DYNAMIC MODELLING AND CORRELATION ANALYSIS OF EMISSION FROM MOTOR VEHICLES IN NIGERIA: A CASE STUDY OF ABEOKUTA OGUN STATE

Johnson Funminiyi Ojo

Department of Statistics,

University of Ibadan, Ibadan, Nigeria.

E-mail: jfunminiyiojo@yahoo.co.uk,

ABSTRACT

High traffic volume and traffic congestion on Nigerian roads have led to increase in the concentration of pollutants in the air and this has posed health risks for human population. In this study, we build appropriate time series models using some vehicular emission data obtained from Abeokuta, in Nigeria. Four pollutants namely, carbon monoxide (CO), carbon dioxide (CO₂), oxygen (O₂) and hydrocarbon (HC) were investigated. Correlation analysis was carried out on each pollutant to see if these pollutants were significant as time progresses. Trend models particularly linear and quadratic were fitted for each pollutants. Time series models were built for these pollutants following model building procedures. The CO, CO₂, O₂ and HC were significant as time progresses with an increasing trend. The fitted linear time series model for these pollutants was Autoregressive Integrated Moving Average (ARIMA) of different order and the non-linear counterpart was ARIMA bilinear(ARIMABL) of order one. ARIMABL performed better than ARIMA with a smaller residual variance and mean square error for forecast. With these models, appropriate measures should be taken by the relevant authorities to curb the danger the emission could cause to humans not only in Abeokuta but in Nigeria.

Keywords: Motor vehicle emission, Trend models, Time series models, Correlation, Nigeria