Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/33908
標題: Development and Application of Grid Rational Algorithm for Predicting Hydrograph (GRAPH) Model
網格式合理化法(GRAPH)應用於集水區流量歷線推估之研究
作者: Chou, Wen-Chieh
周文杰
關鍵字: 網格式合理化法 (GRAPH);Grid Rational Algorithm for Predicting Hydrograph;HEC-GeoHMS;HEC-HMS;數值高程模組(DEM);常態化差異植生指標(NDVI);無因次單位歷線;運動波法;HEC-GeoHMS;HEC-HMS;Digital elevation model;Normalized difference vegetation index;Dimensionless unit hydrograph;Kinematic wave model
出版社: 水土保持學系
摘要: 
Among different empirical methods, rational formula is the most frequently used equation of calculating peak discharge for small watersheds. More recently, rational method has been used to develop a particular hydrograph shape. Furthermore, improving hydrograph simulation for rational method to compare some models widely used like the unit hydrograph concept and kinematic wave model can also extend the view and utilization of rational formula.
According to Soil and Water Conservation Technical Regulations published by the Taiwan's Council of Agriculture, peak flow estimation in a small-ungauged watershed can be calculated by rational formula. A proper rainfall-runoff process and real rational calculation for domestic engineers to design hydraulic structures and manage floodplains is critical in hazard mitigation strategies on the Taiwan Island. A computer program named Grid Rational Algorithm for Predicting Hydrograph (GRAPH) was developed in this study to delineate the watershed information automatically from digital elevation model (DEM). The model developed in this study proposed a raster-based, spatially distributed, hydrologic routing technique so that compatibility with most modern Geographic Information Systems is assured. Results derived in this study can be expressed briefly as follows.
Model establishment
The α value was proposed to modify the time of concentration; the β value was used to adjust the peak flow discharge. Normalized difference vegetation index (NDVI) was used to obtain raster runoff coefficient from satellite imagery. The runoff coefficient (C) can be expressed as linear reverse transformed of NDVI: C=((1-NDVI)/2). Time-varying runoff coefficient was presented to express the dynamic model of infiltration capacity. The recession model was established successfully by equation: , where is the preceding discharge, t is the hourly time after peak, and are watershed coefficients. In general, the GRAPH represented a highly coincident at small headwater watershed.
Model comparison
Data from Studies on The Application of Computerized Hydrologic Files in Taiwan was used to run dimensionless unit hydrograph method. The HEC-HMS (Hydrologic Modeling System) and HEC-GeoHMS developed by Hydrologic Engineering Center, US Army Corps of Engineers were used to manipulate the kinematic wave model. Chia-Tso-Laio Creek watershed was employed to simulate storm events for comparing real data obtained from The 4th River Management Bureau. GRAPH reached the highest accuracy for predicting peak time occurrence and the second highest model coefficient efficiency in three methods.
Flood simulation
Using isochrones of a watershed, time-area technique is evidenced to be applicable up to larger watersheds; so the study area is enlarged to whole Chou-Shui-Hsi River watershed. It showed a strong exponential reverse relationship between average α value and watershed area in an excellent convergent behavior. A larger area is simulated; the lower α value will be obtained during the same storm. The disturbed downstream watershed may decrease the time of concentration. The obtained relationship between average measured peak discharge and average β value is shown as: , with R2=0.8571. In recession limb, it showed a strong relationship between watershed area and value as: , where Area is the subwatershed area in km2 with R2=0.9998. In this study, the debris-dammed and failure type hazard model was established and testified by on-site cases. The calculated flood peak discharge and arrival time can be provided to policy-making authorities for incorporating the hazard mitigation plans.

合理化公式是推估小型集水區洪峰流量的經驗公式中,最早也最常被使用者之一,近年來合理化公式也被發展成特定的流量歷線型態。本研究進一步將流量歷線的模擬與廣泛使用流量歷線觀念的「單位歷線法」與「運動波模式」相比較,更能延伸合理化公式的利用與範圍。而一個適當的降雨-逕流模式,且真正合理的計算方法,以提供國內工程師用於設計水土保持與水利構造物或管理洪泛平原,在台灣的災害防救措施中是非常重要的一環。
本研究中提出一個水文演算理論「網格式合理化法 (GRAPH)」,其利用數值高程模型自動萃取集水區之各類地文及水文資訊,並以此網格資訊為基礎,進行空間分佈型的水文演算技術,以便於配合現有地理資訊系統之發展。獲得的成果可概述如下列三項:
一、模式建置
模式中α 值用於校正集流時間,而β值則用以調整洪峰流量 。網格式逕流係數則由衛星影像產生常態化差異植生指標(NDVI)後,利用反向配置之公式:C=((1-NDVI)/2),以推估網格式地表逕流係數,而動態的逕流係數概念亦被提出,用於呈現入滲容量因時而異的真實狀況。退水段模擬則成功地建立乘冪的方程式: ,其中 為前流量,t為洪峰後小時數, 及 則為集水區特定係數, 可由洪峰流量求得, 則可由集水區面積大小求得。整體而言,在小型上游集水區中,GRAPH可以獲得狀況極佳之模擬。
二、模式比較
由「台灣水文資料電腦檔應用之研究」獲得之資料,用於執行無因次單位歷線法之演算。美國工程師團水文工程中心所發展的水文模式系統(HEC-HMS)及地理水文模式系統(HEC-GeoHMS)則用於處理運動波法演算。濁水溪水系的加走寮溪集水區則用於資料模擬與比對實測資料。模擬獲得的成果顯示出GRAPH具有第一精確的洪峰時間預測,與第二精確的模式效率係數。
三、洪水模擬
利用集水區等集流時間線,時間-面積法的觀念早已被學者證明可將演算面積擴大,故本研究亦利用此技術,將探討面積擴大為濁水溪流域,以進一步瞭解整體集水區之水文反應。結果顯示集水區面積( )與α值呈現反指數函數關係,並具有良好的收斂行為,其顯示同一場暴雨中,集水區範圍越大其α值越小,原因可能為下游較多人為利用與破壞,導致集水區集流時間加速。而在平均實測之洪峰流量則與平均β值呈現二次方多項式關係: 其R2為0.8571。退水段模擬中,集水區面積與 值呈現線性關係: ,其R2為0.9998。
本研究以此洪水歷線,建立河道堰塞再潰壩的土石流概念模式,其計算之洪峰流量與到達時間,可以提供決策當局在擬定減災措施時之參考與進一步之研究方向。
URI: http://hdl.handle.net/11455/33908
Appears in Collections:水土保持學系

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.