數值地形模型網格解析度對自動化河川萃取之影響

在森林學相關領域中，河川網絡模擬之應用範圍日益廣泛，而利用網格式數值地形模型(Digital Terrain Model, DTM)萃取山地集水區河川網絡，較人工測繪快速有效率，而水系萃取分佈與準確度通常直接受到數值地形模型網格解析度影響。本研究以國立台灣大學實驗林區內之北勢溪流域為試區，使用網格解析度5m、10m、20m、40m之數值地形模型，配合3種流向演算法（單流向、多流向、無限流向）與12等級之集流閥值(Threshold Value)進行組合演算，自動化萃取網格式河川網絡，並對照1/5,000地形圖之河流線，分別使用誤授率與漏授率之網格對位誤差檢核，以及人工局部判釋方式，分析其網格空間對位誤差與河川網絡分佈特性。研究結果指出，由於網格解析度增高，涵蓋基準圖層河川網絡面積減少，其相對對位偏移明顯提升。然而經過「再取樣」程序之DTM，隨網格解析度降低，其萃取之河川網絡產生不合理「轉向偏移」與「交會點錯估」比率增加，且其河道彎曲度明顯下降。整體而言，網格解析度對山地集水區河川網絡萃取分佈影響，主要在於河道對位誤差與河川網絡之空間產出特性。但不論以何種網格解析度之數值地形模型配合流向演算法萃取河川網絡，對位誤差仍存在，故僅可適用於河川位置粗估萃取，如擬應用於精確河道對位與河寬推估，仍有諸多問題尚待解決。綜合評估結果，高網格解析度數值地形模型適用於萃取山地集水區河川網絡形狀，凸顯其彎曲度與河道寬度變化；低網格解析度數值地形模型，因為其網格涵蓋面積較大，除可萃取基準圖層之河道外，在位置上亦包含邊坡單元，故適用於同時呈現河川網絡與邊坡區域位置。建議使用者視分析目的選用不同網格解析度數值地形模型。

關鍵字

數值地形模型；流向演算法；網格解析度；河川網絡

並列摘要

Traditionally, stream networks were produced manually in hydrologic analysis. Though certain degree of accuracy can be achieved, the process requires large amount of time. In comparison, deriving mountainous stream networks from grid based digital terrain model (DTM) is more efficient than the manual process, however, the accuracy of the results varies with different digital terrain model resolution. This study selected the Bei-Shih-Shi mountainous watershed as the study site, and utilized digital terrain models with 5m, 10m, 20m, 40m grid size to derive stream networks automatically. Three different flow direction algorithms and twelve threshold values were applied in this study. The resultant stream networks were manually compared to the stream networks found on 1/5,000 topographic map. The results indicated that: (1) The conformity of flow path increased as the grid cell size decreased, however the commission error and omission error increased as well. (2) As the grid size decreased, the sinuosity increased and the variations of stream width became more significant. (3) The commission error always happened in headwaters. (4) The commission error and omission error were more effective on stream width when Multiple-flow direction algorithm and D-infinity flow direction algorithm were applied. (5) Overall, though deriving the streams with different DTM resolution, bias remain existed when comparing to the stream networks shown on the topographic maps. The result obtained from the Bei-Shih-Shi watershed shown that the error rate was high. For accurate estimation of stream network, further investigations are needed.

並列關鍵字

Digital terrain model ； Flow direction algorithm ； Resolution ； Stream networks

國際替代計量

數值地形模型網格解析度對自動化河川萃取之影響

全文下載

主題瀏覽