This study reviews historical crude oil production for 67 countries. These countries supply virtually all the world's crude. We used historical production data with a non-linear least squares curve fitting method to automatically determine the parameters of Hubbert's model.
We arbitrarily classified the goodness of fit of the production trends of all countries into three categories (Good, Fair and Poor). A "Good Fit" occurs when the dimensionless minimum root mean squares (D) is 14.49%, the mean of all D's. Forty-two of the sixty-seven countries exploited production trends fell into the "Good" category. A "Fair Fit" occurs when the D is between the mean and 24%. This category is given to production trends that have more than slight deviations from the Hubbert model. Seventeen countries fell into this category. A "Poor Fit" occurs when the D 24%. Eight countries fell into this category.
An attempt is made to explain deviations of actual data from the Hubbert model. Our observations indicate that political, economic, and exploration policy may be important causes of these deviations. These causes are external to the physical nature of petroleum reservoirs whose behavior is controlled by reservoir mechanics and, thus, difficult to predict with any model.
Hubbert's model determines reserves that are higher than "booked" reserves. The reason for this is that Hubbert model reserves implicitly include the effects of continuing exploration and discovery. The paper ends with a discussion of non-mechanical factors that contribute to the parameters of the Hubbert model. These factors include Sedimentary volume or area, petroleum concentration, exploration efficiency, Economic incentives, political events, and technology.
Our analysis indicated that ultimate world recoverable oil will be about 1.76 trillion bbls. Of this amount, about one trillion bbls remains to be produced at the end of 1995 (i.e. world oil remaining reserves). OPEC has a remaining reserves of 710 MMMSTB which is 71% of world reserves. Therefore, if this estimate is correct and assuming a scenario of 60 MMSTB world constant daily production, then the supply of crude oil will last only 45 years.
There exists a number of different methods to evaluate oil production trends and forecasts of ultimate oil reserves. These approaches differ in the degree of structural details and the weight that they give to geological and economic factors in modeling of oil supply. They range from engineering techniques to econometric, and finally to curve-fitting or statistical procedures using simple models. Among the simple models is the Hubbert model. It has proved particularly popular in forecasting ultimate recoverable oil reserves in the US and elsewhere due to the availability of its required data. However, the use of the Hubbert approach is subject to a number of shortcomings which will be discussed later.
In 1956, M. King Hubbert, then with the United States Geological Survey, made an analysis and forecast of United States oil production trends which proved to be remarkably accurate. His forecast that the US oil production rate would peak in 1970 was remarkably accurate.
In 1956, Hubbert predicted that oil production in the lower forty-eight states would peak between 1965 and 1970 and decline continuously thereafter. Hubbert developed his prediction by fitting historic production data to a symmetric bell-shaped curve. History proved Hubbert's prediction to be remarkably accurate. Production in the lower forty-eight states peaked in 1970 and has since declined fairly steadily. In his 1962 analysis, Hubbert fitted logistic curves to cumulative production and discoveries and estimated that 171 billion barrels of oil would be recovered from fields in the lower forty-eight states which was substantially lower than that made by the US Geological Survey (USGS) and lower than concurrent estimates made by most industry analysts. Their estimates ranged from 590 to 660 billion barrels.