摘要
Abstract
This paper analyzed the performance of a support vector machine(SVM)and a multimodal SVM that considered temperature and humidity in inverting sea surface wind speed from cyclone global navigation satellite system(CYGNSS)data,so as to explore their potential application for ocean wind field monitoring.Sea surface wind speed data inverted from CYGNSS were selected for analysis,and the wind speed data from the fifth generation of European Centre for Medium-Range Weather Forecasts(ECMWF)atmospheric reanalysis(ERA5)were validated,as well as buoy observation data.The results show that the multimodal SVM model significantly improves wind speed inversion accuracy,with R2 values exceeding 0.85 at all tested buoy locations,and prediction errors are notably lower compared to the standard SVM model,particularly under high wind speed conditions.Additionally,spatial distribution characteristics of wind speeds within the study area in spring,summer,autumn,and winter are analyzed using the optimal model.The results indicate significant seasonal variations in the spatial distribution of wind speed,with higher wind speeds in the northern region during spring,concentrated wind speeds in the eastern and central regions during summer,maximum wind speeds in the central region during autumn,and overall maximum wind speeds in winter.The application of the multimodal SVM model in sea surface wind speed inversion can effectively enhance prediction accuracy,making it highly significant for the accurate monitoring of ocean wind fields.关键词
多模态支持向量机(SVM)/旋风全球导航卫星系统/风速/温度/湿度Key words
multimodal support vector machine(SVM)/cyclone global navigation satellite system/wind speed/temperature/humidity分类
测绘与仪器
