SCS模型在山区小流域山洪灾害预报预警中的适用性分析

预报预警是山洪灾害防治中重要的非工程措施,是防灾减灾救灾的关键环节。选取分别代表秦巴山地水资源保护区、西南地震作用区、喀斯特地区、黄土高原超渗产流区、东南沿海台风影响区的湖北省丹江口官山河流域、四川省都江堰白沙河流域、贵州省望谟县望谟河流域、陕西省子洲县岔巴沟流域、广东省高州市马贵河流域这5个典型流域,通过构建改进的SCS模型进行历史山洪模拟,探讨该模型在小流域暴雨山洪预报预警中的广泛适用性。结果表明:基于SCS模型构建的小流域暴雨山洪预报方案中官山河、白沙河、望谟河、马贵河方案均为乙级方案,可用于正式发布预报;岔巴沟为丙级方案,可用于参考性预报。SCS模型在5个典型示范区的洪峰流量预报精度普遍在80%左右,高于我国山洪灾害洪峰流量预报平均水平(40%)。总体来说,SCS模型对各类典型流域的山洪模拟效果较好,能够用于山洪灾害预报预警业务。SCS模型结构简单,参数少,特别适用于无资料地区推广应用。

Applicability of SCS Model in Flash Flood Forecasting and Early Warning

As an important non-engineering measure of flood disaster control, forecasting and early warning is the key link of disaster prevention, reduction and relief. In this paper we select five typical small watersheds to examine the applicability of SCS model. The five watersheds are: Guanshan River Basin of Hubei Province, Baisha River Basin of Sichuan Province, Wangmo River Basin in Wangmo County of Guizhou Province, Chabagou basin in Zizhou County of Shaanxi Province, and Magui River Basin in Gaozhou City of Guangdong Province, representing the water resources protection area in Qinling-Daba mountains,the southwest earthquake action area,the karst area, the loess plateau area, and the southeast coastal typhoon-affected area, respectively. Results reveal that the forecast schemes for Guanshan River, Baisha River, Wangmo River and Magui River are above the second class which can be used in practical flood forecasting, and the scheme of Chabagou basin achieves the third class which can be taken as reference for practical flood forecasting. The prediction accuracy of SCS model in the five typical demonstration areas is generally around 80%, higher than the average level (40%) of flash flood peak forecast in China. Generally speaking, SCS model has good applicability in flash flood simulation for various types of typical basins. It can be used for flash flood forecasting and early warning. SCS model has simple structure with small number of parameters, especially suitable for ungauged basins and easy to be popularized and applied.