基于集合预报的淮河流域洪水预报研究

以湖北省漳河水库为例，选取2008—2017年的降雨和流量数据，分析了新安江模型和SCS模型基于分钟降雨条件下的径流模拟情况，同时结合2020年漳河水库6月27—29日的一次洪水过程，对比分析了不同流域洪水过程中，两个模型的模拟特点。结果表明：（1）新安江模型的效率系数高于SCS模型10%~20%，洪峰相对误差和峰现时差两者相当，但SCS模型可用于无水文资料地区的径流模拟。（2）对漳河水库历史个例的模拟，当时间分辨率低于阈值时，新安江模型和SCS模型计算的效率系数会趋于稳定，且新安江的分辨率阈值要小于SCS模型。在时间分辨率为15 min左右时，两个模型的洪峰相对误差和峰现时差达到最小值。（3）在2020年洪水个例的模拟中，两个模型在效率系数和洪峰相对误差上均存在最优时间分辨率阈值为10 min。

A review of USDA SCS curve number method

Taking Zhanghe reservoir in Hubei province as an example, the runoff simulations of Xin'anjiang model and SCS model under the condition of minute precipitation were analyzed based on the precipitation and discharge data from 2008 to 2017. At the same time, combining with a event at Zhanghe reservoir from June 27 to 29 in 2020, the simulation characteristics of the two models in the extreme flood event in different basins were compared and analyzed. The result show that:(1) The efficiency coefficient of Xin'anjiang model was 10%-20% higher than that of SCS model, and the relative errors in flood peak and the peak time difference are comparable. But SCS model can be used for runoff simulation in areas without hydrological data. (2) In the simulation of the Zhanghe reservoir historical case, the time resolution threshold existed in the efficiency coefficient of Xin'anjiang model and SCS model, and the resolution threshold of Xin'anjiang model was smaller than that of SCS model. When the resolution was about 15 min, the relative error of flood peak and the peak time difference of the two models can reach the minimum value. (3) In the simulation of the flood in 2020, the two models had the optimal time resolution threshold that is 10 min both in efficiency coefficient and relative error of flood peak.