摘要
确定河源位置是基于数字高程模型(DEM)提取水系的关键环节,对数字流域水系的空间拓扑关系和几何特征具有重要影响。选取具有实际勘测河源位置资料的山丘区小流域即美国俄亥俄州东南部的Indian Creek流域,基于高精度DEM(1m)提取河网,系统分析集水面积阈值法、坡度-集水面积阈值法、曲率-坡向变率法和等高线聚类法这4种方法的优劣;利用实际勘测河源位置信息,构建灵敏度、可靠度等评价指标体系。结果表明,集水面积、坡度-集水面积阈值法简单、高效,需人为确定阈值,利用不同阈值提取河网差异十分显著;曲率-坡向变率法、等高线聚类法可避免阈值确定的不确定性,且可靠性及灵敏度均较高。与实际勘测河源位置相比较,等高线聚类法吻合度最高,集水面积阈值法最低。研究成果可为流域水文分析提供参考。
Identifying the locations of river sources based on the digital elevation model(DEM)is the key to channel network extraction,which determines the spatial topological structure and geometric characteristics of channel network.In this paper,a small typical mountainous watershed with field mapped river sources(Indian Creek Basin in southeast of Ohio)was selected as study site.Based on the high-precision DEM(1 m),four methods,namely contributing area threshold,slope-area threshold,curvature and slope-direction change method,and clustering analysis of contour lines,were used to extract the channel networks.The performance indices including sensitivity and reliability were used to evaluate the accuracy and precision of above four methods in combination with field mapped river sources’ location.The results show that the constant threshold methods(e.g.,contributing area and slope-area)is simple and effective;However,it must determine a spatially constant threshold for channel network extraction,and the extracted channel networks were sensitive to thresholds.The curvature and slope-direction change and clustering analysis of contour lines methods can avoid the subjective and uncertain determination of constant threshold,and therefore achieves high reliability and accuracy.The locations river sources extracted using the contour clustering analysis has the best agreement with the mapped river sources,whereas the contributing area threshold method achieves the worst performance.The research results can provide a reference for hydrological analysis.
作者
吴晨晨
吴剑
叶磊
王文丽
周惠成
WU Chen-chen;WU Jian;YE Lei;WANG Wen-li;ZHOU Hui-cheng(School of Hydraulic Engineering,Dalian University of Technology,Dalian 116023,China)
出处
《水电能源科学》
北大核心
2020年第5期22-25,30,共5页
Water Resources and Power
基金
国家重点研发计划(2016YFC0402203)
国家自然科学基金项目(51709033)。
关键词
数字高程模型
河源
河网
分析评价
DEM
river source
river network
analysis and evaluation
