Data Assimilation

In many applications the environment being monitored through the EPCL will be subject to patterns of behavior dominated by diurnal fluctuations and fast transient disturbances set against background longer-term trends over weeks, possibly months. The data are inevitably subject to gaps, outliers, and noise. Our primary means of extracting and interpolating "clean" signals from the high-volume time-series of the data base is through real-time filtering algorithms of the MicroCAPTAIN and TVP (MATLAB®) software packages produced by the Centre for Research on Environmental Systems and Statistics at the University of Lancaster, England. Interpretation of these preprocessed data are the subject of current research, for example, on identifying transport and degradation behavior of a nitrifying biomass from the diurnal signal components extracted from the ammonium-N time-series for crude sewage and clarified effluent. Other studies are progressing towards schemes for assimilating such data (in real-time) into industry-standard simulation models, specifically the IAWQ model of the activated sludge process.