We consider the results of the statistical analysis using the methods of the principal components and canonical coherences applied to the processing of long (1986–2005) time series of hydrogeochemical observations at the flowing wells and springs in Kamchatka. The time-frequency diagrams of the evolution of informative statistics characterizing the collective behavior of multidimensional hydrogeochemical time series are constructed, and the time intervals and frequency bands where the synchronization signals (Lyubushin, 2007) appear are identified. The features of their occurrence are analyzed in comparison to the strong (Mw = 6.6–7.8) local earthquakes. It is found that such signals in the measurements of some multidimensional time series can arise both before and after earthquakes, i.e. these signals have a precursory (P2) and postseismic (P3) character.