188宝金博页面版

J． Lake Sci．(湖泊科学)，2017，29(6):1520-1527DOI 10. 18307/2017. 06242017 by Journal of Lake SciencesBMA 集合预报在淮河流域应用及参数规律初探*刘开磊 1 ，胡友兵 1 ，汪跃军 1 ，王秀庆 2(1:淮河水利委员会水文局(信息中心)，蚌埠 233000)(2:沂沭泗水利管理局水文局(信息中心)，徐州 221000)摘 要:以淮河流域吴家渡水文站作为试验站点，采用基于贝叶斯平均法(BMA)的集合预报模型处理来源于马斯京根法、一维水动力学方法、BPNN(Back Propagation Neural Network)的预报流量序列，通过分析 BMA 的参数以及其预报结果，对各方法在淮河典型站点流量预报中的适用性进行验证与分析． 经 2003—2016 年 19 场洪水模拟检验可知，BMA 模型能够有效避免模型选择带来的洪水预报误差放大效应，可以提供高精度、鲁棒性强的洪水预报结果． 通过进一步比较各模型统计最优的频率与 BMA 权重值之间的相关性，发现权重值不适用于对单场洪水预报精度评定，而适用于描述多场洪水预报中，模型为最优的统计频率;基于大量先验信息，提前获取 BMA 的权重等参数，将是指导模型选择、降低洪水预报不确定性、改进洪水预报技术的有效手段．关键词:集合预报;洪水预报;不确定性;权重;淮河流域;贝叶斯平均法Performance and parameterization of the BMA model applied in the Huaihe Ｒiver BasinLIU Kailei 1 ，HU Youbing 1 ，WANG Yuejun 1 ＆ WANG Xiuqing 2(1:Huaihe Ｒiver Commission of the Ministry of Water Ｒesources，Bengbu 233000，P．Ｒ．China)(2:Water Conservancy Administration of the Yi-Shu-Si Ｒiver Basin，Xuzhou 221000，P．Ｒ．China)Abstract:In this study，the BMA (Bayesian Model Averaging)method is used to deal with forecasts derived from three differentflood routing models，i．e． the Muskingum，the one-dimensional hydrodynamic routing model and the BPNN (Back PropagationNeural Network)． The numerical experiments are processed at the Wujiadu hydrological station in the Huaihe Ｒiver Basin as a testsite． By studying the BMA method parameters and ensemble forecast results the applicability of each method in the Huaihe Ｒiverstation flow forecast is verified and analyzed． According to the results from our experimental tests among 19 flood events，from 2003to 2016，it is concluded that the BMA ensemble forecasting method can effectively avoid the flood forecast error amplificationcaused by uncertainty underlying in model selection and can provide high accuracy and robust flood forecasting result． After compa-ring the statistically optimal frequency and BMA weight value of each model，it is concluded that the BMA weight parameter reflectsthe probability that the model is optimal in the mean sense and is not suitable for evaluating the technicity of the model in one singleflood event． It is an effective method to guide the model selection，reduce the uncertainties of flood forecasting and improve theflood forecasting techniques．Keywords:Ensemble forecast;flood forecast;forecast uncertainty;weight;Huaihe Ｒiver Basin;Bayesian Model Averaging随着水文学理论方法发展和计算机技术的进步，洪水演进模型已经成为水情业务部门开展河道洪水预报的重要技术手段［1-2］ ． 但是受河道洪水汇流过程复杂性以及模型自身结构特点的影响，不同模型在各场次洪水的预报实践中，往往表现出明显差异． 在洪水预报中，如何降低模型选择带来的不确定性，以及如何在保证预报精度与结果可靠性的同时，为防汛调度提供更为丰富的水情信息，成为洪水预报领域亟待解决的问题．目前国内外常采用动力波以及扩散波理论描述河道一维非恒定水流运动． 在动力波理论中，一般以圣* 国家重点基础研究发展计划项目(2016YFC0402703，2016YFC0400909)资助． 2017-01-18 收稿;2017-03-23 收修改稿． 刘开磊(1988～)，男，博士，工程师;E-mail:bbqx88@ 126．com．