Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin

博士 === 國立臺灣大學 === 生物環境系統工程學系暨研究所 === 90 === This study proposed a wavelet-based multi-resolution analysis of hydrological systems and its applications to the flood forecasting of a river basin. There were two kinds of applications of wavelet transform to the modeling of hydrological syste...

Full description

Bibliographic Details
Main Authors: Chien-ming Chou, 周建明
Other Authors: Ru-yih Wang
Format: Others
Language:zh-TW
Published: 2002
Online Access:http://ndltd.ncl.edu.tw/handle/40256241569595290145
id ndltd-TW-090NTU00404009
record_format oai_dc
spelling ndltd-TW-090NTU004040092015-10-13T14:38:19Z http://ndltd.ncl.edu.tw/handle/40256241569595290145 Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin 水文系統之小波多分辨分析及其應用於流域之洪水預報 Chien-ming Chou 周建明 博士 國立臺灣大學 生物環境系統工程學系暨研究所 90 This study proposed a wavelet-based multi-resolution analysis of hydrological systems and its applications to the flood forecasting of a river basin. There were two kinds of applications of wavelet transform to the modeling of hydrological systems at different resolution levels. The one was the application of redundant wavelet transforms to the observed effective rainfall and direct runoff time series to obtain the wavelet coefficients at each resolution level. Then, these wavelet coefficients were applied to the modeling of rainfall-runoff processes using the linear black box model of discrete system and differential hydrological grey models with dual series DHGM(2,2) at each resolution level. In addition, the ARX, NARX, time-varying ARX and NARX models were carried out at each resolution level. Then applying the model groups’ optimization method to decide and keep the best model to improve the accuracy of flow forecasting. The other was the application of wavelet transforms to the observed effective rainfall and hydrological systems to obtain the wavelet coefficients at each resolution level. The ordinary direct runoff time series were regarded as the output of systems. Applying the Haar wavelet transform altered the state vector and input vector of the state space. In this way an overall detail plus approximation described each new state vector and input vector, which allowed the wavelet-based multi-model Kalman filters to simultaneously estimate, decompose state variables and carry out the parameters estimation. This study further applied the discrete wavelet transform to decompose and compress the unit hydrograph and Volterra kernels, thereby generating smooth reparametrizations of the unit hydrograph and Volterra kernels requiring that fewer coefficients be estimated. A wavelet-based linearly constrained least mean square algorithm and Kalman filters were then employed to on-line estimate wavelet coefficients of the unit hydrograph and Volterra kernels to model the time-varying linear and nonlinear rainfall-runoff processes, respectively. To verify the appropriation of the model adopted, the study chose typhoon or storm events occurred in recent decades at the upstream of Kee-Lung River as case studies to show the goodness and accuracy. The results demonstrated that the analytical method is suitable for the modeling of the rainfall-runoff process at each resolution level and can improve the accuracy of the runoff estimation for small watersheds in Taiwan. The new approach deduced in this study can also offer a useful reference and application for the flood forecasting and flood-damage mitigation engineering of river basins in Taiwan. Ru-yih Wang 王如意 2002 學位論文 ; thesis 258 zh-TW
collection NDLTD
language zh-TW
format Others
sources NDLTD
description 博士 === 國立臺灣大學 === 生物環境系統工程學系暨研究所 === 90 === This study proposed a wavelet-based multi-resolution analysis of hydrological systems and its applications to the flood forecasting of a river basin. There were two kinds of applications of wavelet transform to the modeling of hydrological systems at different resolution levels. The one was the application of redundant wavelet transforms to the observed effective rainfall and direct runoff time series to obtain the wavelet coefficients at each resolution level. Then, these wavelet coefficients were applied to the modeling of rainfall-runoff processes using the linear black box model of discrete system and differential hydrological grey models with dual series DHGM(2,2) at each resolution level. In addition, the ARX, NARX, time-varying ARX and NARX models were carried out at each resolution level. Then applying the model groups’ optimization method to decide and keep the best model to improve the accuracy of flow forecasting. The other was the application of wavelet transforms to the observed effective rainfall and hydrological systems to obtain the wavelet coefficients at each resolution level. The ordinary direct runoff time series were regarded as the output of systems. Applying the Haar wavelet transform altered the state vector and input vector of the state space. In this way an overall detail plus approximation described each new state vector and input vector, which allowed the wavelet-based multi-model Kalman filters to simultaneously estimate, decompose state variables and carry out the parameters estimation. This study further applied the discrete wavelet transform to decompose and compress the unit hydrograph and Volterra kernels, thereby generating smooth reparametrizations of the unit hydrograph and Volterra kernels requiring that fewer coefficients be estimated. A wavelet-based linearly constrained least mean square algorithm and Kalman filters were then employed to on-line estimate wavelet coefficients of the unit hydrograph and Volterra kernels to model the time-varying linear and nonlinear rainfall-runoff processes, respectively. To verify the appropriation of the model adopted, the study chose typhoon or storm events occurred in recent decades at the upstream of Kee-Lung River as case studies to show the goodness and accuracy. The results demonstrated that the analytical method is suitable for the modeling of the rainfall-runoff process at each resolution level and can improve the accuracy of the runoff estimation for small watersheds in Taiwan. The new approach deduced in this study can also offer a useful reference and application for the flood forecasting and flood-damage mitigation engineering of river basins in Taiwan.
author2 Ru-yih Wang
author_facet Ru-yih Wang
Chien-ming Chou
周建明
author Chien-ming Chou
周建明
spellingShingle Chien-ming Chou
周建明
Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin
author_sort Chien-ming Chou
title Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin
title_short Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin
title_full Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin
title_fullStr Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin
title_full_unstemmed Wavelet-Based Multi-Resolution Analysis of Hydrological Systems and Its Applications to the Flood Forecasting of a River Basin
title_sort wavelet-based multi-resolution analysis of hydrological systems and its applications to the flood forecasting of a river basin
publishDate 2002
url http://ndltd.ncl.edu.tw/handle/40256241569595290145
work_keys_str_mv AT chienmingchou waveletbasedmultiresolutionanalysisofhydrologicalsystemsanditsapplicationstothefloodforecastingofariverbasin
AT zhōujiànmíng waveletbasedmultiresolutionanalysisofhydrologicalsystemsanditsapplicationstothefloodforecastingofariverbasin
AT chienmingchou shuǐwénxìtǒngzhīxiǎobōduōfēnbiànfēnxījíqíyīngyòngyúliúyùzhīhóngshuǐyùbào
AT zhōujiànmíng shuǐwénxìtǒngzhīxiǎobōduōfēnbiànfēnxījíqíyīngyòngyúliúyùzhīhóngshuǐyùbào
_version_ 1717755528360230912