TRUNCATED SHIFTED PARETO MODEL


     For the purposes of this assessment, as in the previous USGS assessment
(Mast and others, 1989), a model of the size-frequency distribution of the 
population of oil and (or) gas accumulations was assumed. The Truncated Shifted
Pareto (TSP) model describes a "J-shaped" distribution in which ever-increasing
numbers of accumulations occur in successively smaller size classes. The 
distribution is called shifted because it has been statistically moved to have 
its origin at the minimum accumulation size, in this case 1 MMBO or 6 BCFG. The
TSP distribution is referred to as truncated because, for the purposes of
analysis, the distribution is cut off at the size of the largest accumulation 
in the distribution. For a detailed discussion of the TSP distribution, see 
Houghton and others (1993).

     An important use of the TSP distribution in this assessment was to
provide a guide to province geologists in their development of estimates of 
undiscovered accumulations. A TSP distribution was fit to the population of 
accumulations known from each play and, in chronological order of discovery, 
to the first third of the accumulations discovered, the second third discovered, 
and the last third. The results of these fitted populations were provided to
province geologists and review panels as source information regarding the changing 
size distribution of accumulations within the play as a function of time.

     The TSP distribution was also commonly used to model the field-size
distribution of the undiscovered population. Unless the province geologist 
had another specific model in mind, a TSP was fit to the estimated median 
size and to the estimated largest accumulation expected at a 5 percent probability 
within the postulated population of undiscovered accumulations, also considering 
the estimated limiting maximum size. The resulting TSP distribution was used to
determine the remaining fractiles of the size distribution of the undiscovered
population.

     Based on sizes and numbers of accumulations of oil and (or) non-
associated gas estimated as undiscovered in each play, resources of
each of these commodities were calculated using a Monte Carlo simulation 
technique and application of play risk. Estimates of undiscovered resources 
are presented as a range of values corresponding to probabilities of occurrence
in order to express the uncertainty inherent in assessment of unknown quantities. 
The input variables of accumulation sizes and numbers are themselves expressed 
as density functions of uncertain quantities. The resulting cumulative 
probability distributions represent the estimated quantity of undiscovered
resources-from these distributions, various fractiles (including the
low (F95), the high (F5), and the mean estimates) are obtained.

     Resources of gas associated with or dissolved in oil (associated-
dissolved gas) were derived through use of estimated GOR's as applied
to the calculated oil. Similarly, estimates of NGL were separately calculated 
for associated and non-associated gas by applying ratios provided by the 
estimators. Total gas and NGL at the play level were determined through summation.