Chapter Sections

Introduction

Stratigraphy

Controls on Reservoir Occurrence

Discussion

Acknowledgements

References

Last modified: September 11, 2012

Chapter 21
Cretaceous Viking Formation of the Western Canada Sedimentary Basin

Geological Atlas of the Western Canada Sedimentary Basin

Authors:
G.E. Reinson - Geological Survey of Canada, Calgary
W.J. Warters - Geological Survey of Canada, Calgary
J. Cox - Mount Royal College, Calgary
P.R. Price - National Energy Board, Calgary

Introduction

The Viking Formation and equivalent strata are Albian in age and occur within the Cretaceous Colorado Group (Leckie et al., this volume, Chapter 20). Viking and equivalent units (Bow Island Formation, Paddy Member, Pelican Formation, Newcastle Member) extend over practically the entire Western Canada Sedimentary Basin (Figs. 21.1, 21.2). They represent a wedge-shaped coarse clastic interval that prograded from the Cordilleran orogenic belt eastward into the foreland basin. Interbedded, predominantly marine-influenced sandstones and shales make up the Viking and equivalent units, which range in thickness from a few metres on the eastern margin of the basin to over 175 m in southwestern Alberta. The Viking and Bow Island formations, and Paddy Member have proven to be prolific oil- and gas-bearing units (Fig. 21.3) and, as such, have been the target for exploratory drilling for several decades. Recent estimates indicate that Viking and equivalent strata contain in the order of 5 and 8 percent, respectively, of the oil and gas reserves known to exist in Alberta alone.

This chapter deals in a general sense with regional stratigraphic relations in the Viking clastic wedge, primarily in the subsurface foreland basin of the Western Canada Sedimenary Basin south of 60° latitude. The Dunvegan and Cardium clastic wedges are described in succeeding chapters (Bhattacharya, this volume, Chapter 22, and Krause et al., this volume, Chapter 23, respectively), and basin-wide relations of all units are documented in the preceding chapter (Leckie et al., this volume, Chapter 20). Detailed stratigraphic and sedimentological relations within and between the Viking Formation, Bow Island Formation and Paddy Member are also summarized here. It should be noted, however, that with the current explosive interest in the application of sequence stratigraphic principles, theories on the development and correlation of Viking and equivalent units continue to evolve rapidly.

Previous Work

Much has been written about the Viking Formation because of its historical record as an important hydrocarbon-bearing reservoir in Western Canada. The last basin-wide synthesis of Viking and equivalent units was published in the first Western Canada atlas (Rudkin, 1964). This synthesis was built upon the subsurface investigations of earlier workers such as Glaister (1959), Workman (1959), Stelck (1958) and Hunt (1954). Relevant to Viking and equivalent strata are the papers that focus on the paleogeography of the Colorado Group on a basin-wide scale, including Williams and Stelck (1975), Caldwell (1984), and Smith (this volume, Chapter 17). Regional syntheses of Viking-equivalent strata in Manitoba are those of McNeil and Caldwell (1981) and McNeil (1984). Our understanding of Viking stratigraphy in Saskatchewan is based on the classic paper of Jones (1961), which was followed by a report by Simpson (1982) on the entire Lower Colorado of west-central Saskatchewan.

The need to understand more about where Viking and equivalent hydrocarbon reservoirs can be predicted to occur, and how they can be effectively exploited and developed, led to several, detailed area-specific studies that emphasized depositional controls and environmental interpretation. Linear Viking sandstone reservoirs have been interpreted as deep-water turbidites (Beach, 1956; Roessingh, 1959), nearshore and barrier shoreline deposits (DeWiel, 1956; Shelton, 1973; Tizzard and Lerbekmo, 1975), and offshore shelf sandstone bodies (Evans, 1970; Koldijk, 1976; Boethling, 1977; Beaumont, 1984; Hein et al., 1986; Amajor and Lerbekmo, 1990). Prior to the mid-1980s, Viking sandstone geometries were thought of as relatively long, but narrow, northwest- to southeast-trending bodies less than 5 m in thickness. The documentation of a major unconformity and an estuarine valley-fill deposit within the Viking Formation (Reinson, 1985; Reinson et al., 1988) led to the recognition of similar such deposits elsewhere in equivalent strata (Boreen and Walker, 1991; Leckie and Reinson, in press; Pattison, 1991; Leckie and Singh, 1991).

With the rebirth in the 1980s of transgressive-regressive cycle analysis and the tectonism versus eustasy conundrum in the form of modern "sequence stratigraphy" (Walker, 1990), the above-cited papers have documented unconformities, incised valley-fills and sequence-parasequence boundaries in Viking and equivalent strata, as indicators of basin-wide relative sea-level fall, subsequent transgression, and minor sea-level oscillations. The Viking Formation (along with equivalent units) is now one of the most highly-studied intervals in the subsurface of Western Canada. As such, new studies and re-examinations of older works continue to be published, and our basin-wide interpretations remain in a state of flux.

Geological Framework

The Viking Formation and equivalent strata were deposited in the Western Canada foreland basin which, by Late Albian time, was totally flooded by a shallow epicontinental seaway extending from the Arctic Ocean to the Gulf of Mexico (Williams and Stelck, 1975; Caldwell, 1984). The basin fill consists of thick wedges of westerly-derived coarse clastics alternating with thick, marine shale successions (Stott, 1982). The individual clastic wedges, in this case the Viking, record an episode of major tectonic activity along the marginal orogenic belt, resulting in a high rate of sediment availability and the necessary depositional gradient to effect rapid progradational input into the foreland basin. The Viking clastic wedge was terminated by a major, possibly global, rise of sea level, during Late Albian time (Vail et al., 1977; Weimer, 1984), which inundated the entire basin. Details of the major clastic wedges and sequential basin fill are discussed in Chapter 20.

Major structural elements are evident on the structure map (Fig. 21.3). Positive features include the Bow Island (Sweetgrass) Arch situated in southeastern Alberta and the Swift Current Platform in southern Saskatchewan. A positive platform area also occurs in northwestern Alberta-northeastern British Columbia, on the north side of the Peace River Arch. The Peace River Arch began to subside during the Mississippian. It was still subsiding during Paddy-Cadotte deposition (Leckie et al., 1990), and in fact remained a basin from Early Cretaceous to early Tertiary time (Cant, 1988). The isopach map of Viking-equivalent strata in the Peace River Arch area reflects this subsidence, as evidenced by a thickening trend coincident with the arch (Fig. 21.4). In southern Alberta, the thick Bow Island deposits probably reflect high sediment input from the southwest rather than similar collapse of the Sweetgrass Arch during this interval of Cretaceous time.

Stratigraphy

Stratigraphic Nomenclature

Figure 21.2 summarizes the nomenclature and illustrates the stratigraphic position of the Viking Formation and equivalents in northeastern British Columbia, Alberta, Saskatchewan, Manitoba, and the northwestern United States.

The term Viking was first introduced by Slipper (1918) to refer to the Cretaceous gas-bearing sandstone in the Viking-Kinsella Field of east-central Alberta. The gas-bearing sand within this field lies approximately 45 m above the base of the marine Colorado shale. The pre-Viking portion of the Colorado shale was set apart as the Joli Fou Formation by Stelck (1958). The post-Viking portion of the Colorado shale has been referred to as the Lloydminster shale by Tizzard and Lerbekmo (1975). After Slipper's initial usage at Viking-Kinsella, many other workers used the term to refer to similar sandy sequences enclosed in marine shales of the Lower Colorado Group, occurring at roughly the same stratigraphic level throughout central Alberta and southwestern Saskatchewan. The Viking was designated initially as a member of the Colorado Group (Hunt, 1954; Reasoner and Hunt, 1954; Gammell, 1955), but was later raised to formational status by Stelck (1958).

Glaister (1959) and Rudkin (1964) discussed the regional stratigraphic nomenclature of the entire Colorado Group in some detail, (as do Leckie et al., this volume, Chapter 20). Rudkin (1964) considered the Lower Colorado Group in the central Alberta plains to consist of the basal Joli Fou Formation, the Viking Formation and the overlying Lower Colorado shales, extending upward to the Base of Fish Scales Zone. In southern Alberta, the Lower Colorado includes the "Basal Colorado Sands", the Bow Island Formation (equivalent to Viking and Joli Fou) and the overlying shale interval. In west-central Alberta, the Joli Fou Formation thins rather dramatically, such that the Viking Formation merges with the underlying Mannville Group to form the Blairmore Group of the foothills (Mellon, 1967). In northeastern Alberta and Manitoba, units correlative with the Viking are the Pelican Formation and the Newcastle Member of the Ashville Formation, respectively (Fig. 21.2).

Correlations of Upper Albian strata in northwestern Alberta have been the subject of controversy for many years (Stelck and Koke, 1987; Stott, 1982; Oliver, 1960; Workman, 1959). How members of the Peace River Formation correlate with the Viking and Joli Fou formations in the central Alberta plains has been the main contentious issue (Leckie and Reinson, in press). For example, Workman (1959) and Smith et al. (1984) correlate the Joli Fou Formation with the Harmon Member, whereas Stelck (1958) and Jackson (1984) show the Harmon to be older than the Joli Fou. Furthermore, the Viking Formation is illustrated to be stratigraphically equivalent to the Paddy Member by Stelck (1958) and Koke and Stelck (1984), but equivalent to both the Paddy and Cadotte members by Workman (1959).

Many other early workers correlated the Viking Formation with both the Paddy and Cadotte members in the northwestern plains (Rudkin, 1964). Stelck (1958), however, basing his interpretations on micro- and macrofaunal evidence, did not believe that the Viking was equivalent to either the Cadotte Member or the upper part of the Blairmore Group. Stelck suggested that the Viking, in fact, was deposited during an erosional hiatus that occurred between the Upper and Lower Blairmore of the southern foothills, and between the Mountain Park and Blackstone formations of the central foothills. He further proposed that a hiatus lies between the Cadotte and Paddy members and that only the Paddy is correlative in part with the Viking. The relation between the Paddy and Cadotte members was also thought to be unconformable by Wickenden (1951) and Oliver (1960), who further suggested that the Joli Fou Formation onlapped northward onto an eroded Cadotte surface. The above correlative relations between the Viking and Peace River formations have been thoroughly documented in recent papers using biostratigraphy (Koke and Stelck, 1984; Stelck and Koke, 1987; Leckie and Reinson, in press), and are incorporated in the nomenclature chart (Fig. 21.2).

With respect to the conterminous United States, the Viking equivalent in Montana is referred to as the Vaughn-Bow Island Sandstone of the Blackleaf Formation (McGookey, 1972; Dolson et al., 1991). In Wyoming, the Muddy Sandstone and Skull Creek Shale are correlative with the Viking and Joli Fou formations; in the Denver Basin of Colorado, the term J Sandstone is used to refer to the sandstone unit that is equivalent to the Muddy and the upper part of the Skull Creek Shale.

Regional Stratigraphy and Facies Trends

Regional lithostratigraphy, thickness and lithofacies trends, and contact relations of Viking Formation and equivalent strata are summarized in this section, utilizing the structure and isopach maps (Figs. 21.3, 21.4), the reference logs (Figs. 21.5, 21.6, 21.7), the regional cross sections A to G displayed in Chapter 20, and the more detailed cross section of the Lower Colorado Group interval shown in Fig. 21.8 left pane, 21.8 right pane.

Typical geophysical logs of the Viking Formation (central Alberta; Fig. 21.5), Peace River Formation (northwestern Alberta; Fig. 21.6), and Bow Island Formation (southern Alberta; Fig. 21.7) serve as reference logs to emphasize both the thickness and lithological variability of Viking equivalents and their enclosing strata throughout the foreland basin. Cross sections A to G depict stratigraphic relations on a basinwide scale (Leckie et al., this volume, Chapter 20, Figs. 20.5, 20.6, 20.7, 20.8), and cross section J - J' (Fig. 21.8 left pane, 21.8 right pane) depicts the more detailed correlative relations between the Peace River, Viking and Bow Island formations.

Viking Formation

In the central Alberta plains, where the formation was first defined, the Viking ranges from 15 to 35 m in thickness, increasing to over 65 m in southern Alberta where it merges with the Bow Island Formation (cross sections D, F, H of Chapter 20; and Figure 21.8 left pane, 21.8 right pane). In both westward and southward directions, the thickness of the Viking Formation increases at the expense of the underlying shale of the Joli Fou Formation. In fact, because of the disappearance of the Joli Fou westward, the Viking becomes indistinguishable from the Blairmore Group of the foothills.

The Viking Formation consists of interbedded units of mudstone, sandy mudstone, interbedded sandstone-shaly siltstone, fine- to medium-grained shaly and clean sandstone, coarse-grained to conglomeratic sandstone, and conglomerate. In general, less than 25 percent of the total Viking interval is made up of clean, relatively porous sandstones and conglomerates (Fig. 21.4), the lithologies that have potential to be petroleum reservoirs. Over 75 percent of the Viking sequence, although containing significant intervals of coarse clastic detritus, is relatively silty, shaly and not very porous or permeable.

The upper boundary of the Viking Formation is usually easy to pick on logs because there is a sharp break in lithology from shale to sandstone or sandy siltstone (Fig. 21.5). In spite of the sharp upper boundary of the Viking, some confusion still prevails as to what should be picked on wireline logs as the upper contact of the Viking Formation in central Alberta. Various terms such as Viking Zone, Viking Formation and Viking Sand have been used interchangeably. In addition, terms such as Upper Viking (which consists of dense, silty shale) and Lower Viking (which is dominated by sandy strata), also have been utilized (e.g., Koldijk, 1976).

Consistency can be more easily maintained if the top of the Viking is picked where there is a significant deflection in the electric or radiation logs, such as illustrated by Love (1955, 1960). This distinct deflection corresponds to the highest occurrence of a lithofacies containing a significant amount of sand (Figs. 21.5, 21.8 left pane, 21.8 right pane).

In the Red Deer - Calgary region of southwest Alberta (Fig. 21.8 left pane, 21.8 right pane), the top of the Viking Formation is complicated by the presence of one or more conglomerate beds separated from the main Viking sequence by a shale interval. Viking tops published by the Alberta Energy Resources Conservation Board, and well-operator tops, both refer to this conglomeratic interval as the Viking A Sand or the conglomerate zone; the upper part of the main Viking interval is referred to as the Viking B Sand or main sand zone. The detached Viking A Sand pinches out toward the north and merges toward the south with the conglomerates directly overlying the main Viking succession.

The lower boundary of the Viking Formation in central Alberta is not as abrupt as the upper boundary. The lower boundary appears to be gradational, locally over several metres, with the underlying Joli Fou shales (Figs. 21.8 left pane, 21.8 right pane, and cross sections D, H of Chapter 20). Thus it is difficult to maintain consistency in determining the base of the Viking because of this regional gradational trend.

In contrast, to the west and northwest in Alberta, the base of the Viking displays a relatively sharp log deflection (Fig. 21.8 left pane, 21.8 right pane). Beaumont (1984) attributes the sharp basal contact to an unconformity at the base of the Viking and equivalent units, including the Paddy Member. Similarly, the Viking/Joli Fou contact in southwest Saskatchewan is sharp and well-defined, and Evans (1970) and Jones (1961) suggest the presence of an unconformity between the two formations.

Several studies have shown that the Viking Formation exhibits a relatively uniform facies succession throughout many parts of south-central and central Alberta (Tizzard and Lerbekmo, 1975; Reinson et al., 1983; Beaumont, 1984; Reinson and Foscolos, 1986). This succession was termed the regional shelf-to-shoreface sequence by Reinson et al. (1988), and consists of a series of coarsening-upward, bioturbated sandy mudstone to bioturbated muddy sandstone cycles. The bioturbated sandy mudstones are interpreted as offshore-transitional deposits, whereas the bioturbated muddy sandstones represent lower shoreface deposits resulting from progradation of the shoreface sediment wedge. Almost the entire formation is made up of three to four of these cyclic segments. The middle and uppermost cycles are commonly capped by "clean", fine-grained sandstones, or by partly bioturbated and interbedded sandstone-mudstones, both of which are interpreted as shoreface-incipient bar deposits. Each coarsening-upward segment in the regional shelf-to-shoreface sequences imparts a distinct log signature, with the top of each cycle being expressed by a pronounced curve deflection on gamma-ray or SP logs (Fig. 21.5).

From central Alberta, the Viking becomes progressively thinner and shalier to the east and northeast. About 125 km northeast of Edmonton, the Viking is represented almost entirely by silty shale but is still distinguishable on wireline logs (Rudkin, 1964). In northeastern Alberta, the correlative Pelican Formation contains coarsening-upward sequences of bioturbated, predominantly shaly deposits several metres thick. It consists of thin-bedded graded sandstones interbedded with mudstones and shales, and minor fine-to medium-grained glauconitic sandstone beds (Glass, 1990). The Flotten Lake Sandstone, the correlative unit to the Pelican Formation in west-central Saskatchewan, is similar in thickness and lithology to the Pelican (Simpson, 1982). This marine shelf sandstone, along with most of the Viking Formation, shales out almost completely in central Saskatchewan and along the Saskatchewan-Manitoba boundary (cross sections F, G of Chapter 20). In northern Saskatchewan, the Viking attains thicknesses of 40 m toward the west and is characterized by marine, coarsening-upward sequences similar to central Alberta (Fig. 21.5); however, in southeastern Saskatchewan, the Viking succession is much shalier, has a tendency to fine upward, and displays marine shelf characteristics (Simpson, 1975).

In Manitoba, the Newcastle Sandstone Member of the Ashville Formation is correlative with the Viking Formation to the west (McNeil and Caldwell, 1981; see Fig. 21.2). MacNeil and Caldwell described this member as consisting of interbedded, bioturbated, fine-grained sandstones and shales up to 12 m thick. These shallow-shelf deposits were derived from the east and occur in the subsurface of northeastern Saskatchewan and northern Manitoba as lobate sand bodies.

To summarize, the Viking Formation in central, south-central and southeastern Alberta, and southwestern Saskatchewan, is a shallow-marine unit consisting of several coarsening-upward, shelf-to-shoreface successions, or parasequences, that reflect progradation into a shallow, epicontinental sea. The paleogeography depicted by Smith (this volume, Chapter 17, Fig. 17.9) reflects this general scheme. Eastward and northeastward from central Alberta and southwestern Saskatchewan, Viking and equivalent strata become thinner and shalier, reflecting deposition in deeper shelf environments. West of Calgary, strandline and coastal-plain facies in some Viking successions (Hein et al., 1986) reflect a non-marine influence in the foothills. The pattern of fluvial through shallow-marine to shelf environments eastward, concomitant with the eastward-thinning of the Viking clastic wedge (Fig. 21.4), suggests that Viking sediments were derived from the western Cordillera and debauched into the foreland basin.

As mentioned previously, the cyclical shelf-to-shoreface succession ("regional Viking") is the most common and characteristic facies succession of the Viking Formation. Recently, it has been shown that sandstone-conglomerate channel deposits associated with incised estuarine valley fills, also occur in the Viking Formation (Reinson et al., 1988). Such deposits do not conform to the simple progradational clastic wedge model previously espoused. In fact, the valley-fill deposits suggest the occurrence of an unconformity within the Viking Formation; this unconformity and associated facies sequence variability, are discussed in a subsequent section.

Paddy Member - Peace River Formation

In northwestern Alberta and northeastern British Columbia, the Paddy Member of the Peace River Formation is correlative with the Viking Formation in central Alberta (Figs. 21.2, 21.8 left pane, 21.8 right pane; cross sections D, H, J, Chapter 20). The distribution of the Peace River Formation is restricted to the region of the Peace River Arch, which subsided substantially during Albian time (Leckie et al., 1990; Stott, 1982). The Peace River Formation consists, from the base upward, of the Harmon, Cadotte and Paddy members (Fig. 21.6). Correlative units in the Rocky Mountain Foothills to the west are the Hulcross and Boulder Creek formations (this volume, Chapter 20). The Harmon is a marine shale unit whereas the Cadotte is primarily a shoreface sandstone; thus the common boundary between these two members is fairly well defined on wireline logs (Fig. 21.6). The Harmon-Cadotte contact is locally gradational, displaying a coarsening-upward progradational log signature. The boundary between the Paddy and Cadotte members is normally sharp on wireline logs, owing to the presence of an erosional unconformity (Stelck, 1958; Leckie and Reinson, in press; Leckie et al., 1990). The Paddy Member is restricted to the subsiding Peace River Arch region, and according to Leckie et al. (1990), ranges from 90 m in thickness near the Rocky Mountains to less than 5 m in the Interior Plains.

Correlations by Leckie and Reinson (in press) and Leckie et al. (1990) indicate the Joli Fou and Viking to be correlative with the Paddy Member, with the Joli Fou thinning and pinching out at the top of the Cadotte Member in a northwest direction from central Alberta (Fig. 21.8 left pane, 21.8 right pane). However, northwestward and westward thinning of the Joli Fou as a consequence of an unconformity at the base of the Viking, as suggested by Stelck (1958) and Beaumont (1984), is not consistent with the relationship depicted here, where the unconformity is shown to be at the base of the Joli Fou Formation (Fig. 21.8 left pane, 21.8 right pane).

The Paddy Member consists of a heterolithic succession of coastal plain and fluvial to tidally influenced valley-fill deposits (Leckie, 1989; Leckie et al., 1990; Leckie and Reinson, in press). The paleogeography, illustrated in Figure 17.9 (this volume), depicts the occurrence of a linear, east-west-trending barrier backed by a broad, brackish bay. Although this map is generally correct in a regional sense, the recognition and delineation of an extensive unconformity at the base of the Paddy, combined with documented vertical facies succession variability from estuarine to fluvial to coastal plain deposits, indicate a much more complicated depositional history, discussed later in this chapter.

Bow Island Formation

In southern Alberta and southwestern Saskatchewan, the Bow Island Formation is equivalent to both the Joli Fou and Viking formations (Fig. 21.2). As seen on the regional cross sections, the Joli Fou Formation thins toward the south and southwest as the Bow Island Formation thickens (cross sections F, G, H, Chapter 20; and Fig. 21.8 left pane, 21.8 right pane). The Bow Island Formation, in fact, occupies almost all of the interval between the Base of Fish Scales sandstone and the top of the Mannville in southernmost Alberta (Fig. 21.8 left pane, 21.8 right pane). At the type locality near Bow Island, Alberta, in the Bow Island No. 1 well (6-15-11-11W4), the formation is 175 m thick (Glass, 1990). The Bow Island Formation thins and becomes more shaly eastward and northward (Fig. 21.4), eventually passing laterally into the Viking Formation, with the distinctive presence of the underlying Joli Fou Formation. In the southern Alberta Foothills, the thin, partial equivalent of the Bow Island Formation may be the Mill Creek conglomeratic beds of the Blairmore Group (see Chapter 20). The Bow Island Formation is also recognized in Montana, and in the northwest of the state it is considered to be equivalent to the Vaughn Member of the Blackleaf Formation.

The Bow Island Formation as a whole is much sandier than the Joli Fou/Viking interval but consists of similar lithologies: mudstone, sandy mudstone, thinly interbedded sandstone-shaly siltstone, fine- to medium-grained shaly and clean sandstone, coarse-grained to granular sandstone, and conglomerate. As in the "regional" Viking facies succession (Fig. 21.5), these lithological units are arranged in a series of stacked coarsening-upward sequences (Fig. 21.7).

Although most of the coarsening-upward successions are dominated by shaly lithologies, the uppermost parts are locally capped by relatively clean sandstones. Three of these sandstone units capping individual sequences are informally and locally referred to by the Alberta Energy Resource Conservation Board and well operators as First, Second and Third Bow Island Sands. These gas-bearing sandstone units thin dramatically and decrease in grain size to the north and east; such trends are also reflected in the regional lithofacies map (Fig. 21.4).

As in the Viking Formation, several thin bentonite beds form distinctive markers within the Bow Island, and the upper and lower boundaries of the formation are normally quite distinctive (Fig. 21.7). The upper boundary generally poses no problem for correlation purposes, nor does the lower contact with the Mannville to the east and away from the foothills (Fig. 21.8 left pane, 21.8 right pane). However, adjacent to the foothills, the lower boundary with the Mannville can be difficult to recognize on some wireline logs.

Although exploration drilling operators have informally recognized three principal sands, the reference log (Fig. 21.7) and cross sections (Figs. 21.8 left pane, 21.8 right pane, 21.18) clearly show eight to nine coarsening-upward cycles capped by sandstone units. Detailed stratigraphy and sedimentological studies on the Bow Island Formation are scarce, and much work remains to be done. It is proposed here that the Bow Island Formation can be divided into three characteristic facies successions that are readily apparent on wireline logs (Fig. 21.7): a lower unit, consisting of stacked, coarsening-upward, shelf-to-shoreface cycles; a middle shaly and variably sandy coastal plain succession containing paleosols with associated caliche horizons; and an upper unit, consisting of interbedded sandstones, mudstones and shaly sandstones thought to have been deposited under transgressive conditions in a brackish to embayed shallow-marine environment. The three units recognized in Figure 21.7 are genetic rather than lithostratigraphic units. Interpreted in this manner, the prevailing view of the Bow Island Formation as being composed primarily of stacked, shoreface wedges that prograded basinward in response to high sediment input from a large delta complex to the southwest, is rather simplistic. Figures 21.4 and 17.9 both project such a trend, possibly because the cyclical coarsening-upward, shelf-to-shoreface succession is the thickest facies unit in the Bow Island, constituting over 70 percent of the formation.

In fact, similar to the Viking Formation, a large unconformity is thought to occur within the Bow Island Formation at the top of the coastal plain unit. This unconformity can be traced over a wide area in southern Alberta, and is discussed in a sequence stratigraphic context in the following section.

Sequence Stratigraphy and Depositional History

The regional stratigraphic framework of the Viking Formation and equivalent strata, discussed above, is based historically on the establishment of lithostratigraphic rock units that can be correlated across the basin. In the last decade, there has been a renewed interest in the study of rock relationsips within a chronostratigraphic framework, primarily as the result of the work of Vail et al. (1977). This renewed interest has led to the concept of sequence stratigraphy, defined by Van Wagoner et al. (1990) as, "the study of genetically related facies within a framework of chronostratigraphically significant surfaces". Earlier workers such as Wheeler (1958, 1959), Sloss (1963), and Moore (1964) were the forerunners in establishing chronostratigraphic frameworks and time lines to delineate rock sequences. Unfortunately, these earlier works, which formed the foundation for the stratigraphic concepts of today, have been given short shrift in the enthusiasm to rally around the "new" sequence stratigraphic concepts as documented in Wilgus et al. (1988) and Van Wagoner et al. (1990).

The "new" sequence stratigraphic concepts are now being attempted or applied to almost all foreland basin clastic wedges in the Western Canada Sedimentary Basin. While the complicated nomenclature and jargon that have developed in modern sequence stratigraphy are somewhat cumbersome and possibly pointless, the re-examination of rock sequences in terms of transgressive-regressive couplets and tectonic/eustatic controls will undoubtedly aid in greater understanding of the depositional histories of the major clastic wedges in the Western Canada Sedimentary Basin.

With respect to the Viking Formation and equivalent strata, the resolution of the principal sequences throughout the entire basin has not been done. However, in specific regions of the Western Canada Sedimentary Basin, such as the Viking Formation of central Alberta, and Paddy/Cadotte members in northwestern Alberta, sequence stratigraphic models have been proposed. These models are reviewed here, and a preliminary model for the Bow Island Formation also is outlined.

Viking Formation - Central Alberta

In south-central and central Alberta, the Viking Formation has traditionally been interpreted as comprising a single clastic wedge that prograded basinward in response to orogenic activity to the west. The recognition of a major unconformity in the middle of the Viking Formation in the Crystal Field (Fig. 21.9), which is marked by a deeply incised valley filled with estuarine deposits (Reinson, 1985; Reinson et al., 1988), led to a re-thinking of the Viking Formation as a single regressive clastic wedge. The presence of a sequence boundary in the Viking Formation in Crystal Field was refined through ichnological studies (Pemberton et al., 1992), and further stratigraphic-sedimentological work in adjacent areas (Pattison, 1991; Boreen and Walker, 1991; Posamentier and Chamberlain, in press). A model for sequential development of the valley-fill sequence at Crystal is illustrated in Figure 21.10.

Because of the prominent occurrence of a sequence boundary in the Viking Formation at Crystal, an attempt has been made to map the sequence boundary eastward into the basin. Using facies sequence analysis of well cores and interpretation of wireline log signatures combined with sequence stratigraphic concepts, it can be demonstrated that the Viking Formation in central Alberta consists of two principal stratigraphic sequences separated by a major unconformity (Figs. 21.11, 21.12). Cross sections K-K' and L-L' are based on the recognition of major bounding discontinuities in cored sequences; specifically, the differentiation of flooding surfaces from major unconformities or sequence boundaries (Figs. 21.13, 21.14).

The sequence stratigraphic model for the Viking Formation in central Alberta (Fig. 21.15) depicts the two principal sequences: a lower progradational succession (highstand systems tract), and an upper transgressive systems tract (including incised valley-fill). The highstand deposit consists of a series of stacked, coarsening-upward shelf-to-shoreface parasequences. The sequence boundary truncates the highstand deposit and is commonly marked by a superimposed ravinement surface. This transgressive surface delineates the base of the upper succession, which records transgressive depositional conditions punctuated by stillstand intervals. In the distal or central part of the basin, lowstand(?) and transgressive deposits consist of thin, interbedded units of interlayered mudstone-sandstone, mudstone, and granular sandstone-conglomerate (Fig. 21.14). Near the western margin of the basin, antecedent linear valleys are filled with thick, estuarine sediments deposited during the transgression.

Paddy Member - Northwestern Alberta and Northeastern British Columbia

Similar to the Viking Formation in central Alberta, Leckie and Singh (1991), Leckie and Reinson (in press) and Leckie et al. (1990) recognize the presence of a sequence boundary at the base of the Paddy Member in northwestern Alberta (Fig. 21.16). From examination of outcrops and well cores, the Paddy Member was shown to contain estuarine valley-fill deposits as well as fluvial channel and coastal plain facies. Leckie et al. (1990) mapped the sequence boundary in the region of the Peace River Arch and showed that much of the Paddy Member in northwestern Alberta was deposited in a broad, shallow valley, hundreds of kilometres long and tens of kilometres wide. This valley has downcut into, and erosionally truncates, the underlying sandstones of the Cadotte Member (Fig. 21.17). The formation of the sequence boundary and the subsequent sea-level rise are imprinted in the multiple paleosol sequence characteristic of the correlative Boulder Creek Formation in the Rocky Mountain Foothills to the west (Leckie, 1989, Figure 41).

Bow Island Formation - Southern Alberta

Identification and mapping of three facies successions in the Bow Island Formation (Fig. 21.7) was achieved through core analysis and correlation of wireline log signatures. The three facies successions can be correlated over an extensive area of southwestern Alberta (Fig. 21.18) and a sequence boundary is present at the top of the middle unit (coastal plain facies succession).

A sequence stratigraphic model for the Bow Island Formation in southern Alberta is postulated as follows (Fig. 21.19). The lower unit of the Bow Island Formation averages 110 m in thickness and consists of up to eight, coarsening-upward, shelf-to-shoreface parasequences, each separated by minor flooding surfaces (Fig. 21.7). These parasequences thicken to the south and west and ultimately pass into time-equivalent coastal plain deposits of the middle unit (Fig. 21.18). This middle unit (coastal plain succession) is 30 m or more in thickness, and represents coastal plain progradation in a northeastward direction over the shelf-to-shoreface parasequence set that constitutes the lowermost Bow Island. The coastal plain unit and underlying parasequence set together form a progradational succession or highstand systems tract (Fig. 21.19). A sequence boundary truncates the highstand deposit, and is commonly marked by a superimposed ravinement lag gravel deposit. Overlying the ravinement erosional surface are the shallow-marine deposits of the transgressive systems tract. These transgressive deposits are widespread across southern Alberta and consist of thick, interbedded sandstone-mudstone, bioturbated mudstone, shaly sandstone, and clean, fine- to medium-grained sandstone and conglomerate. All of these lithofacies display evidence of marine biotic and/or tidal influence. Thick, coarse-grained clastic deposits can occur in the transgressive systems tract as channel fill, such as in Blood Field (Cox, 1991).

Controls on Reservoir Occurrence

Sequence stratal units and their bounding surfaces directly control the distributional trends, external geometry and internal heterogeneity of hydrocarbon-bearing reservoirs in the Viking Formation of south-central and central Alberta. From an exploration viewpoint, the highly prospective trends appear to occur at erosional sequence boundaries, parasequence boundaries, or other related bounding surfaces (Fig. 21.15).

Viking reservoirs can be divided into progradational shoreface bars, transgressive bars/sheet sands, and channel deposits related to estuarine valley-fill (Table 21.1). Shoreface bar reservoirs form the uppermost lithofacies of prograding parasequences within the highstand systems tract. Joarcam Field (Jardine, 1954) typifies the shoreface bar reservoir. Coarse-grained to granular bar and sheet sand reservoirs form part of the transgressive systems tract that lies above a major sequence boundary present within the Viking Formation. The Gilby-Bentley Field (Koldijk, 1976) is an example of the transgressive bar/sheet reservoir. Estuarine channel and valley-fill deposits, for the most part, lie within the transgressive systems tract, although it is equivocal whether the lower part of some of the deeply incised valleys are in fact lowstand deposits. Crystal Field is an example of the estuarine channel/valley-fill type of reservoir (Fig. 21.9).

Discussion

It is recognized that several, major, sea-level fluctuations affected the Western Interior Seaway during the Early Cretaceous (Hancock, 1975; Kauffman, 1977; Vail et al., 1977; Weimer, 1984; Haq et al., 1987). Caldwell (1984) suggests that two major transgressive-regressive cycles in the interior plains reflect global eustatic sea-level fluctuations. Furthermore, the two major cycles are thought to contain several transgressive-regressive "subcycles" within their transgressive phases, which Caldwell attributes to local and/or regional tectonic controls.

The regional stratigraphic correlations and facies and sequence stratigraphic analyses documented here indicate that the Viking Formation and equivalent strata may contain evidence of several sea-level fluctuations. The effects of these sea-level fluctuations have been preserved in the sedimentary record of the Viking, Bow Island and Peace River formations in the form of sequence boundaries separating regressive deposits (HST) from transgressive deposits (TST).

When the Viking and Bow Island sequence models are compared (Figs. 21.15, 21.19) along with the correlations in the north-south cross section (Fig. 21.8 left pane, 21.8 right pane), it is evident that the sequence boundary within each formation is one and the same. The regional aspect of this sequence boundary is enhanced by the observations of Jones (1961) who argued for emergent conditions during late Joli Fou deposition in west-central Saskatchewan. Cross-sections K-K' and L-L' (Figs. 21.11, 21.12) illustrate truncation and erosion completely through the Viking Formation and into the underlying Joli Fou, leaving one to postulate that base level drop was substantial enough to effect at least very shallow-water conditions overlying previously covered Joli Fou shales.

The middle Viking unconformity (Crystal unconformity) is thought to represent a basin-wide drop in sea level, not only in central Alberta and western Saskatchewan, but also in the western United States. This is because correlative strata ( J Sandstone) in the Denver Basin contain an unconformable incised valley system similar to that in the Viking at Crystal (Weimer, 1984; Reinson et al., 1988). Weimer considered this unconformity to represent a worldwide sea-level lowstand that occurred at approximately 97 Ma (Vail et al., 1977; Hancock, 1975).

The sequence boundary at the base of the Paddy Member (Fig. 21.16) is thought to be equivalent to the major unconformity between the Mannville Group and Joli Fou Formation in the central Alberta Plains (Leckie and Reinson, in press; Leckie et al., 1990). Thus, this unconformity cannot be considered to coincide with the sequence boundary that is postulated to occur basinwide in the Bow Island-Viking Formation, as discussed above.

At present, the relative roles of eustasy versus tectonic factors are still not resolvable in assessing transgressive-regressive cyclicity and stratigraphic sequences in Late Albian Viking and equivalent strata of the Western Canada foreland basin. However, the occurrence of two basin-wide unconformities (intra-Viking and Paddy/Cadotte) within the Middle to Late Albian succession is at least somewhat substantiated by the results presented here. Eustasy must have played a major role in the occurrence of these sequence boundaries, but the overprinting of local and regional tectonics is clearly evident in the parasequences and flooding surfaces prevalent in the regressive (progradational) sequences, or highstand deposits.

This chapter is not intended to answer all questions, but to present "state of the art" knowledge about a very complex and interesting sedimentary package. Much stratigraphic work remains to be done on Viking and equivalent strata in the Western Canada Sedimentary Basin. In particular, the relation between the Viking and Bow Island formations in Alberta is yet to be investigated thoroughly. As well, the Viking and Bow Island formations in southwestern Saskatchewan should be studied in detail within the context of the Alberta stratigraphic model.

With respect to hydrocarbon occurrences, Lower Cretaceous Viking reservoirs in the Alberta Basin provide outstanding examples of depositional control on reservoir continuity and performance. The variable geometries of reservoir sandstone bodies in the Viking Formation can be explained in terms of depositional framework related to an overall sequence stratigraphic model. This depositional framework also governs the complexity of internal facies-controlled heterogeneities, which in turn govern reservoir productivity. It follows that the geometries and reservoir performance of hydrocarbon-bearing sandstone bodies in both the Aptian-Albian Mannville Group and the Turonian Cardium Formation should also be related to facies complexity caused by variations in depositional trends. As in the Viking reservoirs, these trends should be controlled by an overall sequence stratigraphic framework (transgressive-regressive cyclical model).

Acknowledgements

We thank J. Lerbekmo and F. Hein for constructive critical review of the preliminary manuscript. We wish also to acknowledge useful discussions with and input from several colleagues, including D. Cant, D. Leckie, and H. Posamentier. D. Smith generously provided preliminary paleogeographic maps, and R. Campbell and B. Garner assisted with cross section construction and checking of numerous well logs. We are grateful to W. Ell for word processing of the various manuscripts and P. Gubitz for drafting most of the preliminary figures.

References