better results than Huff's, allocation and ranking methods for solving design hyetograph of 24-h duration in this area. The major characteristics of design hyetograph of 24-h duration were 11-h time to rainfall peak and 11.51% rainfall peak ratio. The empirical equations of exponential type to estimate the design hyetograph of 24-h duration from the data of modified ranking method were also derived and showed good agreement. Meanwhile, the modified ranking method to solve design hyetographs from different rainfall recording periods showed more stability than the other three methods. As for the least needs of rainfall recording periods, the modified ranking method was approximately over five years and the Huff's, allocation and ranking methods were approximately over ten years.

並列關鍵字

watershed ； storm pattern ； design hyetograph ； error analysis