An efficient model to store and retrieve watershed boundaries using graph-theoretic approaches is proposed. Our approach utilizes three algorithms and accepts as input standard Digital Elevation Model DEM derived flow direction grids (D8) or stream reach catchment boundaries and outputs derived watershed boundaries. This technique was applied to the $\approx$ 36,000 km² Delaware River Watershed using the National Hydrography Plus Version 2 Dataset. This technique is showed to provide significant reductions in processing (98–99%), storage (81–82%), and retrieval complexity (95–96%) for polygons and (80–86%) for nodes when compared to existing and hypothetical data models that are used to create watershed boundaries.

提出了一种利用图论方法存储和检索流域边界的有效模型。我们的方法利用三种算法，并接受数字高程模型DEM派生的流向网格（D8）或河流到达集水区边界作为输入标准，并输出派生的分水岭边界作为输入标准。此技术已应用于$\approx$36,000 km ²特拉华河流域，使用的是国家水文地理增强版2版本数据集。与现有技术和现有技术相比，该技术可以显着减少多边形的处理（98–99％），存储（81–82％）和检索复杂度（95–96％），以及节点（80–86％）用于创建分水岭边界的假设数据模型。