Research And Implementation Of Parallel Algorithms For Hydrological Analysis Based On Grass GIS

Hydrologic analysis is an important aspect of GIS applications. Hydrological analysis can locate surface runoff pollution, predict the impact of topographic changes on runoff, as well as determine whether to generate floods and its scope, thus it is widely used for disaster prediction, agriculture, forestry, and regional planning. As a major source of hydrological data analysis,DEM(Digital Elevation Model, abbreviated DEM) can provide a wealth of hydrological information.Grass is composed of models with various independent functions and clear divisions. With good portability and scalability, Grass supports command line mode and interface mode,and simple to operate. Additionly with the open source nature of Grass, whose source code can be modified, it is widely applied in GIS research.With the development of acquisition and processing of geospatial data, the data volume of hydrological analysis is increasing, and coupled with the strong continuity of hydrological analysis, hydrological serial algorithm cannot meet the requirements. Taking all these factors, we consider realizing parallelization of hydrological analysis algorithm on Grass GIS.The main work of this paper is organized as follows:Firstly, the theoretical basis and related technologies of this study are introduced, then we summarize existing hydrological analysis algorithms, analysis their advantages and limitations, specify the principle of slope and aspect serial algorithmP & D depressions algorithm and D8 flow algorithm, give the corresponding pseudo-code and processes, introduce technology of parallel computing and discuss Grass GIS systems and data structures.Secondly, for the time-consuming problem when hydrological analysis algorithms process large data, we choose the MPI parallel programming model and propose the parallel algorithms which can be applied on Grass platform. Utilizing the characteristics of raster data,we devide the raster data to subprocesses to shorten the running time of the algorithm and realize the algorithm parallelization.Finally, the implementation of hydrological analysis parallel algorithm on the Grass GIS is introduced. We describe the process and programming of serial and parallel algorithms for hydrology hydrological in detail. The graphics shows the experimental results of serial and parallel algorithms, and the results are summarized and analyzed in detail. Experimental results show that by way of the data divided parallel algorithms for hydrological analysis can achieve good speedup and efficiency.