6.1. Photographs of various taphonomic states. (A) Isolated limb bone. (B) Articulated limb bones. Pocket knife is approximately 7 cm long. (C) Tortoise (T) resting on former soil surface (S) and buried by fining upward flood deposits. Note photo scale next to shell. (D) Gnaw marks (G) on underside of a jaw bone. (E) Scratch marks (M) on a limb bone. (F) A fragment of tortoise shell that was heavily weathered before burial and fossilization. (D), (E), (F), scale in centimeters. Photos by the authors.
PALEOECOLOGY IS THE STUDY OF THE INTERACTIONS, habits, and lifestyles of extinct species and the ancient communities they formed. Paleoecologists, using data from the fossil and geologic record, reconstruct these communities of plants and animals to gain a better understanding of the relationship between members of that community (Shipman, 1981) and how they responded to environmental changes. Before the paleoecology of an assemblage can be interpreted, the paleoecologist must gain a better understanding of the events that intervened between death and fossilization and what effects these events have on the retrieval of information about the past (Shipman, 1981). This process is called taphonomy. Efremov (1940) introduced the concept of taphonomy as the study of embedding or burial. It is the analysis of the transition of organics from the biosphere into the lithosphere to become part of the geologic record. The term taphonomy can be broken down into the Greek words taphos (burial) and nomos (laws).
Clark, Beerbower, and Kietzke (1967) made the first attempt at describing the ancient environments preserved in the White River Badlands. They proposed four different environments of deposition within the Chadron Formation. These included streams represented by channel fills of the Red River and its tributaries; limestone deposits near channels and channel margin facies; channel margin sands and silts from riverbanks; and massive to bedded clays and silts along ancient floodplains. Clark, Beerbower, and Kietzke (1967) proposed four general biotic structures, including aquatic, semiaquatic, river-border forest, and forest–savanna. They divided the paleoecology of the lower nodular zone of the Brule Formation into three different facies: near stream, open plains, and swampy plains. (See chapter 1 for a more detailed description on Clark and his colleagues’ research contributions.) Refinements to geochronology (such as paleomagnetism and radioisotopic dating), the application of paleopedology to deep time records, and an expanded paleobotanical archive have since fine-tuned paleoclimatic interpretations of the Cenozoic and refined our understanding of the paleoecological changes indicated by the fossils in the White River Badlands.
During the uplift of the Black Hills, which began during the later phases of the Laramide Orogeny, Paleozoic rocks were exposed to weathering, and the resulting sediments were transported by streams and rivers flowing from west to east. These sediments are preserved within a slightly northwest–southeast-trending asymmetric basin that makes up much of what we call the White River Badlands (Fig. 4.2). The basin is bound to the north by the Sage Ridge Fault in the Badlands Wilderness Area and to the south along the Sandoz Ranch, White Clay, and Pine Ridge fault system (Clark, Beerbower, and Kietzke, 1967). A significant amount of White River Badlands sediment consists of reworked ash, which was carried via eolian and fluvial processes from volcanoes in eastern Nevada and Colorado. This combination of geologic factors promoted rapid rates of sedimentation, which aided in the fossilization process, allowing carcasses to be buried quickly after death and protecting them from destructive forces (Clark, Beerbower, and Kietzke, 1967; Shipman, 1981).
The fossilization process is a rare event and requires a series of steps to occur. The preservation potential of a fossil is the balance between the factors of preservation and destruction. Chances of preservation tend to increase if organic remains can be buried rapidly in a particular sedimentary environment (Shipman, 1981). Some of the most frequent causes of death associated with White River vertebrates include predation, flood events, starvation, and dehydration.
Regardless of the cause of death, the key ingredient in the process of fossilization is water. Once buried, and with the flesh removed, bones begin to interact with either water that is percolating downward through the soil zone (vadose water), or if the depth of burial is sufficiently deep, the bones are completely saturated and interact with groundwater (phreatic water). Most bones will experience both as they are progressively buried with additional sediment. The bone material, which is a fine-grained carbonate-substituted bioapatite (dahllite: Ca5(PO4,CO3)3(OH)), reacts with water, and it begins to recrystallize into a more stable form of apatite (francolite: Ca5(PO4,CO3)3(F)) by incorporating fluoride, uranium, and rare earth elements such as lanthanum (Metzger, Terry, and Grandstaff, 2004; Grandstaff and Terry, 2009). The proportions of these newly incorporated elements vary as a function of environmental conditions (oxidizing versus reducing and acidic versus alkaline), and can be used to aid paleoenvironmental interpretations and decipher the rate at which fossilization occurs (Grandstaff and Terry, 2009; Drewicz et al., 2011; Drewicz, 2012). In addition to the recrystallization of bone material, additional minerals, such as calcite and quartz, will precipitate into void spaces in the bone over time.
Recent National Park Service paleontological surveys have documented high concentrations of bone, often referred to as bone beds, in distinct stratigraphic layers and regions of the park. It appears that the majority of these bone beds occur in the Scenic Member of the Brule Formation. Within the Poleslide Member, bones can be found throughout the section with increased concentrations in certain layers. The bone bed accumulations of the Scenic Member represent both watering holes and fluvial channel deposits, whereas higher concentrations of bone in the Poleslide Member are associated with slower rates of eolian sedimentation and more likely are time-averaged samples that took a longer period of time to form (Benton et al., 2007, 2009). By adding information gained from paleosols associated with these bone beds, taphonomic models can be developed that provide insight into the paleoecological dynamics of the White River Badlands. Below we discuss two significant fossil sites in which taphonomic mechanisms have been studied in great detail. Both sites are located within the Scenic Member of the Brule Formation.
The characteristics of vertebrate taphonomy within the Badlands change as a function of paleoclimates throughout the Eocene and into the Oligocene, which in turn influenced the types of paleoenvironments, and thus the depositional processes that were possible. The Chadron Formation and the lower part of the Scenic Member of the Brule Formation were dominated by fluvial depositional processes. Overbank flooding and the migration of river channels was common. Episodic flooding events washed fine sediments onto broad, flat landscapes that buried the bones of animals that had been killed and scattered by predators and scavengers and occasionally trapped slower animals or those that were caught unaware.
When we examine the relationship between individual paleosol profiles and vertebrate fossils, a repetitive pattern can be detected. Fossil bones within the floodplain mudstones tend to be resting on the surface of individual paleosols (Fig. 3.5). Whether the fossil is an isolated bone or an articulated skeleton, roots are commonly seen directly beneath the fossilized remains (Figs. 3.5, 6.1). With each successive flooding event, the previous landscape would be buried, thus preserving the bones. This repetitive nature of flooding and burial explains the overall richness of vertebrate fossils within the Badlands, but it also explains why these strata are so rich in fossil tortoises: they were unable to escape the floodwaters and were easily buried by the muds and sediments that fell out of suspension onto the ancient floodplains (Fig. 6.1). On occasion, animals would also be carried by the floodwaters and end up on the banks of the rivers, similar to what is seen today on the Serengeti Plains of Africa when animals drown while attempting to cross rivers during migration. Carcasses are washed downstream and end up collecting in areas of slower water. The likelihood of the skeleton being preserved depends on the amount of time between death and burial and whether burial occurs before scavengers scatter the remains.
The type of paleosol associated with fossil bones is a clue to the relative degree of landscape stability and can be directly associated with the proximity to active river channels. Flood frequency is related to flood size. The larger the flood, the less frequent it tends to be. In modern terms we classify this relationship as 1-, 5-, 10-, 100-, and 500-year flooding events. Regardless of flood magnitude, though, areas directly adjacent to the river will always flood, and new sediments will be deposited and bury any carcasses that have accumulated on the landscape since the last flood.
In soil terms, this relationship between the magnitude of flooding and the amount of land that will be covered translates to the amount of time available for a soil to form before it is buried by new sediments. Areas closer to active rivers are characterized by soils that are weakly developed, commonly entisols, which are stacked one on top of the other as a result of frequent burial (Fig. 3.5). Farther away from the active channel, soils will not be covered as frequently by floods and new sediments, and thus they will be able to develop to a greater degree (e.g., alfisols, mollisols). Any sediment that is added tends to be in thin increments that can be easily incorporated into the active soil. Only massive flooding events will bury these soils with enough sediment to shut off any further soil development.
This relationship between proximity to an active channel and the relative degree of soil formation is also seen in the strata of the Badlands and explains why fossils tend to be concentrated within certain parts of the park. Within the lower part of the Scenic Member a definite change in the style of fluvial sedimentation and degree of paleosol development can be seen from the southwest part of the park near Scenic, South Dakota, toward the northeast into Dillon Pass. Near Scenic, the paleosols are predominantly entisols and inceptisols, and they commonly have articulated to semiarticulated skeletons resting on ancient land surfaces (cf. Fig. 3.5 and Fig. 3.9). These two features suggest a proximal position to former rivers. Moving northeast toward Dillon Pass within this same interval of strata, the paleosols become progressively better developed overall with a greater proportion of inceptisols. Fossils are still preserved at the top of individual soils, but the numbers of fossils preserved and the style of fossil preservation suggest longer periods of time between flooding events (Fig. 3.8). Bones are more commonly scattered, as if modified by scavengers, and tortoises occur in numerous orientations (including upside down and lying on their side), suggesting that they were removed from their original life position after death.
Flooding is only one part of the story. As rivers flow, they migrate across their floodplain via lateral migration: the progressive erosion of the outside bend (cut bank) in a river meander and deposition on the inside of the meander (point bar). As lateral migration proceeds, any bones lying on the active landscape or buried by recent flooding are secondarily incorporated into the channel as the riverbank is progressively eroded. Because of their size and relative density, bones tend to collect within the river channel along with the coarser sediments such as sand or gravel. Over time, large amounts of bone material can be concentrated and buried in the river channel. This process, referred to as time averaging, creates jumbled deposits of individual bones from numerous types of animals that did not necessarily live either in the same habitat or at the same time. Paleosols will not be directly associated with these ancient river channels, although they may overprint the entire channel deposit if it is abandoned by the river as it shifts the position of its channel (avulsion). Otherwise, continued lateral migration will deposit finer channel materials on top of the bones, followed by overbank muds deposited during the next flooding event. These muds can in turn be modified by ancient soil formation.
The Big Pig Dig site was excavated for 15 field seasons and provided an incredible opportunity for park visitors to see paleontological resources and interact with scientists. The site was found by two park visitors, Steve Gasman and Jim Carney, who spotted exposed vertebrae and limb bones in a drainage ditch along Conata Road (County Road 502) southeast of the Conata Picnic Area (Plate 9; Fig. 1.1). Once it was determined that the bones were at risk to theft and erosion, the National Park Service initiated an emergency salvage operation, which turned into a long-term scientific excavation.
The excavation revealed a dense accumulation of semiarticulated to disarticulated bones with no apparent preferred orientation (Plate 10). Bone positions ranged from highly angled to horizontal. The analysis and interpretation of the lateral extent, as well as the geometry of individual sediment layers, supported the concept of a watering hole (Plates 11, 12). Sedimentological analysis suggested that the watering hole had formed from a stream channel or an oxbow lake formed by a meander loop cut off from a channel (Terry, 1996a).
The bones were intimately associated with a green mudstone layer representing anoxic/reducing conditions (Plate 10), with thicker portions of the green mudstone associated with greater concentrations of bone. The bones showed little or no weathering, although some were modified by carnivore/scavenger processing, including puncture and groove marks that were filled with surrounding sediment. Some bones were crushed along their length or on either end, possibly as a result of either burial compaction or processing. Spiral fractures, a type of crack within fresh bone that is normally associated with carnivore/scavenger processing, have also been noted on some fossils (Stevens, 1996a, 1996b).
Initially the fauna recovered from the site was not considered diverse, with a total of five taxa represented, including Archaeotherium (Figs. 5.32, 5.33), Subhyracodon (Fig. 5.50), Mesohippus (Fig. 5.47), Leptomeryx (Fig. 5.46), and cf. Prosciurus. However, by the close of excavations in 2008, a total of 19 genera had been discovered, with the minimum number of individuals for the four original taxa as follows: 8 Subhyracodon, 29 Archaeotherium, 11 Leptomeryx, and 8 Mesohippus (Shelton et al., 2009). A total of 19,290 specimens were eventually collected from the site, and 187 m2 were excavated.
In 1995 one of the authors (D.O.T.) was invited by the National Park Service to complete a sedimentary analysis of the Big Pig Dig to gain a better understanding of the depositional environment of the site. After extensive sampling and detailed analysis, the Big Pig Dig was interpreted as the attritional accumulation of animals around a watering hole during drought conditions (Terry, 1996b). Animals may have died from thirst, with small or weakened animals becoming trapped within the soft substrate of the watering hole while attempting to reach the remaining water. This model can be broken down into five steps that explain many of the sedimentological and paleontological characteristics of the Big Pig Dig:
Step 1: Basin forms. A floodplain basin is initially filled with water. Animals are using the basin as a watering hole.
Step 2: Drying starts. With the onset of drought conditions, animals begin to concentrate around the watering hole. Animals are likely wading into the water. Some animals die, possibly floating on the surface. As animals decompose, portions of the body drop off and settle to the bottom in random orientations. Scavengers and predators are likely exploiting the abundance of dead or weakened animals.
Step 3: Drying continues. Animal mortality increases as the drought intensifies. Water level within the basin drops. Animal remains become more concentrated, possibly contaminating the water. Animals walking through the water step on remains of other animals. This trampling pushes bones down into the soft mud, orients bones at high angles, and mixes muds at the bottom of the lake. Decay of animals within the water likely induces anoxic conditions. Scavengers and predators continue to exploit dead or weakened individuals. Depending on the stability of the substrate, small or weakened animals may become trapped while heading for remaining water in the center of the basin. Some animals die while clustered together into herds (entelodonts).
Step 4: Basin dry. Continued drought eventually leads to the complete drying of the basin. Some bones are completely covered, while others protrude from the surface of the dry lake bed. Mud cracks likely develop. Muds and sediments left over are likely rich in organic matter as a result of abundant animal remains. Anoxic conditions may exist in the remaining muds and sediments.
Step 5: Basin filled and bones covered. The basin is flooded, and the bone-bearing basin is buried by an influx of new sediment.
This model explains many of the sedimentological and taphonomic characteristics of the site. The fine-grained nature of the deposits, with occasional thin stringers of silt and fine sand, suggests that fluvial action was responsible for the deposition of individual bone bed units and the eventual burial of the bones. The lenticular geometry of the bone-bearing green layer (Plate 11), the rarity of soil features within individual horizons, and the preservation of relict depositional fabric are consistent with a small pond or floodplain lake, such as an oxbow. The lower contact of the bone-bearing green layer is undulatory and has a gradual transition over 1 to 2 cm with the underlying lower red layer (Plate 11). The foot traffic of animals would have mixed the bottom of the basin, resulting in a gradational lower boundary between the bone-bearing green and lower red layers. This would also explain the occurrence of swirled zones of relict bedding. Depending on the consistency of the substrate, animals would have made hoofprints or footprints that would protrude downward into the lower red layer (Plate 11).
Trampling by animals could also explain the high-angle orientation of some bones. As animals walked across the area, some long bones or other articulated sections would have been pushed down into the soft substrate of the lake bottom. The next influx of sediment (upper red layer) would cover the bone-bearing green layer, creating a sharp contact and incorporating fragments of the bone-bearing green layer as ripped-up clasts along the contact. Depending on the flow velocity during this next influx of sediment, some bones may have been transported a short distance or partially aligned to the current. The fresh to slightly weathered state of the bones suggests that they were not exposed at the surface for a significant period of time.
The presence of bone processing/modification suggests that carcasses and/or weakened animals were accessible and that scavenging and predation was taking place. Although we cannot be absolutely sure as to what animal or animals were responsible for the bone modification, the entelodont Archaeotherium is a likely suspect (Figs. 5.32, 5.33). Many of the puncture marks on the bones are similar in size and shape to those that could be made by the teeth of Archaeotherium. Scavenging by Archaeotherium as well as any of the many predators that existed at the time could explain the scattered, articulated to disarticulated nature of some bones. The low diversity of the bone bed’s fauna may be indicative of drought-sensitive species, though further preparation of field jackets in the laboratory may yield additional species, especially smaller ones.
Although this particular model is based on the concept of a drought-induced death assemblage, there is no direct sedimentological or stratigraphic data to support the development of drought conditions. The most likely sedimentological evidence of drought is the intricate network of cracks and veins throughout the site (Plates 10, 11). The origin of these features is problematic and should not be taken as proof of desiccation. The best evidence for drought conditions comes from comparisons of the bones within the Big Pig Dig to modern and ancient analogs. Studies of modern ecosystems in Africa suggest that animal concentrations increase as the amount of available water decreases (Western, 1975). During these dry periods, animals will cluster around remaining sources of water. Foster (1965) noted that rhinoceroses died half-submerged in rivers, possibly in the attempt to gain relief from the drought. In some instances animals remained around a drying watering hole even though flowing rivers were available several kilometers away. Foster (1965) suggested that death was due not to a lack of water but to an insufficient amount of forage available to support the animals. For example, Goddard (1970) applied the term “nutritional anemia” to explain the death of large numbers of black rhinos during a prolonged drought as a result of insufficient food sources.
Mead (1994) proposed, on the basis of an analysis of the jaw mechanics and facies associations of Subhyracodon versus Hyracodon (Figs. 5.50, 5.48), that Subhyracodon was a water-dependent organism that frequented riparian habitats, while Hyracodon was water independent and frequented open-range habitats. Subhyracodon is abundant in the Big Pig Dig, while Hyracodon has yet to be discovered. According to Mead (pers. comm.), the presence of Mesohippus in the Big Pig Dig also supports the concept of a water-dependent fauna (Fig. 5.47). Hunt (1990) reports the grouping of chalicotheres within the lower Miocene Agate bone bed of northwestern Nebraska. According to Hunt (1990), the Agate bone bed represents a short-term, attritional accumulation of animals around a watering hole during drought conditions. The presence of segregated animals supports this interpretation. Such segregation would not be expected to develop if the bone bed was the result of catastrophic flooding and mass death. The Agate bone bed also preserves evidence of bone processing and trampling (Hunt, 1990).
The Big Pig Dig bone bed may be one of several within this stratigraphic interval. Analysis of core samples shows that the same relict bedding structures and cracking and veining and colors similar to the bone-bearing green layer occur below the present quarry (Terry, 1996a). An oxbow lake or similar depression on a floodplain would provide a suitable setting for seasonal flooding events, eventual desiccation, and burial of animal remains. Vertebrate remains have been found nearby within greenish layers lateral to the bone-bearing green layer (Kruse, pers. comm., 1995). Several of these greenish layers are repetitive, with reddish layers similar to the upper and lower red layers.
The Brian Maebius site was discovered by and named after an intern employed at Badlands National Park in 1997. The site has since been studied by several institutions, culminating in four master’s theses (DiBenedetto, 2000; Factor, 2002; McCoy, 2002; Metzger, 2004). The site is located in the Badlands Wilderness Area and is stratigraphically positioned in the Scenic Member of the Brule Formation just below the Hay Butte marker. The site is relevant to this discussion because of the dense accumulation of fossil bone and unique preservational environments, which range from severely deteriorated to pristine, concentrated in several stratigraphic layers that have been overprinted by pedogenesis (Factor, 2002).
At the Brian Maebius site, the base of the stratigraphic section is marked by a sandstone layer that fines upward into a mudstone and claystone (Plate 13). The site is interpreted as a former channel that was isolated by avulsion or neck cut off to form an oxbow lake (Factor, 2002). Lateral accretion surfaces suggest migration to the north and northeast with flow direction to the east. DiBenedetto (2000) describes a tree stump felled by flooding and possibly aligned with paleoflow (Plate 7a). The bones appear to have been washed in during periods of high flow, possibly during periods of reactivation of the oxbow during flooding events, in which the oxbow acted as a sediment trap.
According to DiBenedetto (2000), carnivore coprolites are common, and he proposed that the site was near a carnivore den. He based his argument on the observation that individual bones have tooth impacts and lack evidence of transport. Some of these impacts are believed to be small, needlelike tooth marks, suggesting they were made by pups or kits. They also exhibit minimal time of surface exposure. Various fossil rodent and leporid remains were collected and are believed to have been deposited as avian pellet concentrations (DiBenedetto, 2000). Unfortunately, there is no sedimentological evidence to support a carnivore den model. An alternative theory is that the site was a slack-water area during flooding events that concentrated carcasses that were later scavenged (Factor, 2002).
Over time, the river-dominated environments of the lower part of the Scenic Member of the Brule Formation gave way to progressively greater amounts of eolian influence. This switch in depositional environments influenced both the formation of paleosols and the preservation of associated fossil remains. During periods of high amounts of eolian influx, landscapes were not stable but instead were vertically aggraded as the soil incorporated this steady influx of sediment. In the rock record, this resulted in zones of massive volcanic silts that appear to be randomly sprinkled with vertebrate fossils and soil features, such as scattered roots and glaebules (Figs. 3.13, 3.14). During periods of low eolian influx, soils were able to develop to a greater degree, and fossil bones were concentrated along the ancient landscape in conjunction with denser accumulations of fossil roots and better-formed soil horizons. Once eolian sedimentation increased again, these former periods of landscape stability would be buried and preserved by the next massive zone of volcanic silts.