Crude oil reserve estimation: An application of the autoregressive integrated moving average (ARIMA) model

Abstract

The oil industry has used decline curve analysis with limited success in estimating crude oil reserves and in predicting future behavior of oil and gas wells. This paper, therefore, explores the possibility of using the Autoregressive Integrated Moving Average (ARIMA) technique in forecasting and estimating crude oil reserves. The authors compare this approach with the traditional decline method using real oil production data from twelve (12) oil wells in South Louisiana. The Box and Jenkins methodology is used to develop forecast functions for the twelve wells under study. These forecast functions are used to predict future oil productions. The forecast values generated, are then used to determine the remaining crude oil reserves for each well. The accuracy of the forecasts relative to the actual values for both ARIMA and decline curve methods is determined by various statistical error analyses. The conditions, under which one method gives better results than the other, are fully investigated. In almost all the cases considered, the ARIMA method is found to perform better than the decline curve method.