Estimating resources in unconventional assets: Spatial bootstrapping with n-effective

Abstract

Estimating undeveloped reserves in unconventional reservoirs has important practical and legal implications. Unconventional reservoirs lend themselves to the use of statistical workflows which often first determine a distribution of well EURs and then treat the EURs at the undeveloped locations as regionalized variables. This approach was formalized by the Society of Evaluation Engineers in Monograph 3 published in 2010 which presents a workflow to assess proved undeveloped reserves in unconventional assets based on the statistical bootstrap approach. However, the recommended workflow makes the strong assumption of spatial independence amongst well pairs which significantly limits the application of the workflow as this assumption is often violated in practice. Further, applying this workflow to reservoirs that violate the independence assumption can result in the underestimation of uncertainty and thus overconfident estimates. To overcome this limitation, we introduce a geostatistical workflow that honors the spatial context of the data by bootstrap sampling with the effective number of independent data, n-effective, rather than the total number of data, thus generating uncertainty estimates that honor the spatial context of the data and extend the applicability of the workflow to fields with spatially correlated data.