聯合機率分析方法在都會區淹水之模擬分析

The Simulation and Analysis for Joint Probability Method in Urban Area Inundation

王昱閔(Yu-Ming Wang)

指導教授 ： 張倉榮

國立臺灣大學/生物資源暨農學院/生物環境系統工程學研究所/碩士(2016年)

https://doi.org/10.6342/NTU201603480

摘要

本研究以聯合機率法分析方式，以位於新北市的新莊、樹林區之塔寮坑溪與中、永和地區的瓦磘溝為研究區域，並以聯合機率法為主要理論依據，考慮多項水文條件，每個研究區域各有4個水文條件，其中三個為雨量站與一個為出水口水位，並產生各別之累積機率分佈曲線圖，並以拉丁超立方法對累積機率曲線進行等區間隨機取樣，產生100組輸入集合，搭配二維地表漫地流模式與一維河川與雨水下水道進行交互演算，最後將模擬所得結果與同時間所有水文條件之發生機率皆相同之傳統模擬方法在發生機率介於1%至20%之下比較其異同。
本研究選定兩研究區域中各五個定點之最大淹水深與兩研究區域之淹水面積做為比較與分析所用之參數。最大淹水深部份，在同定點相同發生機率下，傳統模擬方法所得之淹水深皆較聯合機率法所得之深度為大，而發生機率介於1%至4%時之差異尤為明顯。而對於淹水面積部份，傳統模擬方法在兩區域所得之淹水面積也較聯合機率法所得之淹水面積大。因此，在相同發生機率下，傳統模擬方法所得之結果不論在淹水深或淹水面積上皆高於聯合機率法所得之結果。

關鍵字

聯合機率法 ； 拉丁超立方法 ； 淹水模式

並列摘要

This research uses the joint probability method as the analysis treatment, and the study areas are Ta Liao Keng and Wayaogou in New Taipei City. The joint probability method (JPM) is the major treatment, and hydrology conditions are considered at the same time. There are four hydrology conditions in each study area, including three precipitation stations and one outlet water surface elevation. The cumulative density functions (CDF) of these conditions are divided into stratums of equal marginal probability, random sampled with the Latin hypercube sampling (LHS), and 100 input condition sets are produced. After running the simulations with the collaboration model of 2D overland flow, 1D river flow and 1D rainwater sewer, the outcomes with probability of occurrence from 1% to 20% are compared with the outcomes of the traditional simulation treatment that all the input hydrology conditions have the same probability of occurrence.
The maximum flood depth of five chosen locations and the flood area of the study areas are conducted analysis and comparison. Under the same probability of occurrence and location, the maximum-flood depths obtained by the traditional simulation treatment are higher than the ones obtained by JPM, especially when the probability of occurrence is between 1% and 4%. Besides, under the same probability of occurrence the flood areas obtained by the traditional simulation treatment are larger than the ones obtained by JPM in the two study areas. In summary, with the same probability of occurrence, the predictions of the traditional simulation treatment are higher than the predictions of JPM in the aspects of maximum-flood depth and flood area.

並列關鍵字

Joint probability method ； Latin hypercube sampling ； flood modeling

參考文獻

1. Bates, P.D., Horritt, M.S., Aronica, G., Beven, K., 2004, Bayesian updating of flood inundation likelihoods conditioned on flood extent data, Hydrological Processes 18, 3347–3370.

2. Hawkes, P.J., 2005, Use of Joint Probability Methods in Flood Management, Defra-Flood Management Division, London, Vol.43(3).

3. Huber, W. C., Dickinson, R. E., 1988, Storm Water Management Model, User’s Manual Ver. IV, U.S. EPA.

4. Hunter, N.M., Bates, P.D., Matthew, S.H., Wilson, M.D., 2007, Simple spatially-distributed models for predicting flood inundation: A review, Journal of Geomorphology, 90, 208-225.

5. Irish, J.L., Resio, D.T., 2013, Method for estimating future hurricane flood probabilities and associated uncertainty, Journal of Waterway, Port, Coastal, and Ocean Engineering, .139(2), 126-134.