水文模型参数选取对模拟径流的年际年内分布影响评估  

Appraising the effect of choosing hydrological model parameters on the inter-annual and intra-annual distribution of simulated runoff

作　　者：张静文 郭家力[2,3] 王敬斌 夏丽丽[1] 岳青华[4] 刘章君 ZHANG Jing-Wen;GUO Jia-Li;WANG Jing-Bin;XIA Li-Li;YUE Qing-Hua;LIU Zhang-Jun(Jiangxi Provincial Institute of Water Science,Nanchang 330029,China;College of Hydraulic and Environmental Engineering,China Three Gorges University,Yichang 443002,China;Engineering Research Center of Eco-Environment in Three Gorges Reservoir Region,Ministry of Education,China Three Gorges University,Yichang 443002,China;Power China Huadong Engineering Corporation Ltd.,Hangzhou 311122,China)

机构地区：[1]江西省水利科学研究院,南昌330029 [2]三峡大学水利与环境学院,宜昌443002 [3]三峡库区生态环境教育部工程研究中心(三峡大学),宜昌443002 [4]中国电建集团华东勘测设计研究院有限公司,杭州311122

出　　处：《气候变化研究进展》2020年第3期325-335,共11页Climate Change Research

基　　金：国家自然科学基金项目(51909112,51779279);三峡库区生态环境教育部工程研究中心(三峡大学)开放基金(KF2019-12,KF2019-13);中国电力建设股份有限公司项目(DJ-ZDZX-2016-02);江西省节水型社会建设试点后评价研究(XM-0000000606)。

摘　　要：使用水文模型进行气候变化对水文水资源的影响评估时,通常假定水文模型参数在历史和未来时期是固定不变的,并未区分水文模型参数受环境变化的影响差异。针对该问题,采用集对分析法以年降水量和年径流量为依据进行丰、平、枯水年的划分,利用SWAT分布式水文模型分别率定不同水平年的模型参数,探究月尺度模型参数的非稳定性对年际年内径流的影响。研究结果表明,分别以年降水量和年径流量划分丰、平、枯水年时,1961—2010年这50年中有70%的年份为同枯或同丰。如果仅选择某一水平年(丰、平、枯)背景下的资料进行建模,将导致模拟精度较丰、平、枯整体考虑时降低,且是以偏大为主。用同样的水文资料与代表不同环境条件下模型参数推求径流过程时,从年际变化角度看,使用某一水平年(丰、平、枯)背景下的参数,总体上模拟径流偏湿润,使得丰水年比重增加、枯水年比重降低;从年内分布看,受不同水平年参数影响,径流的集中度减小,年内分配更加均匀,径流集中期延迟。本研究可为提高未来变化环境下水文模拟可靠性提供参考,对于应对变化环境下的水资源适应性管理具有重要意义。In the assessment of the impact of climate change on the hydrology and water resources by using hydrological models, it is generally assumed that the parameters of hydrological models are stationary through the historical and future periods without distinguishing the differences of the parameters of hydrological models under the influence of environmental changes. To solve this problem, Set Pair Analysis(SPA) method was used to partition wet, medium and dry years based on annual total precipitation and annual average runoff. The parameters of SWAT distributed hydrological model for different years of recurrence interval were calibrated by to explore the influence of parameters instationarity on the intra-annual and inter-annual distribution patterns of runoff. The results showed that when the times series were respectively classified into wet, medium and dry years by annual total precipitation and annual average runoff, 70% of the 50 years were of the same type(wet or dry). If only the data of an individual certain type(wet, medium and dry) are selected for calibration, the simulation accuracy will be lower than that of the overall consideration of wet, medium and dry years, mainly due to the simulated runoff is overestimated. When the same hydrological data and model parameters under different environmental conditions are used to derive the runoff process, from the perspective of inter-annual variation, the parameters of a specific type(wet, medium and dry) are used to generally generate the wetter results, which increases the proportion of wet years and decreases the proportion of dry years. While from the perspective of intra-annual variation, influenced by the parameters of different types, the concentration of runoff decreases which means the distribution tends to be more uniform, and the runoff concentration period is delayed. The results can provide reference for improving the reliability of hydrological simulation under changing environment in the future and are of great significance for the adaptive mana

关 键 词：水文模型参数 集对分析法 非稳定性 集中度 集中期 

分 类 号：P333[天文地球—水文科学]