机构地区: 中国科学院海洋研究所
出 处: 《海洋技术》 2008年第4期63-67,共5页
摘 要: 主要结合P-J模式和全通道AMSR-E卫星数据,利用两种BP神经网络方法对热带海区蒸发波导高度进行了反演研究。(1)利用BP神经网络反演得到的气象参数通过P-J模式计算蒸发高度;(2)利用卫星AMSR-E数据直接反演蒸发波导高度。两种方法得到的与浮标实测参数计算得到的蒸发波导高度之间的相关系数相当,都为0.82左右,均方根差后者比前者小,分别为2.64m和1.89m。利用后一种网络直接反演了南海地区的蒸发波导高度,平均结果与其他文献结果基本一致。 Two methods based on the multi-parameter neural network and P-J mode are proposed to retrieve the evaporation duct height from the AMSR-E data. One is to retrieve the sea surface parameters and then calculate the evaporation duct height, the other is to retrieve the evaporation duct height directly. The correlation coefficients between estimated evaporation duct height calculated by P-J mode and from AMSR-E data and buoy measurement parameters are about 0.82 and root-mean-square errors of them are 2.64 meters and 1.89 meters respectively. The latter is better than the former. The evaporation duct height in South China Sea is retrieved from AMSR-E data using the later method and the averaged result is consistent with other statistical results.
领 域: [天文地球]