基于神经网络的卫星高度计海面风速反演算法  

Neural Network Method for Retrieving Sea Surface Wind Speed from Satellite Altimeter

作　　者：任建奇[1] 严卫[1] 王迎强[1] 

机构地区：[1]解放军理工大学气象学院,江苏南京211101

出　　处：《海洋技术》2007年第2期47-50,共4页Ocean Technology

摘　　要：通过对TOPEX/Poseidon高度计资料与NDBC浮标实测数据进行时空匹配处理,得到同步数据集,利用人工神经网络方法试验得到海面风速反演算法,并与业务运行的M CW算法进行分析比较,指出考虑波浪状态影响因素的神经网络算法在均方根误差和对称性方面的优越性。研究表明利用神经网络方法反演海面风速是可行的。A neural network method considering the impact of sea state is presented to retrieve sea surface wind speed from TOPEX/Poseidon altimeter data. Training data set and test data set come from TOPEX data collocated pairs with NDBC buoy data. Comparing with operational MCW algorithm, this method performs well in RMSD and symmetry. Research results show that the algorithm using neural network method is reliable in sea surface wind speed retrieving.

关 键 词：卫星高度计 海面风速 B-P网络 反演算法 

分 类 号：P425[天文地球—大气科学及气象学]