摘要
本文利用Argo资料对经验正交函数(EOF)和GDEM模式拟合声速剖面的效果进行了比较。结果表明,两种方法都能较好地重构局地声速剖面。EOF方法的拟合精度与选取的EOF阶数有关,阶数越高,精度越高;GDEM模式的拟合精度介于EOF的前5阶到前10阶之间。EOF易受到样本数量和采样深度的限制,难以提供完整的声速剖面;而GDEM模式能够将剖面扩展到更深的深度,与仅有上层海洋采样数据的声速剖面相比,结合深海水团的信息可使拟合结果达到更高的精度。
Based on Argo data, a comparison is made between the results of fitting from the Empirical orthogonal functions（EOF） and the GDEM methods in describing the sound speed profile（SSP）. The comparison shows that both motheds can effectively reconstruct the SSPs, where the precision of EOF increases with the number of modes, while that of the GDEM lies within the range of 5 modes to 10 modes of EOF. Restricted by sampling, the EOF method is incapable of providing a integrated profile, but the result can be improved by the GDEM method, by which a better precision would be achieved with deep sea water mass information.
出处
《应用声学》
CSCD
北大核心
2010年第2期115-121,共7页
Journal of Applied Acoustics
