A challenging aspect of competitor analysis and benchmarking within the major liquids rich plays of Western Canada is the inconsistent reporting of field products within the public production data record. For example, in the Alberta Duvernay play there is a broad understanding that liquids make up a substantial quantity of the field volumes, yet we often find wells or fields with substantial gaps in field liquids reporting. In the Waskahigan and Kaybob South Duvernay fields, over 70% of wells are found to be reporting zero (or very nearly zero) historical field liquids. The remaining 30% of these wells only report a fraction of the liquids volumes that have actually been produced.
Across the Duvernay play, approximately 60% of field liquid volumes (oil + field condensate) have been estimated by GLJ to be underreported (Figure 1). This reporting issue is something we commonly see in Alberta and British Columbia in liquids-rich resource plays including the tight Duvernay and Montney plays but also in other conventional reservoirs such as the Ellerslie Formation. Some call it “Gas Equivalent Reporting”, some call it “Recombined Gas Reporting” and some refer to it colloquially as “Fat Gas Reporting”, but all these terms carry the same meaning. Reported gas volumes, either on the full life or on just a portion of the life of the well, represent both liquids as well as raw gas volumes and therefore should be viewed as unreliable for direct use.
Figure 1: GLJ “Best Estimate Case” adjustments to historical reported liquids volumes (Duvernay, entire play)
Challenges associated with imperfect public production reporting spurred GLJ to construct an automated solution to systematically impute or correct these volumes. In this case study, we look at the Alberta Duvernay liquids rich gas play using GLJ IntelliCasts™(IntelliCasts), a recently launched GLJ subscription product that delivers evergreen and reliable well-level forecasts that seamlessly corrects for under-reported field liquids, while also layering on forecasts for plant natural gas liquids (C2-C5+, by component).
To extract insights from public production data, one must pre-process or adjust publicly reported well production volumes using an informed expectation of production behaviour. This inference-based pre-processing step requires insights that are not always directly accessible in the public record. Since Condensate-Gas-Ratios (CGR) typically decline as a function of producing time it is necessary to apply time dependent adjustments. The adjustment process is complicated further by the fact that not all wells are under-reporting field liquids at every point in their life.
In order to generate relevant results from these pre-processing adjustments, algorithms need to be equipped with both an informed expectation or “analogue model” for a CGR decline profile and have the ability to recognize which type of reporting is likely occurring at every time interval. As is expanded upon in Figure 2, three distinct cases of gas and liquids reporting must be considered:
I: Wells that report all liquids continuously (i.e., public record is reliable as-is)
II: Wells that report liquids sporadically, or at levels that are unreasonably low relative to expectations
III: Wells that report no liquids at all
Figure 2: Inference-based historical production adjustment flow chart
For the Duvernay model within IntelliCasts, the selection of an appropriate “analogue model” for CGR behaviour leans heavily on thermal maturity estimates sourced from GLJ’s extensive reserves database and also utilizes certain liquids data salvaged from the public record. Feeding the thermal maturity estimates on GLJ’s data back end is a pipeline of human gathered petrophysical and geological interpretations as well as other public datasets, most noteworthy of which being the butane isomer ratio, defined as the ratio iC4:nC4 (Figure 3). Derived from gas analysis data, the butane isomer ratio has been shown to be strongly correlated to thermal maturity in the Duvernay and can be used to support geological interpretations in the identification of play areas that may be more liquids-prone. For more information about GLJ’s geological method of liquids estimation in the Duvernay, visit this post.
Figure 3: iC4/nC4 ratio as a proxy for Duvernay thermal maturity
In IntelliCasts, publicly reported gas production is split to provide appropriate allocations to field liquids and raw gas streams, both on a historical and forecast basis. Figure 4 shows an example of production data adjustments on a single Duvernay well that sits within GLJ’s “Tier V” thermal maturity group. The well was detected to be reporting all field liquids in early time (months one to three), but liquids reporting abruptly transitioned to Gas-Equivalent reporting on the fourth month of production. In the fourth month of production, the algorithm detected the change and automatically switched to rely on the analogue CGR profile model, simultaneously imputing liquids for the missing months and shrinking the historically reported gas. On the six plots shown in the figure, dark blue coloured data series are associated with imputed liquids data, where estimates have been made based on an expected CGR type well. Green coloured series originate from publicly reported data.
Figure 4: Historical production adjustment on a Duvernay well reporting predominantly Gas – Equivalent production
If we switch from looking at a single well level up to the play level, the impact of IntelliCasts production adjustments is such that the forecast initial deliverability of the total Duvernay field liquids is estimated to be between 1.7x and 2.4x the last reported public aggregate liquids production rate (Figure 5). The range in liquids deliverability at the aggregate level is the result of uncertainties summed from the well level for each constituent well in the play.
Figure 5: Duvernay product forecasts (field liquids, raw gas)
With adjusted production data in hand, better resource forecasting estimates can be undertaken. Figure 6 showcases a sample of wells in the Kaybob South & Waskahigan fields and how inference-based production pre-processing adjustments can result in Estimated Ultimate Recoverable (EUR) volume forecasts for field liquids that are far higher (and more reliable) than what would be estimated based on unadjusted raw public data alone. GLJ has compared these results against those generated from actual reserves evaluation studies involving proprietary field production data and found them to be very trustworthy approximations.
Figure 6: Duvernay field liquids forecast EUR from GLJ IntelliCasts™; Red indicates liquids production forecasts for a particular well, whereas the green bars indicate GLJ IntelliCasts™ estimation. (Note: when there is no red bar present, zero liquids were reported)
For the assessment of net present value in liquids-rich wells, it is critical to have a good understanding of the historical and future production allocations between field liquids and gas. Due to reporting inconsistencies, the public production record cannot always be counted on to provide this insight. In applications such as asset screening, competitor analysis, and benchmarking, there simply is no proprietary production dataset to fall back on. Technically rigorous and reliable adjustments to the public production record in plays like the Montney and Duvernay is just one of the many reasons why IntelliCasts insights are an indispensable tool in any upstream and midstream asset evaluation toolkit. The direct linkage between GLJ’s internal play data and IntelliCasts means that subscribers can spend their time making confident data-driven decisions for their businesses.
For more information on GLJ IntelliCastsTM, please visit our website: https://www.gljpc.com/glj-intellicasts