Recursive estimation of model parameters with sharp discontinuity in non-stationary air quality data

Recursive method of time series filtering and smoothing based on the state–space concept provides a natural approach to the modeling of non-stationary environmental time series. The flexibility of this stochastic formulation allows for a suitable degree of variability in the estimated components, and in this paper we show how it can be extended for handling sharp changes or discontinuities in the model parameters. The approach is based on the time variable parameter version of the well known linear regression model and exploits the suite of recursive Kalman filtering and fixed interval smoothing (FIS) algorithms. The practical utility of the method is demonstrated by an example of modeling of the RSP levels during an episode event.