Wavelet ridge estimation of jointly modulated multivariate oscillations

Wavelet ridge analysis is a technique for estimating the time-varying properties of a modulated oscillatory signal from a contaminated observation. Here this technique is extended to the multivariate case, that is, to a set of N real-valued time series. The bivariate case is illustrated with an application to a set of freely-drifting oceanographic floats. A freely distributed software package implementing this algorithm is available online at http://www.jmlilly.net.