节点文献

基于汇流模型的黄土坡面微地形空间离散化研究

Research on Spatial Discretization of Microtopographic Loess Slope Based on Confluence Model

【作者】 张鑫

【导师】 张青峰;

【作者基本信息】 西北农林科技大学 , 土地资源与空间信息技术, 2020, 硕士

【摘要】 空间离散化(Discretization)是将地表分割成模型运行的基本地域单元的方法,是分布式水文模型研究水文变量空间分布的核心内容。选择合适的空间离散网格单元建立水文模型是水文模型提高模拟效率及模型得以广泛应用的重要基础。基于DEM开展空间离散化研究已在水文模型中得到初步探索,本研究基于微地形数字高程模型(M-DEM),通过网格离散法将M-DEM网格大小划分为1mm-10mm,分析空间离散化对流域水文特征的影响。在GIS技术、元胞自动机模型方法的基础上,结合分布式水文模型与前人研究成果构建SCS-CA模型与CA汇流模型,对不同雨强(1mm/min、1.5mm/min)、不同坡度(5°、10°、15°)下黄土微地形人工掏挖耕作坡面汇流进行模拟,通过模型汇流模拟结果与实测资料的统计分析,研究空间离散化对水文模型汇流模拟效果的影响并确定模型运行的最佳离散网格单元。本文主要研究结论如下:(1)DEM网格离散化程度对基于DEM提取流域水文特征有着重要影响,其中对水流路径长、主河道长与坡度的数值影响较大,高程受网格离散的影响较小,与此同时,决定虚拟河网结构的集水阈值也发生变化。(2)通过SCS-CA模型与CA汇流模型模拟结果与实测资料的统计分析可知空间离散化对水文模型模拟结果有着重要影响,1mm/min雨强5°坡、10°坡与15°坡SCS-CA模型的最佳离散网格单元大小分别为4mm、5mm和4mm;1.5mm/min雨强5°坡时CA汇流模型的最佳离散网格单元大小为4mm;1mm/min雨强5°坡、10°坡与15°坡CA汇流模型的最佳离散网格单元大小分别为5mm、6mm和6mm;1.5mm/min雨强5°坡时CA汇流模型的最佳离散网格单元大小为6mm。(3)通过对不同离散程度(1mm-10mm)网格模型模拟结果分析可知,在同一降雨条件下,随着网格的增大,模型模拟汇流量减少,并且在构建水文模型时,并不是网格越小,模型的模拟结果越好,模型在一定的网格大小范围内,模拟结果较好,而超出这个网格范围水文模型的模拟误差则非常大。因此在模型应用中,需要研究者根据具体的流域以及不同模型的模拟精度等因素选择最佳的离散网格单元或寻找模型模拟结果较好的离散网格范围。(4)对SCS-CA模型与CA汇流模型模拟结果的相对比增益的指数r~2度量可知,在黄土微地形坡面,人工掏挖耕作措施下,SCS-CA模型的模拟效果优于CA汇流模型。综上可知,本研究基于GIS技术、元胞自动机模型理论与方法以M-DEM为基础构建分布式SCS-CA模型与CA汇流模型模拟汇流,成功地实现了微地形的空间离散化,为基于网格法确定水文模型的最佳离散网格单元提供一个新的研究思路,对水文模型的空间离散化研究具有重要意义。

【Abstract】 The core content of the distributed hydrological model to study the spatial distribution of hydrological variables is spatial discretization.It is an important basis for hydrological models to improve simulation efficiency and be widely used to establish appropriate hydrological models to establish hydrological models.Research on spatial discretization based on DEM has been initially explored in hydrological models.This study is based on the micro-terrain digital elevation model(M-DEM).The grid size is used to divide the M-DEM grid size into 1mm-10mm,and the impact of spatial discretization on the hydrological characteristics of the basin is analyzed.Based on the basic principles of multiple disciplines,GIS technology,and cellular automaton model method,combined with distributed hydrological model to build SCS-CA model and CA confluence model,for different rain intensity(1mm/min,1.5mm/min),artificial digging and cultivating the slope confluence on the micro-topography of loess under different slopes(5°,10°,15°).Through statistical analysis of model confluence simulation results and measured data,the effect of spatial discretization on hydrological model confluence simulation effects is studied and the best discrete grid unit for model operation is determined.The main conclusions of this paper are as follows:(1)The degree of DEM grid discretization has an important influence on the extraction of hydrological features based on DEM.Among them,the numerical effects on the length of the flow path,the length of the main channel and the slope are greater,and the elevation is less affected by the grid dispersion.At the same time,the water collection threshold that determines the structure of the virtual river network also changes.(2)Through the statistical analysis of the simulation results and measured data of the SCS-CA model and CA confluence model,it can be seen that the spatial discretization has an important influence on the simulation results of the hydrological model.The best discrete grid cell size of SCS-CA model for 5°slope,10°slope and 15°slope at 1mm/min rain intensity is 4mm,5mm and 4mm,respectively;The optimal discrete grid cell size of the CA confluence model at a slope of 5°at a rain intensity of 1.5 mm/min is 4 mm;At 1mm/min rain intensity,the optimal discrete grid cell sizes for the 5°,10°and 15°slope CA confluence models are 5mm,6mm and 6mm,respectively;The optimal grid cell size of the CA confluence model at a slope of 1.5°at a rain intensity of 5 mm/min is 6 mm.(3)Through the analysis of the simulation results of the grid model with different degrees of dispersion(1mm-10mm),it can be seen that under the same rainfall conditions,with the increase of the grid,the simulated flow of the model decreases.And when building a hydrological model,it is not that the smaller the grid,the better the simulation results of the model.The simulation results are better within a certain grid size,and the simulation errors of the hydrological model beyond this grid range are very large.Therefore,in the application of the model,researchers need to select the best discrete grid unit or find the range of the discrete grid with better model simulation results according to the specific watershed and the simulation accuracy of different models.(4)According to the r~2 measure of the relative gain of the simulation results of the SCS-CA model and the CA confluence model,the simulation effect of the SCS-CA model is better than that of the CA confluence model on the loess micro-terrain slopes and manual digging and cultivation measures.In summary,based on GIS technology,cellular automaton model theory and method,this study builds a distributed SCS-CA model and CA confluence model to simulate confluence based on M-DEM,and successfully achieves the spatial dispersion of micro-terrain It provides a new research idea for determining the best discrete grid unit of the hydrological model based on the grid method,and is of great significance to the spatial discretization research of the hydrological model.

【关键词】 微地形空间离散化水文模型GIS
【Key words】 Microtopographyspatial discretizationhydrological modelGIS
节点文献中: