Real-time prediction of extreme ambient carbon monoxide concentrations dueto vehicular exhaust emissions using univariate linear stochastic models

Historical data of the time-series of carbon monoxide (CO) concentration wa s analysed using Box-Jenkins modelling approach. Univariate Linear Stochast ic Models (ULSMs) were developed to examine the degree of prediction possib le for situations where only a limited data set, restricted only to the pas t record of pollutant data are available. The developed models can be used to provide short-term, real-time forecast of extreme CO concentrations for an Air Quality Control Region (AQCR), comprising a major traffic intersecti on in a Central Business District of Delhi City, India. (C) 1999 Elsevier S cience Ltd. All rights reserved.