中文摘要：

流域边界的确定是进行山地区域脆弱生态环境信息分析、评价及资源合理配置的重要基础,但提取过程中往往存在流域空间单元图斑破碎、流域空间粒度单一的不足。研究考虑了喀斯特山地的地表环境条件特殊性,综合数字水文地形模型在喀斯特流域边界提取的优势,提出了喀斯特山地流域边界多层次提取的径流树模型。方法基于坡面径流模拟算法对单一径流和集水区的准确划分,构建空间径流树模型,并以径流末端或出水口作为搜索终止条件,搜索判断所有径流的空间相交关系以控制不同层次的流域集水区合并,算法时间复杂度为O（nlogn）。最后,实验验证运用算法对珠江源头流域进行了四个层次的流域提取处理,提取得到的不同层次细节的流域界线,并与综合云南省流域分布图、等高线和人工实测的流域界线相吻合。

英文摘要：

Drainage boundary extraction is the foundation of analyzing and evaluating fragile eco-environment information and reasonably allocating resources in mountain areas. However, in single spatial granularity, some drainage is divided into small pieces with meaningless. Considering the particularity of land surface environment in Karst Mountains, and integrating the advantages of digital hydrology and geomorphology model in Karst drainage boundary extraction, our research proposed a stream-tree model for drainage boundary extraction of multi-hierarchical in Karst Mountains. The overland flow simulation approach is used to receive streams and catchments accurately. Taking the end of streams or outlets as terminal conditions, the algorithm searches and judges the spatial intersection relationship of streams to merge catchments what belong to the same drainage in different hierarchies. These features are within O（nlogn） time, where n is the number of streams. Finally, by the stream-tree model, experiment extracts drainage boundaries with different details in four hierarchies at the source of Pearl River. Results were consistent with drainage boundary which is drawn by the figure of basin distribution, contours, and artificial surveying in Yunnan province