This study presents a variational mode decomposition (VMD)-based approach to analyse wide-area (WA) measurements based signals. The commonly used empirical mode decomposition (EMD) has limitations such as sensitivity to noise and sampling rate. However, VMD is an entirely non-recursive algorithm, where the modes are extracted concurrently. These modes help in extracting dynamic patterns of different power system disturbances. The performance of the proposed scheme is extensively validated on the IEEE-39 bus New England test system. The modes generated and the frequency deviation contours of the disturbances including generation loss, fault and line outage are assessed using VMD and the results provide improved performance in terms of decomposition quality compared with the existing EMD technique. Furthermore, a data-mining model known as decision tree is used to classify different power system disturbances based on intrinsic mode functions generated through VMD. The suggested method shows improved decomposition quality and classification accuracy. Thus, the proposed scheme is a potential candidate for improving WA situational awareness along with a post-mortem analysis of real events occurring in a power system. © The Institution of Engineering and Technology.