APPLICATION OF AUTOREGRESSIVE FORECASTING TECHNIQUE TO PETROLEUM PRODUCTION IN NIGERIA

Abstract

The decision problem concerning forecast of gas and crude oil productions in Nigeria oil industry is the main focus of this research study. Thirty and twenty three years respectively time series of gas and oil production were obtained from NNPC records and analyzed by the autoregressive forecasting technique and multiple linear regressions.

A correlogram developed enabled us to build predictive models for gas and oil (Equation 7 and 8 respectively) for forecasting future production.

The actual gas produced for thirty years fall within range of 12996MCM and 43636MCM while forecasted value fall within the range of 27262.3MCM and 94701.3MCM.

The actual oil produced for twenty-three years fall within range of 450.9Mbbl and 855.7Mbbl while forecasted fall within the range of 64.1Mbbl and 984Mbbl.

The mean of actual gas produced for thirty years was 26896.6MCM and the forecasted value was 30575.3MCM.

The mean of actual oil produced for twenty- three years was 634.4Mbbl and the forecasted value was 620.6Mbbl.

Durbin Watson rests for significance carried out suggests that the models developed are reliable. Each model can be use to predict at 0.5 level of significance.