Trend Analysis of Sea Surface Temperature and Sea Level Pressure Employing a Pseudo-EMD Method

Empirical mode decomposition (EMD) is an emerging signal processing techniques that has found many applications. EMD decomposes a one-dimensional (1D) signal into hierarchical components known as intrinsic mode functions (IMFs) and a residue, based on the local properties of the signal. The first IMF depicts the highest local oscillations, while the residue depicts the trend of a signal/data. Application of interpolation methods causes huge computation time and other artifacts, which limits the use of EMD for many real life signals. This paper introduces a pseudo EMD (P-EMD) method that replaces the interpolation step by direct envelope estimation using order statistics filters to overcome several difficulties. The P-EMD method is utilized to determine and analyze the trend of various signals including sea level pressure and sea surface temperature.