GLJ has undertaken a regional study combining geological and engineering data to better understand Gas Initially in Place (GIIP) estimates and its relationship to deliverability for reservoirs of the Spirit River Formation in the Deep Basin of Western Canada. Understanding this relationship is important for commercial viability, reserves and resource evaluation, and development planning. The project combines published regional geology work, publicly available log, core, and engineering data, petrophysical analysis, and regional mapping.
Before 2010, most production from the Spirit River Formation came from vertical wells completed in numerous Falher cycles of conventional clastic reservoir typified by the Elmworth Field (Figure 1). To the south of Elmworth, the Spirit River Formation covers a large area of the Deep Basin, but the majority of the fairway was not a significant producer from vertical wells due to lower quality reservoir.
In the last 5 years, the Spirit River Formation has become a major gas play as advancements in horizontal well drilling and completion technology have allowed for commercial exploitation of the Spirit River Formation, from Brazeau to Wapiti - an area of approximately 350km by 75km. Multi-stage fracturing operations have allowed for economic gas production from tighter, shalier rock. Horizontal wells drilled into the Spirit River Formation have increased from less than 50 in 2010 to approximately 2,500 in 2017. Recent peak production of 2.6 Bcf/day, from the horizontal wells, makes this one of the most prolific and economic gas plays in Western Canada. As the Spirit River continues to be developed, there is a need to better understand GIIP, both for future well development and for accurate reserve and resource bookings.
Figure 1: Regional Cross section of the Spirit River Formation focusing on the stacked shorefaces in the Elmworth field. Current horizontal well development is to the south highlighted by the blue box (Jackson, 1984).
The Spirit River is divided into three main members, the Notikewin, Falher and Wilrich. There is no formal agreement on further division of the Notikewin, Falher and Wilrich members, which can cause considerable confusion within the industry. GLJ has chosen to subdivide the members into informal sub-members; Falher A to Falher I and Wilrich A to Wilrich C (based on Petrel Robertson’s 2013 stratigraphic study of the Wilrich). The Wilrich A to C sub-members consist of prograding shorefaces to the northwest. The Wilrich shorefaces are typically capped by a coal bed which separates the Wilrich reservoirs. From Brazeau to Wapiti, the Falher, and later the Notikewin were deposited on top of the Wilrich in a series of coastal plain channels and incised valley fills (Figure 2). The spatial progradation of these shorefaces and their overlain channels is depicted in Figure 3.
Figure 2: Diagrammatic Cross Section of the Spirit River Formation, from NW to SE.
Figure 3: Overview of Prominent Spirit River Plays and Trends. The Wilrich shorefaces were first deposited in progradational shoreface sequences to the northwest, and subsequently overlain by coastal plain channels of the Falher and Notikewin.
To calculate and compare GIIP volumes throughout the different areas in the Spirt River in the Deep Basin, GLJ needed a simple and effective way to calculate petrophysical reservoir parameters but which still allow for regional variations. Core data is the foundation of our petrophysical model. We reviewed 63 publicly available cores to calibrate our deterministic petrophysical model. Neutron density cross-plot porosity was found to predictably tie core porosity, due to the ability of the cross-plot to estimate grain densities. Figure 4 is an example well which depicts how grain densities are variable, and how the neutron density cross-plot accounts for this variability. We calculate a volume of shale (Vsh) from a linear conversion of the gamma ray log to correct total porosity (PhiT) to effective porosity (PhiE). The Simandoux equation is used to calculate water saturation, and ties oil-based core water saturations. There is uncertainty in water saturation calculations as there is little to no produced water from the reservoir, and therefore no directly measured water resistivity values. There is also a lack of special core analysis where a, m and n values have been measured. More detailed petrophysical methods can be utilized in limited areas if sufficient core analysis exists. However, using a basic neutron density cross-plot combined with linear Vsh from gamma ray, and Simandoux equation appears to give a good regional perspective.
Using the evaluated wells, net pays, porosities and water saturations are mapped per zone and area to calculate a hydrocarbon pore volume and in turn, GIIP.
Figure 4: Example Petrophysical Evaluation from Edson/Ansell (49-19W5)
Estimated ultimate return (EUR) is calculated for all the Spirit River horizontal wells by decline analysis. The horizontal wells from Falher H to Wilrich C were parsed from the data set, and their EURs were normalized by length. These normalized horizontal EURs were mapped to understand trends within the Spirit River shorefaces (Figure 5). Not surprisingly, production trends have a strong correlation to reservoir quality. The highest production trends are in the thickest, deepest reservoirs. Rates in Wapiti, Kakwa and Resthaven are superior than the rest of the Deep Basin due to their cleaner reservoirs and higher initial pressures. By using this data, we are starting to analyze production variability causes by completion differences, well lengths fracture frequency and tonnage. Some operators have stepped off main producing trends (e.g. in Ansell/Edson and in Wild River/Marsh). Results are varied with off trend wells, but with technological improvements in hydraulic fracturing, results are becoming increasingly promising.
Figure 5: Production Trends in the Spirit River – Falher H to Wilrich C (EUR/100m).
As an example, using GLJ’s EUR estimates, and public completions data, we binned wells with similar completions (tonnage/metre) and plotted their probabilistic distributions of the EURs on probit plots. In general, it was found increased tonnage results in an increased EUR (Figure 6).
Figure 6: Completions vs. Deliverability - All Spirit River Horizontals
We compared two main producing areas: Edson and Kakwa. Even though the producing horizons are different by just comparing the net pay thicknesses and average porosity, reservoirs in Kakwa and Edson appear to be similar at first. Kakwa has substantially higher estimated EUR than in Edson. By plotting distributions of EUR binned by frac tonnage on probit plots, we see distinct differences between Edson and Kakwa.
In Edson, with larger completions, EUR increases, suggesting tighter, shalier reservoir is accessed (Figure 7a). In Kakwa (Figure 7b), EUR distributions of different tonnage overlie each other once you reach a certain tonnage. Optimal tonnage may not have been achieved in Edson as the P50 is still increasing with higher tonnage, whereas in Kakwa an optimal proppant tonnage has been reached as tonnage does not appear to increase EUR. The results of the EUR probit plots do tie back to reservoir properties. In Edson, reservoirs are comprised of shoreface sands, with higher shale content and higher water saturations and lower initial pressures. In contrast, the Kakwa reservoirs are cleaner, higher pressured and bound by coals. As such, it may suggest the whole reservoir is accessed with a medium-sized hydraulic frac, and in Edson, larger fracs are still accessing the shalier, tighter reservoirs. By overstimulating wells in areas such as Kakwa, returns will be diminished. See Figure 8 for a reservoir comparison.
Figure 7: Completions vs. Deliverability - Edson vs. Kakwa
Figure 8: Kakwa vs Edson - Reservoir Differences
To conclude, the Spirit River Formation has a complex stratigraphic framework consisting of multiple zones and facies within those zones. The main regional targets in the Spirit River Formation are shorefaces characterized by Falher H, Falher I, and Wilrich A to C. We have concluded that the best way to regionally estimate volumes in place petrophysically is to use a method consisting of neutron density cross-plot porosity, linear Vsh and the Simandoux equation to calculate water saturation. We believe our petrophysical approach is a good methodology for a regional perspective as the neutron density cross-plot accounts for variations in grain density and the methodology ties core porosity and water saturation. This simple petrophysical model allows us to see regional variations within the Spirit River Formation, and vertically in the stratigraphic column. Mapped GIIP volumes align well with projected EUR values.
Established reservoir pay trends from petrophysical well log evaluations mirror production trends in the Spirit River. There are some variations in EUR within the same facies which are potentially due to differences in completion techniques. In general, with an increase in tonnage, we see an increase in EUR; and in turn, potentially improved economics. Using petrophysical interpretations to estimate GIIP and integrating relationships with EUR and completions allows for the ability to identify future opportunities outside of reserves. The combination of a simple petrophysical model, regional mapping, and engineering data allows us to estimate GIIP, compare and contrast different reservoirs within the Spirit River Formation, and compare different completion practices.
GLJ’s John Hirschmiller will be presenting a talk, and poster on this subject at the CSPG GeoConvention on May 7, 2018. Be sure to check it out! If you are interested in any of the ideas above, contact John, who would be pleased to speak with you.
Please subscribe to be notified when we make a new blog post.