基于一阶谱信息的单站高频地波雷达风速提取  

Extraction of Wind Speed From the First Order Spectrum of Single High Frequency Surface Wave Radar

作　　者：楚晓亮[1] 张晓愉[1] 张荣真 王曙曜 

机构地区：[1]中国海洋大学信息科学与工程学院,山东青岛266100 [2]中船重工鹏力(南京)大气海洋系统有限公司,江苏南京211106

出　　处：《中国海洋大学学报（自然科学版）》2017年第2期122-127,共6页Periodical of Ocean University of China

基　　金：海洋公益性行业科研专项项目(201505002);国家自然科学基金项目(61671166)资助~~

摘　　要：单站高频地波雷达反演风速通常需要利用二阶谱信息,但利用二阶谱信息进行风速反演的距离较小,不能满足远距离风速反演的需求。本文首先对高频地波雷达一阶谱与风速间的相关性进行了分析,在此基础上,基于一阶谱信息,利用极限学习机(ELM)方法对不同站点和不同频率的雷达获得的实测数据进行了风速反演。结果表明,利用极限学习机方法对实测数据获得的风速结果能够体现风速变化的趋势,两批数据反演结果的相关系数分别是0.46和0.42,均方根误差分别为1.93和1.89m/s。尽管其相关系数较低,但均方根误差可以接受,说明利用单站雷达一阶谱信息来进行风速反演是可行的。The information of second order spectrum is needed when the wind speeds are retrieved from the single radar backscatter. However, the requirement of wind speed inversion in the far range is not satisfied because that the detection range of second order spectrum is short. In this paper, the correla- tion between the first-order spectrum from high frequency surface wave radar and the wind speed is ana- lyzed, and wind speed inversion from the high frequency radar first-order spectrum data with the differ- ent frequency and site are carried out based on the extreme learning machine. It is shown that the inver- sion results of wind speeds can show the tendency of changes in wind speed of in-situ, the correlation co- efficients are low but the root mean square errors can be accepted. This shows that it is feasible to ex- tract the wind speed from the monostatic high frequency radar first order spectrum.

关 键 词：高频地波雷达 一阶谱 风速反演 极限学习机 

分 类 号：P716[天文地球—海洋科学]