Development of a data model to facilitate rapid Watershed Delineation

An efficient model to store and retrieve surface watershed boundaries using graph theoretic approaches is proposed. Our approach utilizes three algorithms and accepts as input standard digital elevation models derived stream catchment boundaries. The first is called Modified Nested Set (MNS), which is a generalized depth first graph traversal algorithm that searches across stream reaches (vertices) and stream junctions (edges) labelling vertices by their discovery time, finish time, and distance from the root. The second is called Log Reduced Graphs (LRG), which creates a set of logarithmically reduced graphs from the original data, to store the watershed boundaries. The final algorithm is called Stitching Watershed, which provides a technique to merge watershed boundaries across the set of graphs created in the LRG algorithm. This technique is show to provide significant advantages in processing, storage, and retrieval complexity when compared to hypothetical data models.