臺灣博碩士論文加值系統

本研究的目的在針對台灣地區各主、次要河川，進行河川資料之蒐集與分
析、建立電腦資料檔、河川底床坡型之分類，並於各類河川坡型中，推導
出河道基本參數與坡度因子之迴歸式。為適當表達並簡化河川地形之表現
方法，研究中首先以最小平方法計算，求得最合適之Johnson SB 統計分
佈曲線，來代表無因次化後的河川縱剖面曲線，將原本複雜漸變的地形特
徵，改以Johnson SB 分佈曲線的三個分佈參數(γ,δ,α)代表之。根據
各河川坡型代表之Johnson SB 分佈曲線，結合群集分析技巧，將各類河
川坡型分類。不過此法效果並不理想，因此本研究轉而直接利用無因次之
河川縱剖面曲線作為群集分析之變數，求得四類河川坡型。此外，本研究
亦探討各河道基本參數與河床坡度因子之相關特性，擬提供一套台灣本土
地區完整之河川資料分析流程，以利日後作為合成台灣地區河川資料之參
考。

This study intends to develop a procedure to generate synthetic
data for typical rivers in Taiwan. Four approaches were
involved in development including data collection and analysis,
file conversion to computer format, slope classification, and
regression analysis for slope and the fundamental river
parameters. In fact, the characteristics of a non-dimensional
channel profile can be captured by Johnson SB distribution
curve. Under the least square fitting, the complex channel
profiles were represented by 3-parameter Johnson SB distribution
parameters (γ,δ,α). It is rather straightforward to have
these representative curves classified into categories with
clustering analysis. Therefore, if the slope type is
determined, the synthetic data about the river parameters can be
generated through the above derived regression equations.