Plint.indd Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada: Responses to Tectonic and Sea-Level Controls A. Guy PLINT 1. INTRODUCTION Sedimentary successions record only three possible conditions: deposition, bypass or erosion. Various ques- tions attend each option: if deposition took place, what were the processes, at what rate did they occur, and did the rates change over time? If non-deposition or bypass is indicated, then for how long, and why? If erosion occurred, what was the process, how much sediment was removed, and over what interval of time? These questions could be addressed in an ‘ideal’ stratigraphic succession that contained numerous, accurately-dated layers, and in which physical sequence-bounding surfaces and the main facies enclosed between those surfaces could everywhere be mapped. In reality, these requirements are rarely met in ancient sedimentary suc- cessions. Perhaps one of the best opportunities to investigate the details of a basin’s history is offered by the Meso- zoic strata of the Western Canada Sedimentary Basin. This clastic-dominated basin extends northward into the NW Territories, and southward through the western United States to the Gulf of Mexico. The basin evolved from a divergent margin in the Palaeozoic to a retro-arc foreland basin in the Late Jurassic. In Canada, the Palaeozoic and Mesozoic strata in this basin contain abundant oil and gas reserves which have been extensively investigated through several hundred thousand boreholes. Fair to excellent quality well logs are publicly available for every well, and abundant core has been cut from reservoir intervals. Although seismic sections are numerous, they are not publicly available. These rocks are also extensively exposed along the western margin of the basin in thrust sheets that form the Front Ranges and Foothills of the Rocky Mountains. Correlation of well logs in the basin to outcrop sections in the mountains not only allows logs to be calibrated against exposed lithologies, but also allows isolated outcrop sections to be placed in a regional stratigraphic and palaeogeographic framework. The main deficiency of the stratigraphic record is the relative paucity of well-dated stratigraphic marker hori- zons, although marine strata yield fossils that can, in the Cretaceous part of the section, provide relative age resolution of about 0.5 Ma. Geologia Croatica 56/1 39–68 31 Figs. ZAGREB 2003 Key words: Cretaceous, Cenomanian, Western Cana- da, Foreland basin, Deltas, Sea-level change, Tecto- nics, Eustasy, Clastic sedimentation, Sequence stra- tigraphy. Department of Earth Sciences, University of Western Ontario, London, Ontario N6A 5B7, Canada; e-mail: gplint@uwo.ca Abstract The early–mid Cenomanian Dunvegan Formation represents a large delta complex that prograded at least 400 km from NW to SE. A regional stratigraphy based on marine transgressive surfaces and equivalent subaerial interfluves allows the formation to be subdivided into ten transgressive–regressive allomembers, labelled J to A in ascending order, each with an average duration of < 200 ky. Analy- sis of stacking patterns and facies distributions of parasequences within allomembers allows transgressive, highstand, falling stage and lowstand systems tracts to be identified. Extensive valley sys- tems that average 1–2 km wide and 21 m deep can be traced for up to 320 km across the top surfaces of allomembers H to E. In their lower 20–40 km, valleys are filled with muddy heterolithic tidal facies but this changes to fluvial-dominated multi-storey channel-fills further up-valley. Interfluve surfaces are marked by palaeosols, the character of which indicate a protracted hiatus with extensive physical, chemi- cal and biological modification of the parent material. Changes in flexural subsidence rate are indicated by isopach pat- terns. Allomembers J–F have a sigmoidal prismatic geometry, succes- sively offlapping to the SE. There is no evidence of thickening toward the orogen. In contrast, overlying allomembers E–A show progressive development of a depocentre along the western margin of the basin. The increasing accommodation rate on the updip coastal plain caused marine deltas to be starved of sediment, leading to progressive back- step of shorelines. Simultaneously, alluvial deposits within the depo- centre show an upward increase in the proportion of subaqueous to subaerial facies, culminating in the incursion of brackish and finally marine waters. Thus tectonic subsidence rate had a first-order affect on both the volume of sediment available to build marine deltas and also on the local character of facies that accumulated on the coastal plain. The onset of flexural subsidence in allomember E appears to have resulted in subtle uplift of a forebulge, resulting in dramatic deflection of river systems. Despite the clear tectonic signature, successive transgressions and regressions involved similar horizontal displacements of the shore- line, regardless of subsidence rate. This suggests that modest eustatic changes also influenced the accommodation available. Based on the measured horizontal excursions of the shoreline, the vertical thickness of alluvial strata, and realistic alluvial gradients, an average eustatic excursion of about 24 m is calculated. The incision of valley systems is attributed in part to periods of eustatic fall. However, valleys seem too long to be explained by eustasy alone, and hence secular changes in discharge are postulated as an additional forcing factor. Climatic cycles in the Milankovitch band may have been responsible for both eustatic and discharge variations. 40 Geologia Croatica 56/1 This paper is concerned with the Dunvegan Forma- tion, which is an early to mid-Cenomanian deltaic unit. This 90–270 m thick clastic wedge was deposited on the western margin of the Western Interior Seaway (Fig. 1), during a period of generally rapid subsidence of the North American plate resulting from dynamic and static loading mechanisms due to subduction and crustal shortening (MITROVICA et al., 1989). At least 400 km of NW to SE delta progradation can be demonstrated within the study area, which took place over about 2 My. The Dunvegan Fm. is extensively exposed in thrust sheets for several hundred km along the NW–SE trending Rocky Mountain Foothills, and also for 300 km in undeformed strata exposed in an E–W transect along the Peace River valley (Fig. 1). Over a study area of about 80,000 km2, the stratigraphy of the Dunvegan Fm. was delineated using 2340 well logs and 60 major outcrop sections (PLINT, 2000; Fig. 2). 1.1. Aim of this paper This paper summarizes the main results from previous studies concerned with various aspects of the Dunve- gan Formation (PLINT, 1996, 2000, 2002; PLINT et al., 2001; McCARTHY & PLINT, 1998, 1999, in review; McCARTHY et al., 1999; McCARTHY, 2002; LUMSDON & PLINT, 2003 – in press; PLINT & WADSWORTH, in review). The analysis of the depo- sitional history of the Dunvegan has involved several main steps. 1) Development of a robust allostratigraphy based on marine transgressive surfaces. This allows lateral facies relationships to be determined, and also per- mits isopach maps to be constructed. It is assumed that the bounding surfaces of allostratigraphic units approximate time lines. Unfortunately, there is no means available to determine absolute ages for the bounding surfaces. Fig. 1 Inset maps show location of study area in western Canada, and Cenomanian palaeogeography of the Dunvegan Formation within the Western Interior Seaway. The line of section shows the location of the dip section shown schematically in Fig. 3, and drawn to scale in Fig. 23 (from PLINT & WADSWORTH, in review). 2) Construction of palaeogeographic maps for each allomember showing maximum marine transgres- sive and regressive limits. This provides a picture of the regional shoreline trends and also permits quan- tification of horizontal movements of the shoreline. 3) Construction of isopach maps for each allomember. These show the geometry of the accommodation available for each allomember. Because each allo- member aggraded essentially to sea level, isopach patterns can be interpreted in terms of subsidence patterns. 4) The analysis of sedimentary facies within a palaeo- geographic and accommodation framework allows vertical and lateral facies changes to be interpreted in terms of proximity to the shoreline and to spatial variation in accommodation rate. 5) Mapping of palaeo-valley systems in both subsur- face and outcrop. The resulting maps reveal the influence of various tectonic features (faults, fore- bulge) on drainage systems. 6) Analysis of palaeosols mantling interfluve surfaces between valley systems. This reveals both vertical and lateral variations in the character and evolution of palaeosols that reflect palaeodrainage conditions and rates of clastic flux related to distance from the margins of valleys. 41Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... Fig. 2 Map of study area showing stratigraphic data base, including location of 2340 well logs and principal outcrop sections in the Foothills and Peace River valley (from PLINT, 2000). 42 Geologia Croatica 56/1 7) Sequence stratigraphic interpretation, including defi- nition of systems tracts on the basis of facies associ- ations, stratal lap-out patterns, palaeogeography and palaeogeomorphology. The main aim is to interpret the principal controls (tectonic, eustatic, climate) on changing patterns of deposition and erosion, utiliz- ing elements 1–6 above. 1.2. Collaboration This paper presents a compilation of results from the research of the author on the regional stratigraphy and sedimentology of the Dunvegan Fm. However, very significant contributions to this understanding have been made by my post-doctoral fellows and graduate students whom I wish to acknowledge here. Dr. Jen- nifer WADSWORTH investigated the sedimentology of the valley-fill deposits exposed along the Peace River, was the first to recognize tidal features, and made pre- liminary maps of their subsurface distribution. Dr. Paul McCARTHY undertook palaeopedological investiga- tions of the interfluve surfaces between the valleys, and confirmed the existence of major hiatal surfaces. Bira FACCINI spent part of his PhD project investigating the sedimentology and stratigraphy of the Dunvegan strata in the valley of the Kiskatinaw River, and documented the vertical changes in alluvial facies in response to changes in accommodation rate. Matthew LUMSDON conducted detailed facies analysis of alluvial strata near the orogenic margin of the basin and showed how drain- age conditions and facies successions on the alluvial plain were strongly linked to regional subsidence rates and patterns. The work of these colleagues is indicated at appropriate points by references to our publications. 2. REGIONAL GEOLOGICAL SETTING During the Cretaceous, North America was progres- sively flooded from the south by warm saline waters of the ancestral Gulf of Mexico – Tethys, and from the north by cooler, less saline water of the Boreal Ocean. These marine embayments merged just before the Albi- an–Cenomanian boundary (96 Ma) to form the West- ern Interior Seaway, which continued to widen until mid-Turonian time (90 Ma) (WILLIAMS & STELCK, 1975; KAUFFMAN & CALDWELL, 1993; HAY et al., 1993; Fig. 1, inset). Static and dynamic loading of the lithosphere above the subducting Farallon Plate pro- duced a wide sedimentary basin comprising two main parts. Static loading by the orogenic wedge generated a relatively narrow (< 300 km), rapidly-subsiding flexu- ral foredeep, to the east of which lay a broader region, 1000–1500 km wide, of moderate subsidence that reflected dynamic loading by large-scale mantle flow (MITROVICA et al., 1989; PANG & NUMMEDAL, 1995; CATUNEANU et al., 1997). Thus, the Interior Seaway can be attributed, in part, to static and dynamic loading of the North American Plate. However, eustatic rise, broadly ascribed to accelerated plate spreading rates (KAUFFMAN & CALDWELL, 1993), was also a factor with Turonian sea level estimated to have peaked at 180 m (SAHAGIAN & JONES, 1993) to 300 m (McDONOUGH & CROSS, 1991), above present level. Relative sea level rise at the Albian–Cenomanian boundary led to widespread deposition of marine shales, marking the onset of the Greenhorn Cycle (KAUFF- MAN & CALDWELL, 1993). In Alberta and British Columbia, this transgressive phase was accompanied by rapid flexural subsidence, recorded by a westward- thickening wedge of laminated, organic-rich marine shale of the Shaftesbury Formation (equivalent to the Hasler, Goodrich and Cruiser formations in the NW part of the basin in British Columbia; STOTT, 1982; Fig. 3). The Shaftesbury Formation coarsens upward into deltaic strata of the Early to Middle Cenomanian Dunvegan Formation which represents a major pro- gradation of the shoreline. However, unlike the fluvio- deltaic Mountain Park Fm. strata below the Shaftesbury Formation, which prograded towards the north and NW, progradation of the Dunvegan Fm. was in the opposite direction, towards the SE, parallel to the Cordillera (Fig. 3). This change in regional sediment dispersal direction probably records the northward migration of the centre of maximum Cordilleran uplift as a result of the lateral migration of accreted terranes along the con- tinental margin during the Cretaceous (EISBACHER, 1981; POULTON, 1994). The principal source of sedi- ment for the Dunvegan Formation is considered to have lain in northern British Columbia and southern Yukon, where the formation is dominated by alluvial sand- stones and conglomerates (STOTT, 1982). Palaeocur- rent observations (PLINT, unpubl. data) and mapping of valley systems and shorelines (PLINT, 1996, 2000, 2002; PLINT & WADSWORTH, in review) show that the principal fluvial palaeoflow direction was toward the SE, although several subordinate sources of sedi- ment were scattered along the Cordilleran margin of the basin further to the south. Isopach maps of successive allomembers of the Dun- vegan Fm. (discussed in more detail below), show that flexural subsidence of the foredeep was minimal during deposition of the older five allomembers, J–F, each of which constitute broadly prismatic packages of sedi- ment that show no significant thickening towards the orogen. In contrast, the younger five allomembers, E–A show major thickening towards the orogen, indicating the onset of renewed subsidence during and following allomember E time (PLINT et al., 2001, LUMSDON & PLINT, 2003 – in press). This phase of subsidence led to major backstep of the Dunvegan delta complex in allomember A time as accommodation rate gradually exceeded sedimentation rate, culminating in widespread marine transgression at the base of the overlying Kaska- pau Fm. (Fig. 3). 43Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... 3. STRATIGRAPHY 3.1. Lithostratigraphy The Dunvegan Formation ranges from 90–270 m thick and crops out extensively on thrust sheets in the Alberta and British Columbia Foothills, and also in the Peace River Valley, where undeformed strata are exposed in the vicinity of Fort St. John, and extend for over 300 km to the east (STOTT, 1982; Fig. 2). The Dunvegan Formation comprises a succession of alluvial and shallow marine sandstones, siltstones and shales that separate underlying marine shales of the Shaftesbury Formation from overlying marine shales and minor sandstones of the Kaskapau Formation (Fig. 3). It has long been recognised that the Dunvegan Formation has an interfingering, diachronous relationship with the underlying Shaftesbury, and overlying Kaskapau formations (STELCK & WALL, 1955; STELCK et al., 1958; ALBERTA SOCIETY of PETROLEUM GEOL- OGISTS, 1960; SINGH, 1983), and the boundaries of the Dunvegan Formation are placed at the first and last appearance of sandstone in an otherwise shale-domi- nated succession. For mapping purposes, this definition serves well. However, in order to analyze in more detail the depositional history and palaeogeography of the Shaftesbury, Dunvegan and Kaskapau formations, it is useful to subdivide the rocks into smaller, genetically- related packages on the basis of mappable surfaces. Such an allostratigraphic approach was introduced by BHATTACHARYA (1989). 3.2. Allostratigraphy Allostratigraphy involves the subdivision of the strati- graphic record into packages “defined and identified on the basis of their bounding discontinuities” (NORTH AMERICAN COMMISSION ON STRATIGRAPHIC NOMENCLATURE, 1983). Bounding discontinuities can take several forms, including subaerial erosion surfaces, the upper surfaces of palaeosols, erosional marine ravinement surfaces, and weakly-erosional to non-erosional marine flooding surfaces (Fig. 4). In a subsurface study, BHATTACHARYA (1989) demonstrated that the Dunvegan Formation could be subdivided into seven transgressive–regressive pack- ages, bounded by regional marine flooding surfaces. The seven packages, or allomembers, were designated G to A in ascending order, and each was shown to con- sist of several component, sandier-upward successions, termed ‘shingles’ (broadly equivalent to parasequences; BHATTACHARYA, 1989; BHATTACHARYA & WAL- KER, 1991a). Analysis of facies successions in core led BHATTACHARYA & WALKER (1991b) to interpret the Dunvegan Formation in terms of a series of pro- grading and aggrading delta complexes. Because of their regional extent, allomembers were interpreted to record an allogenic, tectonic control on subsidence rate. Fig. 3 Schematic NW–SE dip section showing the Dunvegan Fm., the stratigraphic terminology applied to the under- and overlying units, and the stratigraphic range of important Zone fossils. The boundaries of the lithostratigraphic Dunvegan Fm., and the allostratigraphic Dunvegan alloformation are also indicated. The former is very diachronous (from PLINT et al., 2001). 44 Geologia Croatica 56/1 In contrast, the more localised distribution of individual shingles suggested an autogenic origin, controlled by delta distributary switching (BHATTACHARYA & WALKER, 1991a). PLINT (1996, 2000) extended the allostratigraphic scheme of BHATTACHARYA & WALKER (1991a), tracing the Dunvegan northward from township 67 to correlate with outcrop sections on the Peace and Beat- ton rivers. An additional three allomembers, J–H were added to the original scheme of BHATTACHARYA (1989), and it is likely that further north, yet more allomembers could be defined, but which downlap and pinch out before reaching the study area. 3.3. Recognition of depositional sequences It can be shown that the marine ravinement and flood- ing surfaces that define allomembers in more down- dip areas can be traced updip into surfaces bounding well-developed interfluve palaeosols (McCARTHY & PLINT, 1998; McCARTHY et al., 1999; PLINT et al., 2001). These palaeosol-bounding surfaces can in turn be traced into erosional surfaces that define a network of valleys incised into the upper surfaces of most of the Dunvegan allomembers (Fig. 4). Therefore, each component allomember of the Dunvegan Formation preserves evidence for relative sea level rise, expressed by sandstones filling valleys, by the abrupt juxtaposi- tion of offshore marine or lagoonal mudstone above sandier, shallower-water deposits separated by a flood- ing or ravinement surface, and by changes in fluvial style in coastal plain areas (McCARTHY et al., 1999). Evidence for relative sea level fall and fluvial incision is provided by extensive valley systems (PLINT, 2002; PLINT & WADSWORTH, in review), interfluve palaeo- sols (McCARTHY et al., 1999), and, in shallow marine areas, by erosive-based delta-front sandstones (PLINT, 1988, 1996; PLINT & NUMMEDAL, 2000). Of the ten Dunvegan allomembers studied, only allomembers J and E show a distinct downward shift of facies and onlap in the delta front area, as is required for the definition of a classical Exxon type 1 sequence (VAN WAGONER et al., 1988). The remaining allo- members show well-developed offlap to subtle onlap of Fig. 4 Diagram showing how allomembers and depositional sequences are defined. Allomembers are bounded by regional marine transgressive surfaces (heavy dashed lines) and their lacustrine equivalents on the coastal plain. Valley fills are included with the underlying allomember. Depositional sequences are bounded on the coastal plain by a subaerial unconformity that defines valleys and interfluves. In downdip areas, this surface continues as the upper surface of the offlap- ping, falling stage systems tract (which may be a palaeosol, or a ravinement surface), and, basinward of the shoreline, as the correlative conformity. This surface is onlapped by valley- filling and marine deposits of the lowstand and transgressive systems tracts. Recognition of the four marine systems tracts relies heavily on regional mapping of shingle stacking patterns (from PLINT et al., 2001). 45Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... parasequences in the delta front area and represent both falling stage and lowstand systems tracts, sensu PLINT & NUMMEDAL (2000), and PLINT et al. (2001). 4. SEDIMENTARY ENVIRONMENTS AND FACIES The Dunvegan Fm. can be subdivided into three main depositional environments: prodelta, delta-front and coastal plain (including valley-fills). The main sedi- mentary characteristics of each of these environments are summarized below. 4.1. Prodelta Prodelta deposits consist mainly of cm to mm-bedded dark grey siltstones and mudstones. Interbedded very fine sandstones are mm to cm thick and become thicker and more abundant upward (Fig. 5). Thicker sandstone interbeds are parallel-laminated or show wave or com- bined-flow ripples. Rare, cm- to dm-scale gutter casts may be present on the base of sandstone beds. In the more offshore facies, gutter casts are oriented parallel to shore, but in more nearshore deposits, they rotate to shore-perpendicular. Bioturbation is generally sparse, and includes Chondrites, Planolites, Teichichnus, and Zoophycos. These sediments are organized in 5–20 m thick, sandier-upward successions interpreted to record progradation of individual delta lobes (Fig. 6). 4.2. Delta-front Delta-front deposits are dominantly fine- to locally medium-grained sandstone and form units from 1 to about 10 m thick. The bases of sandstone bodies may be either gradational with underlying prodelta facies (Fig. 7), or sharp and erosive, commonly with spec- tacular gutter casts (Figs. 8, 9). Swaley cross-strati- fication and parallel lamination typify fine sandstone whereas dm-scale trough crossbedding is present in the upper fine- to lower medium-grained sandstones. The delta-front sandstone bodies are commonly capped by 1–2 m of planar-laminated to massive rooted sandstone interpreted as beach and backshore deposits. Sharp- based delta front sandstones are widely-recognized throughout the formation and are interpreted to record shoreface progradation during relative sea level fall (i.e. during the falling stage systems tract; PLINT & NUMMEDAL, 2000). Delta-front sandstones can be mapped in well- logs, and their maximum progradational extent in each allomember can be traced. For each allomember, shorelines face consistently towards the E to SE and Fig. 5 Detail of prodelta facies consisting of cm-scale interbeds of very fine sandstone and mudstone. Sandstone beds are sharp- based with small grooves and flutes, and upper surfaces are typically ornamented with combined-flow ripples. Beatton River section 1; scale bar = 20 cm. Fig. 6 Three successions of prodelta storm beds grading up into thicker sandstones of the delta front. The main package of sandstone is 4 m thick and is capped by a coal. Flood Creek, Alberta. 46 Geologia Croatica 56/1 can be traced continuously for several hundred km. The palaeogeographic maps of allomembers J to A (Figs. 10–16) show that the delta shoreline was broadly cus- pate to lobate, suggestive of moderate to strong wave reworking of individual delta lobes. Large bulbous delta lobes up to 80 km wide and 100 km long are recognized in allomembers I, D, E, and F, and are interpreted as lowstand deltas deposited at the end of periods of slow relative sea level fall. Fig. 7 Stacked prodelta to delta-front facies successions. The mid- dle succession is 4 m thick; Sulphur River, Alberta. Fig. 8 Delta front sandstone unit resting erosively on dark laminat- ed prodelta mudstone. Sandstone is 7 m thick and is sharply overlain by laminated dark marine mudstone; person for scale. Railway cutting section, Grande Cache, Alberta. Fig. 9 Detail of large canoe-sha- ped gutter casts on the base of the delta front sandstone shown in Fig. 8. Gutter cast is 80 cm. wide. Railway Cut- ting section, Grande Cache, Alberta. 47Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... 4.3. Coastal Plain Coastal plain deposits dominate the NW portion of the delta complex and comprise a complex array of facies broadly representative of a generally poorly-drained environment, traversed by numerous non-migrat- ing river channels that formed ribbon sandstones in a matrix of lacustrine, marsh and poorly-drained flood- plain mudstone, interstratified with crevasse splay and thin lacustrine delta sandstones. The abundance of tracks shows that the area was inhabited by numerous dinosaurs, primarily Hadrosaurs, but with Ankylosaur, Theropod, crocodilian, large avian and possibly pterosa- ur tracks also recognized (M. LOCKLEY, pers. comm., 1997; R. McCREA, pers. comm., 1999; SCOTT, 2000). Coastal plain facies have been grouped into three facies associations that form distinct nonmarine systems tracts (PLINT et al., 2001) that reflect differing sedimenta- tion:accommodation ratios. The facies characteristics of each systems tract are discussed in more detail below. 4.3.1. Channel-dominated, low-accommodation systems tract This systems tract consists of multi-storey, laterally-ac- creted meandering channel sandstones, typically 5–8 m thick, that were only deposited within valleys, the base of which forms the lower boundary of the systems tract. There is very little floodplain mudstone preserved within the valley-fills. Within a valley-fill, mud-free, cross-bed- ded sandstone generally dominates the lower storeys, whereas the uppermost, and most complete storey is sometimes much more mud-rich, with well-developed heterolithic stratification (Figs. 17, 18). This suggests that the latest stage of valley-filling may have taken place under tidally-influenced conditions (cf. SHANLEY et al. 1992; SMITH, 1987; WILLIS, 1997). The uppermost Fig. 10 Palaeogeographic and isopach map of Dunvegan allomember J. Heavy stippled line in this and figures 11–16 indicate the maximum progradational extent of the delta-front sandstone, mapped from gamma ray log signatures. Note how isopach lines trend perpendicular to the orogen in this map and in Figs. 11–14. This pattern is interpreted to reflect deposition at a time when the rate of flexural sub- sidence was negligible. Contour interval in figures 10–16 = 10 m. 48 Geologia Croatica 56/1 1–5 m of the valley-fill is sometimes occupied by a sand- ier-upward succession, commonly highly bioturbated, interpreted to record final filling of the valley by a small bay-head delta (Figs. 18, 19). The bioturbation suggests incursion of brackish water, and this is consistent with the progressive drowning of the valleys and updip mi- gration of marine influence (cf. McLAURIN & STEEL, 2000). The upper boundary of the systems tract is placed at the generally abrupt transition (an expansion surface of MARTINSEN et al., 1999), to lacustrine-dominated facies, that record a major change in the accommodation/ supply ratio. The prevalence of multi-storey sandbodies and scarcity of preserved floodplain mudstone suggests that these deposits reflect a relatively low accommoda- tion rate that permitted protracted reworking of alluvial deposits within the valley (cf. ASLAN & BLUM, 1999; MARTINSEN et al., 1999). Available evidence does not permit us to differentiate sediments deposited during regional base-level rise from possible terraces deposited during the falling stage (cf. BLUM & PRICE, 1998). 4.3.2. Lacustrine-dominated, high-accommodation systems tract This systems tract is characterized by widespread, well-laminated to poorly-laminated and carbonaceous mudstone and coal facies that represent a range of lake and wetland environments. These facies enclose, and are interstratified with ribbon channel sandbodies (Fig. 20), crevasse splays and sandier-upward, lake-fill successions (Fig. 21). Palaeosols are generally poorly developed and hydromorphic in character (palaeosol types 1–3, fig. 8 of McCARTHY et al., 1999). The base of this systems tract is defined in two ways. It may abruptly overlie multistorey channel sandstones filling valleys, where the facies transition records the abrupt expansion of deposition beyond the confines of the val- Fig. 11 Palaeogeographic and isopach map of Dunvegan allomember I. 49Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... leys (cf. MARTINSEN et al., 1999). The prevalence of lacustrine facies indicates that accommodation was gen- erated faster than it could be filled, resulting in standing bodies of water (cf. HAMPSON et al., 1999). Alterna- tively, it may abruptly overlie a sharp, micro-erosional surface above a well-developed palaeosol that defines an interfluve surface (McCARTHY et al., 1999). It is significant that bioturbated lacustrine deposits are consistently observed for a few metres, immediately above sequence-bounding interfluve palaeosols. Fur- ther up section, bioturbation disappears until the next interfluve surface is crossed. This observation suggests that interfluves were initially flooded by brackish-water lagoons, with an attendant burrowing fauna. This inter- pretation is consistent with the observation (McCAR- THY & PLINT, 1998, in review) of barite crystals in mudstones immediately above the top H and top G interfluve surfaces. This mineral is known to precipitate in modern hydromorphic soils associated with saline groundwater (STOOPS & ZAVALETA, 1978) The lacustrine-dominated, high-accommodation sys- tems tract is the principal component of the nonmarine portion of sequences. It is here emphasized that the maximum flooding interval which is stratigraphically equivalent to the time of maximum marine transgres- sion, consistently lies within, and not at the top of the lacustrine-dominated systems tract. This observation suggests that, for a significant interval of time after maximum transgression, accommodation was generated faster than it could be filled. 4.3.3. Palaeosol-dominated, low-accommodation systems tract This systems tract is dominated by blocky mudstones that represent moderately- to well-developed palaeo- sols (palaeosol types 4–5, fig. 8 of McCARTHY et al., 1999). Commonly, several well-developed palaeosol profiles are found in close vertical juxtaposition. These palaeosols are generally interstratified with crevasse- splay sandstones and thin units of lacustrine and wet- land facies. The base of this systems tract is defined where laminated lacustrine mudstones pass upward, typically over 1–2 metres, into a succession consisting of relatively well-developed complex palaeosols that are welded together (McCARTHY et al., 1999). The top of the palaeosol-dominated systems tract is defined by a Fig. 12 Palaeogeographic and isopach map of Dunvegan allomember H (from PLINT et al., 2001). 50 Geologia Croatica 56/1 subaerial erosion surface marking an interfluve (McCA- RTHY & PLINT, 1998; in review; McCARTHY et al., 1999; McCARTHY, 2002) that is expressed as a sharp, but commonly subtle stratigraphic discontinuity having a micro-erosional relief of centimeters to decimeters. This systems tract is interpreted to represent a flood- plain environment in which autogenic processes contin- ued to build and fill minor depositional topography, but where the overall rate of vertical aggradation gradually decreased to nil, permitting progressively more time for pedogenesis. The upper part of each sequence- bounding palaeosol succession represents several tens of thousands of years of sediment starvation and in-situ polygenetic pedogenesis (sensu MARRIOT & WRIGHT, 1993). These palaeosols represent a period of nil to negative accommodation during which pedog- enic overprinting is recorded in the upper part of the interfluve palaeosol complex (McCARTHY & PLINT, 1998; McCARTHY et al., 1999; Fig. 22). The signifi- cance and spatial variability of interfluve palaeosols is discussed in more detail below. 5. PALAEOVALLEY SYSTEMS Extensive valley systems exist at the tops of Dunvegan allomembers H, G, F and E (Fig. 23). Valley-filling deposits are well-exposed along the Peace River val- ley between the Alberta–British Columbia border and Erin Lodge, some 140 km downstream, and also on Fig. 13 Palaeogeographic and isopach map of Dunvegan allomember G (from PLINT et al., 2001). 51Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... the Beatton and Kiskatinaw rivers (Fig. 1; PLINT, 2002). Valley-filling sandstones up to 30 m thick are also present at the top of allomembers D, C and A. However, these younger sandbodies are typically < 1 km wide and are generally impossible to map in subsurface with the available well density: they will not be discussed further in this paper. Up to 4800 wells and 40 outcrops were used to map valley systems incised into the top surfaces of allomembers H to E. Individ- ual valleys can confidently be mapped for distances of 110 km in allomember H, 250 km in allomember G, 200 km in allomember F, and 320 km in allomember E. Valleys maintain a remarkably consistent depth over their entire length, with maximum depths ranging from 31–40 m, with an overall average of about 21 m. Valley width varies from a few hundred metres to a maximum of about 8 km, with most valleys being about 1–2 km wide. Cross-valley depth varies abruptly suggesting the development of a terraced morphology. Limited core in downdip valley reaches and abun- dant outcrop in up-valley reaches permits five main facies to be identified. Up-valley reaches are dominated by fine to medium grained, trough crossbedded sand- stone. Widely-distributed but uncommon tidal bundles defined by carbonaceous drapes, bundle sequences and sigmoidal cross-stratification are present up to 30 km landward of the marine transgressive limit, and indicate tidal backwater effects, probably under a microtidal regime. Rippled and planar laminated sandstone is an accessory facies. Locally-preserved lenses of rooted, Fig. 14 Palaeogeographic and isopach map of Dunvegan allomember F (from PLINT et al., 2001). 52 Geologia Croatica 56/1 blocky to laminated mudstones and coals are interpreted to represent remnants of floodplain deposits preserved within the valleys. Inclined heterolithic stratification on a dm- to m-scale is common at the top of valley-fills up to 110 km landward of the marine transgressive limit; this facies probably records a complex interac- tion between tidal processes and ?seasonal river dis- charge fluctuations. Down-valley reaches, typically within about 50 km of the lowstand shoreline, have a sandstone-dominated lower portion and a mudstone- rich upper portion in which a variety of thinly-bedded heterolithic strata are encountered. Heterolithic facies may be organized in sandier-upward successions that might represent upward-shallowing bay-head deltas, or muddier-upward successions that might represent tidal flats or muddy IHS filling a channel. Ubiquitous sub-mm scale mud drapes and bidirectional ripple cross lamination provide strong evidence for tidal processes (Fig. 19). Soft-sediment deformation and micro-faulting are abundant and might reflect dewatering, slumping, earthquake shock and dinoturbation (PLINT & WADS- WORTH, in review). Within about 20–40 km of the maximum regressive shoreline, valleys in each allomember become unrecog- nizable, with the exception of one major channel in each of allomembers E and F, which obviously continues as a distributary that feeds a major lowstand delta lobe. The most seaward part of the delta plain lacking valleys Fig. 15 Palaeogeographic and isopach map of Dunvegan allomember E. Note the appearance of a semi-circular depocentre in the NW cor- ner of the study area, interpreted to record the onset of renewed flexural subsidence (from PLINT et al., 2001). 53Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... is interpreted as the LST, deposited when sea level was slowly rising and rivers started to backfill their valleys. Valley filling took place primarily during the early TST. In downdip areas, the top of the valley fill is overlain, at a Glossifungites transgressive surface, by a few dm of intensely bioturbated muddy sandstone that consti- tutes a very thin marine TST. Further landward, a thin coal commonly separates the valley-fill from overlying marine deposits; the coal typically extends about 30 km seaward of the maximum marine transgressive limit and records the development of sediment-starved mires behind the transgressing shoreline (Fig. 19). Further landward, in areas not inundated by the sea, valley-fills are typically overlain by laminated lacustrine mud- stones, thin coals or pedogenic mudstones. 6. INTERFLUVE PALAEOSOLS: SPATIAL VARIABILITY AND PALAEOCLIMATIC SIGNIFICANCE The palaeosols that mantle interfluve surfaces are similar to modern Alfisols and each example records (i) aggradation on an alluvial/coastal plain; (ii) a sub- sequent static and/or degradational phase related to valley incision, non-deposition and soil thickening; and (iii) a final aggradational phase related to valley-fill- ing and renewed sedimentation across the coastal plain (McCARTHY & PLINT, in review; Fig. 22). Within this overall template, however, variations in palaeosol thickness, redoximorphic features, illuvial clay content, and geochemistry suggest developmental control by hydrological characteristics that were influenced both by the nature of the underlying alluvial sediments, and Fig. 16 Composite map showing the maximum progradational shorelines of Dunvegan allomembers D, C and A, and also isopachs for the entire D–A interval. Note the prominent depocentre along the western margin of the study area, and progressive rotation of isopach lines from N–S to NW–SE (from LUMSDON & PLINT, 2003). 54 Geologia Croatica 56/1 55Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... by distance from the dissected valley edge. Palaeosols that developed closer (few hundred m to few km) to valley margins are thicker, have greater quantities of illuvial clay, and display characteristics suggestive of well-drained conditions relative to those palaeosols that developed further (> 5–10 km) from valley margins. Variations in drainage and palaeotopography during landscape dissection resulted in different palaeosol development styles on interfluve surfaces that can be shown, on the basis of physical correlation, to have the same geomorphic age. Renewed coastal plain aggradation and syndeposi- tional pedogenesis is marked, at sites a few km from valleys, by an obvious increase in grain size, multiple organic horizons, sedimentary microlamination, and the presence of sphaerosiderite. These features suggest the resumption of periodic flooding of the coastal plain as fluvial channels once again migrated beyond the valley walls. The presence of barite coupled with elevated lev- els of Ca, Mg, Sr, and P suggest that interfluve palaeo- sols adjacent to valley margins were initially flooded with brackish water (McCARTHY & PLINT, 1998; DRIESE et al., 1992). In contrast, an interfluve palaeo- sol located 15 km from the nearest valley is directly capped by a coal, indicating water-table rise, increased accommodation rate, but negligible clastic supply (Fig. 22). The interfluve palaeosols contain abundant pedogenic clay minerals (interstratified illite–vermicu- lite) in the < 0.2 mm fraction that yield δ18O values of -12.9‰ to -11.6‰. These isotopic ratios are considered to record the composition of the Cretaceous rainwater at this palaeo-latitude of about 65°N, and suggest soil temperatures in the range 9–16°C (VITALI, et al., 2002). This temperature estimate is consistent with Fig. 17 Multi-storey valley-fill at Beatton River, B.C. Note how base of valley-fill cuts downward from right to left, eventually intersecting the top of a sheet-like delta-front sandstone. Lower storeys are crossbedded sandstone but the uppermost storey shows well-developed inclined heterolithic stratification, possibly deposited under the influence of seasonal discharge fluctuations. Maximum thickness of val- ley-fill = 23 m (from PLINT & WADSWORTH, in review). Fig. 18 Transverse section summarizing the main characteristics of valley-fills in an up-valley, fluvial-dominated reach (from PLINT & WADSWORTH, in review). 54 Geologia Croatica 56/1 55Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... estimates based on pedological characteristics (McCAR- THY & PLINT, 1999) and on palaeobotanical informa- tion (7–13°C; UPCHURCH & WOLFE, 1993). 7. RECOGNIZING ACCOMMODATION CHANGE Evidence of accommodation changes, both positive and negative, can be inferred from various sedimentological and stratigraphic features of the Dunvegan Fm. Seven main lines of evidence will be summarized: 1) Marine transgressive surfaces; 2) Isopach patterns; 3) Distri- bution and character of nonmarine facies; 4) Valley systems; 5) Interfluve palaeosols; 6) Sharp-based delta- front sandstones, and 7) Palaeodrainage patterns. 7.1. Regional marine transgressive surfaces Regionally-mappable marine transgressive surfaces are among the most recognizable features of the Dunvegan Fm., allowing the formation to be divided into allo- members (Fig. 23). Importantly, the well log response that corresponds to these transgressive surfaces can be traced inland for up to about 150 km across the coastal plain that lay landward of the transgressive limit of the marine shoreline. Tracing these non-marine (lacustrine) flooding surfaces allows the complete geometry of the allomember to be determined, including both the marine and non-marine portions. Transgressive surfaces provide clear evidence of an increase in the accom- modation : sedimentation ratio which led to shoreline transgression. Comparison of maps in Figs. 10–16 shows how the transgressive distance varied within and between each allomember. Figure 24 summarizes the ‘average’ distance that the shoreline moved landward and seaward for each allomember. 7.2. Regional isopach maps and accommodation changes Figures 10–14 illustrate the regional palaeogeography and geometry of the lower 5 Dunvegan allomembers, Fig. 19 Longitudinal section summarizing the main characteristics of valley-fills from the valley mouth to a point about 180 km up-valley (from PLINT & WADSWORTH, in review). Fig. 20 Example of lenticular fluvial sandstone (up to 8 m thick), interpreted to represent an anabranch of an anastomosed river, enclosed in grey pedo- genic and lacustrine mudsto- nes. Coldstream Creek, B.C. 56 Geologia Croatica 56/1 57Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... J to F. Allomembers J to F are markedly sigmoidal in profile, thinning updip onto the coastal plain, and also seaward through downlap onto the condensed section (FSU marker, Fig. 23). The maps for allomembers J to F very clearly illustrate the progressive basinward (i.e. southeastward) shift of the depocenter. Bathymetric deeps (i.e. water-filled space) remaining between delta lobes were filled by particularly thick intervals of the succeeding allomember. This pattern may reflect differ- ential compaction over older lobes, leading to diversion of rivers into the intervening, topographically low areas (cf. BLUM & PRICE, 1998). Isopach lines in Figs. 10– 14 are essentially parallel, and trend NE–SW, perpen- dicular to the orogen. The isopachs show no evidence of thickening towards the Foothills deformed belt. This critical observation shows that during deposition of allomembers J to F, this part of the basin experienced essentially no differential tilting toward the orogen. Figures 15 and 16 illustrate the palaeogeography and geometry of allomember E and allomembers D–A combined (because the bounding surfaces of allomem- bers D, C and B cannot be mapped across the entire study area). The isopach pattern for allomembers E to A shows a dramatic change relative to older allomembers. Allomember E shows a pronounced updip thickening towards the NW with the development of an isopach ‘moat’ that extends about 70 km from the present deformation front. All of this semi-circular depocentre is filled with nonmarine deposits which reach 80 m in thickness in outcrop in the Pine River – Coldstream Creek area (LUMSDON & PLINT, 2003 – in press; Fig. 2). In the SE however, isopach patterns for allom- ember E resemble those of older allomembers in pre- serving the broad NE–SW trend, with the thickest parts of allomember E filling the bathymetric deep fronting allomember F. The semi-circular sedimentary ‘moat’, first rec- ognized in the NW part of allomember E, expands in allomembers D–A to a radius of about 250 km, across which allomembers D–A thicken from about 30 to 100 m (Fig. 16). Note how isopach lines progressively rotate to parallel the orogen, reflecting the increasing importance of flexural subsidence towards the SW. The thickening of allomembers E to A in the NW corner of the study area is interpreted to record the onset of a new phase of flexural subsidence adjacent to the active margin of the basin (cf. ROGERS, 1998). The generation of new alluvial accommodation in the updip portion of the delta system had a pronounced affect on deposition in the marine realm downdip. Allomembers J–E are relatively thick and the delta-front areas are rel- atively sand-rich. In contrast, the deltas in allomembers D–A (located in the SE corner of the study area; Fig. 16), are relatively thin and sand-starved. This is inter- preted to reflect sediment partitioning between alluvial and marine depocentres, with the balance between the two progressively shifting in favour of the alluvial realm, as the orogen-proximal subsidence rate increased through allomembers E to A. 7.3. Distribution and character of nonmarine facies In each allomember, only the lower 2–4 parasequences, which comprise the transgressive and highstand sys- tems tracts, include both marine delta front / prodelta and correlative coastal plain deposits (Fig. 25). Youn- ger parasequences offlap seaward and consist of only marine delta front / prodelta facies (PLINT et al., 2001). The offlapping deltaic parasequences are interpreted to have been deposited during the falling stage systems tract (PLINT & NUMMEDAL, 2000; PLINT et al., 2001). The top of the HST became a non-depositional surface during the FSST. During this time, rivers began to incise valleys, starting at the seaward margin of the highstand systems tract (PLINT, 2002). Rivers incised progressively headward during the FSST, and also con- tinued to erode valleys across the upper surface of the FSST as sea-level continued to fall. Interfluves between valleys were starved of sediment for many tens of ky, during which time they developed thick and mature soil profiles (McCARTHY et al., 1999, McCARTHY & PLINT, in review). Fig. 21 Examples of stacked sandier-upward lake-fill successions typical of much of the Dunvegan coastal plain. Lynx Creek, Alberta; person for scale. 56 Geologia Croatica 56/1 57Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... The restriction of coastal plain deposits to the TST and HST, which have an aggradational to progradation- al stacking pattern, show that alluvial accommodation was only generated during the rising half of the relative sea-level cycle. The fact that the alluvial deposits in allomembers I–F form a gradually updip-thinning sheet suggests that accommodation was generated by eustatic sea-level rise rather than by flexural subsidence. The Fig. 22 Diagram summarizing the developmental history of interfluve palaeosols, and also showing how the history and character of the pal- aeosol varies with distance from the margins of palaeovalleys. The distance from the valley margin strongly influenced the drainage of the interfluve, and also controlled the rate of sediment flux once the valley had been filled (from McCARTHY & PLINT, in review). A B c incision ~ Local soil succession depends on location, topography, parent material, vegetation, etc. Moderately well-drained No truncation Poorly - drained No truncation . == =. c - .-. ..,: . .....; Transgressive - Highstand Poorly-drained coastal plain with frequent flooding. Sediment supply and accommodation rate are both high. Cumulic pedocomplexes develop across the coastal plain. Falling Stage - Lowstand Fluvial incision results in water-table drop near dissected valley edge. Soil profile thickening, clay illuviation and erosion occur close to the valley margin . A high water table persists towards the centre of the interfluve. --"<'""--r~- < 1 km --'''!''''-- few km -~.,,!~.--- :> 10 km ------i Thick cumulic cap Brackish influence cc co • Blocky structure Clay coatings Mottles 1) Root traces ~ Siderite concretions Sphaerosiderite Fe-oxide nodules Organic fragments Transgressive After valley fills, renewed flooding occurs on interfluves, and water table rises. Thick cumulic soils cap the truncated interfluve palaeosol near the valley margin whereas peat accumulates far from the source of clastic supply. A brackish influence is felt adjacent to valleys but groundwater remains fresh toward the centre of the interfluve. o Horizon composed of organic soil material A Mineral soil horizon mixed with humified organic material Bw Subsurface horizon with some colour or structure development 8 9 Gleyed subsurface horizon (contains reduced iron) Bt Subsurface accumulation of silicate clay C Mineral horizon relatively unaffected by pedogenic processes AC Transitional horizon with some A and C properties BC Transitional horizon with some Band C properties 58 Geologia Croatica 56/1 59Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... updip and westward thickening of alluvial deposits in allomembers E–A however, reveals an additional com- ponent of flexural subsidence along the basin margin. Thus, the overall geometry of allomembers (‘sigmoidal- prismatic’ in allomembers J–F, versus wedge-shaped in allomembers E–A) reveals a fundamental difference in the geometry of accommodation generation. For allomembers J–F, tectonic tilting was negligible and eustatic change appears to provide the simplest expla- nation for the parasequence stacking pattern and for the restriction of alluvial deposits to the rising half of the relative sea level cycle. In contrast, updip thickening of alluvial deposits in allomembers E–A records the onset of flexural subsidence which partially overprinted the effects of eustatic changes in the more distal (SE) part of the basin. Allomembers E to A contain semi-circular depo- centres in the NW portion of the study area (Figs. 15, 16). Alluvial rocks filling those depocentres are well- exposed in NE British Columbia. Non-channelised allu- vial facies represent lake, mouth-bar, levee and crevasse splay environments, and organic histosol (coal), poorly- drained, intermediate redox and better-drained palaeo- sols. Siderite, organic debris and dinosaur tracks are Fig. 23 Summary dip section (located in Fig. 1) drafted to scale, with the top of allomember A as a datum, showing the stratigraphic distribu- tion of valley-fills. Only valleys on allomembers H–E can be mapped with the available well density (from PLINT & WADSWORTH, in review). Fig. 24 Scale drawing summarizing the average shoreline displacement for each transgres- sive–regressive cycle of the Dunvegan Fm. Note the similarity in shoreline displacement for allomembers I–C. Only the transgressions at the tops of allomembers C and A show a transgressive distance that greatly exceeds the regressive distance, probably reflecting an accelerating subsidence rate and increasing accommodation:sedimentation ratio (see also Fig. 16). 58 G eologia C roatica 56/1 59 P lint: C lastic S edim ent P artitioning in a C retaceous D elta S ystem , W estern C anada... ubiquitous elem ents of these facies. C hannelized facies are dom inated by non-m igrating (probably anastom o- sed) sandstone channel-fills (LU M SD O N & PLIN T, in press). The non channelized facies form tw o associations w hich are interpreted on the basis of intrinsic facies characteristics, to represent different accom m odation rates. A ssociation A is dom inated by crevasse splays, Fig. 25 D ip section, draw n to scale and datum ed on the dow nlap surface, show ing (below ) the organization of parasequences w ithin allom - em bers H –D . The upper diagram is a chronostratigraphic representation of the low er cross-section, based on the assum ption that each parasequence represents an equal interval of tim e. The chart em phasizes the depositional hiatus on the upper coastal plain that devel- ops during the falling stage and low stand system s tracts (from P LIN T et al., 2001). o Section flattened on regional downlop surface ~=---~6 4 :3 ~ ~ 10 9 m ,,~., "U~·"U ; " Coaslal plain 2:(0 Della - front sandstone _ - _ - Prodeila mudstone ~ Sequence - bounding paleosol Eroded sediments ISTRATIGRAPHIC CROSS-SECtK5N] E - --+ ~ g o d' ill I 60 Geologia Croatica 56/1 61Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... levees, intermediate redox and better-drained palaeo- sols. Association B is dominated by lakes, mouth-bars, organic and poorly-drained palaeosols. Association A, indicating relatively well-drained conditions, dominates allomembers G to the lower part of E. From near the middle to near the top, allomember E is progressively dominated by association B, but the uppermost 5 m shows an abrupt return to better-drained palaeosols (Fig. 26). Allomembers D to A show alternations between associations A and B, although overall, association B predominates upward. The basal units of the overlying Kaskapau Fm. comprise a 40–50 m thick transitional succession dominated by facies of association B, but also including brackish-water lagoonal deposits, over- lain by marine deposits (Fig. 26). When the vertical changes in alluvial facies associa- tions are considered in the context of the geometry of their host allomembers, the relationship between facies and accommodation rate becomes clear. Allomembers H to F thin towards the NW suggesting minimal accom- modation in that direction, and this is reflected in their constituent alluvial facies (association A), which are dominated by relatively well-drained environments. The upward change within allomember E, from association A to relatively poorly-drained association B is interpret- ed to record an upward increase in accommodation rate. This change is independently suggested by the isopach ‘moat’ in allomember E, attributed to renewed flexural subsidence (Figs. 15, 26). A 5 m thick pedocomplex at the top of allomember E suggests a final phase of very low accommodation rate. Allomembers D–A are again dominated by poorly-drained freshwater environments, which pass upward into rocks of the basal Kaskapau Fm. which show progressively increasing marine influence. This facies succession reflects an increasing accommodation rate, which is also suggested by region- al isopach patterns (Fig. 16). 7.4. Valley systems Valley systems can be mapped for up to 320 km, and certainly extend to the NW beyond the study area (PLINT, 2002; PLINT & WADSWORTH, in review). Representative maps of valley systems on allomembers F and E are shown in Figs. 27 & 28. A crucial observa- tion is that, at least over the study area, valleys show no systematic change in depth with distance upvalley. The considerable length, modest depth (average 21 m) and lack of longitudinal shallowing suggests that the valleys might reflect both eustatic and discharge-related forcing mechanisms. The incision of valleys into offlapping deltaic deposits suggests that there was some relation- Fig. 26 Summary stratigraphic section oriented E–W across the northern part of the study area (Fig. 1) showing, to scale, the thickness vari- ations of the Dunvegan allomembers. Flexural subsidence was negligible in allomembers J–F, but became increasingly rapid in allomem- bers E–A as shown by the great expansion of the section towards the west. The increased accommodation rate is reflected in the vertical changes in alluvial facies in outcrop, with an overall upward increase in the proportion of waterlogged and subaqueous facies (associa- tion B) relative to better-drained facies (association A; from LUMSDON & PLINT, 2003). 60 Geologia Croatica 56/1 61Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... ship to relative sea-level fall which triggered erosion, starting at the offlap break fronting the HST. However, valleys seem ‘too long’ for the depth of incision, and the lack of headward shallowing suggests that an additional mechanism operated from upstream to effect incision. It is possible that cyclical changes in the ratio of sediment load to discharge might have been responsible for river incision far upstream (PLINT, 2002). It is not unlikely that changes in river discharge and small eustatic changes were both linked to global climatic oscillations in the Milankovitch band, and in combination, were responsible for most of the changes in accommodation pattern. Intermittently-active flexural subsidence was the third recognizable influence on the system. 7.5. Interfluve palaeosols Interfluve palaeosols are recognized on the basis of abundant clay coatings, Fe oxide nodules and mot- tles which indicate the presence of Bt horizons. The prolonged stability of these surfaces is also indicated by the fragmentation, ageing and assimilation of clay coatings, the intense biological reworking of the matrix, concentration of residual Zr and TiO2, and increased kaolinite content. Minor erosion of the interfluve sur- face is suggested by the presence in the palaeosols of papules, pedorelicts and embedded grain argillans, all indicative of minor remobilization of the palaeosol. The sharp irregular upper boundary of the palaeosol Fig. 27 Palaeogeographic map of allomember F showing the maximum progradational limit of the marine shoreline, the boundary between aggradational (HST) and offlapping (FSST) parasequences, the distribution of valleys, and the maximum transgressive limit of the suc- ceeding allomember E shoreline (from PLINT & WADSWORTH, in review). 62 Geologia Croatica 56/1 63Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... (i.e. the sequence boundary) also indicates subtle ero- sion (McCARTHY et al., 1999). Immediately above the sequence boundary, the sediment tends to be well- laminated and grain size increases, containing abundant silt and very fine sand. This indicates renewed supply of sediment to the surface and low levels of physical reworking. The presence of abundant, partially-oxidized siderite and multiple surface organic layers indicates increased water-logging and flooding in response to a rising, but fluctuating water table. The stratigraphic position and vertical spacing of these very distinctive palaeosols is predictable to within 1–2 m on the basis of the projection of sequence bound- aries from nearby well logs. This gives us confidence in our interpretation of sequence-bounding unconformities on the basis of well log correlations. The characteristics of the interfluve palaeosols indicate very significant depositional hiati, of the order of tens of ky, which cor- roborates the evidence provided by valleys for lengthy periods of negative accommodation across the coastal plain. 7.6. Sharp-based delta-front sandstones Progradation of a delta typically produces a succession in which prodelta mudstone grades up into interbed- ded sandstone and mudstone and finally clean delta front sandstone. This Walthurian succession of facies indicates a gradual change in environmental conditions from deeper to shallower. In contrast to this ‘norm’, ero- sive-based delta front sandstones are widespread in the Dunvegan Fm. Swaley cross-stratified or cross-bedded Fig. 28 Palaeogeographic map of allomember E showing the maximum progradational limit of the marine shoreline, the boundary between aggradational (HST) and offlapping (FSST) parasequences, the distribution of valleys, and the maximum transgressive limit of the suc- ceeding allomember D shoreline (from PLINT & WADSWORTH, in review). 62 Geologia Croatica 56/1 63Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... sandstone of the delta-front rests sharply on laminated mudstone or siltstone of the prodelta (Fig. 8). The contact is generally ornamented with large gutter-casts oriented shore-normal. This succession implies removal of accommodation causing shallow water processes to impinge directly on a muddy, low-energy prodelta sea floor. Eustatic fall or tectonic uplift seem to provide the only reasonable explanation for this facies succession (PLINT, 1988, 1996; PLINT & NUMMEDAL, 2000). It cannot be satisfactorily explained simply by changes in sedimentation rate. Isopach maps provide no evidence for regional tilt- ing, particularly in allomembers J–F, and therefore it is difficult to attribute accommodation loss and relative sea level fall to flexural uplift. In view of this geometric constraint, the pronounced offlap of deltaic parase- quences has been attributed to eustatic fall (PLINT et al., 2001). One consequence of eustatic fall would be accommodation loss at the shoreline, resulting in the development of sharp-based delta-front sandstones. 7.7. Palaeodrainage patterns Observations and computer simulations (GARCIA- CASTELLANOS, 2002) reveals two consistent and characteristic features of the drainage system in fore- land basins. Smaller tributary streams drain perpen- dicular to the orogen but merge into a large river that flows parallel to the orogen. Secondly, the large axial river lies on the orogen-distal side of the basin, close to the forebulge. This pattern is exemplified by the New Guinea foreland basin (Fig. 29) in which the tributaries of the two main rivers, the Fly and Digul flow perpen- dicular to the orogen before turning abruptly to east and west, draining into the Gulf of Papua, and the Arafura Sea respectively. In the south a low but distinct drain- age divide is revealed by the small rivers that traverse the southern margin of the island (Fig. 29) and it seems likely that this divide marks the crest of the forebulge. The interpreted forebulge lies 250–300 km from the deformation front in the New Guinea Highlands; this distance is consistent with the flexural wavelength of the lithosphere, and with observations in other foreland basins. Drilling data are sufficiently dense that the course of the Dunvegan palaeovalleys can be reconstructed with some confidence. Valleys on allomembers H, G and F have a broadly rectilinear pattern that is interpreted to reflect control of the precursor rivers by faults in the Palaeozoic basement. A large SE-prograding lowstand delta is evident in allomember F (Fig. 27). Overall, valleys are oriented perpendicular to the coast. The val- leys on allomember E show a different pattern (Fig. 28) in which all the tributary valleys flowing from the W and SW merge eastward into a single trunk valley that flows SE, parallel to the coast for over 150 km before terminating in a large lowstand delta lobe. This pattern bears a remarkable similarity to that of the Digul and Fly rivers in New Guinea. It is postulated that a subtle forebulge deflected the main valley on allomember E from a shore-perpendicular to a shore-parallel trend. The drainage pattern and main tectonic elements of New Guinea and of allomember E are compared, at the same scale, in Fig. 30. Valleys on allomembers H to F are not deflected shore-parallel. This might be evidence that a forebulge was not a significant topographic ele- ment on the coastal plain. This in turn might reflect a phase of negligible flexural deformation, which is also indicated by the isopach maps. In contrast, the dramatic shore-parallel re-orientation of valleys in allomember Fig. 29 Map of rivers in the central por- tion of the New Guinea foreland basin. The distinct drainage divide close to the southern coast is inter- preted to reflect the position of the forebulge. Note how the Fly and Digul river systems flow initially perpendicular to the fold and thrust belt, but distally, are diverted by and flow parallel to the forebulge. 64 Geologia Croatica 56/1 65Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... E suggests that rivers were at this time influenced by a forebulge. It may be no coincidence that the apparent uplift of a forebulge is linked to the onset of renewed flexural subsidence, as revealed by the isopach maps of allomembers E and A–D. 7.8. Quantification of Eustatic Change Isopach maps (Figs. 10–14) show that in allomembers J–F, flexural subsidence was unimportant. Neverthe- less, transgressions and regressions, and the cutting and filling of valleys continued to occur. Because the geographic limits of transgressive and regressive shorelines, and the seaward limits of valleys can be mapped with some accuracy, it is possible to determine the horizontal extent of transgressive and regressive events in each allomember. By applying geologically realistic gradients to the surfaces over which the shore- line migrated, and including the thickness of vertically- accreted coastal plain deposits, it is possible to make some first-order estimation of the vertical change of sea-level. This approach is summarized in Fig. 31. The maximum transgressive limit of marine shore- lines commonly lies within 10–20 km of the previous HST/FSST offlap break, and rarely inundates the HST coastal plain to landward. This implies a consistent balance between accommodation and sediment supply for each T/R cycle. The width and inferred slope of the FSST (60 km x 1:2:500) implies a slope height of about 24 m (Fig. 31). An additional average of 11 m of coastal plain sediment accumulated during the TST and HST, suggesting a minimum vertical accommodation of about 35 m. Calculation of water and sediment loads on the basis of simple Airy isostasy suggests that a eustatic rise of about 24 m would result, after sedimentation to sea level, in about 35 m of vertical accommodation (PLINT & WADSWORTH, in review). This estimate of eustatic change does not take account of the elastic strength of the lithosphere which would tend to decrease the amount of subsidence for any given sediment and water load. It is therefore probable that eustatic excur- sions were somewhat less than 24 m. It is perhaps no coincidence that the average depth of valleys (21 m) is close to the eustatic change inferred on geometric cri- teria. Modest eustatic fall at the coastline drove FSST delta progradation and also probably effected river incision across the HST and older FSST surfaces. How- ever, the small (<24 m ) and slow (?50–100 ky) eustatic change that is indicated by this analysis seems unlikely to have been capable of driving valley incision for >300 km upstream, as is observed (PLINT & WADSWORTH, in review). We are led to infer an additional, upstream forcing mechanism, perhaps linked to changes in dis- charge:sediment load ratio, controlled by climatic cycles in the Milankovitch-band. 8. CONCLUSIONS A regional stratigraphy based on marine transgressive surfaces and approximately equivalent subaerial inter- fluve surfaces allows the deltaic Dunvegan Formation to be subdivided into ten transgressive–regressive allom- embers with an average duration of <200 ky. Allom- embers prove to be practical and robust stratigraphic units, well-suited to a data-set that consists mostly of well logs. More sophisticated interpretation of chang- ing accommodation patterns is based on analysis of the stacking pattern of component parasequences within allomembers. This approach allows transgressive and highstand systems tracts to be recognized. These contain Fig. 30 Maps comparing the pattern, scale and palaeogeographic relationships of the modern Fly River (flipped and rotated) so as to mimic the orientation of the valley system in Dunvegan allomember E. 64 Geologia Croatica 56/1 65Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... both marine and aggradational coastal plain deposits. Offlapping deltaic parasequences that lack an updip alluvial component are assigned to the falling stage sys- tems tract. Aggradational to onlapping parasequences at the seaward limit of the allomember are assigned to the lowstand systems tract. The seaward termini of val- leys are considered to mark the geographic boundary between FSST and LST. Valley systems that are typical- ly 1–2 km wide, average 21 m deep, and > 320 km long are incised into the top surfaces of most allomembers; valleys have been mapped in detail for allomembers H to E. Valleys are back-filled with muddy heterolithic tidal facies within a few tens of km of the valley mouth. Further updip, fluvial processes dominated, produc- ing multi-storey sandstone-dominated channel-fills. A weak tidal influence is inferred locally, and the effects of semi-diurnal tides extended about 30 km inland from the marine shoreline. Major palaeosols mantle interfluves. A wide range of pedological features in the interfluve palaeosols indicate a protracted hiatus with extensive physical, chemical and biological modifica- tion of the parent material. The drainage of interfluve palaeosols progressively deteriorated with increasing distance from valley margins. Allomembers can be grouped on the basis of isopach patterns. Allomembers J–F have a sigmoidal prismatic geometry, successively offlapping to the SE. There is no evidence of thickening toward the orogen. In con- trast, allomembers E–A show progressive development of a depocentre along the western margin of the basin, presumably in response to renewed flexural subsidence. The increasing accommodation rate on the coastal plain had a profound effect on marine deltas further downdip, which were starved of sediment, leading to progressive backstep of shorelines. Simultaneously, alluvial depos- its within the depocentre show an upward increase in the proportion of subaqueous to subaerial facies, cul- minating in the incursion of brackish and finally marine waters. Thus tectonic subsidence rate had a first-order affect on both the volume of sediment available to build marine deltas and also on the local character of facies that accumulated on the coastal plain. Renewed flexural subsidence appears to have resulted in subtle uplift of a forebulge, resulting in dramatic deflection of river sys- tems in allomember E, and also probably in allomember D. Despite the clear tectonic signature, transgressions and regressions of the shoreline took place over similar Fig. 31 Block diagram summarizing the stratigraphic, geographic and topographic relationships between valleys and various systems tracts of a delta complex. Strata assigned to the TST and HST include aggradational coastal plain deposits. Offlapping deltaic strata of the FSST lack coeval coastal plain deposits and prograde an average of 60 km during slow relative sea level fall. Fluvial incision is inferred to have occurred during this time, and was accompanied by development of very mature interfluve palaeosols. Valley systems gener- ally terminate about 35 km from the maximum progradtional limit of the delta front, and this limit is interpreted to indicate the FSST/LST boundary. Aggradational and onlapping LST strata average 35 km wide, but include major lowstand deltas constructed at the termini of valley systems. On the basis of the assumed slopes and measured geographic dimensions, it is possible to infer cyclic eustatic changes of about 24 m (see text for discussion) (from PLINT & WADSWORTH, in review). 66 Geologia Croatica 56/1 67Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... distances, regardless of whether subsidence was negli- gible (allomembers J–F) or accelerating (allomembers E–A). This suggests that modest eustatic changes also influenced the accommodation rate. On the basis of the measured horizontal excursions of the shoreline, the vertical thickness of alluvial strata, and realistic alluvial gradients, an average eustatic excursion of about 24 m is calculated. The incision of valley systems is attrib- uted in part to periods of eustatic fall. However, valleys seem too long to be explained by eustasy alone, and hence secular changes in river discharge:sediment load ratio are postulated as an additional forcing factor. Cli- matic cycles in the Milankovitch band may have been responsible for both eustatic and discharge variations. Acknowledgements The results presented here build on many years of field and subsurface work by the author, during which the regional stratigraphy of the Dunvegan was established. He is grateful to Joyia CHAKUNGAL, Steve DON- ALDSON, Patrick ELLIOTT, Bira FACCINI, Yuanx- ian HU, Sid LEGGETT, Matthew LUMSDON, Dany MARTINIONI, Paul McCARTHY, Chantale McIN- TOSH, Jennifer McKAY, Joanna MOORE, Annemarie PLINT, Jessica RYLAARSDAM, David ULIČNY and Jennifer WADSWORTH for cheerful and energetic assistance in the field over 18 seasons. I thank the Natu- ral Sciences and Engineering Research Council of Can- ada for long-term support of our research through Oper- ating Grants, and through the LITHOPROBE Project; and also Canadian Hunter Exploration, Home Oil, Pan Canadian Petroleum, Union Pacific Resources, Wasca- na Energy and Petro–Canada (all but the last of which have been consumed by corporate take-overs!), for their generous support of our research on the Dunvegan Formation. I am also grateful to Margot McMECHAN for her help with palinspastic restoration of Foothills sections, Don STOTT for guidance early in this study, and Glen CALDWELL for insight into biostratigraphic problems. Finally, I thank the organizing committee of the IAS Meeting in Opatija for inviting me to contribute this review paper. 9. REFERENCES ALBERTA SOCIETY OF PETROLEUM GEOLOGISTS (1960): Lexicon of geologic names in the Western Canada Sedimentary Basin and Arctic Archipelago.– Calgary, Alberta, 380 p. ASLAN, A. & BLUM, M.D. (1999): Contrasting styles of Holocene avulsion, Texas Gulf Coastal Plain, USA.– In: SMITH, N.D. & ROGERS, J. (eds.): Fluvial Sedimento- logy VI. Int. Assoc. of Sedimentol., Spec. Publ., 28, 193–209. BHATTACHARYA, J.P. (1989): Allostratigraphy and river- and wave-dominated deltaic sediments of the Upper Cretaceous (Cenomanian) Dunvegan Formation, Alber- ta.– Unpubl. PhD Thesis, McMaster University, Hamilton, Ontario, 588 p. BHATTACHARYA, J.P. & WALKER, R.G. (1991a): Allostra- tigraphic subdivision of the Upper Cretaceous Dunvegan, Shaftesbury and Kaskapau formations in the northwestern Alberta subsurface.– Bull. of Canad. Petrol. Geol., 39, 145–164. BHATTACHARYA, J.P. & WALKER, R.G. (1991b): River- and wave-dominated depositional systems of the Upper Cretaceous Dunvegan Formation, northwestern Alberta.– Bull. of Canad. Petrol. Geol., 39, 165–191. BLUM, M.D. & PRICE, D.M. (1998): Quaternary alluvial plain construction in response to glacio-eustatic and cli- matic controls, Texas Gulf Coastal Plain.– In: SHANLEY, K.W. & McCABE, P.J. (eds.): Relative Role of Eus- tasy, Climate and Tectonism in Continental Rocks. Soc. of Econ. Paleont. and Mineral., Spec. Publ., 59, 31–48. CATUNEANU, O., BEAUMONT, C. & WASCHBUSCH, P. (1997): Interplay of static loads and subduction dynamics in foreland basins: reciprocal stratigraphies and the “miss- ing” peripheral bulge.– Geology, 25, 1087–1090. DRIESE, S.G., MORA, C.I., COTTER, E. & FOREMAN, L.J. (1992): Paleopedology and stable isotope chemistry of Late Silurian vertic paleosols, Bloomsburg Formation, central Pennsylvania.– Jour. of Sed. Petrol., 62, 825–841. EISBACHER, G.H. (1981): Late Mesozoic–Paleogene Bows- er Basin molasse and Cordilleran tectonics.– In: MIALL, A.D. (ed.): Sedimentation and Tectonics in Alluvial Basins. Geol. Assoc. of Canada, Spec. Pap., 23, 125–151. GARCIA-CASTELLANOS, D. (2002): Interplay between lithospheric flexure and river transport in foreland basins.– Basin Research, 14, 89–104. HAMPSON, STOLLHOFEN, G.H. & FLINT, S. (1999): A sequence stratigraphic model for the Lower Coal Meas- ures (Upper Carboniferous) of the Rhur district, north- west Germany.– Sedimentology, 46, 1199–1231. HAY, W.W., EICHER, D.L. & DINER, R. (1993): Physical oceanography and water masses in the Cretaceous West- ern Interior Seaway.– In: KAUFFMAN, E.G. & CALD- WELL, W.E.G. (eds.): Evolution of the Western Interior Basin. Geol. Assoc. of Canada, Spec. Pap., 39, 297–318. KAUFFMAN, E.G. & CALDWELL, W.E.G. (1993): The Western Interior Basin in space and time.– In: KAUFF- MAN, E.G. & CALDWELL, W.E.G. (eds.): Evolution of the Western Interior Basin. Geol. Assoc. of Canada, Spec. Pap., 39, 1–30. LUMSDON, M.P. & PLINT, A.G. (2003): Changing alluvial style in response to changing accommodation rate in a proximal foreland basin setting: Upper Cretaceous Dun- vegan Formation, NE British Columbia, Canada.– In: BLUM. M.D. & MARRIOT, S.B. (eds.): Fluvial Sedi- mentology VII. Int. Assoc. of Sedimentol., Spec. Publ., 35, in press. MARRIOTT, S.B. & WRIGHT, V.P. (1993): Paleosols as indicators of geomorphic stability in two Old Red Sand- stone alluvial suites, South Wales.– Jour. of the Geol. Soc. London, 150, 1109–1120. MARTINSEN, O.J., RYSETH, A., HELLAND-HANSEN, W., FLESCHE, H., TORKILDSEN, G. & IDIL, S. (1999): Stratigraphic base level and fluvial architecture: Ericson 66 Geologia Croatica 56/1 67Plint: Clastic Sediment Partitioning in a Cretaceous Delta System, Western Canada... Sandstone (Campanian), Rock Springs Uplift, SW Wyo- ming, USA.– Sedimentology, 47, 235–259. McCARTHY, P.J. (2002): Micromorphology and develop- ment of interfluve palaeosols: a case study from the Cenomanian Dunvegan Formation, NE British Columbia, Canada.– Bull. of Canad. Petrol. Geol., 50, 158–177. McCARTHY, P.J. & PLINT, A.G. (1998): Recognition of interfluve sequence boundaries: Integrating paleopedol- ogy and sequence stratigraphy.– Geology, 26, 387–390. McCARTHY, P.J. & PLINT, A.G. (1999): Floodplain palae- osols of the Cenomanian Dunvegan Formation, Alberta and British Columbia, Canada: Micromorphology, pedog- enic processes and palaeoenvironmental implications.– In: ALEXANDER, J. & MARRIOT, S.B. (eds.): Floodplains: Interdisciplinary Approaches. Geol. Soc. of London, Spec. Publ., 163, 289–310. McCARTHY, P.J. & PLINT, A.G. (in review): Spatial vari- ability of palaeosols across Cretaceous interfluves in the Dunvegan Formation, NE British Columbia, Canada: pal- aeohydrologic, palaeogeomorphologic and stratigraphic implications.– Sedimentology. McCARTHY, P.J., FACCINI, U.F. & PLINT, A.G. (1999): Evolution of an ancient coastal plain: palaeosols, inter- fluves and alluvial architecture in a sequence stratigraphic framework, Cenomanian Dunvegan Formation, NE Brit- ish Columbia, Canada.– Sedimentology, 46, 861–891. McDONOUGH, K.J. & CROSS, T. (1991): Late Cretaceous sea level from a paleoshoreline.– Jour. of Geophys. Res., 96, 6591–6607. McLAURIN, B.T. & STEEL, R.J. (2000): Fourth-order non- marine to marine sequences, middle Castlegate Forma- tion, Book Cliffs, Utah.– Geology, 28, 359–362. MITROVICA, J.X., BEAUMONT, C. & JARVIS, G.T. (1989): Tilting of continental interiors by the dynamical effects of subduction.– Tectonics, 8, 1079–1094. NORTH AMERICAN COMMISSION ON STRATIGRA- PHIC NOMENCLATURE (1983): North American Stra- tigraphic Code.– Am. Assoc. of Petrol. Geol. Bull., 67, 841–875. PANG, M. & NUMMEDAL, D. (1995): Flexural subsidence and basement tectonics of the Cretaceous Western Interior basin, United States.– Geology, 23, 173–176. PLINT, A.G. (1988): Sharp-based shoreface sequences and “offshore bars” in the Cardium Formation: Their relation- ship to relative changes in sea level.– In: WILGUS, C.K., HASTINGS, B.S. KENDALL, C.G.St.C; POSAMENTI- ER, H.W., ROSS, C.A. & VAN WAGONER, J.C. (eds.): Sea Level Changes – An Integrated Approach. Soc. of Econ. Paleont. and Mineral., Spec. Publ., 42, 357–370. PLINT, A.G. (1996): Marine and non-marine systems tracts and fourth-order sequences in the Early–Middle Ceno- manian Dunvegan Alloformation, northeastern British Columbia, Canada.– In: HOWELL, J.A. & AITKEN, J.F. (eds.): High Resolution Sequence Stratigraphy: Innova- tions and Applications. Geol. Soc. Spec. Publ., 104, 159–191. PLINT, A.G. (2000): Sequence stratigraphy and paleogeogra- phy of a Cenomanian Deltaic Complex: the Dunvegan and lower Kaskapau formations in subsurface and outcrop, Alberta and British Columbia, Canada.– Bull. of Canad. Petrol. Geol., 48, 43–79. PLINT, A.G. (2002): Paleo-valley systems in the Upper Cre- taceous Dunvegan Formation, Alberta and British Colum- bia.– Bull. of Canad. Petrol. Geol., 50, 277–296. PLINT, A.G. & NUMMEDAL, D. (2000): The falling stage systems tract: recognition and importance in sequence stratigraphic analysis.– In: HUNT, D. & GAWTHORPE, R.L. (eds.): Sedimentary Responses to Forced Regression. Geol. Soc. Spec. Publ., 172, 1–17. PLINT, A.G. & WADSWORTH, J.A. (in review): Sedimen- tology and palaeogeography of four large valley systems incising delta plains, Western Canada foreland basin: implications for mid-Cretaceous sea-level changes.– Sedi- mentology. PLINT, A.G., McCARTHY, P.J. & FACCINI, U.F. (2001): Nonmarine sequence stratigraphy: Updip expression of sequence boundaries and systems tracts in a high-resolu- tion framework, Cenomanian Dunvegan Formaton, Alber- ta Foreland Basin, Canada.– Am. Assoc. of Petrol. Geol. Bull., 85, 1967–2001. POULTON, T.P. (1994): Chapter 18: Jurassic and lowermost Cretaceous strata of the Western Canada Sedimentary Basin.– In: MOSSOP, G. & SHETSEN, I. (eds.): Geo- logical Atlas of the Western Canada Sedimentary Basin. Canad. Soc. of Petrol. Geol. and Alberta Res. Council, 217–316. ROGERS, R.R. (1998): Sequence analysis of the Upper Cre- taceous Two Medicine and Judith River formations, Mon- tana: Nonmarine response to the Claggett and Bearpaw marine cycles.– Jour. of Sed. Res., 68, 615–631. SAHAGIAN, D. & JONES, M. (1993): Quantified Middle Jurassic to Paleocene eustatic variations based on Russian Platform stratigraphy: Stage level resolution.– Geol. Soc. of Am. Bull., 105, 1109–1118. SCOTT, J.J. (2000): Fossil vertebrate footprints in the Dun- vegan Formation (Upper Cretaceous: Cenomanian) of British Columbia and Alberta.– Unpubl. BSc Thesis, Uni- versity of Saskatchewan, Saskatoon, 115 p. SHANLEY, K.W., McCABE, P.J. & HETTINGER, R.D. (1992): Tidal influence in Cretaceous fluvial strata from Utah, U.S.A.: A key to sequence stratigraphic interpreta- tion.– Sedimentology, 39, 905–930. SINGH, C. (1983): Cenomanian microfloras of the Peace River area, northwestern Alberta.– Alberta Res. Council Bull., 44, 321 p. SMITH, D.G. (1987): Meandering river point bar lithofacies models: modern and ancient examples compared.– In: ETHRIDGE, F.G., FLORES, R.M. & HARVEY, M.D. (eds.): Recent Developments in Fluvial Sedimentol- ogy. Soc. of Econ. Paleont. and Mineral., Spec. Publ., 39, 83–91. STELCK, C.R. & WALL, J.H. (1955): Foraminifera of the Cenomanian Dunveganoceras zone from the Peace River area of Western Canada.– Res. Council of Alberta Report, 70, 81 p. STELCK, C.R., WALL, J.H. & WETTER, R.E. (1958): Lower Cenomanian Foraminifera from the Peace River area, western Canada.– Alberta Res. Council, Geol. Div., Bul- letin, 2/1, 1–35. 68 Geologia Croatica 56/1 STOOPS, G.J. & ZAVALETA, A. (1978): Micromorphologi- cal evidence of barite neoformation in soils.– Geoderma, 20, 63–70. STOTT, D.F. (1982): Lower Cretaceous Fort St. John Group and Upper Cretaceous Dunvegan Formation of the Foot- hills and Plains of Alberta, British Columbia, District of Mackenzie and Yukon Territory.– Geol. Surv. of Canada Bull., 328, 124 p. UPCHURCH, G.R. & WOLFE, J.A. (1993): Cretaceous vegetation of the Western Interior and adjacent regions of North America.– In: KAUFFMAN, E.G. & CALDWELL, W.E.G. (eds.): Evolution of the Western Interior Basin. Geol. Assoc. of Canada, Spec. Pap., 39, 243–282. VAN WAGONER, J.C., POSAMENTIER, H.W., MITCHUM, R.M., VAIL, P.R., SARG, J.F., LOUTIT, T.S. & HARD- ENBOL, J. (1988): An overview of the fundamentals of sequence stratigraphy and key definitions.– In: WILGUS, C.K., HASTINGS, B.S., KENDALL, C.G.St.C; POSA- MENTIER, H.W., ROSS, C.A. & VAN WAGONER, J.C. (eds.): Sea Level Changes – An Integrated Approach. Soc. of Econ. Paleontol. and Mineral., Spec. Publ., 42, 39–45. VITALI, F., LONGSTAFFE, F.J., McCARTHY, P.J., PLINT, A.G. & CALDWELL, W.E.G. (2002): Stable isotopic investigation of clay minerals and pedogenesis in an inter- fluve paleosol from the Cenomanian Dunvegan Formation, N.E. British Columbia, Canada.– Chemical Geology, 192, 269–287. WILLIAMS, G.D. & STELCK, C.R. (1975): Speculations on the Cretaceous paleogeography of North America.– In: CALDWELL, W.E.G. (ed.): The Cretaceous System in the Western Interior of North America.– Geol. Assoc. of Canada, Spec. Pap., 13, 1–20. WILLIS, B.J. (1997): Architecture of fluvial-dominated val- ley-fill deposits in the Cretaceous Fall River Formation.– Sedimentology, 44, 735–757. Manuscript received February 14, 2003. Revised manuscript accepted June 20, 2003.