Predicting the Discoveries and Finding Costs of Natural Gas: the Example of the Scotian Shelf

Predicting the discovery rate and marginal finding costs of natural gas resources requires a well-documented and long statistical history. For partially explored basins, the statistical history is often inadequate. Attempts at avoiding the problem have been made using probabilistic modelling approaches. These are used to estimate the parent population of pools available for discovery and the probable discovery rate. The phenomenon of economic truncation, however, calls into question the precision and utility of such estimates. Furthermore, the exploration process is known to be biased toward larger pools, but no method of determining the extent of the bias has been discussed in the literature to date. To avoid these defciencies, this paper employs the pool size distribution estimates routinely produced by geologists to drive a probabilistic modelling framework taking explicit account of the physical laws of resource depletion. The methodology is discussed and applied to Canada's Scotian Shelf. In order to put the predicted costs for the Scotian Shelf in perspective, the results are then compared to forecasts for Alberta.