Dalton Wells: Geology and significance of debris-flow-hosted dinosaur bonebeds in the Cedar Mountain Formation (Lower Cretaceous) of eastern Utah, USA

Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217 – 245 www.elsevier.com/locate/palaeo

Dalton Wells: Geology and significance of debris-flow-hosted dinosaur bonebeds in the Cedar Mountain Formation (Lower Cretaceous) of eastern Utah, USAB David A. Eberth a,*, Brooks B. Britt b, Rod Scheetz c, Kenneth L. Stadtman c, Donald B. Brinkman a a

Royal Tyrrell Museum of Palaeontology, Box 7500, Drumheller, AB, Canada T0J 0Y0 b Department of Geology, Brigham Young University, Provo, UT, United States c Earth Sciences Museum, Brigham Young University, Provo, UT, United States

Received 26 January 2005; received in revised form 10 August 2005; accepted 4 November 2005

Abstract The Lower Cretaceous Dalton Wells dinosaur locality, north of Moab, Utah, consists of a 2 m thick, stacked succession of four fossil-rich bonebeds with an estimated collective lateral extent of more than 4000 m2. To date, 215 m2 of the bonebeds have been excavated and more than 4200 vertebrate fossil field specimens have been collected. The site occurs at the base of the Yellow Cat Member of the Cedar Mountain Formation, and lies unconformably on Upper Jurassic strata of the Morrison Formation. The age of the assemblage is tentatively accepted as Barremian. Nine dinosaur taxa are recognized, making Dalton Wells one of the richest (abundance and diversity) Lower Cretaceous dinosaur localities in the world. The bonebeds are rare examples of fossiliferous subaerial debris flow deposits. At least four flows originated an undetermined distance up-slope from an Early Cretaceous site of dinosaur bone concentration and flowed through the site carrying bones short distances toward an adjacent area of lakes, ponds and wetlands. Host sediments most likely accumulated on the eastward-dipping slopes of a backbulge in a foreland basin. The overall depositional setting was alluvial–lacustrine in a warm-to-hot and seasonally wet-and-dry climate. The consistent expression and location of lacustrine facies in both the Brushy Basin Member of the Morrison Formation and the Yellow Cat Member of the Cedar Mountain Formation indicate that high rates of subsidence typified the Dalton Wells area both before and after the J–K transition, and further suggest that there was some form of long-term to episodic and localized influence on subsidence. The depositional slope required to generate debris flows may have been established by forebulge crustal flexure, but could also have been established or amplified by the high rates of subsidence that maintained the lacustrine paleoenvironments in the field area. Concretionary deposits in the Yellow Cat Member at Dalton Wells are diagenetic and apparently resulted from groundwater precipitation, likely enhanced by the presence of organics (bones) and lacustrine carbonates. The regional and cross-facies expression of calcretes in the Yellow Cat Member combined with the rarity of pedogenic indicators confirms that they are not paleosols. Mineral precipitation likely occurred as groundwater flow was established in response to foreland basin development.

B The use of the plural bDalton WellsQ is consistent with the name used by the Dalton family and local community for the original Dalton family ranch at this location. The singular bDalton WellQ is a modification of the original name that has been recently employed by some governmentagencies following USGS feature name usage due to an error on the Merrimac Butte Quadrangle. * Corresponding author. Fax: +1 403 823 7131. E-mail address: [email protected] (D.A. Eberth).

0031-0182/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.palaeo.2005.11.020

218

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

It remains unknown under what specific circumstances animals died and carcasses accumulated prior to being reworked by debris flows. However, the stratigraphic association of seasonally wet-and-dry facies throughout the Yellow Cat Member suggests that drought-induced mortality events were probably common. In a seasonally wet-and-dry environment, debris flows may have been triggered by intense rainfall or seismic events. D 2005 Elsevier B.V. All rights reserved. Keywords: Debris flow; Dinosaur; Cretaceous; Bonebed; Cedar Mountain Formation; Utah

1. Introduction The Lower Cretaceous Dalton Wells fossil locality (DW) is located west of Arches National Park and northwest of Moab, Utah (Fig. 1), and consists of a stacked succession of four, sauropod-dominated, multitaxic dinosaur bonebeds. Historically, the site has been regarded as a single bonebed or quarry because the bonebeds have limited stratigraphic and geographic distributions, and there has been no previous documentation of the site’s internal stratigraphy (e.g., Britt et al., 1997; Eberth et al., 1997). The site is extremely rich: more than 4200 dinosaur fossils were collected for preparation and study, and we estimate that it contains many thousands more. Catalogued vertebrate fossils from DW in collections at Brigham Young University (Earth Science Museum) indicate that the site contains a significant diversity and abundance of Early Cretaceous dinosaurs, but only a six fragments from four non-dinosaurian taxa (Table 1). In addition, the host formation (basal-most Cedar Mountain Formation), records episodic sedimentation in a regional setting that was being transformed from an internallydrained basin dominated by low rates of sedimentation and a large amount of reworking (Upper Jurassic, Brushy Basin Member, Morrison Formation) to a foreland basin dominated by higher rates of sedimentation in alluvial, coastal, and shallow marine depositional systems (Dakota Sandstone). Although DW yields an important Early Cretaceous dinosaur assemblage that likely reflects an actual dinosaur community (Britt et al., 1996, 1997), very little is known of the site’s depositional history and taphonomy because of the complexities of it’s depositional history, subsequent overprinting by diagenetic and local-to-regional tectonic events, and modern weathering and scree cover. Thus, study of the Dalton Wells fossil site and its associated facies represents a unique opportunity to (1) examine a rare, but diverse Early Cretaceous terrestrial vertebrate assemblage and its paleoecology, and (2) assess the manner in which tectonic reorganization of Western North America may have influenced vertebrate preservation. Here we present stratigraphic and sedimentologic data that: 1) provide a local-to-regional depositional

context for the site; 2) help explain the site’s origins; and 3) provide further insight into the nature of basin evolution in the region during Late Jurassic–Early Cretaceous. Although we include some taphonomic data from the site, Britt et al. (2004) have interpreted those data more fully. 2. History Dalton Wells has long been known to local residents and it is likely that occupants of the Civilian Conservation Corps camp at Dalton Wells (1935–1941; Louthan, 1993) were aware of the fossils as well. J. bPopQ Leroy Kay showed the locality to J.A. Jensen of Brigham Young University (BYU) in the early 1960s, but there was little interest in excavating at that time, probably because it was thought to be just another Morrison Formation outcrop. Following the discovery of a fragmentary maxilla containing an iguanodont tooth in the early 1970s, scientific interest in the site began to grow. The tooth indicated that the stratum was Cretaceous, not Jurassic, and the specimen was designated the type of Iguanodon ottingeri by Galton and Jensen (1978). Jensen sampled the quarry in 1973, and in 1975 he and one of us (KLS) excavated the site for five days, recovering specimens thought to represent new dinosaur taxa. Based on these results Jensen expanded the excavation, recovering more than 800 specimens. In 1994 a joint BYU and Museum of Western Colorado (MWC) crew reopened the site, and collection and preparation of specimens has since progressed rapidly. Since 1978 more than 4200 specimens representing a total of nine dinosaurian taxa have been recovered (Table 1). 3. Methodology Standard field, quarry mapping, and paleontologicalexcavation techniques were employed. Stratigraphic sections and sedimentary facies were photographed, measured and described at three scales: regionally (over 10s of km’s), locally (over 100s of m to km’s), and within the excavated site. All field localities were mapped on USGS 7.5V quadrangle sheets and/or

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

219

Fig. 1. Location map for Dalton Wells, including locations of measured sections and outcrops examined during this study. Photograph of the south western face of the DW quarry looking toward the east-northeast.

220

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Table 1 Taxonomic list from Dalton Wells (n = number of specimens in cases where the numbers are b 5) Chelonia Family uncertain (n = 2) Choristodera Champsosauridae (n = 1) Archosauria Pterosauria or Aves (n = 1) Crocodylia Crocodylidae (n = 2) Theropoda Dromaeosauridae Utahraptor ostrommaysorum Ornithomimidae? Gen. and sp. new (gracile theropod) Uncertain Nedcolbertia justinhofmanni Uncertain Gen. and sp. new (n = 1) Sauropoda Camarasauridae Gen. and sp. new Brachiosauridae Cedarosaurus weiskopfae Titanosauridae Gen. and sp. new Ornithopoda Iguanodontidae Gen. and sp. new Ankylosauria Polacanthidae Gastonia burgei

1 : 100,000 topographic maps (1927 North American Datum [NAD 27]). GPS data collected using the WGS 84 datum were corrected accordingly. Hand samples collected for lithologic, petrographic, X-ray diffraction, and microprobe studies were prepared and processed at the Geology Department of Brigham Young University and the Sedimentary Geology Laboratory at the Royal Tyrrell Museum. Bone distributions and facies variations in the bonebeds were documented using photographs and two-dimensional mapping techniques (baseline and grid mapping). 4. Stratigraphy and age Comprehensive studies and reviews of the stratigraphy, sedimentology and tectono-stratigraphic history of the Morrison and Cedar Mountain formations are available in Aubrey (1996, 1998), Currie (1997, 1998) and Kirkland et al. (1997, 1999). Although there is no agreement on the history of the tectonic transformation of the basin from intracratonic to foreland, there is general agreement that the complex stratigraphy and

paleogeographic distribution of subunits within the Cedar Mountain Formation reflect sedimentation and sediment-preservation patterns initiated by the Sevier Orogeny and mediated by differential uplift, weathering, subsidence, forebulge development and migration history, and, possibly, local salt-diapir-related tectonics. The most recent and detailed stratigraphic subdivision of the Cedar Mountain Formation in Eastern Utah is that of Kirkland et al. (1997, 1999) who recognized five members (Buckhorn Conglomerate, Yellow Cat, Poison Strip Sandstone, Ruby Ranch, and Mussentuchit). Kirkland et al. (1997, 1999) suggested that the entire formation extends in age from Barremian to Cenomanian and recognized three major unconformities based largely on vertebrate biostratigraphy: 1) the base of the Buckhorn and correlative Yellow Cat members on the Morrison Fm.; 2) the base of the Poison Strip Sandstone on the Yellow Cat; and 3) the base of the Mussentuchit on the Ruby Ranch. In this paper we tentatively adopt the stratigraphy of Kirkland et al. (1997, 1999) with the following modifications (Fig. 2). First, because there is no discrete basal conglomeratic unit (Buckhorn Conglomerate) at Dalton Wells or in the area, we regard the Yellow Cat Member as the lowest stratigraphic unit of the Cedar Mountain Formation in the Moab area. The base of the Yellow Cat Member is easily recognized by its massive green mudstones that yield extraformational granules, pebbles and cobbles, and the presence of discontinuous calcareous and oxide-rich concretions. Our measured sections confirm that the Dalton Wells locality occurs stratigraphically at the base of the Yellow Cat Member of the Cedar Mountain Formation, and above — but in contact with — the Brushy Basin Member of the Morrison Formation (Fig. 2). Thus, the base of the bonebed and the Yellow Cat Member at Dalton Wells rests on the J–K unconformity. Secondly, our observation that the Yellow Cat Member and Poison Strip Sandstone both possess beds of massive to pebbly sandstone (see our measured sections) raises questions about the nature of their contact and how to differentiate the two units. Kirkland et al. (1997, 1999, p. 209) defined the Poison Strip as a stratigraphic zone of laterally continuous and stacked pebbly sandstone bodies throughout eastern Utah. We concur that there is a pronounced up-section transition from green mudstones and minor sandstones (Yellow Cat Member) to prominent stacked sandstone bodies (Poison Strip Sandstone). Such transitions are typical of non-marine successions in foreland basins and reflect times of increased sediment supply and limited accommodation (e.g., Shanley and McCabe, 1994, 1995,

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

221

Fig. 2. Stratigraphy of the Cedar Mountain Formation in eastern Utah modified from Kirkland et al. (1997, 1999). Photograph highlights the interfingering contact between facies of the Yellow Cat Member and Poison Strip Sandstone at location 4-29-97 (Fig. 1). Outcrop is tens of meters wide.

222

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

1998; Eberth et al., 2001). However, exposures in the western corner of our field area (outcrop 4-29-97, Figs. 1 and 2) consist of interfingered sandstone (Poison Strip Sandstone) and massive concretionary mudstone (Yellow Cat Member) and, thus, exhibit a gradational transition between the two units. We propose that increased accommodation in the Dalton Wells area and farther east — perhaps caused by local salt-diapir-related faulting (Doelling, 1988; Johnson and Aubrey, 1994; Aubrey, 1996) — allowed for a gradational transition from the Yellow Cat Member to the Poison Strip Sandstone. This interpretation implies that there is no unconformity between the Yellow Cat and the Poison Strip Sandstone in the DW area. Accordingly, we place the Yellow Cat–Poison Strip contact at the base of the first prominent and stacked alluvial sandstone in a given area or section. Kirkland et al. (1999) recognized a paleosol at the top of the Morrison Formation characterized locally by a dark red color and a variety of paleo-pedogenic features. They inferred that this paleosol represents the J–K unconformity — ranging in age from Tithonian to Barremian. However, this pedogenic redbeds profile is poorly preserved at Dalton Wells and occurs as a small, laterally-limited patch that is only a few centimeters thick. Thus, in the vicinity of the Dalton Wells quarry this marker zone was either removed by Cretaceous erosion, or only poorly developed. Kirkland et al. also recognized a widespread multi-layer calcrete zone at the base of the Yellow Cat Member, but were uncertain as to its origin and chronostratigraphic significance (Kirkland et al., 1999, p. 204; Kirkland, pers. comm., 2003). Here we argue that the calcrete is diagenetic, postdating deposition of the DW bonebeds. The youngest strata of the Brushy Basin Member of the Morrison Formation have been dated radioisotopically at 145–149 mya (Kowallis et al., 1998). Palynomorphs from an organic-rich sample collected at the base of the overlying Poison Strip Sandstone (e.g. species of Concavissimisporites, Trilobosporites and Verrucosisporites) indicate an age range of latest Neocomian (Hauterivian) to Aptian, and probably Barremian (G. Waanders, pers. com., 2003). Kirkland et al. (1997, 1999) also suggested a Barremian age for the Yellow Cat Member based on (1) comparison of the DW and European dinosaur faunas, and (2) the presence of the charophyte (Nodosoclavator bradleyi). However, there are no shared dinosaur genera between the Cedar Mountain Formation and supposedly equivalent strata in Europe, and N. bradleyi is now known to range from the Berriasian through to the Aptian

worldwide (M. Shudack, pers. com., 2003). Our unpublished taxonomic data indicate a close similarity between the titanosaur from Dalton Wells and Janenschia (= Tendaguria, Bonaparte et al., 2002) from the Upper Saurian beds of the Tendaguru of Tanzania. The apparent similarity of the Dalton Wells and African titanosaurs raises the possibility that the Yellow Cat Member could be older than Barremian. This possibility is compatible with the results of Aubrey (1998) and Currie (1998) who partitioned the lowest portions of the Cedar Mountain and Burro Canyon formations using non-marine sequence stratigraphic and tectonic models and concepts, and suggested that these lowest portions may be Tithonian to Neocomian in age. Nonetheless, until independent chronostratigraphic data are available, the most parsimonious age assignment for the Dalton Wells assemblage remains Barremian. 5. Tectonic and depositional setting Currie (1998) studied the tectono-stratigraphic history of the Morrison and Cedar Mountain formations in eastern Utah and western Colorado, and suggested that Cedar Mountain strata were deposited eastward of a crustal forebulge that migrated eastward episodically during the Late Jurassic and Early Cretaceous. In this setting, Buckhorn conglomerates were transported by fluvial trunk systems incising the forebulge. Currie also inferred the presence of a significant depositional hiatus following Buckhorn deposition from a regionally developed calcrete horizon variably occurring (1) at the top of the Buckhorn, (2) within the base of the Cedar Mountain Fm., or (3) within the top of the Morrison Formation. We regard the calcrete as equivalent to that described from DW by us (see below) and from the lower Yellow Cat Member of eastern Utah by Kirkland et al. (1999). In Currie’s interpretation, clastics were subsequently shed from the rebounding foredeep in central Utah and swamped the entire basin by ultimately covering the forebulge by late Early Cretaceous. Although these higher strata of the Cedar Mountain Formation were not correlated with Kirkland’s subdivisions, we speculate that the early stages of this inferred rebound and sediment supply event were responsible for the widespread sheet-style dispersal of the Poison Strip alluvial clastics described by Kirkland et al. (1999). Aubrey (1998, Fig. 4) described the stratigraphic architecture of the Upper Jurassic–Lower Cretaceous section from eastern Utah and southwestern Colorado. He subdivided the Burro Canyon Formation (equiva-

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

lent of the Cedar Mountain Formation southeast of the Colorado River [Stokes, 1952; Craig, 1981]) into a lower, Jurassic conglomeratic succession (correlable with the Buckhorn Conglomerate but included with the Morrison), and an upper, Cretaceous-age shaledominated succession, correlable with the Cedar Mountain Formation elsewhere in Utah and marked by a pronounced calcrete at it’s base. Again, we regard this calcrete as equivalent to that occurring in the Yellow Cat Member of eastern Utah and in the DW bonebed as described by Kirkland et al. (1999) and us, respectively. Thickness trend data from the Yellow Cat Member along a 25 km southwest–northeast oriented transect (Fig. 3) demonstrate a pronounced eastward thickening of the Yellow Cat Member in the Dalton Wells area. This transect defines an asymmetric east-west cross-sectional shape for the Yellow Cat Member in eastern Utah. Accordingly, we hypothesize that during

223

the Early Cretaceous, Dalton Wells was positioned on the western margin of a rapidly subsiding sub-basin of limited size. This interpretation parallels that of Johnson and Aubrey (1994) and Aubrey (1996) who proposed syndepositional half-graben growth displacement as a result of fault and salt-diapir mergers in the Salt Anticline (Cache Valley) area during the Early Cretaceous. Their cross-section (Aubrey, 1998, Fig. 8) is sub-parallel to ours and shows a similar asymmetry. We accept the interpretation of Johnson and Aubrey (1994) that subsidence was syndepositional, and thus, we hypothesize that some degree of eastward-dipping relief was maintained in this area during deposition of upper Brushy Basin Member and Yellow Cat Member strata. From the Tithonian into the Early Cretaceous differential rates of subsidence across the Dalton Wells field area apparently resulted in the establishment of betterdrained alluvial-dominated paleoenvironments to the

Fig. 3. Cross-sectional geometry of the Yellow Cat Member from the Dalton Wells area northeast to Flowing Well — a distance of 28.4 km. Note the thickening of the Yellow Cat Member at Dalton Wells. Stratigraphic datum is the Cedar Mountain–Morrison formational contact.

224

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

west and poorly drained lacustrine-dominated paleoenvironments to the northeast. The fence diagram shown in Fig. 4 is based on the correlation of three measured

sections across the Dalton Wells area (Figs. 5–7). The measured sections include the upper 30–50 m of the Brushy Basin Member of the Morrison Formation and

Fig. 4. Fence diagram showing interrelationship of alluvial, lacustrine and debris flow facies across the Dalton Wells field area for both the Brushy Basin Member of the Morrison Formation and the Yellow Cat Member of the Cedar Mountain Formation. Note the dominance of alluvial facies to the west and the lacustrine facies to the north-east in both units.

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

225

Fig. 5. Measured section and interpreted paleoenvironmental settings at Section 5-2-97 (west of the bonebeds). Insets show the relative location of the measured section (top), percentage of alluvial versus lacustrine facies (middle), and a photograph of typical outcrop (bottom). Symbol code at bottom is used throughout the paper.

the lower 20 m of the Cedar Mountain Formation including the DW bonebeds and their laterally correlative strata. To the west (Fig. 5) alluvial deposits are

dominated by red to variegated paleosols in the Brushy Basin Member and mixed paleochannel and debris flow deposits in the Yellow Cat Member. To the northeast

226

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Fig. 6. Measured section and interpreted paleoenvironmental settings at Dalton Wells fossil locality (Section 4-30-97). Insets and symbols as in Fig. 5.

(Fig. 7), the Brushy Basin Member is strongly dominated by lacustrine beds whereas the basal beds of the Yellow Cat Member comprise a 13.5 m thick succession of mixed lacustrine and debris-flow deposits. The increased occurrence of lacustrine paleoenvironments to the northeast is demonstrated at Gaston Quarry, 19 km northeast of Dalton Wells where differential com-

paction over prominent stacked lacustrine limestones in the Yellow Cat Member has created a local stratigraphic high (Figs. 3 and 8D). In the context of this trend, the section at DW (Fig. 6) represents a paleogeographically intermediate zone with a mixture of alluvial and lacustrine facies in both the Brushy Basin and Yellow Cat members.

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

227

Fig. 7. Measured section and interpreted paleoenvironmental settings at Section 5-1-97 (northeast of the bonebeds). Insets and symbols as in Fig. 5.

In the DW field area, the depositional settings for both the Brushy Basin Member of the Morrison Formation and the Yellow Cat Member of the Cedar Mountain

Formation have been interpreted as alluvial–lacustrine with a warm-to-hot, seasonally-dry climate (Masura and Davis, 1996; Eberth et al., 1997; Demko and Parrish,

228

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Fig. 8. Comparison of Brushy Basin Mbr. and Yellow Cat Mbr. facies. (A and B) Faintly developed paleochannel deposits with vague lateral accretion beds (double headed arrows). Outcrop in B is 6 m thick. (C and D) Sheets of lacustrine limestone interbedded with claystone and siltstone. (E and F) Horizontally-bedded sheets of contorted and bioturbated siltstone interpreted as trampled lake-margin deposits. Outcrop in E is 2.5 m thick. (G and H) Variegated paleosols and root marked siltstones. Scale bars in H are 10 cm.

1998; Aubrey, 1998; Kirkland et al., 1999). Both units share the following facies (Fig. 8): 1) Isolated and rare small-scale, sandstone- to intraclast-filled paleochannels;

2) 100 m-wide, decimeter-thick marls of inferred lacustrine origin; sometimes interbedded with facies 3 and 4 (below); 3) Ripple-laminated to planar-stratified siltstones and sandstones interpreted as lacustrine shoreline depos-

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

its that exhibit mud cracks, dinosaur tracks and trackways, and a variety of well-defined-to-cryptic invertebrate trace fossils; and 4) Laterally-extensive, mono-chromic-to-variegated red and grey-green massive to rooted mudstones interpreted as very poorly-drained paleosols of alluvialto-paludal origin. The overlying Poison Strip Sandstone of the Cedar Mountain Formation is characterized by well-devel-

229

oped, deeply scoured, multistoried, lenticular-to-sheet shaped channel-fill deposits (Fig. 9A). These deposits are dominated by pebble-rich, fine sandstones that exhibit evidence for ephemeral flow conditions (Fig. 9B,C,D,E) including: massive clean sandstones with isolated pebble strings and matrix-supported extraformational pebbles and cobbles; thick sets of upper flow-regime plane beds; meso-to-large-scale crossbeds with topsets and oversteepened foresets; and climbing ripples.

Fig. 9. Sandstone facies of the Poison Strip Sandstone. (A) Gradational transition upward into the Poison Strip Sandstone (YCM, Yellow Cat Member; PSSs, Poison Strip Sandstone). Person for scale. Note the internal scour surface in the middle of the photograph. (B) Upper flow regime plane beds indicating fast and shallow flow. Exposure is 50 cm thick. (C) Fully preserved topsets indicating very high rates of sediment supply. Note chert pebble below pencil. (D) Large scale cross-beds indicating the presence of meter-scale dune forms. JacobTs Staff is 1.5 m long. (E) Typical massive sediments at the base of a Poison Strip Sandstone body (ma, massive; stx, trough cross-beds). Note the overall thinning-upward nature of the sandstones and the trough cross-beds above the massive sandstone.

230

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

The Yellow Cat Member and Poison Strip Sandstone share the following facies: 1) Massive green sandy-to-pebbly mudstone with matrix supported clasts; 2) Fine grained-to-granulitic calcareous sandstone with massive pebbly bedding, trough cross bedding, planar horizontal stratification and ripple lamination; and 3) Invertebrate trace fossil assemblages consisting of feeding traces, galleries and chambers. This group of shared facies confirms that the Yellow Cat Member records the stratigraphic transition of eastern Utah from an internally drained and self-cannibalising basin in the Late Jurassic to an aggradational foreland basin in the Early Cretaceous. 6. Bonebed geology The DW bonebeds crop out along a narrow NW–SE oriented ridge 55 m above the valley floor (Fig. 1). Fossils weather out of the exposed edges of the ridge but overburden and thick scree commonly prevent examination of in situ fossil occurrences. Based on abundant bone fragments that occur below the scree, however, we calculate that the unweathered extent of the stacked bonebeds is about 4000 m2. Of this area, only 215 m2 have been excavated to date. The following descriptions are based on data collected from trenched sections during 1997, 2001, and 2003. 6.1. Descriptive sedimentology Fig. 10 is a measured section through the southeast corner of DW (Fig. 1, bmeasured sectionQ). Fig. 11 is an east–west cross-section through the bonebeds compiled from exposures in a trench that was excavated along the south face of the Dalton Wells ridge. It shows the vertical and lateral distribution of fossil bone, the J–K unconformity, and the location of the measured section in Fig. 10. The fossiliferous host-rock consists of grey-green (10GY 7/2; 10GY 5/2; 5G 7/2), massive-to-vaguely graded, sandy-to-pebbly mudstone. Samples that are soaked in water and processed for a few minutes in a blender break down rapidly into a range of grain sizes. Swelling behavior and XRD analysis of clay separates indicate that illite and very minor amounts of smectite are present (Dana Griffen, pers. comm., 2003). Coarse fractions range in size from very fine sand to pebbles and cobbles. In the coarsest grained samples, the fine sand fraction is dominated by green lithics consisting of

lithified mudstone reworked from the Morrison Formation. A subordinate amount of mature, well-rounded silicate minerals are present and consist of quartz and chert (common), and microcline feldspar (rare). Thin section examination and sediment grain-size analyses indicate that mudstone dominates the matrix through the fossiliferous units (Fig. 12). However, because we were unable to conduct analyses on large volumes of rock containing large fossils and cobbles, it is possible that large clasts dominate at the base of each deposit. Although bounding surfaces between the stacked mudstone deposits are difficult to identify and trace, fossil bone density and granule-to-cobble distributions allowed us to recognize a maximum of four stacked fossiliferous mudstone beds at DW (Units 1–4, Figs. 10 and 11A). Unit 1 is discontinuously exposed and incises the top of the Morrison Formation. It is a maximum of 1 m thick and is truncated by Unit 2, the richest bonebed horizon. Matrix-supported granules, extraformational pebbles and cobbles (4–7 cm in diameter), and isolated vertebrate elements have a patchy distribution along the base of the unit (Fig. 11B,C). Claystone laminae are present near the top of Unit 1 as well as in Unit 4 (Fig. 11D,E). The richest concentration of fossils occurs in Unit 2. It is typically 20–50 cm thick but thickens to nearly 1 m at the SE end of the locality where it exhibits strong erosional relief (Fig. 11A). Pebbles, cobbles and bones in the lower one-half of the bed are normally-graded. The upper one-half of the bed consists of massive, sandy mudstone (claystone-to-siltstone) with sparsely scattered granules and bones, but no pebbles. Wellrounded and polished extraformational chert pebbles and cobbles with maximum dimensions of 7 cm are locally abundant though generally sparse, and are usually found within the lowest 10 cm of the bed. Rounded, extraformational pebbles (up to 3 cm) and granules, and mudstone intraclasts are abundant and occur as matrix-supported clasts, especially through the lower one-half of the bed. At meter 30, near the southeast end of the bonebed (Figs. 10C and 11A), there is an asymmetrical lenticular channel-shaped dfillT that is notably rich in pebbles, cobbles and fossils. It has erosional relief of 40 cm and trends NW–SE. Disarticulated dinosaur fossils are heavily concentrated in the lower one-half of Unit 2 (Fig. 11A) and range in size from complete m-scale limb bones to mm-size broken fragments. Incomplete elements dominate the assemblage (Britt et al., 2004) and the majority of fossils in the quarried areas consist of cm–dcm-size unidentifiable fragments. Sorting is poor and elements are matrixsupported. Most elements are oriented with long axes

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

231

Fig. 10. Detailed measured section through the Dalton Wells fossil locality. Symbols as in Fig. 5. (A) Stacked tabular sandstones and siltstones of the Middle Sandstone. Hammer is 30 cm long. (B) Concretionary development (CaCO3) in units 3 and 4. Scale is 10 cm. (C) Local bone and cobble filled scour on top of Unit 1. Scale is 10 cm.

horizontal to sub-horizontal, but several smaller fragments and ribs were observed with long axes oriented sub-vertically (Fig. 11B). There are two additional massive-to-graded greengrey clayey granulitic to pebbly siltstone units (3 and 4, Figs. 10 and 11). Unit 3 is 20–80 cm thick and flat based. Locally, the lower one-half of Unit 3 exhibits a discrete, massive, granulitic and pebbly concretionary

zone. Above this lower zone, pebbles and granules increase and then decrease in abundance, indicating a reversed- and normally-graded succession. The upper one-third of Unit 3 exhibits finely interlaminated sandy siltstone and claystone similar to those in Unit 1 (Fig. 11E). Unit 4 is up to 1.75 m thick, locally exhibits 20 cm of erosional relief on Unit 3, and fines upward into massive sandy siltstone with locally developed clays-

232

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Fig. 11. (A) W–E cross-sectional geometry through the Dalton Wells bonebeds. Different shades indicate units 1–4. (B) High-angle orientation of rib fragment (f) in Unit 1. Hammer is 30 cm. (C) Massive distribution of pebbles (p) and reworked mudstone clasts in Unit 1. Z, zebra lamination. Scale units are centimeters. (D) Clay laminae (L) in Unit 1. Scale units are centimeters. (E) Alternating laminae of clay and siltstone in the upper portion of Unit 4. Note the occurrence of pebbles and granules in many of the photos.

tone laminae. Vertebrate fossils are generally rare in Unit 4 but occur with greater abundance toward the northwest (Fig. 11A). Trace fossils are common only in Unit 4 (see below). A patchy, oxide- and carbonate-rich concretionary cement is present in all four units at DW (Figs. 10 and 13). Concretionary development is laterally discontinuous and irregularly shaped, but preferentially encases fossil bone, especially in units 2 and 3. Weathered concretions have a brown oxide coating and exhibit local silica replacement, limonite, and manganese staining in the form of dendrites or a local black stain. Concretionary development is less well developed upward through each unit as well as through combined units 1–4. A smaller-scale, horizontally streaked pattern of oxides (zebra lamination) is present locally (Fig. 13B, lower right). The concretionary horizon is widespread throughout the field area, but has variable expression, crossing facies boundaries and extending down, locally, into the top of the Morrison Formation

(Aubrey, 1998). One kilometer northeast of DW the concretionary zone is very massive (Figs. 7 and 13D, measured Section 5-1-97, meter 66), completely cementing many meters of section. At other localities, concretionary development is less well developed than at DW and is recognized by sparsely developed nodules (Figs. 1 and 8B outcrop 4-29-97). No root traces or macroscopic paleosol features were observed in the quarry matrix or lithosomes. Verticalto-horizontal, sub-cylindrical invertebrate feeding traces and dwelling structures are very common in Unit 4 (Fig. 14). Some burrows follow the preserved surface of bones and in some cases burrows occur as grooves in fossil bone (Britt et al., 2004; Nolte et al., 2004). Unit 4 contains numerous vertical burrow fills, about 2 cm in diameter that were infilled by overlying sands (see below). Exposed outer surfaces of these burrow-fills exhibit concentric transverse ridges and cross-sections indicate the burrow filling is meniscate. Rare, horizontal burrows occur near the top of Unit 4

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

233

Fig. 12. Sediment grain size analysis of matrix from Dalton Wells bonebeds. (A) Sample from middle of Unit 1. (B) Sample from the top of Unit 3. Note the high proportion of silt and clay from both samples.

but are smooth walled and up to 6 cm in diameter. Short lateral branches of these burrows often terminate in expanded flask-shaped chamber-fills (Fig. 14A). Bidirectional horizontal orientation (trend) data were collected from 275 long bones and analysed using Oriana software (Kovach Computing Services, 1994–2004, version 2.02a). The average trend is 160– 3408, but the length of the mean vector (r) is 0.08 indicating a very weak concentration of data along this trend (Fig. 15A). Hydraulically reworked skeletal elements often exhibit bimodal orientations (parallel and

perpendicular to current flow) that develop in response to element morphology, water depth, and current strength (e.g., Voorhies, 1969; MacDonald and Jefferson, 1985; Behrensmeyer, 1990; Morris et al., 1996). However, because bimodal data sets often impair the application of statistical analyses designed for unimodal analysis and complicate interpretations of the data (Morris et al., 1996), we have used a combination of statistical methods. We tested the hypothesis that the data are uniformly and, thus, randomly distributed using a Rao’s Spacing

234

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Fig. 13. Different expression of the concretionary development in the Dalton Wells area. (A) Nodular concretions near the top of Unit 4 above the bonebeds. Hammer is 30 cm long. (B) Oxide streaks (zebra lamination) in Unit 1 (below mudstone clast). Scale in centimeters. (C) Laterally variable CaCO3 concretionary development in units 3 and 4 (see also Fig. 10). Scale bar is 10 cm. (D) Jasper-like silica cement (dark patches) in concretionary mass in the Yellow Cat Member at section 5-1-97. Exposed surface is less than 10 cm wide.

Fig. 14. Variety of invertebrate trace fossils at Dalton Wells fossil locality. All samples from Unit 4. (A) Branched crustacean(?) gallery and chambers. Scale bar 10 cm. (B–D) Tubular feeding traces. Scale units in centimeters. Knobby texture in dCT reflects differential weathering around cemented granules.

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

235

Fig. 15. Orientation and paleocurrent data. (A) Long bone trend data collected from Dalton Wells quarry maps. Data are bidirectional. (B) Five paleocurrent directions collected from mesoscale cross-beds and ripple lamination in the Middle Sandstone above the DW bonebeds suggest an eastsoutheast flow direction.

236

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Test, a Rayleigh Test, Watson’s U 2 Test and a Kuiper’s Test (Fig. 15A). The Rao’s Test is a consistently more powerful discriminator of uniformity in bidirectional data sets than a Rayleigh Test, and is not limited to testing theoretical distributions (uniform or von Mises) — as is the case with the Watson’s U 2 and Kuiper’s tests. Only the Rao’s Spacing Test convincingly falsified the hypothesis of uniformity ( p b 0.01), and, thus, supports the interpretation that a significant number of long elements were oriented along a NNW–SSE trend. A uniform data distribution was nearly, but not convincingly, falsified by the Rayleigh Test, the Watson’s U 2 Test, and the Kuiper’s Test, raising the possibility that long bone orientations exhibit a bimodal distribution, characteristic of stream flow influence. However, visual inspection of the rose diagram for DW (Fig. 15A) reveals that the distribution comprises a primary mode (NNW–SSE) and four minor modes. Accordingly, we conclude that the distributional randomness sup-

ported by the Raleigh, Watson’s U 2, and Kuiper’s tests is not the result of current-parallel and -perpendicular orientations, but rather, randomness in element orientations. In summary, our orientation data indicate that a small number of long bones (~30) exhibit a non-random orientation that may reflect debris flow or other orienting influences, but that the vast majority of elements appear to be randomly distributed. Above the stacked succession of fossiliferous pebbly mudstones (units 1–4), there is a 2 m thick, complexly interbedded sandstone–mudstone unit (Middle Sandstone, Figs. 1 and 10) that consists of granulitic sandstone, marl, and grey-green clayey siltstone (Fig. 16). The unit forms the ridge sandstone that caps the Dalton Wells fossil site (Fig 1) and is tabular and flat-bedded throughout, showing a poorly developed overall upward-fining trend. Although this coarse unit occurs below what is generally regarded as the Poison Strip Sandstone (Kirkland et al., 1999) we regard its occur-

Fig. 16. Features of the Middle Sandstone above the Dalton Wells fossil locality. Hammer is 30 cm long. Scale bar is 10 cm. (A) Stacked assemblage of tabular sandstone beds and thinly bedded siltstone. Note abundance of plane beds. (B) Close-up of dAT showing grain size changes and dominance of plane beds, interpreted as upper-flow-regime plane beds. (C) Shrinkage cracks in silty claystone. (D) Lower surface of sauropod track.

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

rence as a further example of the interfingering of Yellow Cat Member and Poison Strip Sandstone facies. Granulitic beds in the lower one-half of the Middle Sandstone exhibit planar-stratification, thin sets of cross bedding, and numerous planar erosion and reactivation surfaces that are highlighted by sharp changes in grain size (Fig. 16A,B). Local current–ripple cross strata are present. Interbeds of micritic marl and siltstone are heavily contorted and bioturbated, and host a variety of linear to u-shaped invertebrate traces, some with meniscate back-fills. The upper and lower surfaces of some of the sandstones are covered with 2–3 mm wide, ~20 mm long, U-shaped burrows. We identify these as Fuersichnus singularis. The top of one of the sandstone beds is an amalgam of oblique to near vertical burrows leading to a common point of origin and is assigned to Fuersichnus communis, a trace made by larval invertebrates, which indicates shallow water lacustrine facies (Hasiotis, 2002). A carbonate cemented, sandy siltstone bed caps the unit and preserves wave-ripples, shrinkage cracks, and sauropod and theropod tracks (Lockley et al., 1999; Fig. 16C,D). In addition, we have observed a small theropod track on a talus boulder assumed to have come from this horizon. The succession can be traced northwestward for approximately 260 m where it pinches out. Overlying the Middle Sandstone is a 3.5 m thick grey-green mudstone succession (Upper Green Mudstone, Fig. 10) that consists of interbedded claystone, calcareous massive bioturbated mudstone with caliche nodules (Fig. 17), and finely laminated claystone and siltstone. The Poison Strip Sandstone sharply overlies this unit (Figs. 9E and 10).

Fig. 17. Possible caliche pebbles (light tone) in Upper Green Mudstone. Scale units are centimeters.

237

6.2. Descriptive petrography Standard petrographic thin sections of bone-rich horizons were examined to assess aspects of the source sediment, kinds and relative timing of diagenetic alteration in the quarry sediments and fossil bones, and paleosol development. Samples from nearby mudrock facies containing dinosaur eggshell, and the overlying calcareous and sandstone deposits were also examined for comparative purposes. Petrographic samples from the bone producing units exhibit massive, contorted, and patchy textures that consist of sand grains, sandy-to-silty claystone, muddy micrite (marl), sparite crystals, and fossil bone, all associated with oxides deposited along grain, crystal, and bone surfaces (Fig. 18). The host claystone is greenish-grey to reddish-pink and birefringent under crossed polars. The assemblage of sand- and silt-size grains is poorly sorted and very mature, consisting of well- to sub-rounded grains of monocrystalline quartz (N50%), chert (N 15%), a small population of polycrystalline quartz, and very rare grains of microcline and unidentified heavy minerals. Intraclasts of rounded to angular sparite, mudstone, and very rare silty micrite are also present and are identified by oxide coatings (Fig. 18A). In situ sparite crystals occur as patches throughout the matrix, in boriginalQ spaces of bone (e.g., haversian canals), and along the outer surfaces of both laminar and trabecular bone. Sparite crystals are often interlaminated with fragmented bone edges where trabecular bone is exposed. In these areas, fragments of trabecular bone appear to be enveloped and floating in a sparite matrix, and separate from the main bone mass (Fig. 18C). In many areas, sparite crystals are etched with jagged or rippled edges and coated with oxide (Fig. 18C,D,E). Opaque, red-brown oxides are common throughout the samples. They coat grains, sparite cleavages, etched and unaltered sparite edges, and internal and external bone surfaces (Fig. 18). Oxides range from coating entire grain margins to occurring as localized and/or partial coatings, are typically excluded between grainto-grain contacts, and are best developed along or adjacent to bone surfaces. Oxides are sometimes interlaminated with etched sparite crystals along bone surfaces (Fig. 18E). In the deeper portions of bone fragments where haversian canals are present, oxides coat the interior walls of the canals and sparite crystals fill the remainder of the internal space (Fig. 18C). In localized patches and horizons, oxides show a well developed horizontally oriented, dendritic or streaky

238

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Fig. 18. Petrographic features of sediments in the Dalton Wells fossil locality area. All scale units in millimeters. (A) Mudrock with poorly sorted assemblage of quartz grains and larger reworked carbonates. Note the streaky oxide staining in the lower one-half of photo. (B) Example of poor sorting and mature clasts in oxide-rich matrix. (C) dExplodedT bone texture in calcite cement. Intact bone extends into the top left of photograph. Exploded fragments to the lower right. (D) Widespread etching of calcite grains (marked by opaque oxides). Note well rounded and large quartz grain at bottom. (E) Etched calcite overlain by laminated oxides. (F) Vague oxide streaks in sample collected near outcrop 4-29-97. Note the possible ped in the lower left.

fabric that is also visible in fresh handsamples (e.g., Figs. 13B and 17A). Petrographic samples from flat, thinly bedded limestones lateral to and above DW are interpreted as lacustrine marls because of their homogenous micritic composition (see below). They contain very sparse trace fossils and thin laminae of very fine sandstone to siltstone grains dominated by quartz. A thin section from a dinosaur eggshell locality (Fig. 1) west of Dalton Wells and 5 m above the Morrison–Cedar Mountain formational contact exhibits rounded, ovalshaped mudstone granules that may be either reworked mudrock clasts or peds (Fig. 18F). Oxide stains are far less common than at the quarry site, but still highlight silt grains and reworked mudstone clasts.

7. Interpretation 7.1. Lower green mudstone (fossiliferous units 1–4) 7.1.1. A debris flow origin The stacked bonebeds and associated facies at Dalton Wells record a complex, multi-event depositional and post-depositional history in a seasonally wet-anddry setting that was characterized by episodic, very high-energy flow events capable of trapping and moving poorly sorted mixtures of clay, sand, pebbles, cobbles and large skeletal elements and partial skeletons. The tabular geometry of these units, vague occurrences of graded pebbles, cobbles and bones, the complete absence of unidirectional, hydraulic-flow-generated sedimentary structures, and, most importantly, the

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

occurrences (though rare) of clay laminae (units 1, 3 and 4; Fig. 11D,E), indicate that the overall massive fabric of these deposits is primary, not the result of dewatering or some form of bioturbation. Tabular–horizontal, flat-based beds of massive sediment that host a matrix-supported and poorly sorted assemblage of granules, pebbles, and cobbles (including bones) are characteristic of high-density, waterladen, debris flows (Allen, 1970; Bull, 1972; Nemec and Steel, 1984, pp. 12–13; Collinson and Thompson, 1982; Vallance and Scott, 1997; Major, 1997, 2000, 2003; Leeder, 1999). Given the absence of any aquatic invertebrate or plant fossils (e.g., ostracods or charophytes), we interpret these as subaerial debris flow deposits. Following Lowe (1979), Nemec and Steel (1984), Collinson and Thompson (1982), Major (2003), and Major and Iverson (1999) we use the term bdebris flowQ in reference to sediment gravity flows that exhibit a plastic rheological behavior. The high mudstone content (Fig. 12) and matrix-supported clasts indicate that flows were cohesive. However, the presence of vague grading, rare claystone laminae, the concentration of large vertebrate elements near the base of each deposit, the predominant horizontal orientation of large skeletal elements, and at least one occurrence (at the base of Unit 1) of a small scale incision suggests that these mudflows were highly saturated and may have occasionally converged on fluidal sediment flow conditions (cf., Nemec and Steel, 1984, Fig. 15). At Dalton Wells, the preponderance of coarse clastics and the co-occurrence of low-density/high surface area bones and high-density/low surface area chert pebbles and cobbles at the base of units 1–4 are incompatible with a gravity settling mechanism explanation for debris flow grading (cf. Fischer and Schminke, 1984; Vallance, 1994), or the induction of grain packing by pore-fluid pressure dissipation (Major, 2000). Instead, the consistent concentration of such large clasts at the base of each deposit more likely resulted from the incorporation of coarse-grained debris into the front of a moving flow, as described by Vallance and Scott (1997) for the cobble and boulder rich Osceola mudflow deposits. In that case, longitudinal size sorting occurred along flow axis as flow advanced. Large blocks and boulders were swept up in front and smaller clasts lagged behind. In this debris-flow wave model (Vallance and Scott, 1997, Fig. 13; cf., Leeder, 1999), deposition is accretionary and does not occur through en masse emplacement of the deposit; coarse-grained debris at the front is deposited first and then overlain by finer sediments as the flow wave advances and attenu-

239

ates. Thus, it is likely that the DW mudflows acted as waves, scouring and assimilating coarse material in their path. Pebbles, bones and cobbles were picked up and incorporated into the base of the advancing mudflow front, and then were subsequently deposited down dip at the bonebed site. As the debris flow wave continued to flow over the bonebed site, finer grained sediments — sourced from the debris flow site of origin — buried the coarser pebbles and bones. Although we recognize four deposits (units 1–4) at DW, Major (1997, 2000, 2003) and Major and Iverson (1999) have pointed out that bounding surfaces between successive accretionary flows may not be preserved or may be modified subsequent to deposition. Accordingly, we cannot rule out the possibility that there were more than four debris-flow events at DW. Modern subaerial debris flows occur in a wide variety of settings (Leeder, 1999). They often result from slope failure in an array of poorly consolidated substrates and can be triggered by rainfall, seismic activity, volcanism, plant denudation (e.g., forest fires), changes in surface and substrate cohesiveness due to altered ground water flow patterns or climate change (Leeder, 1999; Major, 2003). Massive bedding, graded distribution of the coarse component, poor sorting, and absence of current-flow sedimentary structures are typical of modern debris flow deposits in alluvial fans and alluvial environments, especially where there is significant relief or shear is exceeded due to heavy rainfall (e.g., Leeder, 1982). The Dalton Wells debris flow deposits have a significant clay content and we propose that this may have contributed to source-slope instability (cf. Fastovsky et al., 1995). We speculate that the Dalton Wells subaerial debris flows were generated during times of intense seasonal rainfall, and that the flows may have been triggered as the shear strength of the host matrix on a nearby naturally sloping surface was exceeded. The bone concentrates that occur at the base of units 1–4 are very poorly size sorted, show little evidence of preferred orientations, and consist of some elements with very high dip angles. These features are consistent with a debris flow origin. All the elements exhibit evidence for complex, multi-event taphonomic histories that include exposure, trampling, scavenging, and reworking (Britt et al., 2004). These data indicate that the bone assemblage was highly modified prior to the debris flow events. Within the bonebed, locally concentrated remains from single or multiple individuals of one taxon (Gastonia and Utahraptor, Britt et al., 2004) suggest that partially articulated carcasses, or groups of carcasses, may have also been reworked into the Dalton

240

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

Wells site, and that the bonebed may prove to be continuous up-dip with the site of death. In any case, the distance of debris-flow transport was not far. The origins of the claystone laminae near the top of Unit 1 are unclear. These prominent laminae contain matrix-supported sand grains, granules, and small pebbles, and thus, do not appear to have formed from deposition of fines in standing water. These may represent secondarily enriched clay intervals within the deposit, possibly eluviants. Alternatively, Major (2000) has described a process in modern debris flow experiments where pore water pressure dissipates after a flow is immobilized. Dissipation is accompanied by a localized increase in grain packing. The co-occurrence of clay laminae and coarser grains may reflect pore pressure dissipation and increased grain packing as upward flowing pore waters provided localized microenvironments in which clays could settle. In contrast, clay laminae in the upper portions of units 3 and 4 are devoid of clasts and are part of regularly repeating dcmthick successions of clay and sandstone (Fig. 11D). These are indicative of deposition from suspension in standing water, following minor flow pulses, and indicate that the area remained submerged or that the area became submerged between the debris flow events. Root structures are completely absent in the host matrix and although there is a patchy texture, there are no clear indications of in situ peds, slickensides, or cutans. It is possible, however, that reworked oval mudstone grains (sometimes highlighted by oxide staining) are reworked peds. If entisols were developing nearby, due to repeated seasonal wetting and drying events, flooding or bioturbation may have easily reworked these fragmented mudstone surfaces. The apparent absence of paleosols and roots, and the rarity of invertebrate traces in units 1–3, and any evidence of subaerial exposure (e.g., shrinkage cracks) throughout units 1–4 suggest that little time passed between flow events and that this area probably remained saturated between flood events. This interpretation is consistent with the distribution of clay laminae discussed above, and evidence for in situ trampling of some of the elements (Britt et al., 2004). Thus, we propose that these flows were deposited during a single dwetT season, or that the area remained saturated over the course of multiple seasons. In contrast, the presence of well-developed invertebrate traces throughout the much thicker Unit 4 and the occurrence of interlaminated sandstone and claystone throughout its upper one-half indicate that sediments in that unit may have been deposited episodically and possibly during multiple seasons. The presence of crustacean galleries in the

top of Unit 4 indicates that there was possibly a significant time gap between the deposition of this unit and the overlying sandstone, and that the upper portions of Unit 4 were subaerially exposed. 7.1.2. Interpretive petrography The petrologic maturity and high degree of rounding of the sand and silt-sized grains plus the occasional presence of reworked dgreenT mudstone clasts (e.g., Fig. 11C) in the host matrix is indicative of a highly reworked and/or weathered sediment source. The presence of sub- to well-rounded silt-sized grains in alluvial-to-paludal deposits is generally unusual in alluvial systems and particularly in foreland basin settings (cf., Eberth and Hamblin, 1993). Thus, it is unlikely that such mature and well-rounded sediments were derived from a newly uplifted orogen. Given that these sediments lie just above the Morrison/Cedar Mountain unconformity, we propose that they were most likely reworked from more proximal exposures of the Morrison Formation. The pattern of oxide coatings and sparite crystal growth and dissolution is compatible with an interpretation that this bone assemblage and its host sediments were diagenetically altered. Sparite precipitation could only proceed under relatively high pH conditions, whereas sparite etching and oxide precipitation indicate lower pH conditions. The ubiquitous opaque oxide coatings on the external surface and especially within the haversian canals of the fossil bones and the high concentration of oxide in the sediments that surround the fossil bone suggest that buried bone may have played an important role in creating a favorable chemical environment for the precipitation of oxides. This is further supported by examination of thin sections from adjacent sites that yield sparse egg shell fragments and very little bone. In these sections oxide coatings are poorly developed. Alternating low and high pH conditions are indicated by the presence of etched sparite crystals with oxide coatings, that are, in turn, coated by smaller sparite crystals. Although the exact mechanism whereby these alternating chemical conditions were established is unknown, episodic changes in the pH of the groundwater and chemical conditions may have been driven by any number of mechanisms including: seasonal changes in moisture, episodic death or feeding events resulting in the addition of more organic bdetritusQ at this locality, seasonal or episodic changes in groundwater flow related to changing climate, geodynamically-influenced hydrology, or combinations of these. The generally excellent condition of the fossil bone in thin section suggests that the changes in pH indicated

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

by the etched calcite crystals were not extreme. Thus, the relatively lower pH conditions that enhanced the precipitation of oxides were not excessive. Fragments of trabecular bone that appear as dfloatT in calcite crystals indicate that sparite precipitation and crystal growth broke and moved fragments of bone, especially where trabecular bone was exposed to the matrix after deposition. In turn, this explains why matrix does not easily separate from some specimens during preparation. We interpret the concretionary horizons in units 1–4 in Dalton Wells as diagenetic calcretes resulting from groundwater precipitation rather than palesol development. Precipitation was likely enhanced by the presence of organics (bones), lacustrine carbonates, and other early diagenetic precipitates. The most intensely cemented zones correlate with the bone- and largeclast-rich lower portions of units 1–4, which would have been the most permeable portions of the debris flow deposits. The complete absence of reworked calcrete clasts and presence of burial compaction effects in Dalton Wells bones confirm that the concretionary development was not syndepositional, but occurred after significant burial. We reject the interpretation that these concretionary deposits represent paleosols or caliches (e.g., Currie, 1998) based on: (1) the absence of pedogenic features; (2) the horizontally-oriented streaky fabric of the oxides visible both micro and macroscopically that indicate that precipitation occurred during horizontal migration of groundwater through the sediments; (3) the laterally discontinuous and patchy development of the concretion horizons is more characteristic of groundwater precipitates than laterally continuous caliche horizons; (4) the preferential development of concretions in association with vertebrate fossils and beds of marl and micrite (e.g., Gaston Quarry of Kirkland et al., 1999) showing that precipitation was controlled by subsurface chemical environment rather than climate and evapo-transpirative processes; and (5) the presence of etched calcite crystals in contact with massive-to-laminated oxide deposits, that together provide clear evidence for a multi-phase precipitation history under variable pH conditions. We conclude that the calcretes associated with the bonebeds most likely developed as groundwater flow patterns were established in response to the regional tectonic reorganization. The position of the concretionary unit low in the Yellow Cat Member section, but above the Morrison Formation, suggests that the Yellow Cat strata were probably more permeable and porous than underlying Morrison strata, and that the ground water cells that were established as the foreland basin developed, favored passing through this horizon. In this

241

context, the diagenetic concretionary development has limited value in determining the original depositional setting of the bonebed. We suggest that any chronostratigraphic data derived from these calcretes will provide only a minimum age for the bonebed. 7.1.3. Interpretation of extrabasinal clasts Although extrabasinal pebbles and cobbles, up to 7 cm in maximum diameter, are quite common at DW, our search for pebbly mudstones at other locations and sections in the field area indicates that clasts of this size and composition are rare at the base of the Cedar Mountain Formation. Instead, red and green intraformational calcareous granules and small extraformational pebbles (b3 cm maximum diameter) are far more typical. In contrast, polished, extra-basinal chert pebbles and cobbles are common in the higher stratigraphic portions of the overlying (and interfingering) Poison Strip Sandstone, and were obviously transported into the area by turbulent alluvial currents and, possibly, by hyperconcentrated flows (cf., Zaleha and Wiesemann, 2005). Our stratigraphic and sedimentologic data from the Dalton Wells area shows that Yellow Cat Member paleochannels were small and that their fill is dominated by locally reworked and mature sediment, similar to the pattern seen in the underlying Brushy Basin Member of the Morrison Formation. Furthermore, large pebbles to cobbles are also very rare in the lower sandstones of the Poison Strip Sandstone that interfinger with the Yellow Cat strata west of the bonebeds. Thus, we regard it as a reasonable hypothesis, following Stokes (1952), that some of the extraformational pebbles within the DW bonebeds represent reworked and locally concentrated gastroliths sourced from the sauropods whose remains were present in the area. A sauropod source for the extrabasinal cobble and pebbles is further supported by the report of gastroliths within a skeleton of the brachiosaurid sauropod, Cedarosaurus, in the Yellow Cat strata of Utah (Saunders et al., 2001). Gastroliths in that specimen are composed predominantly of chert and range in size from small pebbles to cobbles. 7.2. Middle sandstone and upper green mudstone Although units 1–3 and the lower portion of Unit 4 represent a stacked succession of debris flow deposits, the upper one-half of Unit 4 and the remainder of the Yellow Cat Member at Dalton Wells record the longterm presence of a paludal–lacustrine paleoenvironments in this area. The two-meter-thick Middle Sandstone (Fig. 10, 4–6 m) is interpreted as a marginal-to-

242

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

very-shallow, ephemeral lacustrine succession that was the site of: ! unconfined ephemeral sheet flooding (tabular, nonchanneled geometry; planar stratification; graded bedding; sharp changes in grain-size); ! shallow standing water (wave ripples; marl; trace fossils); and ! subaerial exposure (shrinkage cracks, dinosaur tracks, dinoturbation). Open lacustrine conditions became established at the DW site following the deposition of the Middle Sandstone, and resulted in the deposition of a nearly 1 m thick package of finely laminated lacustrine muds and marls (Fig. 10, Upper Green Mudstone, 6–7 m). These deposits were similar to the lacustrine deposits in the underlying Brushy Basin Member of the Morrison Formation. The remaining mudstone succession (Fig. 10, Upper Green Mudstone, 7–9.5 m) is interpreted as a mixed paludal–lacustrine setting (tabular thin beds of marl, interlaminated siltstone and claystone with caliches [Fig. 17], sandy-to-granule-rich siltstone; invertebrate feeding traces and galleries) that experienced variations in clastic supply, water depth and subaerial exposure, and, possibly, salinity. Granulitic-to-cobbly, alluvial channel deposits of the Poison Strip Sandstone unconformably overlie this unit. 8. Discussion There are only a handful of sites where vertebrate fossils are interpreted as occurring in association with well documented subaerial debris flow deposits (Andrews and Ersoy, 1990; Andrews and Alpagut, 1990; Fastovsky et al., 1995; Loope et al., 1998, 1999; Schmitt et al., 1998; Eberth et al., 2000; Rogers, 2005) and, thus, Dalton Wells represents an important example of a rarely documented preservational mode for bonebeds. The presence of debris flow deposits at the base of the Cedar Mountain Formation suggests that significant topographic relief existed in this area or nearby during the tectonic transformation of the region to a foreland basin. Because of the ubiquitous occurrence of grossly similar deposits of massive, pebbly-toconglomeratic mudstones at the base of the Cedar Mountain Formation across eastern Utah into western Colorado (e.g., Kirkland et al., 1997), we propose that subaerial exposure and erosion of a migrating forebulge (cf. Currie, 1998) is the most likely mechanism whereby relief was created both within the Dalton Wells area and regionally. We regard it as likely that this topo-

graphic relief was further amplified in the Dalton Wells area by either local salt-diapir-related faulting (e.g., Doelling, 1988; Johnson and Aubrey, 1994; Aubrey, 1996) or some other mechanism that increased local rates of subsidence. In this context, Dalton Wells represents an unusual example of a bonebed whose origins were partially mediated by tectonics. Whereas long bone orientation data from Dalton Wells are equivocal as to the direction of flow (see above), five paleocurrent measurements from medium-scale trough cross-beds in the Middle Sandstone Unit immediately above the bonebeds (Fig. 15B) show consistent, eastward-oriented flow directions. If relief was generated by uplift associated with forebulge migration, the location of the Dalton Wells area must have been on the eastern flank of the advancing forebulge, and thus, in the backbulge (cf. Currie, 1998). The paleogeographic location of Dalton Wells updip and west of an ephemeral lake system may account for the abundance of terrestrial vertebrates at the site. We speculate that terrestrial vertebrates — dinosaurs in particular — frequented this setting in search of water and food. Taphonomic analysis (Britt et al., 2004) indicates that the Dalton Wells elements were reworked from a nearby up-dip location where carcasses of sauropod, ornithopod, ankylosaurid and theropod dinosaurs accumulated, decomposed, and were scavenged and trampled prior to debris flow reworking. Although we do not know under what specific circumstances animals died and carcasses accumulated, the stratigraphic association of lacustrine and shoreline facies in both the underlying Brushy Basin and overlying Yellow Cat members indicates that lake levels rose and fell through time in this area. In this context, we also speculate that the assemblage most likely represents the remains of animals that died in a variety of dwater-hole-relatedT mortality events that may have included drought, disease, predation, and other forms of attritional mortality. Bones accumulated on a landscape that consisted of Morrison Formation strata but were not preserved until (1) some were reworked by debris flows and (2) accommodation was created allowing for the preservation of the deposits. In the seasonal wet-and-dry climate that characterized this region during the Barremian, debris flows may have been triggered by intense rainfall or seismic events. 9. Conclusions 1) The Dalton Wells site consists of a stacked succession of four bonebeds in a two-meter-thick stratigraphic interval that occurs at the base of the Yellow

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

2)

3)

4)

5)

Cat Member of the Cedar Mountain Formation on the regionally-expressed Jurassic–Cretaceous unconformity. The deposits and associated fossils are tentatively regarded as Barremian. Bonebed sediments were deposited by subaerial debris flows (cohesive mudflows). The flows reworked Morrison and Cedar Mountain sediments, and swept Early Cretaceous vertebrate remains into the Dalton Wells site. Distance of transport for the bones was not far. Debris flows may have been triggered by intense rainfall or seismic events. At least four debris flow events are recognized. They all may have occurred during a single season or year. Documented examples of bonebeds hosted by debris flows are rare, thus, Dalton Wells represents an important addition to our knowledge about this preservational mode. Bonebeds were most likely deposited in a backbulge setting characterized by alluvial–lacustrine paleoenvironments and a warm-to-hot, seasonally wet-anddry climate. The landscape relief necessary to generate debris flows was probably created initially by forebulge uplift associated with early stages of foreland basin tectonics. Slope was probably maintained or amplified by high rates of local subsidence — possibly salt-diapir-related faulting — during the Tithonian and Barremian–Aptian. Calcretes in the bonebed host matrix originated through diagenesis, not paleosol development. Precipitation most likely occurred as new groundwater flow patterns were established in response to foreland basin development. Calcrete development may have been enhanced by the presence of organics (bones) and lacustrine carbonates. Although it is unknown what killed the animals in the DW assemblage prior to being reworked by debris flows, drought-induced mortality events are a likely cause given the evidence for a seasonally wet-and-dry climate in the Yellow Cat Member of the Cedar Mountain Formation.

Acknowledgements DAE and DBB thank the Royal Tyrrell Museum and RTM Cooperating Society for financial support during this research. DAE thanks Cindy Britt for her hospitality during numerous visits to Utah. Portions of this study were funded by a Dinosaur Society grant to BBB and RDS, who were also supported by the Museum of Western Colorado during the early stages of this project. BBB and KLS thank Brigham Young University’s College of Physical and Mathematical

243

Sciences for financial support of this study. David Tingey of the BYU Geology Dept. ran XRD analyses. Jim Kirkland provided encouragement and numerous insights during the long gestation of this project. We thank Ray Rogers and an anonymous reviewer for their thoughtful and helpful critiques of an earlier draft. References Allen, J.R.L., 1970. Physical Processes of Sedimentation. George Allen and Unwin, London. 248 pp. Andrews, P., Alpagut, B., 1990. Description of the fossiliferous units at Pasalar, Turkey. Journal of Human Evolution 19, 343 – 361. Andrews, P., Ersoy, A., 1990. Taphonomy of the Miocene bone accumulations at Pasalar, Turkey. Journal of Human Evolution 19, 379 – 396. Aubrey, W.M., 1996. Stratigraphic architecture and deformational history of Early Cretaceous foreland basin, eastern Utah and southwestern Colorado. In: Huffman Jr., A.C., Lund, W.R., Godwin, L.H. (Eds.), Geology and Resources of the Paradox Basin, Utah Geological Association Guidebook, vol. 25, pp. 211 – 220. Aubrey, W.M., 1998. A newly discovered, widespread fluvial facies and unconformity marking the Upper Jurassic/Lower Cretaceous boundary, Colorado Plateau. Modern Geology 22, 209 – 233. Behrensmeyer, A.K., 1990. Transport-hydrodynamics: bones. In: Briggs, D.E.G., Crowther, P.R. (Eds.), Paleobiology: A Synthesis. Blackwell Scientific Publications, Oxford, pp. 232 – 235. Bonaparte, J.F., Heinrich, W.-D., Wild, R., 2002. Review of Janenschia Wild, with the description of a new sauropod from the Tendaguru beds of Tanzania and a discussion of the systematic value of procoelous caudal vertebrae in the Sauropoda. Palaeontographica A 256, 25 – 76. Britt, B.B., Stadtman, K.L., Scheetz, R.D., 1996. The Early Cretaceous Dalton Wells dinosaur fauna and the earliest North American titanosaurid sauropod. Journal of Vertebrate Paleontology 16, 24A (supplement to #3). Britt, B.B., Eberth, D.A., Brinkman, D.B., Scheetz, R., Stadtman, K.L., Mcintosh, J.S., 1997. The Dalton Wells fauna (Cedar Mountain Formation): a rare window into the little-known world of Early Cretaceous dinosaurs in North America. GSA Annual Meeting, Salt Lake City, Abstracts with Programs, A-104. Britt, B.B., Eberth, D., Scheetz, R., Greenhalgh, B., 2004. Taphonomy of the Dalton Wells dinosaur quarry (Cedar Mountain Formation, Lower Cretaceous, Utah). Journal of Vertebrate Paleontology 24, 41A (supplement to #3). Bull, W.B., 1972. Recognition of alluvial-fan deposits in the stratigraphic record. In: Rigby, J.K., Hamblin, W.K. (Eds.), Recognition of Ancient Sedimentary Environments. SEPM Spec. Pub. 16, 63 – 83. Collinson, J.D., Thompson, D.B., 1982. Sedimentary Structures. Allen and Unwin, London. 207 pp. Craig, L.C., 1981. Lower Cretaceous rocks, Southwestern Colorado and Southeastern Utah: Rocky Mountain Association of Geologists, 1981 Field Conference, pp. 195 – 200. Currie, B.S., 1997. Sequence stratigraphy of nonmarine Jurassic– Cretaceous rocks, central Cordilleran foreland-basin system: Geological Society of America. Bulletin 109, 1206 – 1222. Currie, B.S., 1998. Upper Jurassic–Lower Cretaceous Morrison and Cedar Mountain formations, NE Utah–NW Colorado: relation-

244

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245

ships between nonmarine deposition and early Cordilleran foreland-basin development. Journal of Sedimentary Research 68, 632 – 652. Demko, T.M., Parrish, J.T., 1998. Paleoclimatic setting of the Upper Jurassic Morrison Formation. Modern Geology 22, 283 – 296. Doelling, H.H., 1988. Geology of the Salt Valley anticline and Arches National Park. In: Doelling, H.H., Oviatt, C.G., Huntoon, P.W. (Eds.), Salt Deformation in the Paradox Region, Utah Geological and Mineral Survey, Bulletin, vol. 122, pp. 1 – 60. Eberth, D.A., Hamblin, A.P., 1993. Tectonic, stratigraphic, and sedimentologic significance of a regional discontinuity in the Upper Judith River Group (Belly River Wedge) of southern Alberta, Saskatchewan, and northern Montana. Canadian Journal of Earth Sciences 30, 174 – 200. Eberth, D.A., Britt, B.B., Brinkman, 1997. Sedimentology, facies architecture, and stratigraphy of the Cedar Mountain Formation (Lower Cretaceous) at Dalton Wells Quarry, eastern Utah, USA. GSA Annual Meeting, Salt Lake City, Abstracts with Programs, A-113. Eberth, D.A., Berman, D.S., Sumida, S.S., Hopf, H., 2000. Lower Permian terrestrial paleoenvironments and vertebrate paleoecology of the Tambach Basin (Thuringia, central Germany). Palaios 15, 293 – 313. Eberth, D.A., Brinkman, D.B., Chen, P.J., Yuan, F.T., Wu, S.Z., Gang, L., Cheng, X.S., 2001. Sequence stratigraphy, paleoclimate patterns and vertebrate fossil preservation in Jurassic–Cretaceous strata of the Junggar basin, Xinjiang Autonomous Region, People’s Republic China. Canadian Journal of Earth Sciences 38, 1627 – 1644. Fastovsky, D.E., Clark, J.M., Strater, N.H., Montellano, M., R. Hernandez, R., Hopson, J.A., 1995. Depositional environments of a Middle Jurassic terrestrial vertebrate assemblage, Huizachal Canyon, Mexico. Journal of Vertebrate Paleontology 15, 561 – 575. Fischer, R.V., Schminke, H.-U., 1984. Pyroclastic Rocks. SpringerVerlag, Berlin. 472 pp. Galton, P.M., Jensen, J.A., 1978. Remains of ornithopod dinosaurs from the Lower Cretaceous of North America. Brigham Young University Geology Studies 25 (pt III), 1 – 10. Hasiotis, S.T., 2002. Continental Trace Fossils. SEPM Short Course Notes, #51. Johnson, B., Aubrey, W.M., 1994. Stratigraphic evidence for Early Cretaceous normal faulting over the Cache Valley salt structure, Paradox Basin, Utah. American Association of Petroleum Geologists, Program with Abstracts. Kirkland, J.I., Britt, B.B., Burge, D.L., Carpenter, K., Cifelli, R.L., DeCourten, F.L., Eaton, J.G., Hasiotis, S., Lawton, T.F., 1997. Lower to Middle Cretaceous dinosaur faunas of the central Colorado Plateau; a key to understanding 35 million years of tectonics, evolution, and biogeography. Brigham Young University Geological Studies 42 (pt II), 69 – 103. Kirkland, J.I., Cifelli, R.L., Britt, B.B., Burge, D.L., DeCourten, F.L., Eaton, J.G., Parrish, J.M., 1999. Distribution of vertebrate faunas in the Cedar Mountain Formation, east-central Utah. In: Gillette, D.D. (Ed.), Vertebrate Paleontology in Utah, Utah Geological Survey, Miscellaneous Publication, vol. 99-1, pp. 201 – 242. Kowallis, B., Deino, A.R., Peterson, F., Turner, C., 1998. High precision radiometric dating of the Morrison Formation. Modern Geology 22, 2356 – 2360. Leeder, M.R., 1982. Sedimentology. George Allen and Unwin, London. 344 pp. Leeder, M., 1999. Sedimentology and Sedimentary Basins. Blackwell Scientific Ltd., Oxford. 592 pp.

Lockley, M.G., Kirkland, J.I., DeCourten, F.L., Britt, B.B., Hasiotis, S.T., 1999. Dinosaur tracks from the Cedar Mountain Formation of eastern Utah: a preliminary report. In: Gillette, D.D. (Ed.), Vertebrate Paleontology in Utah, Utah Geological Survey, Miscellaneous Publication, vol. 99-1, pp. 253 – 257. Loope, D.B., Dingus, L., Swisher III, C.C., Minjin, C., 1998. Life and death in a Late Cretaceous dune field, Nemegt basin, Mongolia. Geology 26, 27 – 30. Loope, D.B., Mason, J.A., Dingus, L., 1999. Lethal sandslides from eolian dunes. The Journal of Geology 107, 707 – 713. Louthan, B., 1993. Dalton Wells CCC camp. Canyon Legacy 19, 5 – 14. Lowe, D.R., 1979. Sediment gravity flows: their classification and some problems of application to natural flows and deposits. In: Doyle, L.J., Pilkey, O.H. (Eds.), Geology of Continental Slopes. SEPM Spec. Pub. 27, 75 – 82. MacDonald, D.I.M., Jefferson, T.H., 1985. Orientation studies of waterlogged wood: a paleocurrent indicator? Journal of Sedimentary Petrology 55, 235 – 239. Major, J.J., 1997. Depositional processes in large-scale debris-flow experiments. The Journal of Geology 105, 345 – 366. Major, J.J., 2000. Gravity-driven consolidation of granular slurries— implications for debris-flow deposition and deposit characteristics. Journal of Sedimentary Geology 70, 64 – 83. Major, J.J., 2003. Debris flow. In: Middleton, G.V. (Ed.), Encyclopedia of Sediments and Sedimentary Rocks. Kluwer Academic Publishers, Dordrecht, pp. 186 – 188. Major, J.J., Iverson, R.M., 1999. Debris-flow deposition: effects of pore-fluid pressure and friction concentrated at flow margins. Geological Society of America Bulletin 111, 1424 – 1434. Masura, J.E., Davis, L.E., 1996. Stratigraphy and sedimentology of the Upper Jurassic Brushy Basin Member of the Morrison Formation near Tidwell Wash, Emery County, Utah, and description of the depositional environment. In: Morales, M. (Ed.), The Continental Jurassic, Museum of Northern Arizona Bulletin, vol. 60, p. 505. Morris, T.H., Richmond, D.R., Grimshaw, S.D., 1996. Orientation of dinosaur bones in riverine environments: insights into sedimentary dynamics and taphonomy. In: Morales, M. (Ed.), The Continental Jurassic. Museum of Northern Arizona Bulletin 60, 521 – 530. Nemec, W., Steel, R.J., 1984. Alluvial and coastal conglomerates: their significant features and some comments on gravelly massflow deposits. In: Koster, E.H., Steel, R. (Eds.), Sedimentology of Gravels and Conglomerates, Canadian Society of Petroleum Geologists, Memoir, vol. 10, pp. 1 – 31. Nolte, M.J., Greenhalgh, B.W., Dangerfield, A., Scheetz, R.D., Britt, B.B., 2004. Invertebrate burrows on dinosaur bones from the Lower Cretaceous Cedar Mountain Formation near Moab, Utah, U.S.A. Geological Society of America, Abstracts with Programs, vol. 36, p. 379. Rogers, R.R., 2005. Fine-grained debris flows and extraordinary vertebrate burials in the Late Cretaceous of Madagascar. Geology 33, 297–300. Saunders, F., Manley, K., Carpenter, K., 2001. Gastroliths from the Lower Cretaceous sauropod Cedarosaurus weiskopfae. In: Tanke, D., Carpenter, K. (Eds.), Mesozoic Vertebrate Life. University of Indiana Press, pp. 166 – 180. Schmitt, J.G., Horner, J.R., Laws, R.R., Horner, J.R., 1998. Debrisflow deposition of a hadrosaur-bearing bone bed, Upper Cretaceous Two Medicine Formation, northwest Montana. Journal of Vertebrate Paleontology 18, 76A (supplement to #3).

D.A. Eberth et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 236 (2006) 217–245 Shanley, K.W., McCabe, P.J., 1994. Perspectives on the sequence stratigraphy of continental strata. American Association of Petroleum Geologists Bulletin 78, 544 – 568. Shanley, K.W., McCabe, P.J., 1995. Sequence stratigraphy of Turonian–Santonian strata, Kaiparowits Plateau, southern Utah, U.S.A.: implications for regional correlation and foreland basin evolution. In: Van Wagoner, J.C., Bertram, G.T. (Eds.), Sequence Stratigraphy of Foreland Basin Deposits, American Association of Petroleum Geologists, Memoir, vol. 64, pp. 103 – 136. Shanley, K.W., McCabe, P.J., 1998. Relative role of eustasy, climate, and tectonism in continental rocks: an introduction. In: Shanley, K.W., McCabe, P.J. (Eds.), Relative Role of Eustasy, Climate, and Tectonism in Continental Rocks, Society for Sedimentary Petrology, Special Publication, vol. 59, pp. iii – iv. Stokes, W.L., 1952. Lower Cretaceous in Colorado Plateau. American Association of Petroleum Geologists Bulletin 36, 1766 – 1776.

245

Vallance, J.W., 1994. Experimental and field studies related to the behavior of granular mass flows and the characteristics of their deposits. Ph.D. Dissertation. Michigan Technological University, Houghton, Michigan. 197 pp. Vallance, J.W., Scott, K.M., 1997. The Osceola Mudflow from Mount Rainier: Sedimentology and hazard implications of a huge clayrich debris flow. Geological Society of America Bulletin 109, 143 – 163. Voorhies, M., 1969. Taphonomy and population dynamics of an Early Pliocene vertebrate fauna, Knox County, Nebraska. University of Wyoming Contributions to Geology, Special Paper No 1, pp. 1 – 69. Zaleha, M.J., Wiesemann, S.A., 2005. Hyperconcentrated flows and gastroliths: sedimentology of diamictites and wackes of the Upper Cloverly Formation, Lower Cretaceous, Wyoming, U.S.A. Journal of Sedimentary Research 75, 43 – 54.