OIL AND GAS 



489 



seabed, have yet to be adequately explored geo- 

 logically (Weeks, 1971). Our techniques for identi- 

 fying hydrocarbon traps, or even those areas that 

 especially favor hydrocarbon accumulation, are very 

 imperfect. On both counts, the results of petroleum 

 exploration to date serve as an inadequate baseline 

 upon which to project results of future exploration. 

 Nevertheless, the results of past exploration provide 

 the most reliable single guide that we have. 



A decision must be made at the outset as to what 

 will be an acceptable minimal sample unit or base- 

 line if one is to attempt to use results of past ex- 

 ploration as a guide to future projections. For 

 example, if one were to estimate the ultimate re- 

 coverable hydrocarbon resources of the United 

 States on the basis of the productivity per square 

 mile of surface area (or per cubic mile of sedi- 

 mentary rock volume) of either the immensely pro- 

 ductive Los Angeles Basin or the totally unpro- 

 ductive Imperial Basin, both Cenozoic basins of 

 Southern California, the results in each case would 

 be grossly in error. In the same way, estimations 

 for unexplored regions would almost certainly be in 

 error if the regional historical statistical data from 

 either the aberrantly rich sedimentary areas or 

 volumes of the Middle East or the exceptionally 

 impoverished ones of Italy and the Adriatic Sea 

 were to be used as guides. Out of ignorance and 

 limited capability, one is tempted in such efforts to 

 select the base unit to be an area as thoroughly 

 explored and as large and diverse as possible, 

 thereby hoping that an average of such extreme 

 results as those from the Los Angeles and the Im- 

 perial Basins will somehow result and give a more 

 meaningful blend. As a practical matter, one works 

 with what is available, recognizing as he does that 

 a fruitful field for further research awaits signifi- 

 cant attack, and knowing that one thing needed is 

 a sound measure of the frequency distribution of 

 hydrocarbon productivity per unit of rock volume 

 for a large enough number of depositional basins 

 to provide a population curve representative for the 

 world. There are few "Los Angeles Basins" in the 

 world, and probably not many "Imperial Basins." 

 Somewhere between these two extremes is the peak 

 of a geologically composite population curve sta- 

 tistically applicable to totally unexplored basins and 

 applicable with measured discrimination to geo- 

 physically analyzed basins or those only partly 

 explored by drilling. 



The conterminous United States ("lower 48"), be- 

 cause of its political and economic unity, its com- 

 paratively long and intensive history of petroleum 

 exploration and exploitation, its relatively large size 



(5.3 percent of the continental area of the earth), 

 its large volume of Phanerozoic rocks (estimated 

 to be eOxlO"* km^ by Gilluly and others (1970) ), and 

 its great geological diversity provides ". . . the most 

 meaningful statistical sample for appraising possi- 

 ble ultimate potential" (Hendricks, 1965, p. 3). By 

 starting with the quantity of petroleum found in the 

 conterminous United States to the year 1961 and 

 "the fraction of the potentially productive rocks 

 that [had] been explored to [that] date," and by 

 estimating the incidence of oil in the unexplored 

 fraction ". . . as determined by comparison with 

 the explored part," Hendricks arrived at an esti- 

 mate of 1,600 billion barrels of oil originally in place 

 for the "lower 48." He estimated further (p. 9) that 

 a total of 1,000 billion of those barrels could be found 

 if economic incentives were sufficient to stimulate 

 additional thorough but profitable exploration, and 

 that 400 billion barrels would be ultimately recov- 

 erable (of which 68 billion barrels had been pro- 

 duced through 1961, and 97 billion today). By con- 

 trast, starting with the same basic historical records 

 and data, but using a method of mathematical analy- 

 sis that is in his words intrinsically ". . . compara- 

 tively insensitive to error in the estimates of [ulti- 

 mate cumulative production] . . . ," Hubbert in a 

 series of papers (1956, 1962, 1966, 1967) concluded 

 that the amount of ultimately recoverable petroleum 

 in the United States is between 150 billion barrels 

 for the conterminous states and about 200 billion 

 barrels for the entire continental United States (in- 

 cluding Alaska and the outer continental shelves). 

 Hubbert's method is one of integral logistic curve 

 fitting. It is related only statistically to rocks and 

 geologic parameters, and to those only through a 

 variety of economic and technologic factors. Sup- 

 posing that human beings will be doing in the future 

 as they have in the past, so far as petroleum ex- 

 ploration is concerned both technologically and from 

 an investment viewpoint, Hubbert's approach is 

 unassailable. 



Hendricks was neither the only nor the first per- 

 son to attempt to estimate undiscovered petroleum 

 resources of the United States through an analysis 

 of unexplored rocks based on analogy with historical 

 results in explored areas. Twenty-three years earlier, 

 Pratt (1942) published estimates of the ultimately 

 discoverable and (presumably) producible petroleum 

 resources of the United States and the world of 

 100 billion and 500 billion barrels, respectively. The 

 Pratt figures came from estimates made by Eugene 

 Stebinger and L. G. Weeks ". . . based on studies . . . 

 made of the records of past discoveries and of the 

 nature and extent of sedimentary basins over the 



