Identifying signals from intermittent low-frequency behaving systems

An approach is presented to recover a true signal from an intermittent regi me-like behaving noisy system given an estimate of the noise covariance C-N . The method is based on minimizing the noise probability distribution on i sosurfaces of the data probability distribution and involves the spectrum o f the matrix CDCN-1 where C-D represents an estimate of the data covariance within a given regime, In single channel, i.e.. one-dimensional, case the intermittent low-frequency time series is split according to the regime lev el. For each regime the obtained subsampled time series is first centred by removing its time mean and then projected onto a higher dimensional space using the method of the delay coordinates. The oscillation for each regime level is then obtained as the leading eigenvector of CDCN-1. In the multi-c hannel case the signal pattern is obtained in the same way locally in each cluster forming the data. The approach is first tested xith the Lorenz syst em yielding the correct oscillation. The method is then applied to the Paci fic 500-hPa geopotential height from a set of multidecadal integrations wit h the General Circulation Model (GCM) of the Hadley Centre, forced with obs erved Sea Surface Temperature (SST), in order to identify the nonlinear atm ospheric response associated with El Nino Southern Oscillation (ENSO).