Abstract

The processes associated with assessing conventional hydrocarbon accumulations are well understood and routinely applied. These assessments can have either a local (prospect) or regional (play) focus, depending on the business need, and rely on risking, sizing, and counting discrete prospects. In contrast, many continuous resource plays are better characterized as a single large hydrocarbon accumulation with highly variable resource density rather than as discrete fields or prospects. Because economic margins may be thin, business success often depends on early identification of favorable areas. Decisions regarding play entry, acreage acquisition, and pilot project design frequently must be made without the benefit of recent well or production data. The assessment methodology for continuous resources must be flexible enough to adapt to situations ranging from little or no well data to thousands of wells.

Play assessment provides a powerful mechanism for integrating geologic insights into business decisions. To incorporate the unique aspects of continuous resources, we apply a hybrid grid and polygon based methodology. The play is subdivided into a series of analysis polygons (subplays or segments) that provide the basis for probabilistic calculations. The polygon boundaries can be dynamic (mapped geologic limits), static (leases), arbitrary, or some combination of these. Polygons should be defined for the purpose at hand and at a granularity that captures but does not overwhelm regional trends. Those geologic properties mapped spatially as grids (e.g. gross thickness, net-to-gross, or porosity) are evaluated within each analysis polygon to obtain inputs for the probabilistic assessment. Geologic properties that are not available as spatially varying estimates can be assigned directly with appropriate ranges of uncertainty. Resource uncertainty can be captured both as ranges around most likely parameter estimates and as multiple geologic or operational scenarios. Risk dependency and volumetric correlation must be defined in order to obtain robust probabilistic aggregations of multiple analysis polygons. Fully probabilistic results lead to better informed business decisions by providing information on high side and low side outcomes.

Concurrent with the probabilistic assessment, deterministic grid-based calculations of in-place and recoverable hydrocarbon volume per unit area elucidate the resolvable geographic variation of resource density. The parallel grid-based approach ensures that results incorporate and are consistent with local geologic understanding and provides a direct basis for calibrating volumetric calculations with well performance data. Ultimately, assessments based on well performance provide the clearest indication of the recoverable resource and economic viability of a play. Unfortunately, well performance-based approaches are not reliable until a high operational efficiency has been achieved and the number of wells with adequate production histories is sufficient to clarify the primary controls on resource density. To provide early prediction of well performance within the context of volumetric assessment, probabilistic ranges of well EURs are estimated for each assessment polygon based on volumetric inputs coupled with assumptions about likely drainage areas. Well drainage area can be entered as a single value or as a range tied to reservoir properties (permeability and rock mechanics), geologic complexity (faulting), and completion scenarios (well spacing, vertical vs. lateral completions, length of laterals, etc.). As performance-based EUR projections become available for wells, they can be compared to the specific recoverable volume prediction from the resource density grid along with the distribution of well EURs predicted for the polygon. Population-based Bayesian statistical techniques provide robust, probabilistic predictions of EUR for producing wells within a play.

Combining analog datasets of initial potential (IP) with well EUR provides a dynamic linkage of the volumetric assessment to well performance. Applying relationships between EUR and IP from analog plays to EUR curves (one per polygon) for the play of interest yields spatially varying families (distributions) of IP’s. By recasting assessed volumes in terms of well productivity, this process supports economic analyses explicitly linked to spatially varying geologic inputs.