Canonical correlation analysis as a statistical method to relate underwater acoustic propagation and ocean fluctuations

Abstract

Numerical models are currently used to understand how environmental fluctuations impact acoustic propagation. Such a process can be tedious in complex fluctuating environments. This letter proposes a complementary approach based upon canonical correlation analysis (CCA) to determine statistical relationships between two sets of observed acoustic and oceanographic variables. It is shown, as an example, how CCA puts forward the impact of external and internal tide on shallow water propagation. Results are consistent with the physical understanding of tide impact on acoustic propagation. They encourage the use of CCA for complex studies, in particular, for environments fluctuating under several environmental phenomena.