基于高低潮的风暴增水人工神经网络预报模型  

Neural network forecast model of storm surge elevation based on high tide and low tide

作　　者：王如云[1] 雷磊[1] 占飞[1] 周钧[2] 

机构地区：[1]河海大学港口海岸与近海工程学院,江苏南京210098 [2]河海大学水文水资源学院,江苏南京210098

出　　处：《海洋预报》2014年第6期23-27,共5页Marine Forecasts

基　　金：中国江苏省水利科技重点项目(2010500312);国家自然科学基金(40906048);中央高校基本科研业务费专项(2014B06314)

摘　　要：针对只有高低潮数据的情况,利用人工神经网络建立起一种预报当前台风时刻后第一个高潮时增水的模型。该模型选取台风在当前时刻、前6 h、前12 h、前18 h的中心经度、纬度、最大风速、中心气压以及当前时刻前第一个高潮时刻的风暴增水为输入单元。台风当前时刻后第一个高潮时刻风暴增水为模型输出单元。利用历史资料形成的规范化后的模式对,对模型进行训练,训练成功后,结合台风因子预报模型,即可用于风暴增水的预报。经过长江口高桥站高低潮实测资料的检验,结果表明该模型提取到了风暴增水效应,说明该模型可用于风暴增水的预报。In view of only high and low tide water data being available, a model applied to predict the first climax after typhoon moment is established by using artificial neural network. The longitude, latitude, maximum wind speed and air pressure of the typhoon centerat the time of current moment, 6 hours before, 12 hours before, 18 hours before and the first climax before the time of typhoon current moment are selected as the input traits of the model, and the first climax after the typhoon current moment is as the output unit. Using standardized modes formed by the historical data, and combining with the typhoon factor prediction model, the established model can predict storm surge after the success of the training. The modeled storm surge agree with the results of the observed data of high and low tide at Gaoqiao station located in the Changjiang River estuary, which indicates that this model can be applied to storm surge forecast.

关 键 词：高低潮 BP人工神经网络 台风因子 风暴增水 

分 类 号：P731.23[天文地球—海洋科学]