NoahMP-RAPID在高海拔山地流域的模拟检验与误差分析  

Evaluation and uncertainty attribution of the simulated streamflow from NoahMP-RAPID over a high-altitude mountainous basin

作　　者：王姝 郑辉[2] 林佩蓉 郑兴 杨宗良 Shu Wang;Hui Zheng;Peirong Lin;Xing Zheng;Zong-Liang Yang(State Key Laboratory of Operation and Control of Renewable Energy and Storage Systems,China Electric Power Research Institute Limited Corporation,Beijing 100192,China;CAS Key Laboratory of Regional Climate-Environment Research for Temperate East Asia,Institute of Atmospheric Physics,Chinese Academy of Sciences,Beijing 100029,China)

机构地区：[1]中国电力科学研究院有限公司新能源与储能运行控制国家重点实验室,北京100192 [2]中国科学院大气物理研究所东亚区域气候-环境重点实验室,北京100029 [3]The University of Texas at Austin

出　　处：《科学通报》2019年第4期444-455,共12页Chinese Science Bulletin

基　　金：国家重点研发计划(2018YFA0606004);国家电网公司总部科技项目;国家自然科学基金(41605062; 41375088)资助

摘　　要：高海拔山地流域水能资源丰富、山洪易发;但产汇流机制复杂、降水数据不确定性大,准确的山地水文过程模拟面临较大困难.本研究采用中国区域地面气象要素数据集CMFD(China Meteorological Forcing Dataset)驱动NoahMP陆面过程-RAPID河网汇流耦合模式,模拟青藏高原东侧大渡河干流逐日流量,并根据铜街子、龙头石流量站观测数据率定RAPID中的波速参数,检验了NoahMP中SIMGM和NOAH两种不同的产流过程参数化方案的模拟能力,评估了CMFD降水驱动数据的系统性偏差.研究发现,基于Philip入渗模型的NOAH方案优于基于TOPMODEL的SIMGM方案,能较为准确地模拟大渡河干流流量逐日变化,相关系数大于0.85、纳什系数约为0.3;对纳什系数的数学分解发现,纳什系数较低主要是由于模拟流量显著偏低造成的,若剔除系统性偏差,NOAH方案的纳什系数可提高至约0.7.模拟流量的系统性偏差主要来自于降水驱动数据;与雨量站观测和反演数据相比,CMFD显著低估了大渡河流域平均降水量,且其系统性偏差与高程有关,在低海拔地区高估,高海拔地区低估.本研究表明,使用NOAH产流方案的NoahMP-RAPID耦合模式对大渡河流域水文过程有较好的模拟能力,可进一步应用于水库调度优化;而提高雨量站密度、降低降水数据产品的系统性偏差是进一步改进高海拔山地流域水文过程模拟的关键.Accurate hydrological simulations in high-altitude mountainous basins are important for optimizing reservoir operations in hydropower production and flood control. However, the simulations experience substantial uncertainties from the sparse precipitation measurements over complex mountainous topography and the imperfect hydrological process representations over the heterogeneous surface. At the Daduhe river basin, which is a high-altitude mountainous river basin located in the eastern slope of the Tibetan Plateau, daily streamflow over the 2-year period of 2014 and 2015 is simulated using the Noah land surface model with multi-parameterizations(NoahMP) and Routing Application for Parallel computation of Discharge(RAPID) coupled model driven by the Chinese Meteorological Forcing Dataset(CMFD). Two different runoff parameterization schemes of NoahMP are used for the simulations: SIMGM, which is derived from TOPMODEL and parameterizes surface runoff using water table depth, and NOAH, in which surface runoff is parameterized using soil moisture content based on the Philip infiltration model. The simulated streamflow is evaluated at two mainstream gauges on the Daduhe river, Tongjiezi and Longtoushi, in terms of Nash-Sutcliffe Efficiency(NSE), correlation coefficient, bias, and standard deviation. The uncertainty of the simulated streamflow is attributed to different runoff parameterization in NoahMP, the flow wave celerity parameter of RAPID, and the forcing CMFD precipitation data. Results show that the NOAH runoff parameterization scheme outperforms the SIMGM scheme. NOAH satisfactorily reproduced the observed streamflow anomaly and flood timing at the two mainstream gauges, and the correlation coefficients between the simulated and observed streamflow are above 0.85. The NSE for NOAH is approximately 0.3. The mathematical decomposition of NSE reveals that the relatively low NSE value is mainly attributed to the significant streamflow bias. Without the bias, NSE can be improved to approximately 0.7. The bias shows

关 键 词：水文过程模拟 不确定性 陆面过程模式 河网汇流模式 径流 

分 类 号：P333[天文地球—水文科学]