Northern Alaska is currently experiencing environmental changes as the result of global warming, which is occurring most rapidly at northern latitudes (Moon et al., 2019). These changes impact the mobility and ecologies of extant megafauna including caribou, moose, and muskoxen (Post and Forchhammer, 2008; Sharma et al., 2009). For example, migratory species are experiencing mismatch in timing of migration and peak resource availability (Post and Forchhammer, 2008), while caribou are losing habitat in the north due to warmer Arctic summers and winters (Sharma et al., 2009). Detailed paleoecological evidence from the remains of past megafauna from this region provides an opportunity to examine how past megafauna lived in this environment (Guthrie, 1989), and therefore help predict responses of living megafauna to present and projected environmental changes.
During the height of the last glaciation (∼28–18 thousand years ago (kya)) (Clark et al., 2009), the North Slope of Alaska was part of an expansive land-mass known as Beringia. Sea levels were ∼130 m lower than today (Lambeck et al., 2014), exposing a shallow continental shelf between northeast Asia and North America known as the Bering Land Bridge (BLB). The BLB extended approximately from the Lena River, Russia, in the west and the Mackenzie River, Yukon, Canada, in the east (Elias and Crocker, 2008) (Fig. 1). The Beringian ecosystem was primarily that of a mammoth steppe, a graminoid-dominated ecosystem that supported a community of large herbivorous mammals, dominated by mammoths (Mamamuthus primigenius), horses (Equus sp.), and steppe bison (Bison priscus) (Guthrie, 2001; Mann et al., 2013; Shapiro and Cooper, 2003; Zimov et al., 2012). The mammoth steppe supported large populations of these herbivores, many of which had larger body sizes than their descendants today at similar latitudes (Zimov et al., 2012). Bison, in particular, had larger body sizes and horns than present-day American bison (Bison bison) (Martin et al., 2018), and were present throughout most of Eurasia and North America in what has been termed “The Bison Belt” (Guthrie, 1989). Steppe bison first arrived in North America ∼195–135 kya (Froese et al., 2017) and their population began to decline around 37 kya (Heintzman et al., 2016; Shapiro et al., 2004).
Modern bison ecology can provide an analog for inferring ancient bison behavior as well as the basis for comparative (paleo) biology and anatomy. Although modern plains bison (Bison bison bison) are often considered grassland grazing specialists (Bamforth, 1987), plains bison in northern habitats (Waggoner and Hinkes, 1986), wood bison (Bison bison athabascae) (Larter and Gates, 1991), and european wisents (Bison bonasus) have been observed to regularly utilize browse in their diet (Kowalczyk et al., 2011). Evidence from macro and micro tooth-wear analysis indicates that steppe bison likely had a broader herbivorous diet and ecological niche that included browsing (Rivals et al., 2010, 2007; Saarinen et al., 2016). The long-distance (>100 km) (Berger, 2004; Hanson, 2015; Plumb et al., 2009) migrations of American bison (Bison bison) across the American Great Plains were legendary and a key component of bison life history (Bamforth, 1987; Flores, 1991). However, isotopic (strontium) analyses of ancient (∼18,500 14C yr BP) bison (Bison priscus) from a site in Ukraine found no evidence of paleomobility (Julien et al., 2012). Analyses of ancient bison specimens can provide opportunities to flesh out the paleoecological life-history of a taxon that shaped Beringian ecosystems (Zimov et al., 2011).
Fortunately for paleoecologists, bones, teeth, and horns of bison are some of the most numerous fossil remains found in Alaska (Guthrie, 1970; Mann et al., 2013). On rare occasions, high sediment deposition rates along with freezing temperatures can result in preservation of virtually complete carcasses or skeletons of past Beringian fauna, revealing vivid paleoecological snap-shots of life in Beringia (Boeskorov et al., 2016; Kirillova et al., 2015; Van Geel et al., 2014; Zazula et al., 2009, 2017). These rare and well-preserved glimpses of past megafaunal paleoecology can emerge from eroding river-banks (Mann et al., 2013; Zazula et al., 2009), during mining operations (Guthrie, 1968), and during construction activities (Zazula et al., 2017). In some instances, soft tissue such as hair, skin, organs, and stomach contents are preserved (Kirillova et al., 2015; Van Geel et al., 2014), as well as associated insects. These remains, along with bones and teeth, can retain chemical clues about an individual’s paleoecology. Isotopic analyses (Kirillova et al., 2015) and analyses of ancient DNA (aDNA) (Zazula et al., 2017) can reveal past mobility patterns and population interconnectivity, contributing to an understanding of past ecosystems, landscapes, and evolution (Froese et al., 2017; Haile et al., 2009; Marsolier-Kergoat et al., 2015; Shapiro and Cooper, 2003; Heintzman et al., 2016; Shapiro et al., 2004).
In this study, we conducted a multi-proxy study, combining isotopic and aDNA analyses, with supporting paleo-forensic analyses, to investigate the sex, age, paleoecology and life history of an exceptionally well-preserved and largely articulated steppe bison (Bison priscus) found on the North Slope of Alaska (Fig. 1 a-c). We further assess peri- and post-mortem events together with the taphonomic history of skeletal remains using a taphonomic analysis of the skeleton and of the plant and insect remains present. Our examination of the physical condition of the remains (bones and teeth) from the specimen provides clues about an individual’s age, sex, and appearance (Fuller, 1959). The taphonomy and geology associated with the specimen provides information about the context surrounding an organism’s death and a possible cause of death. We used insect (Elias et al., 2000) and plant (Bigelow et al., 2013) macrofossils associated with the bison’s remains to provide valuable paleoecological information. Our multi-proxy, paleoecological approach adds to the broader understanding of ancient bison ecology during the Late Pleistocene.
Isotopic analyses have become a popular tool in paleoecology for determining the ecological and life-history traits of ancient fauna (Bocherens, 2003; Britton et al., 2009). Tissues that form in discrete layers over a period of an individual’s life, such as tooth enamel, hair, or horn, can be subsampled to allow inference of inter- and intra-annual paleo-mobility and paleoecology using isotopic analyses of these sample types (Balasse et al., 2001; Britton et al., 2009; Stevens et al., 2011; Zazzo et al., 2012). Intra-tooth strontium isotope ratio (87Sr/86Sr) analysis has been developed as a useful tool to track the mobility of past animals and humans (Balasse, 2002; Britton, 2009; Britton et al., 2011; Hoppe et al., 1999; Hoppe, 2004; Julien et al., 2012; Koch et al., 1995; Pellegrini et al., 2008; Radloff et al., 2010; Viner et al., 2010; Widga et al., 2010). 87Sr/86Sr ratios can vary across landscapes depending in part on the age and rock type of the underlying geology (Bataille et al., 2014; Brennan et al., 2014). These landscape 87Sr/86Sr signatures can enter animals through their diet and drinking water, replacing calcium in tissues such as teeth and bones (Capo et al., 1998). The combination of 87Sr/86Sr and oxygen isotopic (δ18O) values from analyses of animal tissues have proved to be a powerful predictor of geographic location (Britton et al., 2009; Gigleux et al., 2017; Knudson et al., 2009). The δ18O values from a specimen can indicate location based on latitude and distance from the coast (Hoppe, 2006; Lachniet et al., 2016). The δ18O values from analyses of bison teeth have also been used to determine approximate local climate and seasonal temperature variation, because δ18O values in water can be closely related to temperature (Bernard et al., 2009; Hoppe et al., 2006; Scherler et al., 2014). Examining bison mobility has been one of the more common applications of this isotopic methodology, partly due to the abundance of bison remains in the archaeological and paleontological records (Britton et al., 2012; Julien et al., 2012; Widga et al., 2010).
Stable carbon and nitrogen isotope ratios (expressed as δ13C and δ15N values, respectively) from analyses of sub-fossils of animals can add a dietary dimension to a paleoecological reconstruction (Drucker et al., 2008; Stevens and Hedges, 2004). δ13C, along with δ18O, values can be generated from the analysis of inorganic carbon in bones and teeth (Koch et al., 1997). Bison horns, which grow throughout the life of a bison, are a carbon and nitrogen-rich keratin tissue that allows intra- and inter-annual paleoecological inferences. δ13C values can also be generated from analyses of organic carbon preserved as bone collagen and the horn keratin, and these methods also produce δ15N values (Iacumin et al., 2001; Schoeninger and DeNiro, 1984). In order to interpret the δ13C and δ15N values of these analyses, potential sources of variability need to be evaluated. One common source of δ13C variation in the diet of herbivores is variation in the proportional contribution C4 vs. C3 plants. However, C4 plants are exceedingly rare or non-existent in the Arctic (Wooller et al., 2007). Furthermore, a lack of trees and even shrubs in Pleistocene Arctic Alaska, even during some warmer interglacials (Willerslev et al., 2014), also excludes “the canopy effect” caused by the concentration of CO2 in dense forest (Drucker et al., 2008) and differences between herbs and shrubs (Schwartz-Narbonne et al., 2019). Thus, the main source of variability in δ13C values we can expect in Arctic vegetation is between wetter and drier environments (Wooller et al., 2007).
There are several drivers of δ15N variation in animals, which can make it a challenging system to study although it often demonstrates important relationships. For example, δ15N values in ancient megafauna have been found to have a strong relationship with the amount of precipitation during the Pleistocene (Carlson et al., 2016; Drucker et al., 2003; Graham et al., 2016; Heaton et al., 1986; Rabanus-Wallace et al., 2017), which is likely due in part to major changes in soil ecology in response to climatic change (Hobbie and Hobbie, 2006; Stevens et al., 2006). However, for higher resolution δ15N we must consider behavioral and physiological explanations for change found in serially analyzed tissues (Drucker et al., 2010). Behavioral explanations including migration or dispersal could explain some of the differences and δ15N values found in relation to aridity (Barbosa et al., 2009) and altitude (Männel et al., 2007). In northern regions, studies have found consistent difference in plants based on mycorrhizal relationships with the soil (Kristensen et al., 2011). However, these are not sufficient to account for the seasonal differences or major regional differences that would be required to explain some of the magnitude of variation in δ15N values we go on to document in this ancient bison and other modern bison in the region (Funck et al., 2020). Alternatively, physiological changes in δ15N values of animals can occur in animals adapted to extreme conditions. Notably, animals who undergo hibernation (Lee et al., 2012), lack sufficient calories or are fasting (Hobson et al., 1993; Hobson and Clark, 1992; Mekota et al., 2006; Voigt and Matt, 2004) as well as animals managing acute physical stress (Delgiudice et al., 2000; Fuller et al., 2005; Habran et al., 2010; Rode et al., 2016) can exhibit relatively large increases in δ15N values. The effect of dietary stress on δ15N values in wood bison was evident in individuals that had undergone nutritional stress during the winter in Northern Alaska (Funck et al., 2020). For this reason, we will focus our interpretation on nutritional stress as the primary factor of intra-tissue δ15N variation but acknowledge that other behavioral and ecological factors may be involved. From our review of the literature, we have noted that changes in δ15N values in herbivores from the mammoth steppe over time have largely not included the possibility that some of the variation could be driven by physiological stress and even starvation. To some degree this is likely obscured by the fact that a majority of the previous research has focused on δ15N values generated from analyses of bone collagen, which provides a more integrated and essentially a life-time measure. Our analyses of hornsheaths are providing a new a more detailed temporal perspective, which could be translated to the huge abundance of archived bison horn sheaths available. Using a combination of isotopic approaches on different sequentially grown tissues can provide a multi-proxy perspective of an individual’s life-history and eventually put this in the context of larger herds.
Ancient DNA (aDNA) from ancient bison specimens from Beringia has been used to monitor gene-flow across the BLB and through the ice-free corridor that connected eastern Beringia to the rest of the Americas (Froese et al., 2017; Heintzman et al., 2016). It has also been used to estimate past changes in effective population size (Lorenzen et al., 2011), which has been used to identify the decline of steppe bison since ∼37 kya in Beringia (Shapiro et al., 2004). Analyses using aDNA can also add a further perspective on individual steppe bison specimens. They can provide information on an individual’s genomic sex and population affinity, and help constrain age estimates for specimens potentially outside the range or towards the limits of radiocarbon dating. For the latter analysis, one such approach uses Bayesian analysis of aDNA sequences from dated specimens that lived at different time periods to calibrate a molecular clock, and then use this calibration to estimate the age of the undatable individual (Shapiro et al., 2011). The mitochondrial genealogy of Pleistocene bison in North America has been relatively well sampled, both geographically and temporally (Froese et al., 2017; Heintzman et al., 2016; Shapiro et al., 2004), and is therefore likely to be well suited for estimating the age of ancient bison specimens.