TRMM卫星降雨数据在ESSI分布式水文模型淮河流域水文过程模拟中的应用

发布时间：2018-05-20 20:42

本文选题：基于Markov随机场IDW插值 + TRMM卫星降雨　；参考：《南京大学》2014年硕士论文

【摘要】：分布式水文模型为探索流域水文循环机制、评估流域水资源动态变化、管理流域水资源提供了强有力的工具。针对大尺度的、区域气候条件较为复杂的淮河流域中上游流域,采用张万昌、张东提出的具有渗蓄一体化的动态产流机制的ESSI (infiltration Excess and Saturation excess Soil-water dynamic Integration model for hydrology)分布式水文模型,对淮河流域水文过程模拟的适用性及利用TRMM (Tropical Rain fall Measuring Mission)卫星降水数据驱动ESSI分布式水文模型的可行性开展了研究。作为影响水文模型模拟精度的最重要的气象输入参数之一,降雨数据源及其时空特征是影响分布式水文模型模拟效果的最重要的因素。TRMM卫星数据不仅为获取降水数据资料提供了更便捷的途径,而且弥补了传统水文气象监测数据时空分辨率低、不连续的缺点。如何有效利用TRMM卫星数据,提高分布式水文模型精度,是当前分布式水文模型研究的一大热点。本研究的目的旨在检验ESSI分布式水文模型在水文气象及下垫面条件较为复杂的淮河流域水文时空过程模拟的适应性基础上,探索TRMM卫星降水数据驱动该模型模拟流域水文时空过程的可行性,为今后进一步研究利用TRMM卫星降水数据作为流域水文、水资源以及洪涝预警的可替代气象台站降水数据的关键驱动数据提供科学依据。淮河流域处于南北气候过渡带,气候、下垫面条件复杂,水资源相关问题突出。为了进一步提高水文过程模拟的精度,进而为准确地描述流域水文循环过程及水资源空问分布特征,本文以淮河流域中上游流域为研究对象,将高时空分辨率的TRMM卫星降雨数据应用到ESSI模型中,并对比评价了应用TRMM卫星降雨数据前后ESSI分布式水文模型的效率。主要得到以下几方面的研究结果：(1)TRMM卫星降雨资料在淮河流域的精度验证。利用地理信息系统(GIS)技术,首先将TRMM原始产品(3小时时间分辨率,0.25。空间分辨率)处理成栅格大小为0.25。的逐日面降雨数据,然后将淮河流域及其毗邻处的地面气象站降雨数据通过反距离插值法(IDW)、泰森多边形、Kriging内插法、基于Markov随机场的IDW改进方法等传统插值方法得到与TRMM降雨数据的空间分辨率相匹配的面降雨数据,最后开展TRMM降雨数据与不同插值方法下的栅格点降雨量以及流域面降雨量的精度对比评价与分析。结果显示TRMM卫星降雨数据与IDW插值、泰森多边形、基于Markove随机场的IDW网格精度相当,略低于Kriging插值,在淮河流域具有良好的适用性；并且在不同的降雨强度级别上TRMM与实测值分布规律基本一致,但是在日降雨量级别0.01mm/d的降水事件中中反演结果明显偏低。在流域日均面雨量的比较上,TRMM日均面雨量数据与Kriging插值得到的面雨量相关性系数达到0.74；在年面雨量的空间分布上,TRMM降雨数据由南向北、由东向西降雨量依次递减,较为准确地描述了降雨的空间变化趋势。(2)ESSI分布式水文模型在淮河流域中上游流域的应用。结合遥感(RS)和地理信息系统(GIS)技术,获取模型所需的数字高程(DEM)、气象、土壤、土地利用(LULC)等参数,选择合理的长期降雨径流过程模型运行方案和模型参数,分别对2001-2004年、2006-2009年降雨径流进行率定和验证,分析ESSI分布式水文模型在淮河流域中上游流域的日、月、年不同时间尺度上的模拟效率,并对模型精度可能影响的原因进行深入探讨。结果显示了模型在率定期和验证期中的日径流模拟与实测径流曲线相关性达到0.82,并呈现季节变化,枯水时期(11-2月)与实测值接近或略小,在平水期(3-6月)大于实测值,丰水期(7-10月)模拟值比实测值偏小。随着时间尺度的增加,确定性系数、相关性系数精度提高,且相对误差为负值,说明模拟值总体小于实测值。从模拟蒸散发、冠层截留、径流量分配、土壤含水量等变量的空间分布来看,模拟的空间分布形态与淮河流域的客观规律相近。说明了ESSI模型率定得到的参数在研究区具有一定的代表性,较为准确地描述了研究区的水文循环过程和空间变化,并具有很好的适用性。(3) TRMM卫星降雨数据在ESSI分布式水文模型中的应用及验证。应用GIS空间技术将原始空间分辨率的TRMM降雨数据重采样成与ESSI模型中空间栅格输入数据相匹配的1km空间分辨率的栅格降雨量,并通过Visual C++6.0成功将1km空间分辨率的TRMM卫星降雨量集成到ESSI模型中,与原始ESSI模型的模拟结果进行横向比较分析。结果显示了,在日径流模拟中,两者模拟径流曲线一致,各精度评价指标相当；而在率定期月径流模拟中,基于TRMM降雨数据的径流模拟结果明显好于原始ESSI模拟的结果,确定性系数能够达到0.8,明显高于原始模拟的确定性系数(0.72),并且在相对误差百分比的对比中,基于TRMM降雨数据的模拟相对误差值(-7.81%)明显小于初始的-11.79%；而同样在验证期,应用TRMM降雨数据后的模拟效率从0.71提高到0.76,相对误差从-12.80%降到-3.14%。另外,从季节分布来看,TRMM的模拟结果同样较为理想,确定性系数、相关性系数、相对误差百分比均有较大的提高,尤其是在降雨量加大的夏季,相对误差由-45.78%提高到3.56%。从产流类型、地表径流、壤中流、地下径流各水文过程来看,TRMM数据呈现栅格状,这主要与其数据来源栅格形式相关。
[Abstract]:The distributed hydrological model provides a powerful tool for exploring the hydrological cycle mechanism of the basin, assessing the dynamic changes of water resources in the basin and managing the water resources of the basin. For the upper and middle reaches of the upper and middle reaches of the Huaihe basin, which is relatively complex in the regional climate conditions, the ESSI Zhang Dongti has a dynamic ESSI system with the integration of seepage and storage. A distributed hydrological model (infiltration Excess and Saturation excess Soil-water dynamic Integration model for Hydrology) is used to study the applicability of the hydrological process simulation in the Huaihe basin and the feasibility of using the satellite precipitation data to drive the distributed hydrological model. One of the most important meteorological input parameters for the simulation accuracy of the hydrological model, the rainfall data source and its temporal and spatial characteristics are the most important factors affecting the simulation effect of the distributed hydrological model..TRMM satellite data not only provides a more convenient way for obtaining the data of precipitation data, but also improves the spatial and temporal resolution of the traditional hydrometeorological monitoring data. How to effectively use TRMM satellite data to improve the accuracy of distributed hydrological model is a hot spot in the current distributed hydrological model research. The purpose of this study is to test the adaptability of the ESSI distributed hydrological model in the hydrological spatio-temporal process simulation of the Huaihe basin with complex Hydrometeorology and underlying surface conditions. On the basis of this, the feasibility of the TRMM satellite precipitation data to simulate the hydrological spatiotemporal process in the basin is explored. It provides a scientific basis for further research and utilization of the TRMM satellite precipitation data as the key driving number of the precipitation data of the substitutable meteorological stations in the basin hydrology, water resources and flood warning. The Huaihe basin is in the north and South gas. In order to further improve the accuracy of hydrological process simulation, and to accurately describe the hydrological cycle process and the spatial distribution characteristics of water resources, this paper takes the middle and upper reaches of the Huaihe basin as the research object, and makes the TRMM satellite rainfall data with high spatial and temporal resolution. In the ESSI model, the efficiency of ESSI distributed hydrological model before and after the application of the TRMM satellite rainfall data is compared and evaluated. The main results are as follows: (1) the accuracy verification of the TRMM satellite rainfall data in Huaihe basin. Using the geographic information system (GIS) technology, first of all, the TRMM original product (3 hours time resolution, 0.25.) The spatial resolution) is processed into a day by day rainfall data of the grid size of 0.25., and then the spatial resolution of the rainfall data of the Huaihe basin and its adjacent ground meteorological stations is obtained through the inverse distance interpolation (IDW), Tyson polygon, Kriging interpolation, and the traditional interpolation method based on the IDW improvement method of the Markov with the airport, and so on. Rate phase matching surface rainfall data, and finally carry out the comparison evaluation and analysis of the TRMM rainfall data and the grid point rainfall of different interpolation methods and the accuracy of the basin surface rainfall. The results show that the TRMM satellite rainfall data and the IDW interpolation, the Tyson polygon, based on the IDW grid precision of the Markove random field are equivalent, slightly lower than the Kriging interpolation, in Huaihuai. The river domain has good applicability, and the distribution of TRMM and the measured value are basically the same at different rainfall intensity levels, but the inversion results are obviously lower in the precipitation events of the daily rainfall level 0.01mm/d. The daily average rainfall data of TRMM is related to the surface rainfall obtained by the Kriging interpolation. The coefficient of sex is 0.74. In the spatial distribution of annual rainfall, the TRMM rainfall data from south to North and east to West decrease in turn, and the spatial variation trend of rainfall is described accurately. (2) the application of the ESSI distributed hydrological model in the middle and upper reaches of the Huaihe basin. Combined with remote sensing (RS) and geographic information system (GIS) technology, the model is used to obtain the model. The parameters of DEM, meteorology, soil and land use (LULC) are used to select the reasonable operation scheme and model parameters of the long term rainfall runoff process model. The 2001-2004 year and 2006-2009 year rainfall runoff is determined and verified respectively, and the ESSI distributed hydrological model is not simultaneously in the day, month, and year of the middle and upper reaches of the Huaihe basin. The results show that the correlation between the daily runoff simulation and the measured runoff curve reaches 0.82, and shows the seasonal variation, and the period of dry water (11-2 months) is close to or slightly smaller than the measured value. In the flat water period (3-6 months), the model is larger than the measured value. The simulated value is smaller than the measured value. With the increase of time scale, the certainty coefficient and correlation coefficient are improved, and the relative error is negative, which shows that the simulated values are generally smaller than the measured values. The simulated spatial distribution of the simulated evapotranspiration, canopy interception, runoff distribution, soil water content and so on is shown. The state is similar to the objective law of the Huaihe River Basin. It shows that the parameters obtained by the ESSI model have certain representativeness in the study area. The hydrological cycle process and the spatial change in the study area are described accurately. (3) the application and verification of the TRMM satellite rainfall data in the ESSI distributed hydrological model and the application of GI The S space technology resamples the TRMM rainfall data of the original spatial resolution into the raster rainfall of the 1km spatial resolution matched with the spatial grid input data in the ESSI model, and successfully integrated the TRMM satellite precipitation of the 1km spatial resolution to the ESSI model through Visual C++6.0, and transversely related to the simulation results of the original ESSI model. The results show that in the daily runoff simulation, the simulated runoff curves are the same, and the accuracy evaluation indexes are equal, while in the regular monthly runoff simulation, the runoff simulation results based on the TRMM rainfall data are better than the original ESSI simulation results, and the deterministic coefficient can reach 0.8, which is obviously higher than the determinant system of the original simulation. The number (0.72), and the relative error value (-7.81%) based on the TRMM rainfall data is obviously smaller than the initial -11.79% in the comparison of the relative error percentage, and the simulation efficiency after the application of TRMM rainfall data is increased from 0.71 to 0.76 in the verification period, and the relative error is reduced from -12.80% to -3.14%., and the simulation of TRMM is simulated from the seasonal distribution. The result is also more ideal, the coefficient of certainty, the correlation coefficient and the percentage of relative error are greatly improved, especially in summer with the increase of rainfall, the relative error is raised from -45.78% to 3.56%. from the type of production flow, surface runoff, the middle flow of the earth and the hydrological process of the underground runoff, and the TRMM data is raster, this is mainly with its data. The source grid form is related.
【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：P334.92

【参考文献】