One of the most important functions of the petroleum geologist is appraisal of prospects in advance of drilling. Almost invariably this involves predicting the amount of oil that may be discovered using only indirect evidence such as seismic maps and profiles. A widely used appraisal method, the Monte Carlo technique, requires geologic information that cannot be known prior to drilling, necessitating the use of estimates that may be highly uncertain. The results of Monte Carlo analysis may be misleading if the assumed geologic parameters, such as reservoir thickness, structural closure, and percent of fill, are not accurate. Simple statistical procedures based on regressions between seismic characteristics that are discernible before drilling and the volumes of oil and gas subsequently discovered in prospects provide an alternative objective basis for prospect appraisal. From a regression, an estimate can be made of the probability distribution of oil or gas field sizes associated with prospects of specified characteristics. The resulting distribution can be used directly for prospect appraisal purposes, or can be combined with engineering and financial data to yield an expected monetary value which incorporates elements of risk and uncertainty.

To demonstrate the method of statistical appraisal, structural prospects were evaluated in part of the Pleistocene trend of offshore Louisiana and Texas, an area of about 3 million acres (1.2 million ha.). Highly significant statistical relationships were found between structural properties measured on prospects shown on regional seismic reconnaissance maps, including area of structural closure, and the volumes of oil and gas subsequently discovered by drilling these prospects. A comparison with conventional Monte Carlo evaluations of tracts in the Gulf Coast outer continental shelf suggests that statistical procedures based on seismic information may be more than twice as effective as Monte Carlo methods employing subjectively estimated reservoir characteristics. It may be possible, however, to improve the performance of Monte Carlo methods by replacing the estimates of reservoir characteristics with statistically predicted reservoir volumes.

First Page Preview

First page PDF preview
You do not have access to this content, please speak to your institutional administrator if you feel you should have access.