| In order to make reliable long-term predictions of human activity on our environment, one must first understand the natural variations of our climate. Here we are able to disentangle the atmospheric signals through the use of the Empirical Mode Decomposition (EMD) method. It is a recently developed nonlinear and non-stationary time series analysis tool. It is used and extended here to investigate the time evolution of signals in the stratosphere and troposphere. The timeseries of atmospheric changes in the stratosphere and troposphere are ideal for this analysis because the data contains quasi-periodic, non-stationary and small amplitude embedded signals. Using the EMD method we decompose the geopotential height in the stratosphere and the troposphere to obtain 5 modes and a trend. The changes in these modes can be traced through the stratosphere and the troposphere and here we pay particular attention to the solar cycle mode. We demonstrate that it has a robust and statistically significant signal throughout the lower atmosphere. The second mode, which is related to the equatorial Quasi-Biennial Oscillation (QBO), is also of interest. Using an independent type of analysis, with the Northern Annular Modes (NAMs) as a spatial filter, it is shown that an equatorial QBO signal can be found at the extratropical surface. This is consistent with the EMD results. In addition, the vertical connection seen in the NAM between stratospheric and tropospheric anomalies is explored analytically using the linear quasi-geostrophic potential vorticity equations. |
