摘要
以伊通河流域为研究区,运用实测数据建立SWAT水文模型。在LH-OAT参数敏感性分析的基础上,应用Bootstrap法对SWAT水文模型敏感参数的不确定性进行了分析,获取了5个敏感性参数的不确定区间及边缘分布,结合变异系数对参数的不确定性程度进行了评估。结果表明:SOL_AWC呈负偏态分布,其余4个参数近似服从正态分布;在5个敏感性参数中,CN2、ESCO的参数不确定性程度最高。进一步分析了参数不确定性对模型模拟结果不确定性的贡献,结果表明:参数的不确定性对模拟结果的不确定性贡献较大,58%的月径流观测数据落入模拟的95%置信区间内。然而在降雪期参数的不确定性对模拟结果的不确定性贡献较小,有8.3%的观测数据落入模拟的95%置信区间内。研究成果可为bootstrap法应用于水文模型参数不确定分析以及SWAT模型在伊通河流域水文模拟方面提供参考。
Working in Yitong Basin, we set up SWAT model by using observed data. Based on parameter sensitivity analysis by LH- OAT, we analyze uncertainty of SWAT model sensitivity parameters using Bootstrap method. The uncertainty ranges and marginal distributions of 5 sensitivity parameters are obtained. The calculated variation coefficients estimate parameter uncertainty. The re- suits show that the marginal distribution of SOL_ AWC is a negative skewed distribution, and others are close to normal distribu- tion. Among 5 sensitivity parameters, CN2 and ESCO have the greatest uncertainty. Analyzing the effects of parameter uncertainties on model simulation further reveals that the parameter uncertainties have great impacts on the results of the model simulation 58% of the monthly observed runoff data fall inside the 95% simulation confidence interval. However, the parameter uncertainties have little impacts on the outputs of the model simulation in snowy period. 8. 3% of runoff observations is within 95% confidence inter- val. The research results can provide references for parameter uncertainty of hydrological model using bootstrap method and hydro- logical simulation based on SWAT model in Yitong watershed.
出处
《中国农村水利水电》
北大核心
2016年第10期95-99,共5页
China Rural Water and Hydropower
基金
国家水体污染控制与治理科技重大专项(2012ZX07201-001)
