NAM模型参数率定方法比较研究

发布时间：2018-01-20 13:17

本文关键词： NAM模型扰动分析法敏感性分析单纯形法遗传算法　出处：《黑龙江八一农垦大学》2017年硕士论文　论文类型：学位论文

【摘要】：我国农业水环境时空分布十分不均匀,降雨和径流是形成农业水环境的重要因素,对粮食产量及品质影响较大,潜在影响流域、湿地等生态系统。此外,降雨是径流形成的主要原因,地表径流对降雨的变化极为敏感,短时间内急剧增加易导致洪涝灾害,对农业影响巨大,对设施农业破坏巨大。降雨径流模型是模拟流域内由降雨形成地表径流过程的概念性水文模型,是通过水文循环的物理结构与半经验公式相结合,利用历史水文资料进行各参数率定,近似地模拟流域内径流过程。该模型已广泛应用于水旱灾害防灾减灾、水资源开发与利用、水环境和生态系统保护等领域。为了研究流域内降雨-径流关系,提高降雨径流模型模拟精度,寻求适合模型的率定方法,研究中选择NAM降雨径流模型,对模型参数进行敏感性分析,分别以单纯形法和遗传算法为率定方法,对美国伊利诺伊州Upper Sangamon流域和我国东北乌裕尔河流域进行日径流模拟。主要开展了如下工作:(1)介绍NAM降雨径流模型参数和结构,确定具体率定方法,给出模型评价标准,对研究区域、数据来源和数据处理方法的基本原理进行了说明。(2)选取NAM模型中的9个主要参数,以伊利诺伊州Upper Sangamon流域Fisher观测站的径流数据和气象数据为研究对象,以拟分析参数改变下流域径流变化量为研究目标,利用扰动分析法分析了9个重要模型参数对于模拟结果的敏感性,得到了各参数的敏感性等级,为下一步率定参数提供依据。(3)介绍单纯形法和遗传算法的率定策略,在模型参数敏感性分析的基础上,以浅层蓄水层最大含水量Lmax、地表蓄水层最大含水量Umax、地表径流系数CQOF、地下径流计算临界值TG作为主要率定参数,采用两种率定方法分别得到Upper Sangamon流域和乌裕尔河流域NAM模型的中间参数并进行日径流计算,得到了不同方法不同流域的日径流过程,验证了两种算法的正确性和可靠性。(4)综合考虑两种率定方法在不同条件流域的适用性,对得到的日径流结果进行比较分析,得到了两种率定方法的适用条件,可依据所掌握流域初始条件来选择NAM模型的率定方法,为后续提高模型模拟精度奠定了基础。
[Abstract]:The spatial and temporal distribution of agricultural water environment in China is very uneven. Rainfall and runoff are important factors in the formation of agricultural water environment, which have a great impact on grain yield and quality, and potentially affect watershed, wetland and other ecosystems. Rainfall is the main reason for the formation of runoff. Surface runoff is very sensitive to the changes of rainfall. A sharp increase in a short period of time can easily lead to flood disaster, which has a great impact on agriculture. Rainfall runoff model is a conceptual hydrological model which simulates the process of surface runoff formed by rainfall in a watershed. It combines the physical structure of hydrological cycle with semi-empirical formula. This model has been widely used in flood and drought disaster prevention and mitigation, water resources development and utilization. In order to study the relationship between rainfall and runoff and improve the precision of rainfall runoff model, the NAM rainfall runoff model is chosen in the study. The sensitivity of the model parameters is analyzed using simplex method and genetic algorithm respectively as the rate determination method. The daily runoff simulation of the Upper Sangamon Basin in Illinois and the Wuyur River Basin in Northeast China is carried out. The main work is as follows: 1). The parameters and structure of NAM rainfall runoff model are introduced. Determine the specific rate of the method, give the model evaluation criteria, the study area, data sources and data processing methods of the basic principles of the explanation. 2) selected nine main parameters of the NAM model. The runoff data and meteorological data of Fisher station in Upper Sangamon basin in Illinois were studied, and the runoff variation under the change of parameters was taken as the research object. The sensitivity of nine important model parameters to the simulation results is analyzed by perturbation analysis method, and the sensitivity grade of each parameter is obtained. This paper introduces the strategy of simplex method and genetic algorithm. Based on the sensitivity analysis of model parameters, the maximum water content of shallow aquifer Lmax is used. The maximum water content Umax, the surface runoff coefficient CQOF and the critical value TG of underground runoff are the main parameters. The intermediate parameters of the NAM model of the Upper Sangamon basin and the Wuyue River basin were obtained by two rate-determination methods and the daily runoff was calculated. The daily runoff process of different watershed with different methods is obtained, and the correctness and reliability of the two algorithms are verified. The results of daily runoff are compared and analyzed, and the suitable conditions of the two methods are obtained. According to the initial conditions of the basin, the NAM model can be selected. It lays a foundation for further improving the accuracy of model simulation.
【学位授予单位】：黑龙江八一农垦大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S161.6

【相似文献】