【作者】 胡志瑞；

【导师】 杨勤科；

【作者基本信息】 西北农林科技大学 ， 地图学与地理信息系统， 2009， 硕士

【摘要】 区域水土流失模型是在借鉴国外坡面和小流域土壤侵蚀模型、大尺度水文模型研究成果基础上,利用GIS空间分析和计算机软件工程建立起来的。本文以甘肃天水市藉河示范区为研究区,对改进后的区域水土流失模型进行了敏感性分析。建立了区域水土流失模型所需参数数据库,并对参数进行了质量评价和不确定性分析,力图为模型输入精准的参数。分别以不同栅格尺寸和不同时间段运行模型,分析其变化对模型模拟结果所产生的影响,根据结果确定了适用于本研究的栅格尺寸和时间段。在此基础上,用扰动分析法分析了模型参数的敏感性,验证了模型的稳定性和通用性,为完善模型提供了依据。本研究主要结论为:(1)基于ANUDEM5.1生成了研究区Hc-DEM,结果表明Hc-DEM可以为水文和土壤侵蚀模拟提供更有力的支持,尽可能的降低了DEM对模型带来的不确定性。解译TM遥感影像并反演得到叶面积指数。在对土壤稳渗速率空间插值不确定性分析的基础上,基于ESDA工具分析了各数据,以反距离权重法完成抗冲系数空间插值、以普通克里格法完成土壤稳渗速率空间插值、以简单克里格法完成降雨量空间插值;(2)通过在不同栅格尺寸下运行模型,分析了栅格大小变化对模型模拟结果的影响。结果表明:随着DEM栅格尺寸变小,坡度随之变缓,最大洼地拦蓄水量、径流速度、携沙能力也随之减小,导致输沙量减小。输沙量减小的趋势与坡度减小的趋势一致。模型效率受栅格尺寸的影响较大,栅格尺寸越小所需时间越长。在栅格尺寸为75m处模拟值有一明显的变化,通过分析初步确定适合本研究的栅格尺寸为75m,模型运行时间与栅格尺寸的关系式为:y=53.773e-0.4125x;(3)通过在不同时间段下运行模型,分析不同时间步长对模型模拟结果的影响。随着时间段增多,模型模拟值单元格径流深、剥蚀量、输沙量、径流量都呈减小趋势。时间段从1到30,模型运行时间增大,且模拟结果在时间段为4时有明显的改变。模型运行时间与时间段的关系式为:y=0.4966x+6.3266,综合考虑模型运行效率和模拟结果,初步选定模型时间段为4;(4)以扰动分析法分析了降雨量、叶面积指数、抗冲系数和土壤稳渗速率等四个参数的敏感性。通过四象限分析法分析,结果表明,四个参数对输沙量影响程度顺序为:降雨量﹥抗冲系数﹥土壤稳渗速率﹥叶面积指数;四个参数对径流量影响程度的顺序为:降雨量﹥土壤稳渗速率﹥叶面积指数﹥抗冲系数。降雨量对输沙量和径流量而言是最敏感的,且为正相关。抗冲系数对输沙量的模拟影响小于降雨量,对径流量无影响。土壤稳渗速率对径流量的影响大于对输沙量的影响。叶面积指数对模拟输沙量和径流量影响相对较为微弱。更多还原

【Abstract】 Based on GIS spacial analysis and computer software engineering, the regional soil erosion model was established by using study results of foreign slope and small watershed soil erosion model and large-scale hydrology model as a source of reference. Taking Ji river demonstration zone as a study area in Tianshui city Gansu province, this paper analyzed the sensitivity of the improved regional soil erosion model. The Geodatabase for regional soil erosion model was established and parameters’quality and uncertainty were analyzed in order to make the parameters more accurate. This paper analyzed the effects of the simulation result caused by changes of different grid size and time, the result simulated by the model respectively, and then determined the suitable grid size and time period for this study. Based on this, the sensitivity of the input parameters was analyzed through disturbance method, the stability and commonality of the model was validated, which provided foundation for perfecting the model. The main results are:(1) Hydrologically correct DEM(Hc-DEM) was set up based on ANUDEM5.1 software. The result showed that Hc-DEM could provide powerful support for hydrology and regional soil erosion simulation, which decreased the model uncertainty resulted from DEM as much as possible. On the basis of analyzing the uncertainty of spatial interpolation for soil steady infiltration rates, and each data using ArcGIS Exploratory Spatial Data Analysis(ESDA). Spatial interpolation for erosion-resisting coefficient was completed with inverse distance weighting method, soil steady infiltration rates using ordinary Kriging method and rainfall using simple Kriging method.(2) By running regional soil erosion model the results caused by changes of grid size parameters was analyzed. The results showed that the slope became slower with the decrease in DEM grid size. Meanwhile, maximum depression storage, runoff velocity and transport capacity also reduced, which lessened sediment discharge. The trend of decrease in sediment discharge is consistant with the decrease in slope. The model efficiency was greatly influenced by grid size. The smaller the grid size, the longer time it required. As a result, there was a clear change in 75m, which was the appropriate size for this study.(3) The impacts of different time lengths on simulation model was analysed under different time period. The result showed that with time increasing, the value of the runoff depth, sediment discharge, denudation quantity, runoff was on the decrease. When the time period was from 1 to 30, the time for finishing running model became longer, and the simulation result had a significant change in the 4, the linear relationship between the running time of model and time period is y=0.4966x+6.3266. To consider the efficiency of model running and simulation result, 4 time period is the appropriate time period for the model.(4) The sensitivity of input parameters of was analyzed through antu-disturbance method. The experiment showed that rainfall is the most sensitive parameter to the sediment discharge and runoff, the erosion-resisting coefficient is less sensitive than rainfall for sediment discharge, but no influences on runoff, steady water infiltration rates is less effective to runoff. Comparatively speaking, the leaf area index had a weak impact on the sediment discharge and runoff.更多还原