Detecting system state transitions in environmental time-series using non linear time series analysis

Environmental systems and time series emanating from such systems present a particular interest. System state transitions can occur in time and/or in space and the detection of such transitions can be particularly useful in the design related to such systems. Since the majority of physical systems present non linear behavior the use of appropriate tools is necessary. Recurrence Plots (RPs) and Recurrence Quantitative Analysis (RQA) along with Cross Recurrence Plots (CRPs) are some tools that permit to extract the underlying system dynamics [1-3]. In the present work we analyze time series from two different environmental systems. In the first case we study records of daily values of the Nestos river (Greece) water level at various measurement stations. Transitions from “periodicity” to “chaos” and “chaos” to “laminarity” are identified in time. In the second part we analyze temperature fluctuations in a horizontal round heated turbulent jet where instantaneous temperature time series were recorded at several points on the jet cross section. The temperature time series are analyzed in a first place using RQA. The variation of RQA measures is related with and interpreted via the transitions of the physical state of the fluid from the fully-turbulent flow near the jet centerline to the transitional flow near the boundary of the jet [9]. In a second phase the CRP analysis reveals correlations between the various parts of the jet.