Secular changes in Cenozoic arc magmatism recorded by trends in forearc-basin sandstone composition, Cook Inlet, southern Alaska
Published:December 28, 2018
- PDF LinkChapter PDF
Kenneth P. Helmold, Marwan A. Wartes, Robert J. Gillis, David L. LePain, Trystan M. Herriott, Richard G. Stanley, Michael D. Wilson, 2018. "Secular changes in Cenozoic arc magmatism recorded by trends in forearc-basin sandstone composition, Cook Inlet, southern Alaska", Tectonics, Sedimentary Basins, and Provenance: A Celebration of the Career of William R. Dickinson, Raymond V. Ingersoll, Timothy F. Lawton, Stephan A. Graham
Download citation file:
A robust set of modal composition data (238 samples) for Eocene to Pliocene sandstone from the Cook Inlet forearc basin of southern Alaska reveals strong temporal trends in composition, particularly in the abundance of volcanic lithic grains. Field and petrographic point-count data from the northwestern side of the basin indicate that the middle Eocene West Foreland Formation was strongly influenced by nearby volcanic activity. The middle Eocene to lower Miocene Hemlock Conglomerate and Oligocene to middle Miocene Tyonek Formation have a more mature quartzose composition with limited volcanic input. The middle to upper Miocene Beluga Formation includes abundant argillaceous sedimentary lithic grains and records an upward increase in volcanogenic material. The up-section increase in volcanic detritus continues into the upper Miocene to Pliocene Sterling Formation.
These first-order observations are interpreted to primarily reflect the waxing and waning of nearby arc magmatism. Available U-Pb detrital zircon geochronologic data indicate a dramatic reduction in zircon abundance during the early Eocene, and again during the Oligocene to Miocene, suggesting the arc was nearly dormant during these intervals. The reduced arc flux may record events such as subduction of slab windows or material that resisted subduction. The earlier hiatus in volcanism began ca. 56 Ma and coincided with a widely accepted model of ridge subduction beneath south-central Alaska. The later hiatus (ca. 25–8 Ma) coincided with insertion of the leading edge of the Yakutat terrane beneath the North American continental margin, resulting in an Oligocene to Miocene episode of flat-slab subduction that extended farther to the southwest than the modern seismically imaged flat-slab region. The younger tectonic event coincided with development of some of the best petroleum reservoirs in Cook Inlet.
Bill Dickinson’s legacy in the field of tectonics and sedimentation is difficult to overstate. Among his many contributions was the recognition that careful examination of sandstone composition yields important insight into tectonic history (e.g., Dickinson and Suczek, 1979). Many of his key data sets were from the circum-Pacific (Dickinson, 1982); we extend his research approach to southern Alaska by examining how forearc sandstone composition records changes in arc magmatism.
Forearc basins lie between a magmatic arc and an accretionary prism, thus occupying a unique position within the convergent-margin trinity that allows the sedimentary basin to be a sensitive recorder of subduction processes (Dickinson and Seely, 1979; Dickinson, 1995). Southern Alaska experienced a complicated subduction history during the Cenozoic, including early Eocene ridge subduction (e.g., Bradley et al., 2003; Haeussler et al., 2003; Cole et al., 2006) and Oligocene to present collision and subduction of a thickened oceanic plateau, referred to as the Yakutat microplate or terrane (e.g., Plafker et al., 1989; Eberhart-Phillips et al., 2006; Haeussler, 2008). The stratigraphy and structure of the Cenozoic Cook Inlet forearc basin chronicle these events, representing an important example of how the upper plate responds to anomalous subduction modes (e.g., Finzel et al., 2011, 2015, 2016; Haeussler and Saltus, 2011; Ridgway et al., 2012; LePain et al., 2013; Finzel and Enkelmann, 2017). This paper adds to previous work by presenting a large data set on the composition of Eocene to Pliocene forearc sandstone, providing new constraints on the tectonic evolution of southern Alaska.
COOK INLET BASIN STRATIGRAPHY
Cook Inlet is a long-lived, northeast-trending forearc basin in south-central Alaska that extends over 500 km (300 mi) from Matanuska Valley in the northeast to Shelikof Strait in the southwest (Fig. 1). The Cenozoic sedimentary fill is entirely nonmarine and locally reaches more than 7600 m (25,000 ft) thick (Fig. 2; Calderwood and Fackler, 1972; Kirschner and Lyon, 1973; Boss et al., 1976; Shellenbaum et al., 2010). Cenozoic strata overlie a regional unconformity on top of Mesozoic plutonic, metamorphic, volcanic, and sedimentary rocks (Fig. 2; Gregersen and Shellenbaum, 2016). Regional correlation of formation-level Cenozoic stratigraphy of Cook Inlet is complicated by limited exposures along the periphery of the basin (Magoon et al., 1976). In addition, all the lithostratigraphic units likely have highly time-transgressive contacts (LePain et al., 2013).
The oldest documented Cenozoic strata in Cook Inlet are assigned to the middle Eocene West Foreland Formation (Fig. 2). Facies in the West Foreland Formation are highly variable, ranging from alluvial-fan conglomerate along the northwestern basin margin (LePain et al., 2013) to tuffaceous siltstone and minor coal in axial parts of the basin (Calderwood and Fackler, 1972). The age of this unit is now well constrained by extensive new geochronological data from interbedded volcaniclastic facies in the Tyonek quadrangle (Gillis et al., 2018). The Hemlock Conglomerate is generally assigned an Oligocene age based on leaf fossils (Wolfe and Tanai, 1980), although recent studies along the northwest margin of the basin indicate the unit locally ranges from middle Eocene to early Miocene (Fig. 2) and is likely time transgressive (Swenson, 1997, 2003; LePain et al., 2013; Gillis et al., 2018). This broader age range is based on palynological data (Zippi and Loveland, 2012) that have been reassigned due to revised formation criteria and map relations (Gillis et al., 2018). The Hemlock Conglomerate is dominated by conglomerate and pebbly sandstone interpreted as fluvial facies deposited in a low-accommodation setting (LePain et al., 2013). The Tyonek Formation is largely early to middle Miocene in age, although it ranges into the Oligocene in some areas (Fig. 2; Wolfe and Tanai, 1980). Tyonek facies include a thick succession of sandstone, siltstone, and thick coal beds, interpreted as fluvial and associated environments in a high-accommodation setting (Hite, 1976; Flores et al., 2004; LePain et al., 2013). The middle to upper Miocene Beluga Formation (Fig. 2) is typified by mud-rich fluvial deposits and associated coals (Flores et al., 2004; LePain et al., 2009). The Beluga Formation is transitional with the overlying upper Miocene to Pliocene Sterling Formation (Fig. 2), which is composed of sandstone, siltstone, and lignitic coal deposited in a low-sinuosity sandy fluvial system (LePain et al., 2009, 2013).
DATA SETS AND METHODS
The data presented here are part of an ongoing Cook Inlet basin-analysis program led by the State of Alaska. Some of this work has appeared in preliminary agency reports, particularly Helmold et al. (2013).
In total, 177 modal analyses and 137 routine core analyses were performed on outcrop samples of Cenozoic siltstone and sandstone collected from five geographic areas (Capps Glacier, Tyonek, Clam Gulch, Homer, and Chinitna Bay) in upper and lower Cook Inlet (Fig. 3). An additional 61 modal analyses and 60 routine core analyses were performed on conventional core samples from the Shell Middle Ground Shoal A43-11 well. Clast counts were performed in the field on 36 conglomerate beds, and eight additional clast counts were conducted for conglomeratic intervals in the Shell Middle Ground Shoal A43-11 well. Sample details, including reference numbers keyed to a regional map (Fig. 3), geographic location, latitude, longitude, and formation assignment, are available in GSA Data Repository Table S1.1
Sandstone and Siltstone Modal Analyses
Modal analyses were obtained by counting a minimum of 300 grains via the traditional point-counting method (Ingersoll et al., 1984; Decker and Helmold, 1985; Dickinson, 1985) to determine the composition of the framework (detrital grains) and intergranular components (matrix, cement, and porosity). The raw petrographic data are available in Table S2, and reduced summary data are given in Table S3 (see footnote 1). A second count of 200 points (including matrix) was conducted for grain size; a summary of the data is available in Table S4 (see footnote 1). It is important to recognize that our data were collected using the traditional point-counting method (Decker and Helmold, 1985), while the data Dickinson used to construct his provenance fields (e.g., Dickinson, 1985) were collected using the Gazzi-Dickinson point-counting method (Gazzi, 1966; Dickinson, 1970; Ingersoll et al., 1984). Our data were converted to an equivalent Gazzi-Dickinson format by making geologically reasonable assumptions concerning the composition of the plutonic rock fragments using the procedure of Wilson (2002). Based on petrographic examination of the plutonic rock fragments, felsic variants were assigned the composition Q40P30K30A0, where Q is quartz, P is plagioclase, K is potassium feldspar (K-feldspar), and A is amphibole/pyroxene. Similarly, intermediate plutonic rock fragments were assigned the composition Q10P60K10A20, and mafic plutonic rock fragments were assigned the composition Q0P70K0A30. The plutonic rock fragments were then reapportioned to their monocrystalline components.
Routine Core Analyses
Routine core analyses were performed by Weatherford Labs, Inc., in their Anchorage, Alaska, and Casper, Wyoming, facilities to ascertain the porosity and permeability of sandstone samples, which are key factors in evaluating their reservoir potential. The analytical data are available in Table S5 (see footnote 1). Many of the outcrops sampled have not been deeply buried and hence are relatively unlithified. This is of critical importance to routine core analyses because loosely consolidated and disrupted samples can yield anomalous data, particularly for permeability measurements. Initial attempts to acquire good-quality, undisrupted samples employed the use of portable electric drills, conduit pipes, juice cans, and electrical boxes, but all failed for various reasons. A procedure was ultimately developed to acquire intact samples that entailed excavating bread-loaf–sized samples from the outcrop (Figs. 4A) that were immediately wrapped in several layers of plastic wrap (Fig. 4B). These samples were later placed in plastic bags and covered with a thin coating of epoxy resin prior to shipping. Thin sections from these friable samples indicated that most of them remained intact, with only minor disruption of the rock fabric in a few samples.
X-Ray Diffraction Analyses
X-ray diffraction (XRD) analyses of whole-rock and clay fractions of 20 sandstone samples were conducted by K/T GeoServices, Inc., Gunnison, Colorado, to determine the mineralogy of silt- and clay-sized components, principally authigenic clay and zeolite cement. The reported values are available in Table S6 (see footnote 1). XRD patterns from air-dried and glycol-solvated clay-fraction samples were qualitatively analyzed to determine clay types. Determinations of mixed-layer clay ordering and expandability were done by comparing experimental diffraction data from the glycol-solvated clay mounts with simulated one-dimensional diffraction profiles generated using the program NEWMOD. Semiquantitative estimates of whole-rock mineral amounts were acquired using Jade Software (Materials Data, Inc.) with the “Whole Pattern Fitting” option. All data, including clay-mineral abundances, were derived from the whole-rock pattern. A diffraction model was fit to the measured pattern by nonlinear least-square optimization, in which parameters were varied to improve the fit of the model to the observed data. Modeling parameters included background, profile parameters, and lattice constants.
Conglomerate Clast Counts
Clast counts consisted of determining the lithology of 50–100 pebble or coarser clasts in conglomerate beds; a few analyses included fewer than 50 clasts. Identifications were aided by hand-lens examination of fresh surfaces where possible, with many of the lithology classifications being confirmed by multiple geologists. Initial clasts were selected for analysis at random, with subsequent clasts being the nearest neighbor, thereby tracing a semilinear path along the outcrop. Data reduction consisted of consolidating the counts into several broad groups with similar lithologic affinities, which were eventually combined into five categories: volcanic, plutonic, sedimentary, metamorphic, and vein quartz. These data are available in Table S7 (see footnote 1). The Sterling and Beluga Formations generally lack conglomerate facies and are each represented by a single clast count. The Tyonek Formation is represented by 10 clast counts, and the Hemlock Conglomerate and West Foreland Formation each have 16 clast counts.
West Foreland Formation
Sandstone of the middle Eocene West Foreland Formation is distinctly volcaniclastic, with average modal composition of Qt10F23L67, Qm6F23Lt71, Qm22P64K14, and Qp7Lvm91Lsm2 (Tables 1 and 2; Fig. 5; Table S3 [see footnote 1]; Helmold et al., 2013). The average grain size is 0.44 mm (upper medium sand), with an average Folk (1980) sorting of 1.49 (poor; Table S4 [see footnote 1]). Volcanic rock fragments are the dominant framework component (Figs. 6A and 6B) and vary in composition from abundant felsic and intermediate varieties (Fig. 6D), including pumice and bubble-wall glass shards (Fig. 6C), to less common mafic variants. Plagioclase is the principal feldspar (average P/F of 0.79), and exhibits varying degrees of alteration and locally extensive zoning. Monocrystalline quartz is a minor component and is likely of volcanic origin, as suggested by the dearth of inclusions and straight extinction. Hornblende and basaltic hornblende are ancillary heavy minerals that display varying degrees of alteration. Intergranular volume, defined as the sum of intergranular pore space, intergranular cement, and depositional matrix (Paxton et al., 2002), indicates the extent of sandstone compaction. West Foreland intergranular volume averages 30%, due largely to abundant detrital matrix (average 9%) and cement (average 18%), which inhibit compaction. Intergranular clay, consisting of authigenic pore-lining and pore-filling clay (Fig. 6B) and detrital matrix, averages over 16% of the bulk rock. Based on optical properties and XRD analyses (Table S6 [see footnote 1]), authigenic clay consists largely of mixed-layer illite/smectite with a high proportion of smectite layers. The zeolites heulandite and clinoptilolite are common (average 10%) as pore-filling cement (Fig. 6F) and alteration of volcanic rock fragments (Fig 6E; Table S6 [see footnote 1]).
Sandstone of the middle Eocene to lower Miocene Hemlock Conglomerate is fairly quartzose, with average modal composition of Qt60F12L28, Qm25F12Lt63, Qm67P13K20, and Qp65Lvm18Lsm17 (Tables 1 and 2; Fig. 5; Table S3 [see footnote 1]; Helmold et al., 2013). The average grain size is 0.45 mm (upper medium sand), with an average Folk (1980) sorting of 0.90 (moderate; Table S4 [see footnote 1]). Monocrystalline quartz is the major framework grain (Fig. 7A); polycrystalline quartz and chert (Fig. 7D) are slightly less common, yielding an average C/Qt of 0.22 (Table S3). Potassium feldspar (predominantly orthoclase) is more abundant than plagioclase, with average P/F of 0.39. The K-feldspar is relatively fresh (Fig. 7B), whereas plagioclase has sustained variable alteration and dissolution. Rock fragments consist of felsic plutonic rock fragments (Fig. 7C), felsic to intermediate volcanic rock fragments, quartz–mica schist, phyllite, and mudstone/shale. Detrital mica, particularly muscovite, is common and typically has undergone ductile grain deformation (Fig. 7E). Intergranular volume averages 12%, indicating the rock has experienced significant compaction, due in part to limited cementation (average 6%). Pore-filling kaolinite resulting from feldspar alteration is locally abundant and has a detrimental effect on reservoir quality, particularly permeability (Fig. 7F).
Sandstone of the Oligocene to middle Miocene Tyonek Formation is fairly quartzose, with average modal composition of Qt50F18L32, Qm17F18Lt65, Qm51P28K21, and Qp58Lvm23Lsm19 (Tables 1 and 2; Fig. 5; Table S3 [see footnote 1]; Helmold et al., 2013), and it likely shares a common provenance with Hemlock sandstone. The average grain size is 0.45 mm (upper medium sand), with an average Folk (1980) sorting of 0.66 (moderate; Table S4 [see footnote 1]). Monocrystalline quartz, polycrystalline quartz, and chert occur in roughly equal amounts, yielding a C/Qt of 0.28 (Table S3). Together, these siliceous grains comprise the main type of framework grain (Figs. 8A and 8B). Plagioclase and K-feldspar (predominantly orthoclase) are present in similar proportions, with average P/F of 0.52. As in Hemlock sandstone, K-feldspar is relatively fresh, while plagioclase has undergone variable alteration and locally extensive dissolution (Fig. 8E). Major rock fragments include felsic plutonic rock fragments, felsic to intermediate volcanic rock fragments, quartz–mica schist, phyllite, and mudstone/shale (Fig. 8D). Detrital muscovite is an accessory mineral and commonly exhibits ductile grain deformation (Fig. 8C). Intergranular volume averages 21%, which is considerably greater than for Hemlock sandstone, indicating less compaction due to shallower burial. Cements (average 5%) are not a prime agent for porosity reduction, although carbonate-cemented zones are locally significant. Pore-filling kaolinite is a common alteration product of feldspars, and locally occludes intergranular pores (Fig. 8F) and replaces grains. Although it has little effect on porosity, kaolinite decreases permeability, particularly where it obstructs pore throats.
Sandstone of the middle to upper Miocene Beluga Formation is enriched in argillaceous sedimentary and metasedimentary detritus (Levinson, 2013), with average modal composition of Qt28F10L62, Qm8F10Lt82, Qm53P40K7, and Qp25Lvm14Lsm61 (Tables 1 and 2; Fig. 5; Table S3 [see footnote 1]; Helmold et al., 2013). The average grain size is 0.26 mm (lower medium sand), with an average Folk (1980) sorting of 1.25 (poor; Table S4 [see footnote 1]). The Beluga Formation exhibits compositional trends that are transitional with the overlying Sterling Formation (Rawlinson, 1979, 1984; Helmold et al., 2013). Sandstone from the lower and middle parts of the Beluga Formation is composed largely of argillaceous sedimentary and metasedimentary rock fragments, including mudstone/shale (Fig. 9E), siltstone (Fig. 9D), argillite, and slate, resulting in a high structural-clay content (Figs. 9A and 9B). Monocrystalline quartz, feldspar, and chert (Fig. 9F) are minor components of the rock, together comprising less than one quarter of the framework fraction. In contrast, sandstone from the upper part of the Beluga Formation contains plagioclase and volcanic rock fragments in addition to the argillaceous detritus. Intergranular volume for all Beluga sandstone averages 20%, indicating moderate compaction (Fig. 9A). Cementation is not significant, with total cements averaging 2% of the bulk rock. Due to high clay content, Beluga sandstone is susceptible to ductile grain deformation (Figs. 9B and 9C).
Sandstone of the upper Miocene to Pliocene Sterling Formation is largely volcanogenic, with average modal composition of Qt20F20L60, Qm8F20Lt72, Qm28P64K8, and Qp18Lvm65Lsm17 (Tables 1 and 2; Fig. 5; Table S3 [see footnote 1]; Helmold et al., 2013) and P/F of 0.89. The average grain size is 0.29 mm (lower medium sand), with an average Folk (1980) sorting of 0.57 (moderate; Table S4 [see footnote 1]). The composition of the Sterling Formation is transitional with the underlying Beluga Formation (Rawlinson, 1979, 1984; Helmold et al., 2013). In sandstone from the upper part of the Sterling Formation, felsic and intermediate volcanic rock fragments (Fig. 10A), including pumiceous (Figs. 10B and 10C), vesicular vitric (Fig. 10D), and vitrophyric (Figs. 10D and 10E) variants, are the dominant type of framework grain and are virtually unaltered. Hornblende and basaltic hornblende are common and show little alteration or dissolution (Fig. 10F). Twinned plagioclase is the principal feldspar and is largely unaltered; K-feldspar is an accessory component. Monocrystalline quartz is a minor framework constituent and is probably of volcanic origin, as indicated by its water-clear, relatively unstrained character and local euhedral sides. By comparison, sandstone from the lower part of the Sterling Formation contains argillaceous sedimentary and metasedimentary detritus in addition to the volcanic rock fragments and plagioclase. The average intergranular volume of all Sterling sandstone is 30%, indicating the rock has not undergone significant compaction. Cements average 1% of the bulk rock, indicating negligible porosity loss through cementation.
Conglomerate of the West Foreland Formation is dominated by clasts of volcanic origin, with limited secondary input of plutonic clasts (Fig. 11; Table S7 [see footnote 1]). Volcanic clasts are largely of intermediate to mafic lavas and tuffs; plutonic clasts are primarily diorite and granodiorite. The volcanic composition of the conglomerate is reflected in interbedded volcaniclastic sandstone (Fig. 5).
Conglomerate of the Hemlock Conglomerate and Tyonek Formation is similar, consisting of roughly equal proportions of volcanic and sedimentary lithologies and vein quartz (Fig. 11). Plutonic and metamorphic lithologies together comprise less than 10% of the clasts. The quartzose nature of this conglomerate is mirrored in interbedded mineralogically mature sandstone (Fig. 5).
Vein quartz and sedimentary clasts of argillaceous rocks, chert, siltstone, and sandstone are the dominant varieties in the single conglomerate analyzed from the Beluga Formation (Fig. 11). Volcanic and plutonic clasts are present in lesser quantities. The argillaceous composition of the sedimentary clasts is echoed in interbedded sandstone, which consists largely of mudstone, claystone, and argillite grains.
The single conglomerate analyzed from the Sterling Formation (from the eastern edge of the basin) was deposited by braided streams associated with an alluvial fan (LePain et al., 2009). It consists largely of sedimentary clasts of very fine-, fine-, and medium-grained lithic sandstone (Fig. 11). Less common metamorphic clasts consist solely of quartzite. Diorite clasts are rare but conspicuous; an unpublished 40Ar/39Ar analysis from one of these clasts produced an Eocene age (LePain et al., 2009).
Forearc basins are not generally recognized for having significant hydrocarbon potential; only three giant petroleum provinces are known from forearc regions, including the Cook Inlet forearc basin of southern Alaska (Dickinson and Yarborough, 1978; Carmalt and St. John, 1986; Dickinson, 1995). The basin has produced substantial volumes of oil and gas and is important to the economy and energy landscape of the region. The basin is the second most prolific petroleum province in the state, with cumulative production of 1.3 billion barrels of oil from eight oil fields and 7.4 trillion cubic feet of gas from 29 gas fields (Hite and Stone, 2013). Organic-rich rocks of the Middle Jurassic Tuxedni Group are the most likely source of oil in Cenozoic reservoirs (Magoon and Claypool, 1981; Magoon, 1994a, 1994b; Lillis and Stanley, 2011). Oil is produced predominantly from Hemlock (~69%) and Tyonek (~29%) reservoirs, with only minor production from the West Foreland (2%) and Jurassic (0.01%) formations (Hite and Stone, 2013). Nonassociated gas in the Cenozoic reservoirs is almost exclusively of microbial origin, being sourced from coal-bearing Miocene to Pliocene strata (Claypool et al., 1980; Magoon, 1994b). Most gas production is from Sterling sandstone (65%), with secondary production from the Tyonek (27%) and Beluga (8%) Formations (Hite and Stone, 2013). Mean estimates of conventional undiscovered technically recoverable resources are 600 million barrels of oil and 13– 17 trillion cubic feet of gas (Stanley et al., 2011a, 2011b; Hite and Stone, 2013).
Forearc basins are commonly rich in volcanogenic detritus, resulting in labile framework compositions prone to compaction (Pittman and Larese, 1991) and zeolite cementation (Boles and Coombs, 1977). Reservoir quality of Cook Inlet nonmarine sandstone is variable throughout Cenozoic strata and is controlled primarily by original detrital-grain composition, burial history, and depositional environment (Helmold et al., 2013).
Sandstone from the West Foreland Formation has the poorest reservoir potential of the Cenozoic succession, with routine measured porosities of 8%–37% and permeabilities of 0.01– 41 md (Fig. 12A; Table S5 [see footnote 1]). The higher end of the porosity range reflects samples with significant ineffective microporosity (porosity that does not contribute to fluid flow) associated with authigenic clay cement and detrital matrix. Most West Foreland thin sections have limited macroporosity, with average thin-section porosity of 3% and a range of 0%–11% (Fig. 12B). Reservoir quality is graphically illustrated in a porosity-framework-cement (ΦFC) diagram (Fig. 13), a technique originally proposed by Franks and Lee (1994). Virtually all West Foreland sandstone is characterized as nonreservoir (Fig. 13). The poor reservoir quality is due to a combination of labile mineralogy and relatively deep burial. The diagenetic alteration of volcanic rock fragments and plagioclase associated with moderate to deep burial resulted in extensive clay and zeolite cementation. In addition, volcanic rock fragments, particularly highly altered grains, have reduced mechanical strength and are prone to extensive ductile grain deformation (Pittman and Larese, 1991).
Hemlock Conglomerate is the main reservoir in many of the oil fields in Cook Inlet (Hite and Stone, 2013). Sandstone in the Hemlock Conglomerate exhibits a broad range in reservoir quality, with porosities exceeding 20% and permeabilities approaching 1 darcy in the most prospective sandstone (Fig. 12A; Table S5 [see footnote 1]). This results from a combination of quartzofeldspathic mineralogy that is physically and chemically stable, and burial depths generally less than 3000 m. This sandstone is characterized by a well-developed intergranular pore system that is modified by compaction and authigenic clay cementation. Depositional environment also affects reservoir quality. The coarser-grained sandstone and conglomerate typical of channel facies form better reservoirs; siltstone and mudstone of overbank and other fine-grained facies are generally of much poorer quality and typically form reservoir seals. This is illustrated by the wide range in observed permeability and porosity values (Figs. 12A and 12B), and their characterization as both compacted reservoir and compacted nonreservoir (Fig. 13). Hemlock sandstone has significant potential to retain reservoir quality at greater burial depth due to its physical and chemical stability.
Tyonek Formation sandstone comprises oil and gas reservoirs in many of the producing fields in Cook Inlet (Hite and Stone, 2013). Reservoir quality is variable, with porosities exceeding 25% and permeabilities of several darcies in the most prospective sandstone (Fig. 12A; Table S5 [see footnote 1]). They are largely characterized as preserved and compacted reservoir (Fig. 13). This results from a combination of quartzofeldspathic mineralogy and moderate burial depth, as is the case in Hemlock sandstone.
The Beluga Formation sandstone comprises productive gas reservoirs in Cook Inlet, with typical porosities exceeding 25% and permeabilities of several-hundred millidarcies (Fig. 12A; Table S5). The effective porosity of the rock may be significantly lower than the measured porosity due to the prevalence of microporosity in argillaceous lithic fragments (Fig. 12B). Because of small pore throats associated with micropores, the migration of liquids, particularly liquid hydrocarbons, may be inhibited. The argillaceous nature of the sedimentary rock fragments makes them susceptible to ductile deformation with burial, and there is considerable evidence of ductile grain deformation in Beluga sandstone (Figs. 9B and 9C). They are largely characterized as compacted reservoir and compacted nonreservoir (Fig. 13). Deeply buried Beluga sandstone in the central part of the basin possibly has degraded reservoir quality due to intense ductile deformation.
Sterling Formation sandstone comprises the most prolific gas reservoirs in Cook Inlet (Hite and Stone, 2013). It has the greatest reservoir potential of all Cenozoic sandstone in the basin, with porosities exceeding 30% and permeabilities of several darcies (Fig. 12A; Table S5 [see footnote 1]), and it is largely characterized as preserved reservoir (Fig. 13). The excellent reservoir potential is due to the young age (5–1 Ma) and shallow burial (<1500 m) of the sandstone, despite modal compositions (abundant volcanic rock fragments, plagioclase, and amphibole/pyroxene) that are highly susceptible to extensive diagenetic modification. The abundance of unaltered hornblende and vitric volcanic rock fragments attests to the limited diagenetic alteration these rocks have sustained. Sterling sandstone subjected to higher temperatures in deeper parts of the basin probably has undergone some diagenesis, likely resulting in lower reservoir quality.
Pioneering work by Dickinson and Suczek (1979), Ingersoll and Suczek (1979), Dickinson (1982), Dickinson et al. (1982, 1983), Dickinson (1985), and Suczek and Ingersoll (1985) examined the influence of plate tectonics on provenance by evaluating the composition of modern sand and ancient sandstone. This work demonstrated that accurate determination of detrital sandstone modes, presented in a series of ternary diagrams, facilitates assessment of provenance and aids in distinguishing tectonic settings of sedimentary basins. The QtFL diagram (Fig. 14A) displays total grain populations, with all quartzose grains apportioned to the Qt-pole, thereby emphasizing grain stability and hence weathering, provenance relief, transport mechanism, and source rock (Dickinson and Suczek, 1979). The QpLvmLsm diagram (Fig. 14B) displays only partial grain populations, but it highlights the nature of the polycrystalline components of the sandstone framework.
Middle Eocene sandstone of the West Foreland Formation contains abundant volcanic rock fragments and plagioclase that suggest derivation from both undissected and dissected volcanic arc terranes (Fig. 14A). Many samples contain pyroclastic material, including pumice and glass shards (Fig. 6C), that record coeval explosive volcanism. Interbedded conglomerate of the West Foreland Formation contains significant granitoid clasts, in addition to diverse volcanic clasts (Fig. 11), implying derivation from dissected-arc terrane. The sandstone and conglomerate probably have a mixed provenance, including both undissected and dissected segments of the Alaska–Aleutian Range arc (Fig. 14A). This interpretation agrees with detrital zircon data indicating most grains were derived from the nearby arc (Gillis et al., 2018). Detrital zircon data of Finzel et al. (2015) are also consistent with this interpretation, although we ascribe one of their two West Foreland samples (their sample 5) to the Hemlock Conglomerate, which may partly overlap deposition of the upper West Foreland Formation, but which reflects a more integrated regional provenance (Gillis et al., 2018).
Middle Eocene to lower Miocene sandstone from the Hemlock Conglomerate and Oligocene to middle Miocene sandstone from the Tyonek Formation have broadly similar compositions that reflect a mixed provenance. Their quartzofeldspathic nature signals a recycled-orogen and/or dissected-arc provenance (Fig. 14A). The Upper Jurassic to Upper Cretaceous Kahiltna assemblage in the Talkeetna Mountains and southern flank of the central Alaska Range is part of a Mesozoic collisional orogen (Eastham and Ridgway, 2000; Hampton et al., 2007; Kalbas et al., 2007), which probably shed quartzose sediment southward into Cook Inlet (Trop et al., 2003; Trop and Ridgway, 2007; Finzel et al., 2011, 2016; Finzel and Enkelmann, 2017). By early Miocene time, exhumation of the western and eastern segments of the Alaska Range (Haeussler et al., 2008; Benowitz et al., 2011) led to dissection of granite, granodiorite, and tonalite plutons, as indicated by the quartzofeldspathic mineralogy with moderate to low P/F. The presence of volcanic and metamorphic rock fragments suggests supracrustal rocks were possibly associated with the plutons. Paleogene closure of the Matanuska segment of the forearc basin (Trop and Ridgway, 2007) may have shed quartzose sediment southward into the upper Cook Inlet segment of the basin as well. The wide variety of possible source regions is consistent with the diverse detrital zircon spectra that include Cenozoic, Mesozoic, Paleozoic, and Precambrian grains; the pre-Mesozoic grains were likely derived in part from the Yukon Tanana terrane (Finzel et al., 2015, 2016).
Middle to upper Miocene sandstone from the Beluga Formation contains abundant argillaceous sedimentary and metasedimentary rock fragments (i.e., mudstone, siltstone, argillite, slate) that likely originated from muddy/silty components of trench-fill and/or abyssal-plain turbidites incorporated in subducted terranes, a variant of recycled-orogen provenance (Figs. 14A and 14B). Sandstone from the lower and middle parts of the Beluga Formation consists largely of argillaceous sedimentary and metasedimentary rock fragments with only minor quartz, and virtually no volcanic detritus. Sandstone from the upper part of the Beluga Formation has a mixed petrology, consisting of argillaceous sedimentary and metasedimentary rock fragments, volcanic rock fragments, plagioclase, and amphibole/pyroxene. The argillaceous grains were likely derived in part from the accretionary prism to the east, which consists largely of fine-grained deep-sea sediments rafted into the trench, accreted, and subsequently uplifted in the Miocene (Buscher et al., 2008). Alternatively, the lithics could be sourced from the heterolithic Kahiltna assemblage, which also contains abundant argillaceous material (e.g., Helmold and LePain, 2015). The volcanic component in the upper part of the Beluga Formation suggests a growing contribution from the magmatic arc to the west.
Upper Miocene to Pliocene sandstone of the Sterling Formation contains abundant volcanic rock fragments and plagioclase that suggest substantial sediment contribution from magmatic arc terranes (Figs. 14A and 14B). Sandstone from the lower part of the Sterling Formation consists of a mixture of argillaceous sedimentary and metasedimentary rock fragments, volcanic rock fragments, and plagioclase, and it is similar in composition to sandstone from the upper part of the Beluga Formation. In contrast, sandstone from the upper part of the Sterling Formation is volcanogenic, with abundant volcanic rock fragments, high P/F, significant detrital amphibole, and common pumice (Fig. 10B) and unaltered vitric fragments, probably of pyroclastic origin. Such grain compositions generally suggest a magmatic-arc provenance. Detrital hornblende is fresh and unaltered (Fig. 10F), in contrast to altered and partially dissolved hornblende in the West Foreland Formation. The volcanogenic composition and pristine appearance of detrital grains suggest that most of the sand was derived from only slightly older and/or coeval arc-related volcanic deposits; significantly older epiclastic volcanic detritus would be highly altered. The transition upward to arc-dominated sources culminated around 4 Ma (early Pliocene), when the argillaceous sediment source, likely metasedimentary rocks of the accretionary prism to the east, was greatly reduced, and detritus was instead principally derived from the active Alaska–Aleutian Range volcanic arc and other sources unroofed from the growing Alaska Range to the north and west of the basin.
Important compositional trends are revealed in the Beluga and Sterling sandstone data (Rawlinson, 1979, 1984; Helmold et al., 2013). They are best illustrated by dividing the two formations into four stratigraphic intervals (Fig. 15): middle Beluga Formation (middle Miocene), upper Beluga Formation (late Miocene), lower Sterling Formation (late Miocene to early Pliocene), and upper Sterling Formation (early to late Pliocene). For middle Beluga through upper Sterling sandstone, there is a steady increase in volcanic detritus and corresponding decrease of sedimentary lithics (Fig. 15A). This provenance transition is clearly reflected in the detrital modes (Fig. 15B), with monocrystalline quartz + K-feldspar signaling plutonic provenance, plagioclase + volcanic rock fragments indicating volcanic provenance, and sedimentary rock fragments + metamorphic rock fragments signifying source terranes in the accretionary prism and/or Kahiltna assemblage. In middle Beluga through upper Sterling strata, the abundance of sedimentary rock fragments + metamorphic rock fragments systematically decreases from 82% to 11% (normalized mean values), whereas the amount of plagioclase + volcanic rock fragments correspondingly increases from 8% to 79% (Fig. 15B). Over this stratigraphic interval, the quantity of monocrystalline quartz + K-feldspar remains relatively constant at 10%–18%. We interpret that beginning around 10–8 Ma (late Miocene), the Alaska–Aleutian Range volcanic arc became a more prominent sediment source for the basin complex, yielding both epiclastic volcanic (sensu Fisher and Schmincke, 1984) and coeval pyroclastic material. Since then, an increasing proportion of sediment was derived from this western volcanic provenance.
Finzel and Enkelmann (2017) offered a different interpretation for Sterling Formation provenance, favoring a principal eastern source based on detrital zircon U-Pb and hafnium data. However, as Finzel and Enkelmann (2017) pointed out, all their Sterling samples have a notable detrital zircon peak ca. 40–30 Ma, which their data suggest is reflective of a northern provenance. Furthermore, the highly variable ɛHf(t) values (–5.2 to +13.9) in 110–90 Ma detrital zircons from the Sterling Formation do not uniquely signal an eastern source. Based on available data from the western Alaska Range, zircons with more-evolved hafnium compositions appear to reflect interactions with older basement (Todd et al., 2016) and could still be sourced from the north and west. Clearly, more data are required to resolve the alternate interpretations.
There is evidence in the conglomerate clast composition along the eastern basin margin that at least some of the Sterling Formation was derived from the accretionary prism to the east (Fig. 11). The dominance of sedimentary clasts in this conglomerate probably reflects local transverse drainages tapping the Chugach accretionary complex. This interpretation is supported by the Eocene age of a diorite clast (LePain et al., 2009) that is similar in composition and age to near-trench plutons exposed in the Kenai Mountains (Bradley et al., 2003). As such, these data from along the eastern basin margin may not be representative of Sterling conglomerate in the basin center, which is likely more reflective of the pervasive influence of the magmatic arc, as indicated by the Sterling sandstone compositional data.
Flat-Slab Subduction, Arc Magmatism, and Sandstone Composition
The collision and subduction of the Yakutat microplate strongly influenced Cenozoic tectonics, sedimentation, and magmatism across south-central Alaska (Plafker et al., 1994; Haeussler, 2008; Haeussler and Saltus, 2011; Ridgway et al., 2012). Insertion of this relatively buoyant material beneath the North American continental margin resulted in a broad region of flat-slab subduction that has been imaged in several seismological studies (e.g., Eberhart-Phillips et al., 2006). The anomalously low-angle subduction precluded partial melting of subducted material and disrupted the mantle wedge, resulting in a broad “volcanic gap,” where the modern arc is not active above the Yakutat slab (e.g., Nye, 1999).
Observations of the influence of flat-slab subduction on arc magmatism can be used to reconstruct the earlier history of the margin. Forearc-basin strata represent an archive that faithfully records the relative contribution from igneous sources through time. Recent studies have used detrital-zircon U-Pb age data from modern drainages and Eocene to Pliocene forearc strata to elucidate changes in provenance through time (e.g., Finzel et al., 2011, 2015, 2016; Herriott et al., 2014; Lease et al., 2016; Finzel and Enkelmann, 2017). A compiled subset of these data that includes only grain ages 75 Ma and younger (Fig. 16B) generally reproduces the pattern of Finzel et al. (2015). This robust collection of 3490 grain ages can be viewed as a proxy for arc activity in the region. There are three main peaks spanning the intervals from Late Cretaceous to Paleocene, middle Eocene to early Oligocene, and late Miocene to present (Fig. 16B). Each peak is interpreted to represent a major pulse of normal subduction and attendant development of arc magmatism. The youngest peak is assumed to be muted, in part due to sampling bias; grains of this age can only be sourced from young strata, resulting in a smaller relative population. Equally conspicuous are the lulls in detrital zircon abundance, which are recorded in the early Eocene, and again in the late Oligocene to late Miocene. These are interpreted as episodes where normal subduction and arc activity were interrupted. The Eocene hiatus is widely ascribed to ridge subduction that affected southern Alaska (e.g., Bradley et al., 2003). Although several studies have documented the cessation of arc magmatism at ca. 56 Ma (Wallace and Engebretson, 1984; Moll-Stalcup et al., 1994; Todd et al., 2016), widespread dike swarms of intermediate to mafic composition continued to be intruded in the western Alaska Range until ca. 51 Ma, and are interpreted as slab-window magmatism, rather than arc magmatism (Haeussler et al., 2013). These early Eocene intrusive rocks likely account for the ca. 56–51 Ma detrital zircons that partly overlap with the episode of ridge subduction (Fig. 16B). As noted in previous studies, the younger arc hiatus likely reflects cessation of arc magmatism associated with initiation of Yakutat flat-slab subduction (Fig. 16C; e.g., Finzel et al., 2011, 2015, 2016; Finzel and Enkelmann, 2017).
The detrital geochronologic data are well complemented by the sandstone compositional data. The relative abundances of plagioclase and volcanic rock fragments are interpreted to record variations in arc volcanism since the middle Eocene (Figs. 16D and 16E; Wartes et al., 2015). The volcanogenic West Foreland and Sterling Formations correspond, in part, with zircon evidence for coeval igneous activity. The quartzofeldspathic Hemlock Conglomerate and Tyonek Formation overlap a major lull in zircon production and are interpreted to reflect a significant, prolonged hiatus in arc magmatism ca. 25–8 Ma (Fig. 16C). The most likely driver behind the arc shut down is flat-slab subduction associated with the early subduction history of the Yakutat microplate. This interpretation suggests that the area affected by low-angle subduction extended farther to the southwest during the late Oligocene to early Miocene than previously recognized (Fig. 16A; see also Finzel et al., 2011). This is an example of how the upper-plate stratigraphic record provides a critical archive of convergent-margin processes, which is particularly important because aspects of the lower plate are lost to subduction.
Temporal variations in arc magmatism are well recorded in the framework-grain composition of forearc sandstone in Cook Inlet basin. The relative contribution from igneous sources exerts a first-order control on the subsequent resistance to compaction and susceptibility to diagenetic alteration, particularly by zeolites. The most favorable compositions for preserving reservoir quality at depth are found in the more quartzose Oligocene to Miocene strata. Thus, pore space for hydrocarbons is fundamentally tied to plate-margin processes, particularly type and angle of subduction, and associated waxing and waning of arc magmatism.
A comprehensive suite of new sandstone and conglomerate-clast compositional data from the Eocene to Pliocene Cook Inlet forearc basin in southern Alaska documents key changes in provenance through time. The basal Eocene unconformity is the result of uplift caused by subduction of buoyant crust of a spreading ridge that entered the trench and interrupted normal subduction. Middle Eocene strata deposited on the unconformity are dominated by plagioclase and volcanic detritus derived from arc magmatism associated with the return to normal subduction. In contrast, middle Eocene to middle Miocene strata are enriched in quartz and K-feldspar and contain comparatively sparse volcanic contributions. This phase is interpreted to reflect the insertion of the leading edge of the Yakutat slab beneath the upper plate, resulting in dramatic reduction in local arc magmatism. Middle Miocene to Pliocene strata record progressive upward increase in volcanic contributions, consistent with resumption of normal subduction and arc magmatism. Variations in sandstone petrology through time corroborate available detrital-zircon data, and, collectively, these data sets render a robust framework for secular changes in Cenozoic arc magmatism of southern Alaska. This study demonstrates the enduring value of sandstone petrology as a tool for deciphering large-scale tectonic processes.
Variable reservoir quality in the Cenozoic sandstone is controlled primarily by detrital-grain composition, burial history, and depositional environment. Mineralogically mature sandstone, enriched in quartz and K-feldspar, has good reservoir potential even at burial depths of 2600 m (8500 ft). In contrast, mineralogically immature sandstone, consisting largely of plagioclase and volcanic detritus, has low reservoir prospectivity—one exception being volcanogenic sandstone of the Sterling Formation, which has excellent reservoir quality due to its young age and shallow burial depth. Secular changes in arc magmatism played a key role in determining the reservoir potential of Cenozoic sandstone by dictating its original mineral composition. The middle Eocene to middle Miocene hydrocarbon reservoirs of Cook Inlet are a direct consequence of the arc hiatus resulting from abnormal flat-slab subduction.
Funding for this work came from the State of Alaska. The samples analyzed were collected over numerous field seasons by many geologists from several agencies, often working in remote settings. We gratefully acknowledge those geologists, including Nina Harun, David Mauel (Alaska Division of Geological & Geophysical Surveys); Paul Decker, Shaun Peterson, Laura Gregersen, Diane Shellenbaum, Meg Kremer (Alaska Division of Oil & Gas); and Paul McCarthy, Jake Mongrain, and Keane Richards (University of Alaska Fairbanks). We thank Dave Doherty and Bob Swenson for introducing us to parts of Cook Inlet basin and encouraging our early research. Our ideas benefited from discussions with Emily Finzel, Ken Ridgway, Jeff Trop, Peter Haeussler, Jeff Benowitz, Jamey Jones, Erin Todd, and Richard Lease. We thank Apache Corporation for permission to incorporate data from the MGS A43-11 well. Special thanks go to Merlin “Spanky” Handley of Pathfinder Aviation for his professionalism, affable demeanor, and skill in shepherding us in the field. Critical reviews by Sue Karl, Emily Finzel, Ray Ingersoll, Peter van de Kamp, and Shaun Peterson greatly improved the manuscript. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Finally, the senior author gratefully acknowledges the support and mentorship of Bill Dickinson, who served as his dissertation advisor at Leland Stanford Junior University.
Figures & Tables
Tectonics, Sedimentary Basins, and Provenance: A Celebration of the Career of William R. Dickinson
CONTAINS OPEN ACCESS
- clastic rocks
- continental margin
- Cook Inlet
- fore-arc basins
- Hemlock Conglomerate
- igneous rocks
- North America
- plate tectonics
- reservoir rocks
- secular variations
- sedimentary basins
- sedimentary rocks
- Tyonek Formation
- United States
- volcanic belts
- volcanic rocks
- Yakutat Terrane
- Sterling Formation
- West Foreland Formation