For migratory populations of caribou (Rangifer tarandus), a key characteristic of their ecology and population biology is the annual journey to areas where they give birth in the spring: their calving grounds (Griffith et al., 2002; Nicholson et al., 2016; Cameron et al., 2020). Migrations to and from the calving grounds represent significant energy expenditures, but have multiple hypothesized benefits that increase calf survival and maintain the population. These include calving in areas with reduced predator pressure, abundant early-greening spring vegetation, and that are in close proximity to habitats used for insect avoidance later in the spring/summer (Griffith et al., 2002; Russell and Gunn, 2019). Thus, maintaining access to spring calving grounds and associated migration routes are top management and conservation priorities that are generally considered critical for maintaining herd viability (Griffith et al., 2002; Gunn et al., 2009; Russell and Gunn, 2019; Joly et al., 2021a). However, migratory patterns of caribou and their seasonal landscape use are sensitive to a variety of factors including annual weather patterns, larger-scale shifts in climate, changes in herd size, development of roads and industrial infrastructure, and a variety of other ecological and anthropogenic influences (Skoog, 1968; Russell et al., 1993; Cameron et al., 2005, 2020; Hinkes et al., 2005; Vistnes and Nellemann, 2008; Vors and Boyce, 2009; Joly et al., 2011). Over the last several decades of climatic and ecological changes across the Arctic, herds such as the Porcupine Caribou Herd (PCH) have demonstrated high annual variability in calving ground geography (Griffith et al., 2002; Cameron et al., 2005; Caikoski, 2020). As biologists and wildlife agencies establish conservation and management goals designed to insulate herds from future ecological stresses, several fundamental questions remain unanswered. How long have herds (or caribou more broadly) utilized particular calving grounds and how well do contemporary patterns of seasonal landscape use and caribou migration correspond to those under different regimes of climate and anthropogenic influences? While the timescales inherent to these questions far outstrip those available from traditional ecological datasets, antlers shed from female caribou provide useful proxies for assessing historical calving ground geography (Miller and Barry, 1992; Miller et al., 2013, 2021). This is possible because female caribou shed their antlers around the time they give birth (Espmark, 1971; Bubenik et al., 1997) and those antlers can persist for hundreds to thousands of years on Arctic landscapes (Sutcliffe and Blake, 2000; Stewart and England, 2008; Le Moullec et al., 2019; Miller and Simpson, 2022). Here, by radiocarbon dating antlers collected from tundra surfaces in Alaska, we test the long-term persistence of calving activity on the modern calving grounds of the PCH. In so doing, we document calving activity on the Coastal Plain of the Arctic National Wildlife Refuge across more than 3,000 years.

Shed female caribou antlers lying on landscape surfaces provide evidence of past calving activity. This is possible because female caribou, like their male counterparts, annually grow and shed antlers (Espmark, 1971; Bergerud, 1976). However, the schedules of antler growth and shedding are different between males and females (Bergerud, 1976; Bubenik et al., 1997). Females generally grow antlers in the spring/summer and maintain their antlers until shedding them within days of giving birth in the spring (Espmark, 1971; Bergerud, 1976; Whitten, 1995). Males, on the other hand, shed their antlers in the late fall and early winter (Bergerud, 1976). Because calving is highly synchronized among caribou and migratory caribou travel in large herds to the calving grounds (10³ to 10⁴ individuals), annual inputs of shed female caribou antlers can accumulate over time to produce large concentrations of shed antlers (~1,000 antlers/km², Cameron and Whitten, 1979; Griffith et al., 2002; Joly et al., 2011; Miller et al., 2013). As antlers slowly decay across decades to centuries or more, they become increasingly weathered (Meldgaard, 1986; Sutcliffe and Blake, 2000; Miller et al., 2021; Miller and Simpson, 2022). Because progressive changes in weathering provide useful information about their time-since-shed, accumulations of shed antlers offer opportunities to evaluate changes in seasonal landscape use, including calving activity, across space and through time (Behrensmeyer, 1978; Miller, 2012; Miller et al., 2013, 2021).

Today, the Coastal Plain of the Arctic Refuge primarily serves as a calving ground for the PCH (Figure 1; Griffith et al., 2002). The Coastal Plain is a flat region of tundra between the north side of the Brooks Range and the coast of the Beaufort Sea (Griffith et al., 2002; Jorgenson et al., 2002). The PCH calve throughout most of the Coastal Plain of the Arctic Refuge and Ivvavik National Park, Canada, immediately east of the Arctic Refuge (Fancy and Whitten, 1991; Griffith et al., 2002). The PCH is currently one of the largest caribou herds in the world (~218,000 individuals in 2017; Caikoski, 2020) and has one of the longest annual migrations of any terrestrial mammal (~1,350 km straight-line round-trip distance; Teitelbaum et al., 2015; Joly et al., 2019), which takes them between their winter range and spring calving grounds. The PCH has been studied for decades by academic, state, territorial, and federal wildlife biologists and an important aspect of that work has been assessing variability in calving ground geography through time (Russell et al., 1993, 2005; Griffith et al., 2002; Russell and Gunn, 2019; Caikoski, 2020). Evaluating the drivers behind annual variation and/or long-term consistency in calving geography has been a research focus due, in part, to industry and political interests in developing portions of their calving ground for hydrocarbon extraction (Whitten, 1994; Griffith et al., 2002; Gunn et al., 2009; Russell and Gunn, 2019). Referred to as the “1002 Area,” this portion of Coastal Plain was set aside by the U.S. Congress in 1980 for potential development (Udall, 1980). Petroleum lease sales were eventually sold in 2020 following a congressional mandate (Brady, 2017), though as of 2022, infrastructure to support hydrocarbon extraction has not been developed.

Calving ground geography for migratory caribou is a tale of two temporal scales. Survey data often show high annual variability in calving locations, with the centers of annual calving grounds differing from year to year across dozens to over 100 km (Griffith et al., 2002; Baltensperger and Joly, 2019; Cameron et al., 2020). At the same time, those annual shifts occur within a broader area that defines an overall “calving ground,” which is often stable across decadal timescales (Griffith et al., 2002; Cameron et al., 2020; Joly et al., 2021b). But what about longer timescales? The vast majority of data on calving grounds of migratory caribou are from the 1980s onward, with sparser data from the 1970s, 1960s, and earlier (Skoog, 1968; Whitten et al., 1984, 1992; Fancy et al., 1989; Fancy and Whitten, 1991; Bergerud, 1996; Griffith et al., 2002; Burch, 2012; Porcupine Caribou Management Board [PCMB], 2016; Benson, 2019). Consequently, these records overlap periods of increasing climate change and anthropogenic pressures. This issue is highlighted by the Central Arctic Herd, which mostly calves to the west of the Arctic Refuge. The Central Arctic Herd was recognized as a herd only in the mid-1970s, but its geographic range had been under development for petroleum resources (including the Trans-Alaska Pipeline) since the 1960s (Cameron and Whitten, 1979; Fancy, 1983; Cameron et al., 1992, 2005; Lenart, 2015; Nicholson et al., 2016; Johnson et al., 2020). Caribou ecology and seasonal landscape use are also influenced by shifts in climate and associated environmental changes, which have been particularly severe in the Arctic (Griffith et al., 2002; Arctic Climate Impact Assessment [ACIA], 2004; Joly et al., 2011; Pearson et al., 2013; Rantanen et al., 2022). Yet, again, observations of calving geography, particularly more standardized monitoring programs that started in the 1970s and 1980s, were initiated after Arctic climates had already started changing. In some cases, records from early European explorers and hunters provide glimpses into historical migrations and calving ground geographies, but such insights are generally based on limited observations with poor spatial resolution (Burch Jr., 2012; Benson, 2019). Oral histories and Traditional Knowledge of Native Alaskans and Canadians are another source of historical insight, though groups such as the Gwich’in traditionally stayed away from the calving grounds and so have restricted knowledge of their historical geographies (Benson, 2019). Although challenging to establish, expanding the timescales across which we study caribou calving geography and migration can help us evaluate the biological significance of modern calving grounds and test for changes associated with recent climatic and anthropogenic perturbations (Miller et al., 2021). Here, we use ancient antlers to expand the time series with which we assess calving ground geography for the PCH and to test how well contemporary patterns of seasonal landscape use are reflected in the deeper past.

Antlers used for this study came from broader sampling efforts to study historical caribou landscape use (Miller et al., 2013, 2021; Miller and Simpson, 2022). Shed antlers reported here were found exposed on landscape surfaces on the Coastal Plain of the Arctic Refuge in the summers of 2011, 2012, and 2018. Two of the antlers were discovered during standardized surveys of well-drained, cobble-rich tundra habitats with abundant Dryas, a low-growing evergreen dwarf shrub (i.e., Dryas river terraces, sensu Jorgenson et al., 2002; Jorgenson and Walker, 2018). One of the antlers came from a gravel bar that was discovered while walking between surveys. To sample Dryas terraces across the Coastal Plain of the Arctic Refuge, we traveled along major river systems that bisect the Coastal Plain from south to north (Miller et al., 2013, 2021). Targeting Dryas terrace habitats, we conducted 50 antler surveys near the Canning, Katakturuk, Hulahula, Jago, Aichilik, Kongakut, and Turner Rivers (Figure 1). Using each river, we sampled from the northern foothills of the Brooks Range to as close to the Beaufort Sea as was logistically possible. Across the region, we sampled areas that are currently within the PCH calving grounds as well as areas not currently known to support calving. Survey sites were accessed by a combination of fixed-wing aircraft, helicopter, or by rafting down rivers to survey locations. Because vegetation on Dryas terraces is sparse and/or low-growing (with few tussocks, willows, or other vegetation that would be visually obstructive), surveyors had clear views of the ground and any antlers in the area (Miller et al., 2013, 2021). All antlers evaluated here were at least partially exposed on landscape surfaces. Surveys and collections were done with permission from, and in collaboration with, the Arctic National Wildlife Refuge (United States Fish and Wildlife Service). The antlers are accessioned into the University of Alaska Museum Earth Sciences Collection (UAMES) at the University of Alaska Fairbanks.

Shed antlers can be discriminated from antlers introduced due to death by the presence of an exposed antler-skull attachment surface (Miller, 2012; Miller et al., 2013). That “pedicle attachment surface” becomes smooth and rounded as the antler detaches from the skull during the cell death processes that lead to shedding (Lincoln and Tyler, 1994; Bubenik et al., 1997). To differentiate female antlers from male antlers, we took advantage of known differences in surface areas between the pedicle attachment surfaces of males and females (females are generally smaller, Miller et al., 2013, 2021). Previous work using museum specimens with known sex has shown that while there is some overlap between pedicle surface areas of older females and younger males, such overlap is limited. Thus, surface area of the pedicle attachment surface offers a useful method for differentiating antler sex for most caribou antlers found on the Coastal Plain (Miller et al., 2013, 2021). Pedicle attachments are approximately elliptical, thus we estimated surface area using the long axis (A₁) and short axis (A₂) of the pedicle attachment surface (measured using IP67 Absolute Mitutoyo digital calipers) and the standard formula for estimating surface area of an ellipse: π * (½ * A₁) * (½ * A₂).

Based on the known size frequency distributions of pedicle surface areas of male and female antlers, we used a previously developed Monte Carlo method to calculate the probability (mean and 95% confidence interval) that each antler evaluated here was female (Miller et al., 2013, 2021). The surface areas of all antler pedicle attachments in this study were quite small, indicating the antlers were definitively female. We measured the length of each specimen by following the curvature of the antler fragment using a flexible plastic measuring tape.

We used Accelerator Mass Spectrometry (AMS) radiocarbon dating to estimate when each antler was grown, which is roughly synchronous with when the antler was shed. Radiocarbon dating measures the amount of ¹⁴C in a sample relative to that of atmospheric radiocarbon in 1950 and can be used to estimate the ages when biological specimens formed across roughly the last 50,000 years (Hajdas et al., 2021). To radiocarbon date the antlers, we sampled bone from the proximal portion of the main antler beam. This region has dense cortical bone and should be most resistant to diagenetic processes and infiltration of younger carbon (Miller et al., 2021). Antlers were cut using a low-speed Dremel with a diamond cutting wheel. Samples were broken into small chips and sent to the Keck-Carbon Cycle AMS Facility at the University of California, Irvine or the Center for Accelerator Mass Spectrometry at Lawrence Livermore National Laboratory for chemical pretreatment and AMS radiocarbon dating. Collagen was prepared using standard treatments for demineralization and humic acid removal (Beaumont et al., 2010; Crowley et al., 2010), ultrafiltered, and lyophilized. For the one sample sent to the Keck-Carbon Cycle AMS Facility (UAMES 52867), lipids were removed by sonicating antler chips in a 2:1 chloroform/methanol solution for 30 min prior to demineralization. This was repeated with a fresh chloroform/methanol solution until the solution was no longer cloudy. The sample was then sonicated for 30 min in methanol and 30 min in Milli-Q water. The sonicator bath was cooled to keep the temperature below 60°C. For the two antlers dated at Lawrence Livermore, lipids were not removed prior to dating. To account for isotopic fractionation, the δ¹³C of a sample must first be normalized to a standard prior to calculating its radiocarbon age (Stuiver and Polach, 1977; Hajdas et al., 2021). We previously reported no difference in δ¹³C values for antlers with or without lipid removal (p = 0.81, Miller et al., 2021), suggesting differences in preparation should be of negligible concern for specimen radiocarbon dates and their comparisons. To evaluate the quality of collagen preservation, and thus the reliability of the radiocarbon dates, we measured the atomic C:N ratios of each specimen, which are generally between 3.1 and 3.5 for well-preserved specimens (DeNiro, 1985; Ambrose, 1990; van Klinken, 1999). We calibrated radiocarbon dates to calendar years using IntCal20 (Reimer et al., 2020) and the “rcarbon” package (version 1.4.2, Crema and Bevan, 2021) in R (version 4.1.1, R Core Team, 2021). Calibrated radiocarbon dates produce a probability distribution of possible ages. We summarized each antler’s calibrated age as the weighted-median of its age distribution along with the range (minimum and maximum ages) of the 2-sigma 95% confidence interval of that distribution (Telford et al., 2004; Crema and Bevan, 2021).

Antler specimens described below came from near three rivers: Canning, Katakturuk, and Turner. Each specimen had pedicle attachment surfaces that were intact and calibrated radiocarbon dates older than 1,000 calendar years before present (BP; where “present” is standardized to 1950). C:N ratios (ranging from 3.18 to 3.27; Table 1) are all within the expected range for well-preserved collagen (DeNiro, 1985; Ambrose, 1990; van Klinken, 1999). Each antler is described below.

UAMES 52867 (Figure 1, large filled circle; Figure 2A) was recovered during a standardized antler survey on a Dryas terrace in a slightly upland region near the Katakturuk River. The antler fragment extends approximately 151 mm from the pedicle attachment surface and includes part of the brow tine. The pedicle attachment surface is small and confidently female (93.2%; Table 1). The pedicle attachment surface is well defined and smooth, indicating that the antler was shed during the life of the animal. When discovered, it was more than half buried in mineral-rich soils that were adjacent to the mound of a ground squirrel (Urocitellus parryii). The antler’s exterior surfaces are well-weathered and have a fibrous texture (Figure 2A). Portions of the antler show signs of chemical maceration and dissolution. Lichen development on the antler is minimal, though moss was attached to the antler along many of the surfaces that were in contact with the soil. The weighted median calibrated date is 3,157 cal years BP (2-sigma range: 3,218–3,072 cal years BP; Table 2).

UAMES 52866 (Figure 1, large open circle; Figure 2B) was found during a standardized antler survey on a Dryas terrace in an upland (near foothills) region near the Canning River. This antler fragment was discovered lying entirely exposed on the tundra. However, it was adjacent to a region that had been dug into (likely by a grizzly bear; Ursus arctos horribilis), so previous and perhaps relatively recent exhumation of the specimen cannot be excluded. Surfaces of the antler are acid etched with mm-scale curvilinear grooves consistent with damage by roots. No observable lichens were present. The antler fragment is 149 mm long and there is no indication of either brow or bez tines. UAMES 52866 has a small antler attachment surface and is confidently female (94.5%; Table 1). The antler includes a small portion of the cranial pedicle, indicating this antler was not shed during life. However, the pedicle attachment surface is well delineated (Figure 2B, dashed line) and portions of the pedicle are cleanly removed from the attachment surface (Figure 2B, arrow). Some areas at the pedicle attachment surface are not pristinely preserved and it is difficult to evaluate some of this specimen’s anatomy. The median calibrated date for UAMES 52866 is 2,581 cal years BP (2-sigma range: 2,714–2,494 cal years BP; Table 2).

UAMES 52865 (Figure 1, filled square; Figure 2C) was discovered while the team was walking between surveys along the Turner River. This specimen was found on a sand bar near the active portion of the channel. The antler fragment is 113 mm long and includes portions of a brow tine and the main antler beam. The antler is moderately rounded, indicating exposure to fluvial processes. The antler is also hollowed-out in some regions, which is also likely due to exposure to fluvial processes. Hollowing occurs along the main beam, brow tine, and into the base of the antler near the pedicle attachment surface. While the pedicle attachment surface is river-rounded, it is generally well-preserved and well-defined, indicating the antler was naturally shed. The pedicle attachment surface is small and confidently female (92.3%; Table 1). No observable lichens are present. While the original shedding location for this antler is unknown, the low gradient and sinuous nature of the Turner River suggests it was unlikely to have been transported over long distances. The Turner is also short (<20 km straight-line distance between its head waters and the Beaufort Sea) and the entire river courses through what are currently the calving grounds of the PCH. Thus, any degree of fluvial transportation is unlikely to substantially modify the spatial accuracy with which this specimen is used as a calving ground indicator. The calibrated radiocarbon date for this specimen is 1,629 cal years BP (2-sigma range: 1,704–1,549 cal years BP; Table 2).

Today, the PCH gives birth to their calves on the Coastal Plain of the Arctic National Wildlife Refuge (Alaska) and Ivvavik National Park (Yukon, Canada). While Traditional knowledge of the Gwich’in people and historical records of early European explorers and trappers indicate similarity in landscape use across at least hundreds of years (Burch Jr., 2012; Benson, 2019), the long-term continuity of contemporary PCH calving grounds remains unknown. Female caribou shed their antlers within days of giving birth and serve as a useful proxy for historical calving activity (Espmark, 1971; Whitten, 1995; Miller et al., 2021). Radiocarbon dates of antlers found on the PCH calving grounds reveal shed antlers as old as 1,629 and 3,157 cal years BP. The antiquity of these shed antlers provides the first physical evidence of calving activity from previous millennia and substantiates the long ecological legacy of the Coastal Plain as a caribou calving ground. Some of the ancient antlers evaluated here also come from periods when summer temperatures were higher than recent decades, indicating that the PCH calving ground may be viable for at least the next few stages of projected climate warming. Millennial-aged antlers have been recovered from multiple high latitude settings in the northern and southern hemispheres (Miller and Simpson, 2022), highlighting the extended ecological histories available from bone accumulations in cold-weather environments. Extended persistence of antlers on Arctic caribou calving grounds also indicates that recycling nutrients from caribou antlers (phosphorus, calcium) is highly inefficient. Paired with large annual inputs of antlers to the calving grounds (10⁵ antlers per year), we also find that caribou calving grounds are an underappreciated sink for nutrients that are important to both plant and animal communities. Evaluating antlers (and other bones) on modern landscapes provides important opportunities to evaluate historical states of living populations and to gain new insight into the biological intricacies of even well-studied ecosystems.

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

JM designed the study, conducted fieldwork, prepared the specimens, analyzed the data, and wrote the manuscript. EW conducted fieldwork and edited the manuscript. PD provided logistical support, curated the specimens, and edited the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by funding from the U.S. Fish and Wildlife Service (Arctic National Wildlife Refuge), The National Geographic Society (9133-12 to JM), a Christine Stevens Wildlife Award from the Wildlife Society (JM), the University of Cincinnati Office of Research, and the National Science Foundation (DEB-2135479 to JM).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.