A facies model for the deposits of tidal bores (upstream-propagating hydraulic jumps associated with the flood tide in estuarine rivers) has not yet been developed, despite the publication of data from laboratory experiments and some modern estuaries. Moreover, only one example of tidal-bore deposits has hitherto been identified in the rock record. Herein, we document evidence for tidal bores in an Upper Pennsylvanian (Virgilian, Gzhelian) incised-valley fill in northeastern Kansas, USA. Across the flanks and crest of a bank-attached bar remnant in this fill are thin (< 10 cm) lenses of muddy, massive sandstone and massive, plant-debris-rich sandstone bounded by erosion surfaces exhibiting symmetrical wave ripples and irregular scouring. This interval is enclosed by sandstones with bidirectional cross-stratification, rhythmites, mud drapes, flaser bedding, and related sedimentary structures that together record appreciable tidal influence. The erosion surfaces bounding the massive sandstone lenses, and the massive sandstone lenses themselves, are evidence for scour and sediment suspension by upstream-propagating bores followed by deposition from suspension after their passage. Not only are these deposits the first sedimentary record of tidal bores to be documented in the Pennsylvanian of the USA, they are only the second example of such deposits to be documented in the global rock record. On the basis of our observations, we propose a provisional facies model for tidal-bore deposits that can be refined by future work.
Tidal bores are upstream-propagating hydraulic jumps that episodically form the leading edge of flood tides in upstream-narrowing, gently sloping, coastal rivers which experience high tidal ranges (6 m+, typically; Lynch 1982; Bartsch-Winkler and Lynch 1982). They attain 9 m in height (Qiantang River, China) and can penetrate more than 100 km inboard of a shoreline (Kampar River, Indonesia, Garonne River, France). Inherent dangers render difficult the direct study of modern tidal bores, but recent field studies illustrate their sedimentological consequences (Chanson et al. 2011; Furgerot et al. 2013; Fan et al. 2014) as well as physical experiments (Chanson 2005; Koch and Chanson 2008, 2009). Published accounts classify bores into undular (tidal-bore Froude number F > 1.0) and breaking types (F > 1.7–1.8; Henderson 1966; Docherty and Chanson 2012), the former entailing well-defined undulations behind a front and the latter manifested as a single, breaking front. Many bores also display a series of irregular “whelps” that follow behind the front. Both direct observations and experiments indicate that shear and scour, together with strong turbulent mixing, occur at the fronts of undular and breaking bores, and also in association with the undulations and whelps that follow the front (Wolanski et al. 2004; Koch and Chanson 2008, 2009; Docherty and Chanson 2012). The concentration of suspended sediment increases for several minutes after the passage of the bore, and upstream advection of sediment occurs over periods of tens of minutes thereafter (Chanson et al. 2011; Furgerot et al. 2013), leading to significant sediment reworking. Tessier and Terwindt (1994) and Greb and Archer (2007) have also suggested that soft-sediment deformation structures in some intertidal-flat sediments may have been the product of tidal bores. Recently, Fan et al. (2014) have provided detailed descriptions of tidal-bore deposits from the Qiantang estuary of eastern China. There, bore deposits comprise a basal, planar to undular erosion surface overlain by massive sand and in turn by sand with planar stratification, all of which are modified by soft-sediment deformation to varying degrees.
We are aware of only one other published study that has documented persuasive evidence of tidal bores in the ancient record. Martinius and Gowland (2011) attributed rare, discontinuous (< 10 m long), erosionally based beds of massive to diffusely stratified sandstone in the Jurassic of Portugal to the action of tidal bores. These 0.05–0.15 m thick beds are enclosed by cross-bedded sandstones that are the product of bidirectional (i.e., tidally modulated) paleoflows in fluvial channels. The undulating basal erosion surfaces of the same beds show asymmetric and irregular excavations < 0.3 m in width and 0.15 m in depth. These features are attributed by Martinius and Gowland (2011) to channel-bed scour beneath upstream-propagating tidal bores. The sand drape that fills these scours and overlies the basal surface is attributed to fallout from suspension after the passage of the tidal bore.
The data presented by Martinius and Gowland (2011) argue strongly in favor of their interpretation of Jurassic tidal bores, and they are entirely consistent with the observations of Fan et al. (2014) on modern bore deposits. Herein, we document deposits from the Pennsylvanian of northeastern Kansas, USA. These deposits preserve a suite of characteristics consistent with, and complementary to, those documented by Martinius and Gowland (2011) and Fan et al. (2014) and attributed to the action of tidal bores. We use this dataset to develop criteria for the recognition of ancient tidal-bore deposits, and we highlight its implications in understanding of the Pangean paleotropical environment of the Pennsylvanian.
Pennsylvanian Incised Valley Fills of the Midcontinent
Mudrock and carbonate-dominated Pennsylvanian cyclothems in Midcontinent USA preserve several sandstone bodies with incised basal surfaces and linear plan geometries. These bodies, although they are frequently lacking in aspects of stratigraphic context, are chiefly interpreted as incised-valley fills (e.g., Archer and Feldman 1995; Archer and Greb 2012) (Fig. 1). Valleys cut during Pennsylvanian drawdowns of eustatic sea level induced by Gondwanan glaciations were partially filled by fluvial deposits and then backfilled by transgressive estuarine deposits during subsequent postglacial sea-level rises (Archer et al. 1994). Paleocurrent data, paleodrainage patterns, and other criteria indicate that the incised fluvial systems represented by these bodies drained southwestward toward lowstand shorelines in the Arkoma and Anadarko basins of Oklahoma (Archer et al. 1994; Archer and Feldman 1995; Feldman et al. 2005). Many physical and biogenic sedimentary structures indicate that tides, probably with ranges > 6 m, were important in these ancient estuaries (Archer and Greb 2012). Wells et al. (2007) generated numerical models of Pennsylvanian tides for an early transgressive scenario in the East Kansas Embayment (EKE), a southwestward-broadening, elongate corridor within which valleys were repeatedly incised and backfilled during Pennsylvanian sea-level drawdowns and rises, respectively (Fig. 2). Some runs of these models generated a mesotidal to macrotidal, diurnal tidal regime for the EKE because of resonant amplification of diurnal tidal constituents. Stratigraphic data (Wells et al. 2007) indicate that the EKE opened onto the transgressive coastline at around what is now the Kansas–Oklahoma border, some 200 km to the southwest of the study site in NE Kansas (Fig. 2B).
We have studied multiple Virgilian (Gzhelian) incised-valley fills in northeastern Kansas, southeastern Nebraska, and adjacent areas (e.g., Fischbein et al. 2009). Some of these fills preserve sedimentary features similar to those attributed to the action of tidal bores by Martinius and Gowland (2011) and by Fan et al. (2014). The subject of the present study is a multistory sandstone body at Echo Cliff Park in Wabaunsee County, Kansas (38° 56′ 58.27″ N, 95° 57′ 53.27″ W). This putative composite valley fill comprises strata of the Wood Siding and overlying Onaga Shale formations (Moore et al. 1944; Mudge 1956) (Fig. 1).
The Echo Cliff Sandstone Body
The Echo Cliff sandstone body comprises at least three stories and possibly more since its basal contact is not exposed (Fig. 2). The lower and middle stories (Stories 1 and 2) consist principally of trough cross-bedded and ripple cross-laminated fine- to medium-grained sandstone. The upper parts of these stories also contain laterally variable components of interlaminated siltstone and fine-grained sandstone (heterolithic) and dark gray siltstone. Stories 1 and 2 have been ascribed to the Wood Siding Formation (Caneyville Limestone of Mudge 1956, and references therein). The upper story (Story 3) preserves heterolithic facies underlain by a basal conglomerate containing clasts eroded from the Brownville Limestone (Mudge 1956), suggesting that it is genetically associated with the overlying Towle Shale (Fig. 1). Large load casts of sandstone subtend from the boundary of Story 3 into underlying heterolithic facies of the uppermost Story 2. Overall, cross-bedding and ripple cross-lamination in the sandstone body show clearly bimodal directional distributions with modes to the northeast and southwest (Fig. 2F).
A facies analysis of the exposure at Echo Cliff is summarized in Table 1. Seven lithofacies are recognized, and their distribution across the Echo Cliff outcrop is shown in Figure 2. The exposed part of Story 1 is composed of the heterolithic Facies 6, as is that of the uppermost Story 3 (Fig. 2). Much of Story 2 comprises stratified sandstones of Facies 2–4 (Fig. 2). Contained within the lower part of Story 2, however, is a single interval of massive, muddy sandstone (Facies 5), which is the focus of this paper. Although thin, this interval is persistent across the extent of the outcrop, and is anomalous in the context of the stratified sandstones that enclose it.
The basal erosion surface of Story 2 shows relief of at least 2 m immediately adjacent to the cliff line (Fig. 3B). The overlying sandstone bedset, which attains a maximum thickness of 2.50 m, preserves a series of convex-upward stratal surfaces with a paleorelief of 2.0 m (Fig. 3A). This macroform unit is draped and overlain by more planar bedsets. Most of the sandstone beds within the convex-upward macroform are ripple cross-laminated and also contain abundant flaser bedding. Flat lamination and locally symmetrical wave ripples are also present. There are discontinuous horizons of imbricated, platy, siltstone intraclasts.
The massive, muddy sandstone unit (Facies 5) is located near the top of the macroform. It can be traced from its crest down to its downlapping termination against the base of the middle story, thinning gradually downflank (Fig. 3A). The base of this interval is defined by the top of the underlying, ripple cross-laminated and flaser-bedded sandstone bed, which is scalloped into a form set of symmetrical (to locally asymmetrical), round-topped ripples that exhibit bidirectional cross-lamination (Fig. 3D, E). The unit overlying the aforementioned surface is referred to Facies 5 and comprises one or more beds of (1) massive, structureless and somewhat muddy fine-grained sandstone, and/or, more commonly, (2) structureless, plant-debris-rich, fine-grained sandstone. These beds drape and fill topography created by the ripple form set on the basal surface. In at least three places, a lens of the muddy massive sandstone has eroded a lens of the plant-debris-rich sandstone (Fig. 3C–E). The lenses of muddy, massive sandstone attain 0.10 m in thickness and 10 m in width. Locally, the top of this bed is also scalloped into a form set of symmetrical and asymmetrical ripples (Fig. 3E).
Current-ripple cross-lamination structures from within the beds below and above the interval of interest, and ripple orientations from the form sets at the basal and upper bounding surfaces, show a bipolar distribution with northeast and southwest modes (Fig. 2F). Imbricated platy siltstone clasts from below the bed indicate northeastward sediment transport.
Multiple attributes of the Echo Cliff sandstone body are compatible with a strong tidal influence on an ancient fluvial system: (1) bimodal to bipolar paleocurrents in Story 2, and (2) an overall abundance of rhythmites, mud drapes, and flaser bedding, structures associated with tidal environments (Davis and Dalrymple 2012). We interpret the sandstone body as the fill of a tidally influenced river channel (Davis and Dalrymple 2012). Numerous authors have noted similar facies from other fluvial–estuarine incised-valley fills in the Pennsylvanian of the central USA (Archer et al. 1994; Feldman et al. 1995; Feldman et al. 2005; Fischbein et al. 2009; Archer and Greb 2012, among others). Given that the down-paleoslope dispersal direction was towards the southwest, then the southwestward paleocurrent mode seen in the Echo Cliff body represents downstream fluvial currents and the northeastward mode must record backflow during flood tides (Fig. 2B, F). We surmise that the channels were 2–10 m deep on the basis of the basal erosional relief of Story 2 as a minimum, and on the thicknesses of stories overall as a maximum value.
The steep-sided erosion surface at the base of the Story 2 (Fig. 3B) suggests that the line of the cliff is located close to a paleochannel cutbank. The convex-upward macroform bedset in the lower part of the Story 2 (Fig. 3A) which is direct contact with this erosional edge is, therefore, interpreted as the remnant of a bank-attached bar of some kind (such as an alternate bar, point bar, or unit bar). The nearly parallel-to-paleoflow orientation of the cliff in which it is exposed and the broadly symmetrical nature of bedding surfaces within the ancient bar indicate that it grew chiefly by aggradation (Fig. 2).
The presence of erosionally based lenses of mixed, muddy and plant-debris-rich sandstone (Facies 5) within a single stratigraphic horizon, however, demands a more specific explanation. The admixed fabric and a lack of physical sedimentary structure are inconsistent with deposition from dilute current flows like those evidenced by features in the enclosing sandstone beds. Rather, the existence of these admixed fabrics, together with the erosional bases of the sandstone lenses, are sedimentologically compatible both with the interpretations of Martinius and Gowland (2011) and with studies of modern tidal bores (Wolanski et al. 2004; Chanson et al. 2011; Fan et al. 2014). Two critical features described from modern deposits (Fan et al. 2014) and Jurassic deposits (Martinius and Gowland 2011) are directly comparable with features in the Echo Cliff sandstone body: (1) scalloped, stepped basal surfaces and (2) a draping, massive sandstone facies.
Alternative interpretations of the Echo Cliff deposits are formation by either: (1) a mass-flow event following nearby bank collapse, (2) a tsunami, or (3) a hyperconcentrated flow during a flood event. A bank-collapse origin can readily be discounted, however, because there are no clasts of fine-grained bank material in Facies 5, even though it lies close to a paleochannel bank. A tsunami origin cannot be entirely discounted, since tsunami-induced bores are known to travel for tens of kilometers upstream from an affected coastline (Tanaka et al. 2012; Chanson and Lubin 2013). However, four points argue against a tsunami origin: 1) the strong association of the described beds with features attributed to tidal activity, 2) their similarity to accounts of tidal-bore deposits (Martinius and Gowland 2011; Fan et al. 2014), 3) their location far from an ancient shoreline (cf. Wells et al. 2007), and 4) the comparative rarity of tsunami in modern settings, relative to the occurrence of tidal bores in 50 to 100 days out of a given year. A hyperconcentrated flood flow almost certainly could not have simultaneously formed the symmetrical and bidirectional wave ripples that are preserved on bounding surfaces in the deposit. Accordingly, deposition in association with tidal-bore dynamics is the most parsimonious explanation of Facies 5.
We interpret the interval comprising several mutually crosscutting lenses of Facies 5 to be the deposits of multiple bores that passed over the developing bank-attached bar in a tide-influenced paleoriver. The scalloped erosion surface with a ripple form set that constitutes the basal bounding surface of the interval is difficult to reconcile with passage of a tidal bore that was also responsible for depositing the overlying Facies 5 sandstone, since the wavelength of the ripples is much less than the likely wavelength of a tidal-bore front, and the actual wavelength of the Facies 5 lenses. Accordingly, we suggest that the basal surface (and perhaps the similar top bounding surface) were fashioned by whelps that formed the declining phase of a bore's passage. The overlying lenses of massive sandstone (Facies 5) can then be attributed to erosion and subsequent deposition from suspensions formed by the passage of later tidal bores. Erosion prior to deposition of the sand lenses occurred beneath the crest of the tidal bore front (cf. Wolanski et al. 2004; Chanson et al. 2012; Furgerot et al. 2013), which suspended large volumes of sediment that were then entrained in the upstream-propagating wave train. By analogy with the data of Furgerot et al. (2013), sand was then deposited from suspension over periods of tens of minutes after passage of the bore front from upstream-directed current motion. Moreover, the basal erosional relief and bed thickness of the ancient deposits described herein (10 cm) is exactly equivalent to the thickness of sediment eroded and advected upward by the modern tidal bore described by the same authors. The two distinct but texturally similar variants of Facies 5 are interpreted to reflect local, bed-scale variations in the character of river bottom sediment, with some zones rich in detrital plant debris. The occurrence of a series of putative bore deposits, bounded by scalloped erosion surfaces, suggests preservation facilitated by the clustering of bores over a short time period.
Discussion and Conclusions
Similarities between modern bore deposits from China (Fan et al. 2014), a putative Jurassic tidal-bore deposit (Martinius and Gowland 2011), and the Pennsylvanian Echo Cliff sandstone body are remarkable, and they strongly suggest a common mode of emplacement. Although the Echo Cliff examples lack the upstream-dipping cross-bedding described in the Jurassic case, they show the scalloped basal erosion surface to be molded into round-topped, asymmetrical and symmetrical ripple forms with bidirectional internal lamination. The sense of asymmetry in some of the ripples at Echo Cliff indicates upstream sediment transport. Textural differences between the Jurassic example, the deposits described here, and for that matter bore deposits described from the modern by Furgerot et al. (2013) and Fan et al. (2014) can be attributed to the nature of river bottom sediment at the different study sites. Specifically, the abundance of coaly traces in some of the bore deposits at Echo Cliff likely reflects the temporary storage of plant debris at certain sites on the paleo-river bed.
We suggest that the association of erosionally based, discontinuous beds of massive (muddy, or plant-debris-rich) sandstone (Facies 5; Table 1) with evidence for upstream sediment transport, in tidally influenced fluvial deposits is indicative of deposition from tidal bores (Fig. 4). These volumetrically minor, but interpretationally critical, facies are fundamentally different from the cross-laminated and cross-bedded sandstones that enclose them. We submit that the sedimentological characteristics listed above, and summarized in Figure 4, are robust criteria for identification of tidal-bore deposits in the rock record. Although we describe but a single case, its distinctiveness renders it a significant discovery.
The presence of tidal-bore deposits so far inland (c. 200 km; Wells et al. 2007) from the interpreted lowstand shoreline during the late Pennsylvanian suggests that: (1) these rivers were influenced by tides more than 200 km inboard of the coast, and (2) tidal ranges at this time were > 6 m near the coastline, likely decreasing upstream. Such a great upstream distance for propagation of tidal effects is not excessive in comparison with modern large tidal embayments (Archer and Hubbard 2003), and it is also consistent with the modeling results of Wells et al. (2007) for the Pennsylvanian EKE. Although the Wells et al. (2007) model is designed to account for incised-valley fills somewhat lower in the regional upper Pennsylvanian succession (Tonganoxie and Ireland incised-valley fills: Archer et al. 1994; Feldman et al. 1995), the location and orientation of the body described here are coincident with the EKE, and so a similar paleogeography can be assumed. We hope that our data stimulate further discoveries of tidal-bore deposits in the Pennsylvanian record of the Midcontinent and elsewhere.
We thank B. Tessier and an anonymous individual for their detailed and thoughtful reviews of the submitted manuscript, and Editor J. MacEachern, Associate Editor S. Dashtgard, Managing Editor M. Lester, and Copy Editor J. Southard for their reviews and advice. Andrew Hutsky assisted with drafting diagrams.