Lithofacies-dependent Rock Physics Templates of an Unconventional Shale Reservoir

✅ Paper Type: Free Essay	✅ Subject: Geography
✅ Wordcount: 8329 words	✅ Published: 18th May 2020

Reference this

Share this: Facebook Twitter Reddit LinkedIn WhatsApp

Lithofacies-dependent rock physics templates of an unconventional shale reservoir in North Slope Alaska

Abstract

Organic-rich shale has become an increasingly important hydrocarbon resource around the globe due to rapid depletion of conventional reservoirs. Successful exploration and production schemes for shale source rocks should base on reliable identification of major organic components (kerogen in particular) and their hydrocarbon-generating potential. There is a growing need to identify organic content in terms of quantity (Total Organic Carbon or TOC) and quality (hydrogen index or HI, which determines kerogen type) in promising shale formations through indirect seismic data, which is usually the only available source of information in frontier settings. We delineated different seismic lithofacies in the Alaska North Slope Alaska in terms of elastic, seismic and petrophysical properties. From this characterization process, rock physics templates of inverted seismic parameters (Acoustic Impedance, or AI, versus P-wave over S-wave ratio, or V_p/V_s are constructed for each lithofacies to assess pore fluid distribution and lithology. We proposed useful correlations between source rock attributes (TOC, Hydrogen Index or HI) and petrophysical properties (bulk density, porosity, gamma ray, sonic velocities V_p/V_s) of major lithofacies. These correlations, together with facies-specific rock physics templates, assist in mapping organic richness and reservoir properties from seismic-derived attributes. Existing shale petrophysical models, calibrated and constrained to North Slope geology, are then verified by this training dataset to observe their applicability in the basin.

Introduction

Alaska North Slope (formally the Colville Basin) is estimated to contain approximately 40% of total US undiscovered, technically recoverable oil and natural gas resources, the bulk of its resources coming from Northern Alaska with more than 30 billion barrels of oil and nearly 200 trillion cubic feet of natural gas (Bird, 2001). Shale oil interest is gaining a lot of traction because of the development of advanced technologies over the past 20 years. Petroleum exploration on the North Slope is limited to the region near the Beaufort Sea coast located between the National Petroleum Reserve Alaska (NPRA) and the Arctic National Wildlife Refuge (ANWR). Few wells have been drilled outside of Prudhoe Bay region, resulting in sparse information for proper formation evaluation and lithofacies classification in shales. Traditionally, formation evaluation and production planning of shale plays pose considerable challenges due to their complex lithology and significant lateral and vertical variation of petrophysical properties. Therefore, a key issue for future exploration of the North Slope is the lateral variability of source rock away from known hydrocarbon accumulations.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!
Find out more about our Essay Writing Service

A seismic lithofacies is not necessarily a single rock or formation but rather a collection of geologically similar rocks that span a comparable range of petrophysical and seismic properties (Avseth, 2010). A seismic lithofacies shares characteristic sedimentologic and rock physics properties, thus serving as a major force in controlling reservoir geometry and porosity. This study attempts to characterize petrophysical, geochemical, and elastic properties of shale lithofacies of the North Slope and build a reliable training dataset (P and S-wave velocities, bulk density) for classification. Previous rock classification techniques introduced in organic shale formations are strongly dependent on many core measurements to reasonably capture shale heterogeneity (Gupta, 2013). These are both time-consuming and expensive to acquire. Due to sparse core information in the study area, well logs are a viable candidate for rock classification as they provide relatively high sampling resolution in the vertical dimension, continuous interval properties, and information available in real-time. Cross-validation and proper calibration of log-derived properties with core data are regularly performed throughout this study.

Geological Setting

Four major source rock units have been identified in the Alaska North Slope. These are the Hue Shale, pebble shale unit (formally the Kalubik Formation), Kingak Shale, and Shublik Formation (Figure 1). The Highly Radioactive Zone (HRZ), also called the Gamma Ray Zone (GRZ) in Figure 2 at the base of the Hue Shale is separated based on petrophysics from the Hue Shale because of its different signature. The most relevant geological features, depositional history, and source rock characteristics of each lithofacies are discussed here.

Figure 1: Generalized stratigraphic column for Alaska North Slope, emphasizing petroleum source rocks and their relative ages. GRZ, Gamma Ray Zone. The Lower Cretaceous Unconformity (LCU) lies under the pebble shale unit (Bird, 2001).

Shublik

The Triassic Shublik Formation of the Ellsemerian sequence is relatively thin (less than 300 feet), regionally extensive, and lithologically heterogeneous, consisting of limestone, sandstone, siltstone, phosphatic nodular shale, and calcareous shale (Parrish, 1987). Shublik facies south of the Barrow Arch is of particular economic interest because it is the principal source of oil and gas in the North Slope, accounting for more than 90% of the recoverable crude oil and 82% of the recoverable gas (Bird, 2001). It is organically rich (TOC ranges from 0.5 to 13.1%), ranging from a strongly oil-prone Type I kerogen to a more gas-prone Type III kerogen (Robinson, 1996).

Figure 2: Map showing major tectonic features of Northern Alaska. ANWR, Arctic National Wildlife Refuge; NPRA, National Petroleum Reserve-Alaska; PB, Prudhoe Bay (Bird, 2001).

Kingak

The Jurassic-Lower Cretaceous Kingak Shale comprises the bulk of the Beaufortian sequence that was deposited during rift opening of the Arctic Ocean (Hubbard, 1987). Kingak Shale on the southern passive rift flank is a mud-dominated succession of prograding shelf deposits characterized by multiple transgressive-regressive sequence sets (Houseknecht, 2004). Kingak Shale contains a mixture of marine and terrigenous organic matter deposited in a marine siliciclastic setting (Peters, 2006). The lower part of the Kingak is typically the most organic-rich interval with an average TOC of more than 5%.

Uplift and erosion of the rift margin produced the regional Lower Cretaceous Unconformity (LCU). This unconformity progressively truncates all older units northward onto the Barrow Arch. It plays an important role in many of the largest oil fields in northern Alaska via development of enhanced porosity in sub-unconformity reservoirs, provision of a migration pathway for hydrocarbons, and juxtaposition of overlying marine mudstone source and seal rocks, such as pebble shale unit and HRZ of the Hue Shale (Bird, 2001).

Pebble

The pebble shale unit was deposited during a south-to-north marine transgression in response to subsidence of the rift margin. It is characterized by a small but distinctive proportion of pebbles and well-rounded frosted sand grains scattered through the shale (Molenaar, 1987). Pebble shale unit differs in its organic characteristics: being oil-prone in some areas and gas-prone in others. Despite its relatively high TOC 1.5-3.8 wt.%, petroleum-generative potential of the pebble shale unit varies because of differences in primary productivity, clastic dilution, and preservation (Keller, 2001).

Hue

The Hue Shale is the distal-deltaic condensed section of the Brookian sequence and was deposited in a deep-water basin plain environment. The upper part of the Hue Shale is thicker but has considerably less generative potential (lower TOC and HI) than the lower part because of more proximal deposition and greater clastic dilution. The lowermost part of the Hue Shale is easily marked on well logs by a characteristic high gamma ray signature. This organic rich interval has a range of TOC from 1.9 to 3.9 wt.% (Keller, 2001).

General Classification

In the ternary diagram commonly used for shale classification, shale can be divided into argillaceous shale (rich in clay minerals), calcareous shale (rich in calcite), and siliceous shale (rich in biogenic and detrital quartz/feldspar). Based on limited X-ray diffraction (XRD) analysis and geological background, Hue Shale is classified as siliceous mudstone whereas Shublik Formation is classified as siliceous marlstone (yellow circles in Figure 3).

Figure 3: Ternary diagram shows shale classification of Hue Shale and Shublik Formation (blue triangles) based on XRD analyses. Other notable shale plays are also presented in the diagram.

Methodology

In this study, regional geology, standard triple combo logging suites, petrophysical, and geochemical analyses of core plugs are basic inputs to obtain facies definition, which is the first step of a more comprehensive statistical rock physics evaluation workflow (Figure 4).

Figure 4: Diagram showing quantitative seismic interpretation workflow with integration of geochemical data. In this study, we focus on the parts of the workflow that are related to the construction of a reliable elastic and geochemical training dataset of each pre-defined lithofacies.

Quantitative seismic interpretation (Figure 4) demonstrates how rock physics can be applied to predict reservoir parameters such as lithology, pore fluid, and source rock character from seismically derived attributes. Based on available logs, cores, and geology, we identify major seismic lithofacies by observing cluster separation in cross plots of different properties. Rock physics aids in converting geologic and wireline logging information into elastic properties (Vp, Vs and bulk density ρ_b). Geochemical parameters are integrated into the workflow by establishing correlations between elastic and source rock properties. After performing proper scale calibration of inverted seismic data in the area of interest, we use this dataset to classify lithology and source rock character to detect best producing intervals.

In this study, we focus on the parts of the workflow that are related to the construction of a reliable elastic and geochemical training dataset of each predefined lithofacies. Well log data (density, gamma ray, resistivity, and sonic wave velocities) are extracted for exploratory cross-plots and quantitative assessment. (The use of cross-plots between relevant log-derived properties to separate lithofacies proves to be a fast and simple process that can also be applied at the wellsite.) Based on the top and base depths of various lithofacies in the well, we delineate and build a log-based training data of each facies. Preliminary quality checks are performed to remove anomalous log readings due to equipment errors. Calibration of logging data based on available core data is also performed. Neutron log cannot be used in radioactive shale intervals because cross-validation shows erroneously higher values of neutron porosity compared to core values.

A challenge of this study is the lack of petrophysical and geochemical data in the same subset of core plugs due to different labs conducting experiments at different times. Therefore, existing correlations in the literature to expand the available dataset are utilized. Intrinsic variability of rock properties within a single lithofacies presents the biggest challenge of quantitative seismic interpretation, especially when an observed attribute change indicates a significant change across facies rather than a minor fluctuation within facies (Avseth, 2010).

Dataset

Two vertical wells located along the Trans-Alaska Pipeline System are shown in Figure 5. These wells, Merak-1 and Alcor-1, were drilled by Great Bear Petroleum in 2012 and cored the Hue and Shublik intervals extensively (Scheirer et al., 2014 and Scheirer et al., 2017). In addition to the standard log suite, available data include dipole sonic log and spectral gamma ray log. The two wells of interest are 1.5 miles apart and have shown excellent correlation in terms of petrophysical properties and source rock character. Vertical Seismic Profiling and 3-D seismic are also available for future study. Available core analyses include: porosity, permeability, oil/gas saturation, X-ray Diffraction (XRD), and computed tomography scans. In addition, geochemical data are available for core and cuttings samples. Cuttings measurements are not included in this study. Geochemical data include Leco TOC, programmed pyrolysis, and vitrinite reflectance (R₀).

Figure 5: Focus area is located between the NPRA and ANWR. The area of interest shows locations of two vertical wells of interest. The blue dashed line indicates the area of available 3-D seismic data.

Seismic lithofacies definition

Logging analysis

Common logging tracks are plotted to verify several key signatures of each lithofacies (Figure 6). The density of the Hue-HRZ interval is relatively constant. However, the HRZ has significantly higher gamma ray, as expected, and lower sonic velocities than the overlying Hue Shale because of smaller clastic dilution (more clay content) and less proximal deposition. A spike at 8700 feet in the density log in the Merak-1 well is due to a change in equipment after setting intermediate casing. Due to these reasons, the Hue Shale and HRZ will be separated into two separate lithofacies.

The pebble shale unit has a wide range of density values due to the pebbles and well-rounded sand grains in its fine-grained matrix. In terms of radioactivity level and acoustic properties, Kingak Shale is a relatively homogeneous interval. Nevertheless, the density values of the Kingak Shale vary considerably due to its depositional history, a mud-dominated succession of prograding shelf deposits characterized by multiple transgressive-regressive sequences. The Shublik Formation has abrupt high gamma ray bands interbedded with lower gamma ray intervals. Spikes in both the gamma ray and density tracks indicate different amounts of clay and carbonate, respectively, throughout the Shublik interval. It also has much higher velocities of both P and S waves compared to other facies because its matrix has a greater amount of carbonate.

Figure 6: Diagram showing the log response of the organic rich shales investigated in this study. From top to bottom: red (Hue), green (HRZ), blue (pebble shale), black (Kingak), and pink (Shublik). This color code is used throughout this study.

Cross-plots of P and S-wave velocities versus bulk density show some degree of separation between different shale units. Whereas Hue and HRZ display considerable fluctuations of acoustic velocity within a small range of density values, pebble shale and Kingak shear velocities show relative independence of bulk density. Despite the small distance between the Merak-1 and Alcor-1 wells, lateral variability in lithofacies’ properties is demonstrated by the data clusters for the pebble shale and Kingak intervals in Alcor-1, specifically, the absence of low-density components in Alcor-1 (Figure 7).

Figure 7: Vp and Vs (m/sec) versus bulk density gm/cc of different shale lithofacies in two wells: Merak-1 (top panel) and Alcor-1 (bottom panel). Graphs are of similar scale for comparison. The wells are plotted separately to observe variation between them.

Another useful cross-plot is Vp versus Vs (Figure 8). Shublik and Hue are readily separated from other clusters. Pebble shale, Kingak, and HRZ clusters overlap. Dashed blue lines represent lines of constant Vp/Vs, which have been suggested to be a good indicator of organic-rich shale (Vernik, 2011). In several published datasets compiled by Vernik as well as in core and log data from Bossier, the Woodford and Bakken shale plays fall within a relatively narrow Vp/Vs range regardless of the wide observed range of saturation, porosity, and effective stress. These parameters seem to be secondary in controlling the reduced velocity ratio typical of organic shales as compared to their inorganic counterpart. In Alaska North Slope, the spread of velocity ratio spans 1.6 to 2.4, significantly broader compared to other shale plays (Vernik et al. 1996). The organically richer Shublik has the narrowest spread and lower average value of Vp/Vs ratio compared to other lithofacies, which supports the inverse correlation suggested by (Yan, 2012) between TOC content and Vp/Vs.

Figure 8: Vs versus Vp from dipole sonic log of Merak-1 and Alcor-1. Blue dashed lines represent constant Vp/Vs ratio. Plots are of similar scale for comparison.

To create a bridge between geochemical data and petrophysical data, dense geochemical data are required. Computation of TOC from available logs, in this case resistivity log, is necessary to supplement the more sparsely distributed geochemical data. In addition to low resolution, resistivity measurements in logging devices are strongly dependent on thermal maturity. Oil generation results in an increase in resistivity whereas expelled gas (products of oil cracking at higher maturity) decreases resistivity (Mann, 1986). Low resistivity therefore can indicate both immature and overmatured oil source rocks as well as gas-only source rock. Hence, resistivity alone is not sufficient for TOC calculation. A widely popular method to calculate TOC from logs in the industry is Passey method (or Δlog(R) technique). The method involves overlaying a properly scaled porosity log (or transit time log) on a resistivity curve (ideally from a deep reading tool). The separation between the two tracks results from two effects: the transit time curve responds to the presence of low-density, low-velocity kerogen and the resistivity curve responds to the formation fluid in pore spaces (Passey, 1990). Generation and expulsion of hydrocarbon from source rock contribute to the increasing resistivity in organic-rich intervals because of the replacement of electrically conductive pore water with non-conductive hydrocarbon (Tran, 2014). In this study, superposition of deep resistivity and sonic transit time logs on a predefined scale 50 μm/feet to one resistivity cycle in log scale) shows good separation in source rock intervals (Hue, Shublik, Kingak) and decent overlap in inorganic intervals. The Miluveach Sandstone (a non-source inorganic rock) is picked to be the baseline interval as the two curves run parallel and overlap in this interval. Values of baseline resistivity R_baseline and baseline transit time in the Miluveach Sandstone Δt_baseline, as well as resistivity and transit time of layers of interest, are input to calculate TOC (Equation 1):

	$∆ \log R = \log (\frac{R}{R_{baseline}}) + 0.02 * (∆ t - ∆ t_{baseline})$	(1)
$TOC = ∆ \log R * 10^{2.297 - 0.1688 * LOM}$			(2)

In Equation 2, LOM is the level of maturity and is determined separately for each source rock. For type II and III source rock, the cross-plot of programmed pyrolysis S₂ peak versus TOC of core plugs is used to determine the LOM value of Hue/HRZ. In Merak-1, the LOM is 8.5 and in Alcor-1, it is 9.5. The LOM in both the Kingak and Shublik in both wells is 12 (Figure 9).

Figure 9: S₂ peak mgHC/gRock versus TOC (wt.\%) of core plugs in geochemical dataset. Black lines indicate different Level of Maturity LOM as defined by (Passey, 1990).

Spikes in the TOC logs might be attributed to anomalies in the deep resistivity log. Cross-validation with geochemical core data shows a reasonable agreement in organic-rich intervals in the Merak-1 well (especially in the Shublik and Hue intervals). Only a small portion of the Kingak Shale is matched because we do not have enough core of this thick interval. In Alcor-1, the Shublik Formation is also sparsely sampled so this method cannot guarantee the match for the whole interval.

Core analysis

This study lacks a complete set of core plugs with geochemical, acoustic, and petrophysical data. Due to time constraints, data of different scales (well log versus core plug) is cross-correlated. Preliminary quality check shows that bulk density of log and core at similar depths are in reasonable agreement.

P-wave and S-wave velocities (extracted from sonic logs at corresponding depths) are plotted against different bulk density (log, dry core plug and as-received core plug) (Figure 10). Log values of bulk density of the Shublik and HRZ show very good consistency with core measurements; thus, no correction is necessary. However, other factors may obscure the value of bulk density log such as varying pyrite concentration and natural fracture system. Log values of density in the Kingak are lower than core values possibly due to sampling bias of core plugs towards pyrite-free and unfractured intervals. Presence of heavy minerals, like pyrite (less than 10% in XRD analysis), could be ignored for the sake of simplicity.

Figure 10: P and S-wave velocities m/sec versus bulk density gm/cc. Log value is denoted as circle, as-received core as diamond, and dry core as star. Saturation of as-received core does not change bulk density much because North Slope shale units have low porosity. Kingak log values of density are lower than core values possibly due to sampling bias of core plugs towards pyrite-free and unfractured intervals.

Cross-plots of Vp vs Vs and TOC vs hydrogen index (HI) show good separation between different lithofacies. A simple correlation between geochemical and petrophysical parameters is not easy to deduce because log response in shale intervals is complex and affected by not only the organics but also mineralogical and pore fluid properties of the rock (Tran, 2018). Looking closer at a single lithofacies, the correlation is stronger, but it is not as profound as the velocity-density relationship. Acoustic analysis in other notable shale plays (Bakken, Bazhenov and Niobrabra) is compiled by (Vernik, 2011), showing that Vp increases as HI decreases, except in high porosity shale where Vp is better correlated with porosity (or density).

A statistically well-defined evaluation requires a comprehensive geochemical analysis of extensive core sets, which is time consuming and expensive. Bit cuttings do not always reflect the correct lithology due to caving and contamination by organic mud additives (Mann, 1986). Therefore, wireline log data, which offers continuous profile of stratigraphic sections of interest with relatively high resolution, proves to be the best alternative. This is where the TOC logs established earlier come in handy. In the Shublik Formation, TOC and acoustic velocities show a strong directly proportional correlation. Hue and HRZ clusters are significantly overlapping, as are pebble shale and Kingak (Figure 11).

Figure 11: P and S wave velocity versus log-derived TOC values for Merak-1. TOC and acoustic velocities show a strong directly proportional correlation in Shublik.

Rock physics template

Rock physics models link seismic properties to geologic properties. Expanding on the earlier rock physics diagnostics, rock physics templates (RPTs) of two selective seismic parameters, Acoustic Impedance (AI, which is the product of bulk density and P-wave velocity) and Vp/Vs ratio, for each lithofacies in Alaska North Slope are created. Geologic trends (pressure variation, pore fluid, sorting, and cementation) also play a role in constraining rock physics models. If we can predict the expected change in seismic response (or seismic-derived attributes such as AI or Vp/Vs as a function of depositional environment or burial depth, we will increase our ability to predict hydrocarbons in organic-rich shale (Avseth, 2010). This RPT approach allows a rock physics analysis not only on well-log data but also on elastic inversion results of seismic data. RPT facilitates prediction of porosity/density as well as discrimination of different pore fluid and pressure scenarios in the area of interest. XRD mineralogy is available in the HRZ and Shublik Formation in Alcor-1 (Table 1). To simplify the matrix composition, only minerals that are of significant amount and critical inputs in existing rock physics models in the literature (quartz, clay, and carbonate) are considered. Note that pyrite is also prevalent in HRZ core plugs (around 10% volume percentage) but will be ignored for the sake of simplicity. Illite is the main clay component in both shale units.

Table 1: Simplified composition for HRZ, Kingak and Shublik to use as inputs of soft sediment model

	Clay (Illite)	Calcite	Quartz	Kerogen
HRZ	0.3	0	0.4	0.3
Kingak	0.3	0	0.5	0.2
Shublik	0.05	0.35	0.4	0.2

Table 1 presents the simplified lithology of the HRZ, Kingak, and Shublik intervals for the rock physics soft sediment template. The soft sediment model uses Hertz-Mindlin contact theory to calculate high-porosity end members at critical porosity and the modified lower Hashin-Shtrikman bound (Mindlin, 1949 and Hashin, 1963) to interpolate to low-porosity end members. The zero-porosity end member is a pure mineral mix of quartz, clay, and calcite, assuming that other minerals only appear as trace amounts in the matrix composition. The RPT used here requires several inputs (effective pressure, volume composition) to calculate shale elastic properties (acoustic velocities at different saturations, bulk density) (Tran, 2018). Pressure data is not available in the two wells studied here, so standard lithostatic and pore pressure gradients are assumed (1 and 0.433 psi/ft respectively) for calculation of effective pressure. Therefore, the effective pressure gradient is 0.567 psi/ft. Other inputs of the soft sediment model are mineral and fluid bulk/shear modulus and critical porosity (0.7 for shale).

This model calculates shale elastic properties and yields a Vp/Vs versus P-wave impedance trend superimposed onto log-derived data points. The soft sediment model examines expected changes of these seismic attributes with regard to changes in pore fluid, pressure, clay content, and mineralogy (blue arrows in Figure 12). This step also serves as a checkpoint to ensure log quality. The cross-plot of AI versus Vp/Vs reveals the trend of RPT-related property change due to shaliness/clay content in the Hue/HRZ (marked by blue arrow 2 in Figure 12). The sub-branches in the trend represent expected change during pore fluid substitution as gas displaces water in pore spaces (Sw varies from 0 to 1). Fluid substitution has to be used with caution because shale lithology (clay minerals) defy the assumptions of Gassmann’s formula (Smith, 2003). The effects of organic content and hydrocarbon-filled pore space will deviate the clusters of each lithofacies away from the main trend lines. The soft sediment model does a decent job of matching bulk density of low-porosity (or high-density) members. Despite the inclusion of low-density kerogen in the model, low-density members (blue points) are not well-positioned as they fall into a higher density zone. This is likely because the soft sediment model does not account for effective pressure anomaly along the interval. Also, the Hertz-Mindlin elastic contact theory, which is based on the behavior of an elastic sphere pack subject to a confining pressure, is more applicable to sand than to shale (Avseth, 2010). Another explanation is that the logging device directly measures a layer of low-density organic material at those depths corresponding to dark blue data points in the RPTs.

Figure 12: A rock physics template (RPT) of Hue and HRZ intervals presented as cross-plots of Vp/Vs versus AI includes a rock physics model locally constrained by depth (i.e., pressure), mineralogy, critical porosity, and fluid properties. The template includes porosity trends for different fluid saturation (from fully water-saturated Sw=1 to fully gas-saturated Sw=0) assuming uniform saturation. Color bar indicates the range of bulk density. Input parameters are given at right. Blue arrows indicate conceptual geologic trends: (1) decreasing porosity (or increasing bulk density), (2) increasing shale content, (3) increasing gas saturation.

To match bulk density of high porosity (or low bulk density) members, the model needs further modifications of its inputs (shear reduction factor, coordination number in Hertz-Mindlin model, kerogen composition, and petrophysical properties). The cross-plot of AI versus Vp/Vs for the pebble shale unit does not show much density dependence. Figure 13 shows the RPT for the Kingak Shale, in which density proves to be the principal driving force of Vp/Vs -AI trend; clusters of various density magnitude clearly separate from each other. Figure 14 shows that the soft sediment model works well in the Shublik Formation to predict bulk density because the range of bulk density accurately matches density values of data points. In the RPT for the Shublik, high density members fall in the lower density range because of the absence of high-density pyrite in the model.

The model is limited to interchangeable substitution of two fluids (in this case, water and gas). The predicted saturation of the soft sediment model shows slight overestimation of gas saturation compared to wet core plugs (at corresponding depths of log data points). This is most likely due to an inadequate fluid preservation process of core plugs or the omission of oil in the fluid substitution recipe in the soft-sand model.

Figure 13: Rock physics template (RPT) of Kingak Shale presented as cross-plots of Vp/Vs versus AI. The template includes porosity trends for different fluid saturation (from fully water-saturated Sw=1 to fully gas-saturated Sw=0) assuming uniform saturation. Color bar indicates the range of bulk density. Input parameters are given at right. Blue arrows indicate conceptual geologic trends: (1) decreasing porosity (or increasing bulk density), (2) increasing shale content, (3) increasing gas saturation.

Figure 14: A rock physics template (RPT) of the Shublik Formation presented as cross-plots of Vp/Vs versus AI. The template includes porosity trends for different fluid saturation (from fully water-saturated Sw=1 to fully gas-saturated Sw=0) assuming uniform saturation. Color bar indicates the range of bulk density. Input parameters are given at right. Blue arrows indicate conceptual geologic trends: (1) decreasing porosity (or increasing bulk density), (3) increasing gas saturation.

There are several challenges in modeling the composition of organic-rich shale and the porosity effects on their velocities. Porosity is not easily determined from log data due to complication in lithology and ambiguity in measurement accuracy, such as neutron tools in the log suite or ultra-low permeability plugs. Therefore, bulk density is used instead of porosity in the RPTs. Additionally, fluid effects on acoustic properties are more problematic because shale lithology defies the main assumptions of Gassmann theory (widely used for clean sandstone rocks) due to rock (clay minerals) and fluid interaction.

conclusion

Major shale lithofacies in the Alaska North Slope can be qualitatively delineated in terms of elastic and petrophysical properties using simple cross-plots. This is especially true for the Shublik Formation, the Hue Shale, and the HRZ. Gamma ray proves to be a better candidate than bulk density to qualitatively separate seismic lithofacies. Cross-plots between elastic properties and TOC or HI show good separation among different shales, but little useful correlation is obtained. Weak inverse correlation between Vp/Vs and TOC is observed in North Slope lithofacies. Organic material is not the sole driving force controlling the velocity-density trend because mineralogy and fluid properties also play a part. Clay content plays a key role in the velocity-density trend of the Kingak Shale, assuming that it is directly related to gamma ray.

Existing shale petrophysical models can be applied if it is properly calibrated to specific regional geology of the North Slope. The soft sediment model is applied to produce rock physics templates, which result in good agreement for bulk density, especially for high density members. These templates show how various geological trends (pressure, saturation, clay content, mineralogy) affect seismic-related attributes (acoustic impedance and velocity ratio Vp/Vs.

A training dataset of elastic properties (P and S wave velocities, bulk density) has been built to advance the statistical rock physics workflow. There is a need to account for different physical scenarios across the area that is prospective for unconventional exploration that might not be present at the wells. A possible solution is to use correlated Monte Carlo simulation to expand the training dataset to account for natural variability. Substantially more core analyses will improve the quality of the training dataset and thus add value for more reliable correlations.

acknowledgments

We wish to thank Great Bear Petroleum for providing the dataset needed for this study. We also appreciate Ken Bird and Les Magoon for providing insightful comments during this study.

References

Avseth, P., Mukerji, T. and Mavko, G., 2010. Quantitative seismic interpretation: Applying rock physics tools to reduce interpretation risk. Cambridge university press.
Bird, K.J., 2001. AAPG Memoir 74, Chapter 9: Alaska, A Twenty-First-Century Petroleum Province.
Gupta, N., Rai, C.S. and Sondergeld, C.H., 2013. Petrophysical characterization of the Woodford shale. Petrophysics, 54(04), pp.368-382.
Hashin, Z., and Shtrikman, S., 1963, A variational approach to the elastic behavior of multiphase materials. J. Mech. Phys. Solids, 11, 127-140.
Houseknecht, D.W., and K.J. Bird, 2004, Sequence stratigraphy of the Kingak Shale (Jurassic–Lower Cretaceous), National Petroleum Reserve in Alaska: AAPG Bulletin, v. 88, p. 279– 302.
Hubbard, R.J., Edrich, S.P. and Rattey, R.P., 1987. Geologic evolution and hydrocarbon habitat of the ‘Arctic Alaska microplate’. Marine and Petroleum Geology, 4(1), pp.2-34.
Keller, M.A., and Macquaker J.H.S., 2001, High resolution analysis of petroleum source potential and lithofacies of Lower Cretaceous mudstone core pebble shale unit and GRZ of Hue Shale, Mikkelsen Bay State #1 well, NSA, in D.W.Houseknecht, ed., NPRA Core Workshop: Petroleum plays and systems in the National Petroleum Reserve-Alaska: SPEM Core Workshop 21, p. 37-56.
Mann, U., Leythaeuser, D., Muller, P.J., 1986, Relation between source rock properties and wireline log parameters: An example from Lower Jurassic Posidonia Shale, NW-Germany: Advances in Organic Geochemistry. v. 10, p. 1105-1112.
Mindlin, R.D., 1949, Compliance of elastic bodies in contact. J. Appl. Mech., 16, 259-268.
Molenaar, C.M., Bird, K.J., and Kirk, A.R., 1987, Cretaceous and Tertiary stratigraphy of northeastern Alaska, in I. Tailleur, and P. Weimer, eds., Alskan North Slope Geology: Bakersfield, California,, Pacific Section, Society of Economic Palentologists and Mineralogists and Alaska Geological Society, p. 513-528.
Parrish, J.T., 1987. Lithology, geochemistry, and depositional environment of the Triassic Shublik Formation, northern Alaska, p. 391-396.
Passey, Q.R., Creaney, S., Kulla, J.B., Moretti, F.J., and Stroud J.D., 1990, A practical model for organic richness from porosity and resistivity logs: AAPG Bulletin, v.74, p.1777-1794.
Peters, K.E., L.B. Magoon, K.J. Bird, Z.C. Valin, and M.A. Keller, 2006, North Slope, Alaska, Source rock distribution, richness, thermal maturity, and petroleum charge: AAPG Bulletin, v. 90, p. 261-292.
Robison, V.D., L.M. Liro, C.R. Robison, W.C. Dawson, and J.W. Russo, 1996, Integrated geochemistry, organic petrology, and sequence stratigraphy of the Triassic Shublik Formation, Tenneco Phoenix #1 well, North Slope, Alaska, U.S.A.: Organic Geochemistry, v. 24, p. 257-272.
Scheirer, A.H., Magoon, L.B., Bird, K.J., Duncan, E. and Peters, K.E., 2014. Toward successful petroleum production from unconventional and conventional reservoirs in the central Alaska North Slope. Unconventional Resources Technology Conference (URTEC).
Scheirer, A.H., Magoon, L.B., Bird, K.J. and Duncan, E.A., 2017, September. Evaluating the Shublik Formation as an Unconventional Resource Play on the Alaska North Slope. In Unconventional Resources Technology Conference, Austin, Texas, 24-26 July 2017 (pp. 3258-3264). Society of Exploration Geophysicists, American Association of Petroleum Geologists, Society of Petroleum Engineers.
Smith, T.M., Sondergeld, C.H. and Rai, C.S., 2003. Gassmann fluid substitutions: A tutorial. Geophysics, 68(2), pp.430-440.
Tran, M.T., 2014. Formation evaluation of an unconventional shale reservoir: Application to the North Slope Alaska (Master’s dissertation, Stanford University).
Tran, M., Aminzadeh, F. and Jha, B., 2018, August. Effect of Coupled Flow and Geomechanics on Transport of a Fluid Slug in a Stress-sensitive Reservoir. In 52nd US Rock Mechanics/Geomechanics Symposium. American Rock Mechanics Association.
Vernik, L. and Milovac, J., 2011. Rock physics of organic shales. The Leading Edge, 30(3), pp.318-323.
Vernik, L., and Landis, C., 1996, Elastic anisotropy of source rocks: Implications for hydrocarbon generation and primary migration: AAPG Bull., 80: 531-544.
Yan, F., Han, D., Yao, Q., 2012, Oil shale anisotropy measurement and sensitivity analysis: SEG Las Vegas Annual Meeting.