利用河川生態指標評估水文合成序列模型重現歷史觀測資料能力之研究

The Use of Indicator of Hydrologic Alteration in the Assessment of the Reproducibility of Synthetic Flow Models

覃炳晧(Bing-Hao Thum)

指導教授 ： 游景雲

國立臺灣大學/工學院/土木工程學研究所/碩士(2012年)

https://doi.org/10.6342/NTU.2012.02944

摘要

完整而可靠的水文記錄資料，乃是水資源規劃及運用之重要憑藉，然而由於資料的缺乏，在水利工程之相關規劃上常使用合成水文序列、由歷史資料經由分析推衍出合成流量序列來進行規劃。合成水文序列可保持歷史流量序列的統計參數及分佈特性，以作為工程規劃之依據，基本上合成水文序列皆能滿足工程之需求。近年來，由於對於生態環境的日益重視，許多工程規劃上都將生態需求納入考量，這些規劃多以傳統工程角度及工具來進行，而未能檢驗是否能反映生態之特性，合成水文序列即是一個相當典型的例子，許多工程師都運用以往用於工程設計之合成流量序列，進行設定相關生態設施之設計條件，然而，卻在水文上之合成水文序列是否能適當呈現生態特性卻仍然有待討論。因此，針對此一議題，本研究欲探討合成水文序列是否能適當呈現生態特性，生態特性主要反應在日流量上，相較於月流量之水文合成序列、但目前為止之技術水準並無一固定之日流量合成水文序列之生成方法，故本研究將評估文獻中較具代表性之數種方法，以其所生成之合成流量序列，來評估是能重現生態特性，作為將來相關生態需求規劃之依據。
本研究採用兩大類合成水文模式，逐日生成模式和流量分配模式，第一部分之序率合成模式包含Shot noise method和Storm-based method兩種，第二部分則分為生成大尺度之模擬流量方法以及分配模式，其各別利用ARMA、PARMA以及Modified k-NN method來分別生成年、月流量，再利用Shot noise和k-NN分配以生成日流量。將其各模式配合水文特性檢驗以及水文改變指標測試並藉由各模式之流量延時歷線來看。逐日生成流量模式shot noise和storm-based模式，由於誤差累積較多，只能重現部分生態特性。然而，結果最好為年分配成月之模式，水文和生態重現性都較佳，但其變化性小，只能依照觀測資料之比例生成。故本研究認為適合之生成模式為月流量分配模式中的k-NN方法，由於此方法為月流量分配為日流量，保持較具意義之月流量特性，而分配流量也能重現生態性，模式運算效率也高。

關鍵字

合成水文序列 ； 水文改變指標

並列摘要

The use of synthetic streamflow is particularly useful for water resources engineering. In the past decades, successful application of synthetic hydrologic sequences has been recognized for traditional engineering purposes, such as reservoir operation, water supply and flood control. With the increasing attention for the ecosystem, hydraulic engineering has to appropriately consider the ecological needs. However, the use of synthetic streamflow for ecological purpose has never been carefully examined. Hence a question is raised, could traditional synthetic streamflow approach provide reasonable result for determining ecological conditions?
For this reason, this study addresses the reproduction of ecological patterns by synthetic streamflow generation model. The Tanshui River in Taiwan is used as a case study. We consider several combinations of different synthetic flow methods to examine this issue. Following approaches are considered, (1) Pure synthetic simulation model (2) Long term simulation and daily disaggregation models. The long term simulation models used in this study are SAMS-2007 and Modified k-NN bootstrap non-parametric approach. These models are applied to generate annual or monthly flow data according to historical record. Then Shot noise and k-NN based disaggregation models are applied to generate the daily flow. According to daily flow data generated, the Indicators of Hydrologic Alterations (IHA) program is used to quantify ecological patterns. By comparing the difference of 32 IHA indicators between historical and generated flow data, we evaluate the reproduction of different daily flow generation models in ecological characteristics. For the perspective of ecological needs, this study discusses the advantage and disadvantage of synthetic streamflow generation models and makes suggestion for their application for facilities design process.

並列關鍵字

synthetic streamflow ； IHA

參考文獻

63. 張騌麒,孫建平(2009), 結構方程模式於建立生態水質水文指標之研究. 農業工程學報 第55卷 第3期, 75-87.

2. Aksoy, H., and Bayazit, M. (2000). A model for daily flows of intermittent streams. Hydrological Processes, 14(10), 1725-1744.

3. Arthington, A. H., Bunn, S. E., Poff, N. L. R., and Naiman, R. J. (2006). The challenge of providing environmental flow rules to sustain river ecosystems. Ecological Applications, 16(4), 1311-1318.

5. Aksoy, H. (2003), Markov chain-based modeling techniques for stochastic generation of daily intermittent streamflows, Advances in Water Resources, 26(6), 663-671.

7. Aksoy, H. (2004), Using Markov Chains for Non-perennial Daily Streamflow Data Generation, Journal of Applied Statistics, 31(9), 1083-1094.