RISKING UNCONVENTIONAL SHALE PLAYS: A DIFFERENT APPROACH

In several ways, the methods used to risk and evaluate conventional resources do not apply to unconventional shale resources. Risking of conventional plays depends on assigning values to a standardized checklist of parameters, with each assigned a probability of success. However, with an unconventional play, some of these parameters (such as migration) are usually inconsequential, while others (such as reservoir quality) might usefully be subdivided. Because conventional plays do include migration, assessment asks only if the parameters are met somewhere within the play area. Unconventional plays, which are less concerned with achieving threshold conditions and more with finding the best possible combination of parameters (the “sweet spots”) must be based on the areal distribution of independent variables. Unconventional shale resources do not depend on simply reaching minimum conditions somewhere, but on the distribution and quality of “superproducing zones”.

One approach would be to base the risk/assessment variables on those parameters known to contribute to success in analogous situations or even early data from an emerging play. With factor analysis, it would be possible to weight the significance of the various factors; those that had the greatest impact on success would be weighted accordingly, while those that were not important would be minimized or discarded. Such an analysis would have the added bonus of focusing exploration on those parameters critical to success, while methods that contributed only minor increments to success could be bypassed, resulting in cost savings.

To be successful, such an approach needs to begin with a broad evaluation of potential variables. The idea would be to identify the critical variables (which will not be the same in all plays), and to minimize those that contribute little to success. This also helps to identify valid analogs to help model gaps in information. Improving success comes from testing depositional and diagenetic models against observations; risk analysis needs to be based on effective models.