Difficutleis in Estimating Resources and Reserves in a Probablistic Framework
David MacDonald
BP Exploration Operating Co. Ltd, Middlesex, United Kingdom
Industry debate during the preparation of the 2007 Petroleum Resources Management System, jointly sponsored by the Society of Petroleum Engineers (SPE), the World Petroleum Council (WPC), the American Association of Petroleum Geologists (AAPG), and the Society of Petroleum Evaluation Engineers (SPEE), raised the intriguing question of whether deterministic or probabilistic techniques should be used in the estimation of resources and reserves. Responses from industry reviewers ranged from a championing of one technique at the exclusion of the other, to a seeming indifference to whether a deterministic or probabilistic approach was applied.
A common recommendation is to do both deterministic and probabilistic estimates as a check that the assessments are consistent. The mean or P50 is compared to the base case then the P10 and P90 are compared to the upside and downside cases. However there is no valid reason for expecting the results of this deterministic look at the probabilistic results to match the deterministic estimate. The value of a probabilistic estimate is in maintaining the distribution of results throughout the entire analysis. How to choose a single representative value is not straightforward, will be case dependant, and not unique from one analyst to another.
Ultimate recovery from any project depends not just on the description of the subsurface, but also on the development plan, and will be directly affected by the choices made during the operation from the surveillance data available for learning and optimisation. Probabilistic cases do not easily allow intervention plans to be tested, nor account for optimisation through learning. Probabilistic analysis of resources and reserves is further challenged as the produced volumes and profiles cannot be easily linked to a specific activity set.
Furthermore, if the definition of Proved Reserves has a commercial element (which it does in most regulatory frameworks), then we need more than just a recovery volume - we need production, price and cost profiles. Two cases which have the same expected ultimate recovery may have very different economics and opportunities for mitigation of potential downsides. Probabilistic techniques can and should be used to inform the technical and commercial understanding of the base case and upside and downside cases.
This presentation will give recommendations for including the insights of probabilistic estimates in deterministic estimation of resources and reserves.
AAPG Search and Discovery Article #90072 © 2007 AAPG and AAPG European Region Conference, Athens, Greece