论文部分内容阅读
流域编码是以子流域划分进行流域相关研究的重要内容。Pfafstetter流域编码以编码唯一、顾及流域拓扑关系及编码效率高等优点而被广泛采用。本文在流域相关研究的分析范围不断增大、数据精度越来越高的需求背景下,以Pfafstetter编码为基础,对流域编码并行化方法进行研究。首先,分析了Pfafstetter编码不全面和码位不一致的问题,改进了Pfafstetter编码规则;然后,从数据并行的角度,讨论了并行计算环境下的数据划分及并行化策略,进而设计了流域编码并行算法;最后,利用长江中上游流域SRTM数据,在集群系统上对流域编码并行算法的正确性和并行性能进行了测试。实验结果表明,本文设计实现的流域编码并行算法可获取与实际较为一致的计算结果,且提高了编码计算效率,可为基于子流域划分的流域分析并行化提供参考。
Watershed coding is an important part of watershed-related research by sub-watershed. The Pfafstetter catchment code is widely used because of its unique encoding, taking into account the topological relations of the watershed and the high coding efficiency. Based on the Pfafstetter coding, this paper studies the code-parallelization method of watershed based on the increasing demand of watershed-related research and data accuracy. Firstly, the problems of incomplete Pfafstetter coding and inconsistent code bits are analyzed, and Pfafstetter coding rules are improved. Then, the data partitioning and parallelization strategies are discussed from the aspect of data parallelism. Finally, using the SRTM data in the upper and middle reaches of the Yangtze River to test the correctness and parallelism of the parallel algorithm of watershed coding in the cluster system. The experimental results show that the parallel algorithm of watershed coding designed and implemented in this paper can get more consistent results with the actual ones, and improve the coding efficiency. It can provide a reference for parallel analysis of watershed based on sub-watershed.