Stable isotope compositions of herbivore teeth indicate climatic stability leading into the middle Miocene Climatic Optimum, in Idaho, U.S.A

Journal Pre-proof Stable isotope compositions of herbivore teeth indicate climatic stability leading into the middle Miocene Climatic Optimum, in Idaho, U.S.A

Elisha B. Harris, Matthew J. Kohn, Caroline A.E. Strömberg PII:

S0031-0182(19)30805-3

DOI:

https://doi.org/10.1016/j.palaeo.2020.109610

Reference:

PALAEO 109610

To appear in:

Palaeogeography, Palaeoclimatology, Palaeoecology

Received date:

30 August 2019

Revised date:

15 January 2020

Accepted date:

16 January 2020

Please cite this article as: E.B. Harris, M.J. Kohn and C.A.E. Strömberg, Stable isotope compositions of herbivore teeth indicate climatic stability leading into the middle Miocene Climatic Optimum, in Idaho, U.S.A, Palaeogeography, Palaeoclimatology, Palaeoecology (2020), https://doi.org/10.1016/j.palaeo.2020.109610

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier.

Journal Pre-proof Stable isotope compositions of herbivore teeth indicate climatic stability leading into the middle Miocene Climatic Optimum, in Idaho, U.S.A.

Department of Biology and Burke Museum of Natural History and Culture, University of

ro

a

of

Elisha B. Harris a,*, Matthew J. Kohn b, Caroline A.E. Strömberg a

Department of Geosciences, Boise State University, 1910 University Drive, Boise, ID 83725,

re

b

-p

Washington, 253 Life Sciences Building, Seattle, WA 98195, United States, (206) 221-6724

United States, (208) 426-1631

lP

Corresponding author e-mail: [email protected]

Jo

ur

na

*

Declarations of interest (for all authors): none

Abstract The effects of global climate change are manifested at the regional level; consequently, evaluation of links between palaeoenvironmental change and turnover of past biotas must use regional-scale climate data. Here we test whether climate change influenced faunal and floral

1

Journal Pre-proof patterns leading up to, and during, the mid-Miocene Climatic Optimum (MMCO) based on datasets collected in the Railroad Canyon section (RCS), central-eastern Idaho, U.S.A. Specifically, we use isotope compositions of herbivore tooth enamel to investigate how Mean Annual Precipitation (MAP), Mean Annual Temperature (MAT), Cold Month Mean Temperature (CMMT), and temperature seasonality varied through the RCS and how this local record compares to global climate reconstructions. Isotope compositions of teeth from the fossil

of

equid Merychippus and rhinocerotid Diceratherium were compared across five time bins from c.

ro

23 to c. 15 Ma. Bulk δ18O values of both taxa indicate that meteoric water compositions were

-p

unchanged through the study interval (consistently c. -15.6‰) but were c. 1.9‰ higher than

re

current regional meteoric water. This difference point to warmer climates in central-eastern Idaho during the Miocene, post-middle Miocene topographic uplift in western North America,

lP

and/or shifts to air mass trajectories. Serial sampling of enamel shows seasonal fluctuations in

na

water compositions and temperature. Bulk δ13C values indicate an average MAP of 190 (range: 10-510 mm·yr-1), with no significant change through the study interval, comparable to today’s

ur

MAP (236 mm·yr-1). The reconstructed semi-arid, seasonal, warm climate for the RCS agrees

Jo

with phytolith assemblage data from the same strata indicating dominantly open, grassy habitats. However, it is inconsistent with previously held ideas of warm-wet mid-Miocene climates in the northern Rocky Mountains (NRM) based on paleosol analyses. Our inferred, stable climate conditions for the MMCO of the NRM contrast with coeval records from Europe and Asia indicating sustained or enhanced warm and wet climates. These differences point to a decoupling of regional climate from global trends and highlight the necessity of studying regional variation to understand the biotic impacts of global climate change.

2

Journal Pre-proof Keywords Mid-Miocene Climatic Optimum, Paleoecology, Enamel, Climate, Seasonality

1. Introduction From a global perspective, the middle Miocene was one of the last intervals of overall warming to temporarily reverse a long-term trend of climatic cooling through the Neogene

of

(Zachos et al., 2008). This warming event, known as the mid-Miocene Climatic Optimum

ro

(MMCO; initial warming beginning c. 18 Ma and peak warming c. 17–14.75 Ma), has been

-p

recorded in many marine and terrestrial records (Böhme, 2003; Flower and Kennett, 1994;

re

Holbourn et al., 2015; Mosbrugger et al., 2005; Zachos et al., 2008) and likely resulted in global temperatures that were 3–6°C warmer than today (Flower and Kennett, 1994; You et al., 2009).

lP

Paleobotanical remains (pollen, wood, and megafloras) from Europe indicate increasingly warm

na

and humid climates associated with widespread warm-temperate evergreen broadleaved and mixed forests at mid-latitudes during the early-middle Miocene (e.g., Ivanov et al., 2011;

ur

Jiménez-Moreno et al., 2008; Mosbrugger et al., 2005; Pound et al., 2012). The MMCO also

Jo

broadly coincided with other important global-scale ecological transformations as temperate grasslands expanded in geographic range across many continents and mammalian browsers were replaced by grazers (e.g., Badgley et al., 2008; Barnosky et al., 2003; Blois and Hadly, 2009; Jacobs et al., 1999; Strömberg, 2011; Tedford et al., 2004); however, causal links between these ecosystem changes and global climate remain unclear. Testing whether middle Miocene biotic changes were driven fully or in part by the MMCO requires continuous, long-term and temporally resolved records of local climate coupled with plant and animal fossils. However, to

3

Journal Pre-proof date no studies have been possible because of the scarcity of such linked, highly resolved local records. The Railroad Canyon section (RCS) of central-eastern Idaho, USA (Fig. 1), offers a valuable sequence with which to evaluate abiotic drivers of biotic evolution because it preserves a nearly continuous record of fossil plants, animals, and climate spanning nearly 8 million years leading into peak mid-Miocene warming (c. 22.9 – 15.2 Ma; Fig. 2; Barnosky et al., 2007; Harris

of

et al., 2017a, 2017b; Retallack, 2007, 2009). Multiple lines of evidence suggest that water

ro

availability decreased through time in the Northern Rocky Mountains (NRM) as the local climate

-p

became increasingly arid during the early–middle Miocene, including sedimentology (e.g.,

re

occasional gypsum and halite deposits in the Renova Formation; Fields et al., 1985), faunal data (Barnosky et al., 2007, 2003; Barnosky and Carrasco, 2002), and new biosilica assemblages from

lP

the RCS that show a decrease in diatoms and increase in dry-adapted pooid, open-habitat grasses

na

(Harris et al., 2017b). In contrast, paleosol data have been interpreted as showing a warmer and wetter, sub-humid climate in the NRM during the middle Miocene (Retallack, 2007). However,

ur

these paleosol-based MAP reconstructions can be questioned given the lack of definitively

Jo

authigenic pedogenic carbonates preserved throughout the RCS (Harris et al., 2017b). To help resolve these contradictory climatic interpretations, we use carbon and oxygen stable isotope data from herbivore tooth enamel as well as palm phytolith data to establish a record of precipitation, temperature, and seasonality for the RCS. We hypothesize that the Railroad Canyon region experienced warming (our temperature hypothesis herein referred to as Htemp1) and more equable temperature seasonality (referred to as Htempseason1) during the MMCO, similarly to other middle Miocene sites in, for example, Europe and Asia (e.g., Mosbrugger et al., 2005; Utescher et al., 2011); and that lowered seasonal water availability in the NRM due to

4

Journal Pre-proof reduced MAP (referred to as Hprecip1) caused a decrease in diatoms and increase in dry-adapted pooid open-habitat grasses (Harris et al., 2017b). We also explore whether local climate change coincided with faunal change in the NRM and compare the MMCO climate at RCS with different regions throughout the world.

2. Background

of

2.1 Geologic setting and biotic change

ro

In the history of western North America, the Miocene was an especially active epoch,

-p

both tectonically and biologically. Initiation of crustal extension resulted in development of the

re

Basin and Range province by c. 17.5 Ma, creating dramatic topographic relief and increased area for habitat development in the NRM through the middle to late Miocene (Atwater and Stock,

lP

1998; Chamberlain et al., 2012; McMillan et al., 2002), which arguably helped drive large

na

mammal speciation (Kohn and Fremd, 2008). Columbia River Basalt flow deposits in eastern Washington, northeastern Oregon, and western Idaho indicate increased regional volcanism from

ur

c. 17.5 to 6 Ma (Liu and Stegman, 2012; Swanson et al., 1979). Furthermore, the Yellowstone

Jo

Hotspot initiated intense volcanic activity in northern Nevada c. 16 Ma, and, in the course of migrating towards its present location in northwestern Wyoming, created the Snake River Plain (Pierce and Morgan, 1992). Additionally, North American ecosystems were transforming during the early to middle Miocene as increasingly open, grass-dominated habitats spread through western North America and the Great Plains (e.g., Harris et al., 2017b; Leopold and Denton, 1987; Retallack, 2007; Strömberg, 2011, 2005) and as C4 grass abundance increased in grassland communities in the latest Miocene (c. 8 Ma; e.g., Edwards et al., 2010; Hyland et al., 2018; Strömberg and McInerney, 2011). Furthermore, decreases in species richness and abundance of

5

Journal Pre-proof closed-habitat mammalian taxa have been attributed to aridification associated with the expansion of grasslands as well as concurrent climatic cooling (e.g., Figueirido et al., 2012; Janis et al., 2004; Samuels and Hopkins, 2017). It was during this period of sustained tectonic activity and ecosystem evolution that the sediments and fossils of the Railroad Canyon were deposited. The RCS of central-eastern Idaho consists of a c. 380 m thick, early–middle Miocene sedimentary sequence belonging to the Renova and Six Mile Creek Formations (Barnosky et al.,

of

2007; Harris et al., 2017a). A transition between the two formations occurs at c. 70 m in the

ro

composite section and is thought to correspond to a depositional hiatus called the early Miocene

-p

unconformity (Harris et al., 2017a). Recently, the RCS, whose age was previously debated, was

re

firmly dated to 22.9–15.2 Ma using U/Pb dating of zircons. This new age has led to important changes in local chronostratigraphy, including pushing back the age of the early Miocene

na

lP

unconformity in central-eastern Idaho (now c. 21.4–21.5 Ma; Harris et al., 2017a).

2.2 Floral and faunal record in the NRM

ur

The NRM hosts a rich fossil record of both plants (e.g., Graham, 1999; Strömberg, 2011)

Jo

and mammals (e.g., Barnosky et al., 2007, 2003; Barnosky and Carrasco, 2002; Janis et al., 1998) documenting biotic change during the Cenozoic. Regional floral evidence mainly from phytoliths shows that vegetation was relatively open and dominated by C3 pooid grasses, but that potential C4 PACMADs had spread across the NRM, making up a small component of grass communities by the early Miocene (c. 21 Ma; Chen et al., 2015; Cotton et al., 2012; Harris et al., 2017b; Hyland et al., 2018; Strömberg, 2005). This floral record can be compared to NRM faunal records, which show an increase in herbivores with adaptations linked to grazing, with oreodont-, camel- and rhino-dominated faunal communities in the early Miocene (c. 22.5 – 18

6

Journal Pre-proof Ma) transitioning to equid-dominated communities in the middle Miocene (c. 16 Ma; Barnosky, 2001; Barnosky and Carrasco, 2002). Faunas in this region were also affected by immigration events that took place c. 18.5 Ma and c. 17.2 Ma (e.g., Woodburne and Swisher, 1995), and notable changes in ungulate diversity in the NRM coincide with an increase in the adaptive radiation of ungulates in the Great Plains (c. 19 – 16 Ma; Janis et al., 2004). In addition, NRM faunas are characterized by increasing endemism, presumably linked to the growing isolation of

of

basins within this tectonically active region (c. 16 – 15 Ma; Barnosky, 2001; Barnosky and

ro

Carrasco, 2002).

-p

Many recent studies have used regional faunal databases to explicitly test for temporal

re

correlation of faunal change, tectonism, and climate change during the early–middle Miocene across the NRM region (e.g., Barnosky, 2001; Barnosky and Carrasco, 2002; Kent-Corson et al.,

lP

2013; Kohn and Fremd, 2008). Barnosky (2001) suggested that a peak in mammalian species

na

richness during the late early–middle Miocene (c. 14.5 Ma) was driven by the MMCO. Specifically, he argued that an increase in mammals adapted to semiarid- to arid environments

ur

(i.e., heteromyid rodents, which today occupy deserts, and equids with grazer morphology, today

Jo

feeding in grasslands) was consistent with an increase in regional temperatures. Others have pointed to increased endemism within the NRM, perhaps as a result of increased topographic heterogeneity over short distances, as a driver of increased faunal diversity c. 16 – 15 Ma (Barnosky and Carrasco, 2002; Kohn and Fremd, 2008). The RCS itself captures an important phase in NRM ecosystem evolution during the early–middle Miocene, documented in multiple, linked records (Harris et al., 2017a). The RCS paleosol record was described by Retallack (2009) as containing mostly Aridisols but later revision by Harris et al. (2017b) indicates that most soils are quite immature (Entisols or

7

Journal Pre-proof Inceptisols) and thus of limited use for climatic reconstructions. Nonetheless, Retallack (2007, 2009) interpreted the record as showing an expansion of short sod grasslands during the early– middle Miocene in an otherwise sagebrush-dominated habitat characterized by warm-wet climates. As described earlier, local phytolith records and stable carbon isotopic data from the RCS contradict this scenario by indicating the landscape was an open-habitat, C3 grassland with a mosaic of woodland patches and a low abundance of C4 grasses in an increasingly water-

of

limited habitat (Harris et al., 2017b).

ro

Fossil mammals have been collected from throughout the RCS by R. Nichols (1979,

-p

1976), A. Barnosky and colleagues (2007), as well as E. Harris and colleagues from the

re

University of Washington, resulting in large collections at the University of Montana Paleontology Center (UMPC), University of California Museum of Paleontology (UCMP), and

lP

University of Washington’s Burke Museum of Natural History and Culture (UWBM). For each

na

locality, Barnosky et al. (2007) placed taxa into a composite lithostratigraphic sequence (Fig. 2). Vertebrate fossils are not uniformly distributed, rather most derive stratigraphically from

ur

between 146 m and 255 m (referenced to 0 m at the base of the section), and the Snowfence East

Jo

and West field sites are the most productive (Fig. 2). Among the biostratigraphically significant taxa from the RCS are Merychippus and Hypohippus (equids), Aepycamelus (camelid), Diceratherium (rhino), and Merychyus (oreodont; see Barnosky et al., 2007). As equid and rhino remains are particularly abundant throughout the RCS, our sampling of material for isotopic analyses focused on these taxa.

2.3 Carbon isotopic composition of mammalian tooth enamel

8

Journal Pre-proof The carbon isotopic composition of mammalian herbivore tooth enamel reflects the isotopic composition of ingested vegetation, with a known offset due to metabolic enrichment (Cerling and Harris, 1999; Koch, 1998; Lee-Thorp and van der Merwe, 1987; Passey et al., 2005). The carbon isotopic composition of plants is influenced by CO2 fractionation during photosynthesis. Different plants employ different photosynthetic pathways, leading to distinctly different carbon isotopic compositions (Ehleringer et al., 1991; O’Leary, 1988; Vogel, 1978).

of

Modern C3 plants (typically trees, shrubs, and cool-growing-season grasses) have an average

ro

δ13C value of -28.5‰ and range between -20‰ and -37‰, whereas C4 plants (warm-growing-

-p

season grasses and sedges) have an average δ13C value of -13‰ and range between -9‰ to 19‰ (Cerling et al., 1997; Farquhar et al., 1989; Kohn, 2010). Metabolic and biomineralization

re

processes fractionate ingested carbon resulting in 13C enrichment between plant material and

lP

herbivore enamel that is 14.6 ± 0.3‰ for ruminants and c. 14‰ for perissodactyls (Cerling and

na

Harris, 1999; Hoppe et al., 2004; Passey et al., 2005). Ultimately this leads to a correlation between δ13C of diets and enamel (δ13Cenamel) such that animals that eat C3 vs. C4 plants will have

ur

enamel that is relatively depleted vs. enriched in 13C (e.g., Cerling and Harris, 1999). Many

Jo

studies have used this correlation to infer the diets and habitats of ancient mammals based on the carbon isotopic composition of their teeth (e.g., Kohn et al., 2005; MacFadden, 2000; Palmqvist et al., 2003). δ13C values in plants and animals have also been proposed as a means to estimate Mean Annual Precipitation (MAP) in the fossil record. Many studies have observed a negative relationship between δ13C values of C3 plants and MAP (excluding plants living in dense, closed forest understories; e.g., Diefendorf et al., 2010; Kohn, 2010; Stewart et al., 1995), resulting from changes in C3-plant water use efficiency in increasingly arid environments (Farquhar et al.,

9

Journal Pre-proof 1989). Because the δ13C of ingested plants can be calculated from the δ13Cenamel (after correcting for metabolic enrichment), the δ13Cenamel from herbivores in pure C3 ecosystems can be used to infer MAP (Kohn, 2010). This approach is especially valuable in times when C4 plants were not present or abundant, such as the early–middle Miocene of Idaho (Harris et al., 2017b). In this way, MAP may be estimated at specific levels in the RCS and changes to MAP through the sequence may be compared with climate records, e.g. from paleosol, floral, or marine climate

of

proxies. For the RCS, we expect to see increasing δ13Cenamel values through time, indicating a

-p

ro

lowering of MAP (Hprecip1 above).

2.4 Oxygen isotopic composition of mammalian tooth enamel

re

The oxygen isotopic composition of tooth enamel (δ18Oenamel) in large-bodied,

lP

mammalian herbivores is affected by many factors, including the δ18O of ingested water,

na

fractionation of oxygen isotopes between body water and enamel during mineralization, and the physiology of a particular taxon (D’Angela and Longinelli, 1990; Delgado Huertas et al., 1995;

ur

Kohn, 1996; Kohn et al., 1998, 1996; Longinelli, 1984; Luz and Kolodny, 1985). The δ18O of

Jo

ingested water is largely dependent on the source, whether primarily taken in through drinking or through eating plant material (Luz and Kolodny, 1985). Modern observations and modeling results indicate that enamel δ18O in obligate drinkers (e.g., modern horses, rhinos, and bovines) is primarily dependent on the δ18O value of ingested water (Kohn, 1996). In contrast, the enamel of drought-resistant species (e.g., modern goats and gazelles), which derive the majority of water from ingested plant tissues, is relatively enriched in 18O (Ayliffe and Chivas, 1990; Balasse et al., 2003; Kohn, 1996; Levin et al., 2006; Luz et al., 1990). The reason for this is that plant leaf water tends to be enriched in 18O due to fractionation by evaporation from leaf surfaces, and

10

Journal Pre-proof enrichment increases with increasing aridity and insolation (e.g., Yakir, 1992; Yakir et al., 1990). Therefore, animals that rely on plants as their main source of water are expected to have higher δ18Oenamel values relative to obligate drinkers from the same areas. The isotopic composition of surface water reservoirs is primarily controlled by the composition of precipitation plus any evaporation. At mid- to high-latitudes, the δ18O of precipitation (δ18Oprecipitation) correlates strongly with temperature (Dansgaard, 1964; Rozanski

of

and Araguás-Araguás, 1992), where higher temperatures correspond with more positive

ro

δ18Oprecipitation values and lower temperatures correspond with more negative δ18Oprecipitation values

-p

(due to Rayleigh fractionation and local air temperatures; Dansgaard, 1964). These changes in δ18Oprecipitation are reflected in the δ18Oenamel values of large-bodied, obligate drinkers, so the

re

δ18Oenamel values of these animals can be used to estimate local surface and meteoric water

lP

compositions (e.g., for bovines and equines see Kohn and Dettman, 2007; Kohn and Fremd,

na

2007), which can then be used to infer general changes in temperature through time. Tooth enamel forms over the course of a few months to 1–2 years, so any changes in food

ur

or water δ18O values (e.g., seasonal variations in water compositions) or inputs/outputs in

Jo

mammals (e.g., switching between different diets or water sources) during tooth formation will result in isotopic changes in the composition of enamel, resulting in isotopic zoning in teeth (e.g., Fricke et al., 1998; Fricke and O’Neil, 1996). Studies of isotopic zoning in teeth have shown that variation in δ18O values within a tooth closely tracks seasonal changes in the oxygen isotope composition of precipitation and plants, normally with high values in the summer and low values in the winter (e.g., Fricke et al., 1998; Fricke and O’Neil, 1996; Kohn, 1996; Kohn et al., 1998). Seasonality signals are preserved as sinusoidal patterns in δ18Oenamel curves reconstructed from serial sampling along the length of a tooth. However, there are many factors that can dampen

11

Journal Pre-proof climatically-induced isotopic zoning signals within teeth (e.g., Fricke et al., 1998; Kohn and Cerling, 2002). For example, large bodies of water are isotopically averaged relative to seasonal δ18O values of precipitation, whereas non-zero residence times of isotopes mean that the δ18O value of animal body water cannot immediately change as environmental δ18O fluctuates (Kohn and Cerling, 2002). These factors therefore reduce environmental δ18O extremes and induce a time lag between environmental change and compositional change reflected in enamel.

of

Ultimately, enamel of drought-resistant species will experience greater dampening and time lag

ro

in seasonal δ18O signals than water-dependent animals (Kohn and Cerling, 2002). Lastly,

-p

dampening can occur over the course of enamel precipitation and maturation (Passey and

re

Cerling, 2002), although the magnitude of such dampening is far smaller than originally predicted (Trayler and Kohn, 2017). Sampling teeth with relatively thin enamel can also

lP

minimize any dampening effects (Kohn and Cerling, 2002). Despite these caveats, many studies

na

have successfully used fossil teeth to infer changes in climate seasonality (e.g., Dettman et al., 2001; Fricke et al., 1998; Kohn et al., 2015; Nelson, 2005; van Dam and Reichart, 2009; Zanazzi

ur

et al., 2007). For the RCS, we sampled large-bodied ungulates that were likely obligate drinkers;

Jo

thus, we would expect tooth oxygen isotope compositions for RCS ungulates to parallel precipitation values, which would likely increase with increasing temperature (Htemp1 above). We would also expect changes in the amplitude of isotope zoning during any climate changes, such that a higher amplitude would reflect an increase in climate seasonality. Specifically, we predict a lowering in seasonal amplitudes during the MMCO (Htempseason1 above).

2.5 Fidelity of stable isotopes in fossil tooth enamel

12

Journal Pre-proof Many studies have tested the susceptibility of stable isotopes in tooth enamel to diagenetic alteration, primarily through evaluation of tissue heterogeneity/homogeneity, expected isotopic differences among sympatric species, consistency of fractionations among different tissues, and crystallinity. These tests show that, while protein-rich bone and dentine are especially susceptible to isotopic alteration, tooth enamel is quite resistant (see summary of Kohn and Cerling, 2002), probably because it is already highly crystalline, with low porosity and

of

extremely low organic content. Clumped isotope (47) analysis provides an especially sensitive

ro

test of diagenetic alteration, and has shown that fossil equid enamel retains an original biogenic

-p

temperature of precipitation (Eagle et al., 2010). Last, while microbes can process phosphatic

re

tissues and change isotopic compositions (e.g., Blake et al., 1998; Zazzo et al., 2004), fossil tooth enamel commonly retains fine-scale, primary, biogenic features, such as decussate enamel fibers,

lP

incremental lines, surface polish, etc., suggesting minimal post-depositional alteration. The

Jo

3.1 Fossil samples

ur

3. Materials and methods

na

enamel analyzed in this study retained such features.

Isotope compositions were obtained from fossils gathered in the RCS from published vertebrate fossil localities (see Barnosky et al., 2007; Harris et al., 2017b for site information) mostly as float during the summers of 2013 and 2014, as well as specimens from the UMPC collections. Specimens were collected from ten sites within the composite RCS and four nearby sites from Mollie Gulch, which, based on lithological similarities, including presence of the early Miocene unconformity, correlates with the lower 80 m of the RCS (Fig. 2). A total of 71 teeth (59 equids, 12 rhinos; Table S1) were isotopically analyzed and all are housed at either the

13

Journal Pre-proof UWBM or the UMPC (Tables S2-S3). Note that because of the rarity of teeth overall, our sample included a mix of molars and premolars. Fossils at UWBM and UMPC were often identified to genus using fossil reference material at respective museums and newly collected material was, insofar as possible, placed taxonomically using UWBM collections; in addition, identifications were cross-checked with published biostratigraphic ranges (e.g., Janis et al., 1998; Tedford et al., 2004). Because of the limited number and distribution of individual teeth collected from fossil-

of

bearing sites in the RCS, we grouped our data into five time bins to increase sample sizes for

ro

each time interval (Table S1). Time bins represent the following age ranges: TB1 represents c.

-p

22.9 – 21.5 Ma (c. 1.4 million years), TB2 represents c. 21.1 – 19.5 Ma (c. 1.6 million years), TB3 represents c. 19.5 – 18.8 Ma (c. 0.7 million years), TB4 represents c. 18.8 – 17.1 Ma (c. 1.7

lP

model of Harris et al. (2017a; Fig. 2).

re

million years), and TB5 represents c. 17.1 – 15.2 Ma (c. 1.9 million years), following the age

na

Bulk tooth enamel samples were analyzed from 42 Merychippus, three Parahippus, three Miohippus, 11 unidentified equids, and 12 Diceratherium (Table S2). Serial sampling of enamel

ur

was conducted on 15 Equidae and eight Rhinocerotidae specimens (Table S3). Additionally, all

Jo

individuals were assumed to be obligate-drinkers based on the inferred, large body sizes of these organisms and the drinking requirements of their modern relatives (i.e., modern perissodactyls employ hindgut fermentation, which require large amounts of ingested water for digestion; McNab, 2002).

3.2 Stable isotope geochemistry All teeth were sampled according the method established by Koch et al. (1997) for bulk and serial sampling of tooth enamel; preparatory work and analyses were performed in the Stable

14

Journal Pre-proof Isotope Laboratory, Department of Geosciences, Boise State University. Approximately 5–10 mg of enamel powder was collected from each tooth using a 0.5 mm inverted cone carbide drill bit attached to a Dremel™ drill. Bulk enamel samples were collected along a single groove carved down the length of each tooth parallel to the growth axis, whereas serial samples were drilled perpendicular to the growth axis (e.g., see Fricke and O’Neil, 1996), which results in maximally non-smeared isotopic signatures (Trayler and Kohn, 2017). Enamel powder was

of

treated with 30% hydrogen peroxide overnight to remove organics. Samples were then decanted,

ro

washed with deionized water, and soaked in a 1M buffered calcium acetate – acetic acid solution

-p

overnight to remove any diagenetic carbonate. The following day, samples were again decanted,

re

washed repeatedly with deionized water, and dried overnight at 50°C in an oven. After pre-treatment, samples were dissolved overnight in 100% phosphoric acid at 90°C,

lP

and the isotopic composition of the carbonate component of enamel (δ13CCO3, δ18OCO3) was

na

analyzed in a 2010 ThermoFisher Delta V Plus continuous flow mass spectrometer coupled with a GasBench II. Two samples of NIST120c (phosphorite rock) were prepared with each batch of

ur

unknown enamel samples following the guidelines above. Six to eight NBS-18 and NBS-19

Jo

calcite samples were analyzed alongside the unknown enamel samples to verify mass spectrometer performance and reference gas calibrations. All sample gases were repeatedly subsampled 15 times and standard deviations of the replicates were calculated for each sample. All stable isotope analyses presented in this study are from the carbonate component of tooth enamel and are reported in permil relative to the Vienna Standard Mean Ocean Water (V-SMOW) for δ18O, and the Vienna PeeDee Belemnite (V-PDB) for δ13C. Mean values for NIST standards were δ18O = 28.3 ± 0.5‰ (2σ) and δ13C = -6.5 ± 0.2‰ (2σ), which is within ranges reported in

15

Journal Pre-proof (Kohn et al., 2015; 28.5±0.3‰, and -6.55±0.25‰ respectively). Errors include intra-run and dayto-day variations in sample preparation and analysis.

3.3 Modeling MAP Mean Annual Precipitation was calculated following the method outlined in Kohn (2010) and Kohn et al. (2015), which is based on the δ13C of modern C3 plants:

of

𝑀𝐴𝑃 = 10 ^[Δ13 𝐶 − 2.01 + 0.000198 × 𝑒𝑙𝑒𝑣 − 0.0129 × |𝑙𝑎𝑡| / 5.88] − 300

(1)

1000

)

(2)

-p

δ13 𝐶𝑙𝑒𝑎𝑓

Δ13 𝐶 = (δ13 𝐶𝑎𝑡𝑚 − δ13 𝐶𝑙𝑒𝑎𝑓 )/(1 +

ro

where elev is elevation in meters, lat is latitude in degrees, and

re

where δ13Catm and δ13Cleaf are the carbon isotope compositions of atmospheric CO2 (-5.63‰ on

lP

average for 22.9–15.2 Ma; Tipple et al., 2010) and leaves (estimated from δ13Cenamel), respectively. δ13Cenamel values were converted to δ13Cleaf values by correcting for the isotopic

na

enrichment factor between diet and enamel (Cerling and Harris, 1999; Hoppe et al., 2004; Passey

ur

et al., 2005). Values between 14.1±1.0 (2; Cerling and Harris, 1999) to 14.6±0.6 (2; Passey et al., 2005) are commonly assumed for large herbivores. For perissodactyls specifically, the best

Jo

constrained observations imply mean values of 14.3±1.6 to 14.7±1.1. From these data, we assume a value of 14.5±1.0, leading to the transfer function: 14.5

δ13 𝐶𝑙𝑒𝑎𝑓 = (δ13 𝐶𝑒𝑛𝑎𝑚𝑒𝑙 − 14.5)/(1 + 1000)

(3)

For application in the RCS, we calculated MAP from average δ13Cenamel values from all herbivores in each time bin with a pooled uncertainty in δ13C given by the 2 s.e. All herbivores are assumed to have pure C3 diets based on habitat modeling of RCS herbivores (Harris, 2016) and habitat reconstruction indicating a C3, grass-dominated open-habitat in the RCS (Harris et al., 2017b). The altitude of the RCS locality was estimated as c. 2,000 m during the early–middle 16

Journal Pre-proof Miocene. This conservative estimate was established by considering the site’s modern elevation (2,342 m), the fact that nearby isotopic records suggest only modest uplift (<1 km) after 15 Ma in southwestern Montana, and that regional paleoelevation was likely over 2,000 m by the late Miocene (Chamberlain et al., 2012; Hyland et al., 2018). This estimate accounts for the fact that regional uplift was ongoing during and after deposition of the RCS (Chamberlain et al., 2012; McMillan et al., 2002). Importantly, uncertainties in elevation do not strongly affect calculated

of

MAP (Kohn, 2010): a ±500 m uncertainty propagates to a ±30 mm·yr-1 uncertainty in MAP.

ro

Calculations of MAP from tooth enamel-based carbon isotope compositions presume that

-p

an average composition adequately represents plant compositions across a landscape. When only

re

one or two taxa are analyzed, dietary or niche selectivity can bias isotope compositions to higher or lower values, in turn biasing MAP calculations. Equids are often assumed to have occupied

lP

the more open, arid parts of the landscape, thus potentially recording positively biased isotopic

na

ratios. However, recent work suggests that no such significant bias is evident (Drewicz and Kohn, 2018), among large herbivores with presumed differences in feeding habits. Specifically,

ur

this work shows that the mean and median disparity of equid and rhino δ13C values is c. +0.1‰

Jo

and -0.2‰, respectively, with a standard 2-error uncertainty of c. 0.5‰ (data summarized in Drewicz and Kohn, 2018). Consequently, we calculated leaf δ13C and MAP both for the mean δ13C values of rhinos plus equids (0‰ offset), assuming a maximum 0.5‰ lower value for leaf δ13C.

3.4 Modeling meteoric water compositions

17

Journal Pre-proof Calibrations of surface water compositions (δ18OH2O) to tooth enamel compositions are referenced to δ18OPO4 values, so we first converted mean δ18OCO3 values for each time bin to δ18OPO4 values (Iacumin et al., 1996): δ18 𝑂𝑃𝑂4 (𝑉𝑆𝑀𝑂𝑊) = δ18 𝑂𝐶𝑂3 (𝑉𝑆𝑀𝑂𝑊) × 0.98 − 8.5

(4)

Kohn (1996) showed that diet and physiological adaptations can have significant impacts on the δ18O values of biogenic phosphate (δ18OPO4). However, Kohn and Fremd (2007) argued that the

of

strong correlation between δ18O values of modern equid phosphate and local water reflects a

ro

digestive strategy that requires large amounts of water (i.e., equids are quite water dependent).

-p

Because rhinos and equids share plesiomorphic digestive strategies with similar diets and a

re

strong need for water, the dependence of tooth enamel isotope composition on local water is probably similar. Within this context, we estimated local surface water compositions (δ18OH2O)

δ18 𝑂𝑃𝑂4 (𝑉𝑆𝑀𝑂𝑊) −23.5 0.83

(5)

na

δ18 𝑂𝐻2𝑂 (𝑉𝑆𝑀𝑂𝑊) =

lP

using the Kohn and Fremd (2007) equation for modern equids:

ur

where δ18OPO4 is determined from δ18OCO3 via Equation 4. Calculated δ18OH2O values are assumed to approximate meteoric water compositions. Use of the generalized herbivore

Jo

expression of Kohn (1996) yields results to within c. 0.2‰ of Equation 5.

4. Results 4.1 Stable isotope results from bulk enamel sampling Enamel carbon isotope compositions for all taxa ranged from -11.3‰ to -7.2‰ with an average of -9.1 ± 0.2‰ (Table S2). Mean δ13C values are -9.0 ± 0.3‰ for equids and -9.5 ± 0.5‰ for rhinos. MAP estimates based on the δ13C of all taxa and rounded to the nearest 10 18

Journal Pre-proof -1 -1 +360 +250 mm·yr-1 are 130+320 −160 mm·yr (mean ± 2 S.E.) for TB1, 210−190 mm·yr for TB2, 100−120 -1 -1 +400 mm·yr-1 for TB3, 170+330 −170 mm·yr for TB4, and 130−200 mm·yr for TB5 (Table 1), where

errors are based on compositional variability alone. Decreasing inferred leaf δ13C values by 0.5‰ assuming equids inhabit drier niches would increase mean MAP values across all time bins -1 -1 +400 from c. 190+320 −170 mm·yr to almost c. 310−230 mm·yr . Calibration uncertainties would

contribute systematic errors of approximately ±120 mm·yr-1 (Kohn and McKay, 2012). There are

of

no statistically significant differences in MAP across time bins (ANOVA, p > 0.05; Fig. 3).

ro

Bulk enamel oxygen isotope compositions for all taxa ranged from 15.9‰ to 25.5‰

-p

(Table S2). Mean δ18OCO3 values for equids are 19.5 ± 1.6‰ and for rhinos are 19.0 ± 2.9‰.

re

Conversions of δ18OCO3 to δ18OH2O following Kohn and Fremd (2007) yield similar average δ18OH2O values for each time bin of c. -16‰ (Fig. 4). δ18OH2O values averaged through all of the

lP

RCS are -15.4‰ for equids and -16.1‰ for rhinos. δ18OH2O values reconstructed from equids and

na

rhinos show no statistically significant change through time (ANOVA, p > 0.05), although there

ur

are so few data for TB5, an increase (or decrease) in δ18OH2O relative to TB1 is possible.

Jo

4.2 Stable isotope results from serial enamel sampling δ18OCO3 values for the equid teeth range from 16.3‰ to 23.2‰ and show on average 2.7‰ intra-tooth isotope variation (Table S3). δ18OCO3 values for the rhino teeth range from 15.9‰ to 27.6‰ and show on average 2.3‰ intra-tooth isotope variation. The magnitude of δ18O seasonal variation for the majority of sampled teeth was less than c. 3.5‰ with the exception of specimens UMVP 4453E (Merychippus, 4.1‰; Fig. 5P), UMVP 5723 (Merychippus, 3.6‰; Fig. 5E), UMVP 5724 (Merychippus, 3.9‰; Fig. 5D), UMVP 4452A (Merychippus, 4.8‰; Fig. 5L), UWBM 100086 (Rhino, 6.5‰, Fig. 5G), and UWBM 100082

19

Journal Pre-proof (Diceratherium, 3.7‰, Fig. 5W). In addition, equid teeth tended to show sinusoidal seasonal variation more often than rhino teeth. In particular, equid specimens UMVP 5723 (Fig. 5E), UMVP 5724 (Fig. 5D) and UMVP 4452A (Fig. 5L) show clear seasonal patterns, whereas seasonality in UMVP 9426 (Fig. 5U) and UWBM 100066 (Fig. 5M) appears more subdued. The remaining equids do not preserve markedly seasonal patterns or have ‘dampened’ seasonal signals (e.g., Fig. 5A). Of the rhino teeth sampled, specimens UMVP 4482E (Fig. 5Y) and

of

UWBM 100082 (Fig. 5W) appear to preserve a seasonal signal, although the amplitude is small

ro

compared to some of the equids. The remaining rhino specimens are more difficult to interpret

-p

and appear to preserve highly dampened seasonal signals. Additionally, there are no significant

re

changes over time in the mean magnitude of δ18O seasonal variation recorded by either equids or

lP

rhinos (ANOVA; p > 0.05).

na

5. Discussion

5.1 Mean Annual Precipitation estimates for the RCS

ur

Modern MAP for Leadore, Idaho (located 13 miles south of the RCS) is 236 mm·yr-1

Jo

(U.S. climate data, 2016), which is within the range of errors estimated for MAP through the early–middle Miocene in the RCS (Fig. 3). This suggests that MAP in the RCS was roughly the same as today (c. 300 mm·yr-1) but could have been higher (c. 500 mm·yr-1). In principle, there might have been some C4 consumption by the herbivores inhabiting this area. However, δ13Cenamel records from North America and around the world do not suggest definitive C4 plant consumption until the late Miocene–Pliocene (Cerling et al., 1997); therefore, we suggest these low MAP estimates are indicative of a relatively dry climate in the RCS. Likely the RCS was characterized by a semi-arid environment during the early–middle Miocene, which is consistent

20

Journal Pre-proof with phytolith vegetation reconstructions suggesting dry, C3-dominated, open habitats (Harris et al., 2017b). Additionally, our isotope-based MAP proxy-record, which reflects a nearly invariant (δ13Cenamel – δ13Catm) further suggests that the observed decrease in diatoms and C3, waterloving PACMAD grasses through the RCS are not a result of a change in Mean Annual Precipitation, rejecting our Hprecip1. Instead, we propose these changes signal a decrease in permanent lakes and standing water on the landscape linked to a change in depositional regime

of

in the basin. That is, these data may represent a transition from low-energy, floodplain deposition

ro

to alluvial fan deposition in increasingly upland environments, as suggested by sediment grain

-p

size coarsening up-section through the RCS (Harris et al., 2017b).

re

In contrast to our new precipitation estimates, independent MAP estimates based on the chemical index of alteration without potash (CIA-K index) of paleosol Bt horizons in the RCS

lP

generally suggest higher MAP for the RCS, namely ranging from c. 345 ± 182 mm·yr-1 (within

na

the range of our data) to 883 ± 182 mm·yr-1 (outside the range of our data) with a marked increase between c. 22.5 and 21.5 Ma (based on revised age model from Harris et al., 2017a)

ur

(Retallack, 2007). Unfortunately, the most comprehensive re-analysis of the correspondence

Jo

between soil chemistry and MAP (Stinchcomb et al., 2016), relies on concentration of ZrO2, in addition to the more standard oxides (SiO2, Al2O3, etc.). The analytical methods used by Retallack (2007) did not include ZrO2 (Retallack, pers. comm., 2018), so recalculations using the spline approach of Stinchcomb et al. (2016) are not possible. As discussed in Harris et al. (2017b), we believe that the significantly wetter conditions inferred from paleosol chemistry are inaccurate for two main reasons: first, the CIA-K index methodology does not apply to soils formed on significant topographic relief (e.g., hillslope or montane soils; Sheldon et al., 2002). As discussed above, at levels above 70 m in the composite

21

Journal Pre-proof section, the RCS depositional environment represents an increasingly steep environment (e.g., alluvial fan deposit; Harris et al., 2017b). This means that the majority of RCS paleosols were likely formed in or near hillslope settings, which are inappropriate for paleosol MAP reconstruction. Second, the CIA-K index is based on the modern relationship between degree of soil weathering and MAP for well-developed Alfisols, Mollisols, and Ultisols. Such a correlation has not been established for poorly developed soils such as Entisols and Inceptisols, which are

of

the abundant paleosol types preserved throughout the RCS. For these reasons, we challenge

ro

previous MAP estimates based on paleosols (Retallack, 2007) and propose that our isotope MAP

-p

estimates are a more accurate reflection of RCS climate during the early–middle Miocene than

re

are previous MAP estimates based on paleosols (Retallack, 2007).

lP

5.2 Meteoric water compositions and temperature estimates in the RCS

na

The nearly indistinguishable average δ18O values of RCS equids and rhinos suggests these taxa had similar water dependencies (i.e., water-dependent digestive physiologies) during

ur

the early–middle Miocene and, thus, did not fractionate 18O significantly differently. This

Jo

supports our earlier view that, for our data, it is appropriate to model local water compositions for both taxa using the modern correlation equation between horse phosphate δ18O and local water compositions (Kohn and Fremd, 2007). Modern meteoric water δ18O values are -17.5‰ VSMOW near the RCS in Heise, Idaho (Coplen and Kendall, 2000). Meteoric water in the RCS is estimated to have been -15.6‰ on average across all taxa during the early–middle Miocene (Table S2), or c. 2‰ higher than today, and did not change significantly during this time (Fig. 4). The difference between modern and Miocene water compositions is slightly larger when ice volume is taken into consideration:

22

Journal Pre-proof modern ice volume has increased seawater δ18O by c. 0.5‰ relative to the Miocene (Pekar et al., 2006), so relative to a modern ocean, the δ18O of Miocene precipitation would have been c. 15.1‰. Relatively high δ18O (for the interior western U.S.) during the Miocene could in part reflect topographic evolution of western North American landscapes. Basin and range extension was occurring during the early-middle Miocene (Atwater and Stock, 1998; Chamberlain et al., 2012; McMillan et al., 2002) and the development of complex basin systems could have resulted

of

in different rain-out patterns that led to depleted δ18OH2O values. Alternatively, or additionally,

ro

higher temperatures during the Miocene as well as changes in source water could have

-p

contributed to higher local δ18OH2O values. A lack of independent, local paleotemperature

re

estimates precludes explicit testing of this hypothesis; however, given the wealth of terrestrial and marine records across the globe showing higher temperatures during the MMCO (e.g.,

lP

Hobbs and Parrish, 2016; Holbourn et al., 2015, 2007; Mix and Chamberlain, 2014; Mosbrugger

na

et al., 2005; Utescher et al., 2007; Zachos et al., 2001), increased MAT may be the most likely explanation for high δ18O values in the NRM.

ur

Regional differences in δ18OH2O in the Pacific Northwest also provide insight into

Jo

potential topographic barriers during the early–middle Miocene. For example, estimated δ18OH2O for the RCS in eastern Idaho is c. 2.2 – 5.6‰ lower than δ18OH2O estimated from the Mascall formation of central Oregon c. 15.1 Ma (-10 to -13.4‰; Drewicz and Kohn, 2018; Kohn and Fremd, 2007). The continental effect, in which isotope values of precipitation tend to decrease away from coastlines (Dansgaard, 1964), could explain 1–2‰ of this difference. The remaining disparity of 1.2 – 3.6‰ likely indicates higher elevations either at the RCS or between the RCS and central Oregon (or both) during the early–middle Miocene. Central Oregon likely had lower elevations compared with these more inland sites during deposition of the RCS as indicated by

23

Journal Pre-proof data from Drewicz and Kohn (2018) which imply consistently high median δ18OH2O estimates of -13.4‰ for the John Day Formation (c. 28 Ma; Retallack, 2004) and -13.3‰ for the Mascall Formation (15.1 Ma; Maguire et al., 2018). In addition to stable isotope data, fossil plants from the same strata can be used to glean information about climate during the early–middle Miocene. Harris et al. (2017b) reported consistent, albeit low (< 2%) relative abundances of palm phytoliths preserved throughout the

of

RCS. Because modern palms (especially seedlings) are typically frost-intolerant, they can serve

ro

as indicators of minimum winter temperatures (e.g., Archibald et al., 2014; Greenwood and

-p

Wing, 1995; Reichgelt et al., 2018). The natural distribution of modern palms is usually limited

re

to regions with Cold Month Mean Temperature (CMMT) >5ºC, although some palms can withstand brief intervals of freezing (Greenwood and Wing, 1995; Larcher and Winter, 1981;

lP

Reichgelt et al., 2018). This tolerance is further reduced under high atmospheric CO2

na

concentrations (>800 ppm increases CMMT estimates by +1.5ºC to +3ºC; Royer et al., 2002). Compilations of Mean Annual Temperature (MAT) and CMMT climate profiles for a range of

ur

modern cold tolerant palm clades suggests that the lowest MAT threshold for palms is c. 11ºC

Jo

and the CMMT is typically >5ºC (Archibald et al., 2014; Reichgelt et al., 2018). Assuming similar climatic tolerances of palms in the early–middle Miocene, MAT and CMMT for the RCS were likely above 11ºC and 5ºC, respectively. Consistently warm MAT in the RCS is also supported by salinization index-based MAT estimates from paleosols which place MAT at c. 10.3°C during the early–middle Miocene in the RCS (Retallack, 2007). All of these estimates are much warmer than current MAT and CMMT for Leadore, Idaho (3.8°C and -8.4°C, respectively; U.S. climate data, 2016), supporting a scenario with considerably warmer early-middle Miocene

24

Journal Pre-proof climates in the NRM. Nonetheless, we find no support for an upward trend in temperatures during this time (our Htemp1).

5.3 Seasonality Serial sampling of equid and rhino teeth potentially constrain climatic temperature seasonality at the RCS during the early–middle Miocene. A temperature signal reflects the

of

correspondence between monthly temperature and δ18O observed in modern datasets (e.g.,

ro

Rozanski et al., 1993). Our oxygen isotopic data indicate that there was seasonality in either

-p

temperature or source of precipitation in RCS, but that this did not change substantially from the early to middle Miocene, allowing us to reject our Htempseason1.

re

The differences in magnitude and shape of seasonal δ18O variation between equids and

lP

rhinos could be due to taxon-specific differences in tooth mineralization or migration. Large-

na

bodied herbivores that migrate (e.g., modern horses) typically consume different water sources and can exploit temporally available resources throughout the year, which results in lower intra-

ur

tooth variability in δ18O and higher intra-tooth variability in δ13C compared with non-migratory

Jo

herbivores (e.g., Feranec et al., 2009; Julien et al., 2012). Non-migratory, resident herbivores are known to display high δ18O variation (associated with seasonal changes in temperature or water source or both) with little to no variability in δ13C due to consistency of available resources in a given geographic area (Bryant et al., 1996; Hoppe et al., 2004; Julien et al., 2012). RCS herbivores display intra-tooth isotopic variability characterized by high δ18O variation and no δ13C variation (Harris, 2016) most similar to what we expect from non-migratory herbivores. Therefore, we believe it unlikely that migration was a main driving force for δ18O variation between equids and rhinos. High variation in strontium isotopes (87Sr/86Sr) can be associated

25

Journal Pre-proof with migratory behavior (Hoppe et al., 1999) if 87Sr/86Sr varies substantially across an area. Although we did not measure strontium isotope compositions of RCS herbivore teeth (in part because of concerns about diagenetic alteration; Hinz and Kohn, 2010; Kohn et al., 1999), future analysis might provide additional support to our hypothesis that RCS herbivores were nonmigratory and that we can use their teeth as indicators of local climate. Another reason for heterogeneity in seasonal δ18O range among sampled specimens could

of

be that herbivores were drinking water from different sources, each with different compositional

ro

averages. For example, we would expect to see little to no seasonal changes in the enamel of

-p

herbivores that drank water from (or ate plants growing nearby) lakes or other large bodies of

re

water, as these tend to have less isotopic variability than seasonal precipitation (see discussion in Kohn and Cerling, 2002). Additionally, as mentioned earlier, due to the scarcity of fossil teeth

lP

recovered from the RCS, we could not limit our analysis to a single tooth position (e.g., M3)

na

among individuals sampled. Because different teeth erupt and form over the course of a few months to about 1 year, each tooth records a slightly different (although at times overlapping)

ur

sinusoidal seasonal isotopic signature (e.g., Fricke et al., 1998; Fricke and O’Neil, 1996).

Jo

Therefore, it is possible that the differences in δ18O variation that we observe between samples could reflect different molar positions, which preserve different portions of the seasonal δ18O cycles over which the enamel formed (Kohn et al., 1998).

5.4 Did climatic events in the RCS coincide with Miocene faunal change in the NRM? Our climate data indicate relative stasis in MAT, MAP, and possibly temperature seasonality in the RCS during the early–middle Miocene. Together with the inferred, relatively stable plant community composition and habitat structure throughout the RCS, our data suggest

26

Journal Pre-proof that many aspects of the environment remained constant in the RCS during this time, even as the depositional regime appears to have shifted (Harris et al., 2017b). This environmental reconstruction for the RCS basin implies that the major faunal changes recorded for the early– middle Miocene of the NRM were not likely driven by changes in local climate and vegetation. Instead we propose that these faunal changes could have been driven by increased immigration or even increased endemism due to tectonic factors in this mountainous region (Barnosky and

of

Carrasco, 2002; Woodburne and Swisher, 1995). Faunal trends are addressed further in Harris

-p

ro

(2016).

re

5.5 Regional climatic differences leading up to and during the MMCO Our record of relatively stable climates leading up to and across the MMCO contrasts

lP

with global patterns inferred from marine cores, which indicate overall warming, and commonly

na

cited terrestrial data pointing to increased precipitation in the window of time captured by the RCS (e.g., Hinojosa and Villagrán, 2005; Mosbrugger et al., 2005; Retallack, 2007; Zachos et

ur

al., 2001). Our seemingly anomalous data raise the question, how uniform was the manifestation

Jo

of this climatic event at other sites of comparable (mid-)latitude? To answer this question, we compiled available records of climate (MAT, MAP) and vegetation reconstructions spanning 26 mid-latitude (30-50º N) regions across the globe for which there is information (semiquantitative or quantitative, see Table S4 and Supplementary material) for the period leading up to and across the MMCO (Fig. 6 and Table S4). We focused primarily on (A) relative direction of change, that is, from warmer to cooler (or vice versa) or wetter to drier (or vice versa), regardless of the absolute background values or the amplitude of the change, in order to study the overall climatic trends. However, the relatively narrow latitudinal band allowed us to also

27

Journal Pre-proof meaningfully compare reconstructions of (B) absolute values—and changes through time—in both temperature and precipitation. Our relatively coarse surveys indicate that the regional manifestation of global climates varied substantially at mid-latitudes. A common trend (54% of the records) is that warm and wet climates were either maintained leading into and during the MMCO or shifted to warm and wet climates during the MMCO, followed by cooler and dryer climates after the MMCO, broadly

of

consistent with global temperature trends (e.g., Miller et al., 1987; Shackleton and Kennett,

ro

1975; Zachos et al., 2001). This pattern is evident on all continents studied; however, several

-p

other patterns occur, sometimes in areas in relative proximity, suggesting that regional or local

re

factors played a major role in climate evolution at the landscape level. For example, the timing of post-MMCO cooling appears to vary, with several studies (25%) indicating continued warm

lP

climates until at least c. 10 Ma (Fig. 6). Precipitation records are even more variable, with some

na

studies (23%) showing aridification leading into the MMCO or either continually dry (17%) or wet (13%) climates throughout the Miocene. For most areas, there are no consistent trends in the

ur

geographic distribution of these patterns. An example is China, where studies of different, or

Jo

even the same, proxies provide contradictory results over short distances (Fig. 6). This inferred, intra-continental climatic variation may suggest that regional, or even local factors were important for determining how global climatic changes were manifested at these smaller spatial scales. It seems likely, for example, that the contrast between the dry climates reconstructed herein for RCS and the much higher rainfall regime of the Pacific Northwest during the MMCO (e.g., Leopold and Denton, 1987; Wolfe, 1994) (Fig. 6), reflects the continental effect, as discussed above. However, there are other explanations that may interfere with robust interpretations of the diverse climate records. First, it may be that the discrepancies

28

Journal Pre-proof in part reflect proxy-specific biases, whereby different sources of data record distinctive parts of the landscape that are responding independently to regional climate change. In Turkey, for example, the contrasting climate inferences (wet throughout the early-early late Miocene vs. relatively dry throughout the same time period; Fig. 6), may be the result of such differences. Thus, at least some of the phytoliths assemblages studied are from “upland” environments, whereas macrofossils and palynofloras tend to reflect, respectively, local lakes/swamps or a

of

vegetation across a larger region, respectively (Akgün et al., 2007; Akkiraz et al., 2011; Kayseri-

ro

Özer, 2017; Strömberg et al., 2007). In addition, unlike NRL and physiognomic approaches to

-p

palynofloras and macrofloras, ‘traditional’ phytolith assemblage analysis does not provide

re

detailed, quantitative estimates of climate but rather can be used to constrain temperature- and precipitation inferences based on other, more climatically informative proxies (e.g., isotopic data,

lP

the phytolith undulation index; Dunn et al., 2015; Harris et al., 2017b; Kohn et al., 2015) or point

na

to broad trends in openness (often linked to rainfall; Strömberg and McInerney, 2011). Given this uncertainty in climate inferences from phytolith assemblage data, the phytolith-based climate

ur

reconstruction from Turkey (Strömberg et al., 2007), needs to be further refined and tested using,

Jo

for example, isotopic data. Another problem with directly comparing across studies is that several of them by necessity sample from across a larger geographic range (e.g., Fox and Koch, 2004; Strömberg et al., 2007; Syabryaj et al., 2007), thereby potentially confounding temporal and spatial patterns. Because of these uncertainties in various inter-proxy biases (e.g., spatial averaging and location, temporal averaging, information provided) and clear gaps in our understanding of MMCO climates at mid-latitudes (e.g., Australia) we refrain from trying to further infer what factors may have influenced regional differences across mid-latitudes, and from summarizing

29

Journal Pre-proof patterns in particular regions (e.g., China, southern South America). Likewise, we cannot at this stage conduct global tests for vegetation responses to regional climate changes during the leadup to and across the MMCO. This is because many of the existing climate inferences for the Miocene (Table S4) use nearest living relative arguments, making it hard to compare vegetation and climate reconstructions without risking circularity. For example, abundant grasses and shrubs signal open habitats (because they are low-statured), but they also indicate dry climates

of

simply because habitat openness is linked to aridity (e.g., Dunn et al., 2015). Therefore, not

ro

surprisingly, in our dataset, which consists dominantly of paleobotanical proxy data, wetter time

-p

intervals are associated with forested habitats, whereas dryer intervals are associated with more

re

heterogeneous to open, grass-dominated habitats (Table S4). This observation emphasizes the need to, as we have, use independent records of climate to test the link between climate change

lP

and vegetation alterations in the fossil record (see also Kohn et al., 2015).

na

The complex climatic pattern evident in Figure 6 highlights how crucial it is to study many different regions to gain a full understanding of the local-regional impacts of climate

ur

change in deep time. Specifically, data from the RCS provide insight into the response of a high-

Jo

elevation site located in the continental interior to global warming during the MMCO. Given that inland, high-elevation sites are generally under-represented in the fossil record, the RCS is an important field area that has the potential to provide a unique perspective on the climate, ecology, and evolution of ecosystems during the early–middle Miocene.

6. Conclusions Stable isotopes from herbivore teeth suggest the RCS was characterized by a semi-arid, seasonal climate during the early–middle Miocene with warmer Mean Annual Temperature

30

Journal Pre-proof (>7°C) than central-eastern Idaho today. These local climate conditions remained stable and essentially unchanged during the lead-up to peak warming during the middle Miocene. Our findings contrast with previously held ideas that the northern Rocky Mountains, and western North America in general, shifted to warm and wet conditions during the middle Miocene, and suggest that this region of the Northern Rocky Mountains was climatically buffered from the long-term effects of global climate forcing leading into the MMCO. Geographically variable

of

manifestations of global climate change are unsurprising in light of current knowledge of how

ro

the ongoing global warming differentially affects local ecosystems across the Earth. Nonetheless,

-p

the discrepancy between local/regional and global climatic trends in RCS and other areas

re

highlights the importance of investigating geologic intervals of extreme global climate change at local scales instead of relying on studies at the regional or global scale. This is especially vital

lP

when testing hypotheses about biotic responses to potential abiotic drivers during major climate

na

change. Out of necessity, paleontologists often rely on comparing patterns of vegetation or faunal change at the basin level to global climate curves, however, such an approach can lead to

ur

erroneous conclusions about ecological or evolutionary processes involved in biotic turnover.

Jo

The upland environment of the RCS is not typically represented in paleontological surveys; therefore, the RCS adds a key piece to our understanding of ecological and evolutionary change during long-term climatic alteration.

31

Journal Pre-proof Acknowledgements We would like to thank C. Bitting, A. Padgett, and A. Vilhena for field assistance with fossil collecting during this project. Additional thanks to S. Evans and R. Traylor for assistance with tooth sampling and isotope analyses as well as K. Moore and G. Stanley for access to fossil collections at the University of Montana Paleontology Center. Thanks to C. Crifò, E. Hyland, and D. Vilhena for comments and suggestions on the manuscript. Lastly, thanks to Idaho State

of

Lands, the Bureau of Land Management, and the National Forest Service for administering

ro

permits for collecting vertebrate fossils. Funding for this project was provided by National

-p

Science Foundation grants number EAR-1024681, EAR-1253713, and EAR-1349749 to CAES,

re

EAR-1349749 to MJK, an Evolving Earth Foundation grant to EBH, a UW Biology Iuvo Award, Sargent Award, and Frye-Hotson-Rigg Writing Fellowship to EBH, a Geological Society of

lP

America Grant in Aid of Research to EBH, as well as funding from the Burke Museum of

Jo

Data Availability

ur

na

Natural History and Culture.

Datasets related to this article can be found at http://dx.doi.org/10.17632/h5w65yh7t4.1, an open-source online data repository hosted at Mendeley Data (Harris et al., 2019).

32

Journal Pre-proof References Akgün, F., Kayseri, M.S., Auiraz, M.S., 2007. Palaeoclimatic evolution and vegetational changes during the Late Oligocene-Miocene period in Western and Central Anatolia (Turkey). Palaeogeogr. Palaeoclimatol. Palaeoecol. 253, 56-+. Akkiraz, M., Akgün, F., Utescher, T., Bruch, A., Mosbrugger, V., 2011. Precipitation gradients during the Miocene in Western and Central Turkey as quantified from pollen data.

of

Palaeogeogr. Palaeoclimatol. Palaeoecol. 304, 276–290.

ro

Archibald, S.B., Morse, G.E., Greenwood, D.R., Mathewes, R.W., 2014. Fossil palm beetles

-p

refine upland winter temperatures in the Early Eocene Climatic Optimum. Proc. Natl. Acad.

re

Sci. U. S. A. 111, 8095–8100. doi:10.1073/pnas.1323269111 Atwater, T., Stock, J., 1998. Pacific-North America plate tectonics of the Neogene southwestern

lP

United States–An update. Int. Geol. Rev. 40, 375–402.

na

Ayliffe, L.K., Chivas, A.R., 1990. Oxygen isotope composition of the bone phosphate of Australian kangaroos: Potential as a palaeoenvironmental recorder. Geochim. Cosmochim.

ur

Acta 54, 2603–2609. doi:10.1016/0016-7037(90)90246-H

Jo

Badgley, C., Barry, J.C., Morgan, M.E., Nelson, S. V, Behrensmeyer, A.K., Cerling, T.E., Pilbeam, D., 2008. Ecological changes in Miocene mammalian record show impact of prolonged climatic forcing. Proc. Natl. Acad. Sci. U. S. A. 105, 12145–9. doi:10.1073/pnas.0805592105 Balasse, M., Smith, A.B., Ambrose, S.H., Leigh, S.R., 2003. Determining sheet birth seasonality by analysis of tooth enamel oxygen isotope ratios: the Late Stone Age site of Kasteelberg (South Africa). J. Archaeol. Sci. 30, 205–215. Barnosky, A.D., 2001. Distinguishing the effects of the Red Queen and Court Jester on Miocene

33

Journal Pre-proof mammal evolution in the northern Rocky Mountains. J. Vertebr. Paleontol. 21, 172–185. Barnosky, A.D., Bibi, F., Hopkins, S.S.B., Nichols, R., 2007. Biostratigraphy and magnetostratigraphy of the Mid-Miocene Railroad Canyon Sequence, Montana and Idaho, and age of the Mid-Tertiary Unconformity west of the Continental Divide. J. Vertebr. Paleontol. 27, 204–224. Barnosky, A.D., Carrasco, M.A., 2002. Effects of Oligo-Miocene global climate changes on

of

mammalian species richness in the northwestern quarter of the USA. Evol. Ecol. Res. 4,

ro

811–841.

-p

Barnosky, A.D., Hadly, E. A., Bell, C.J., 2003. Mammalian response to global warming on

re

varied temporal scales. J. Mammal. 84, 354–368. doi:10.1644/15451542(2003)084<0354:MRTGWO>2.0.CO;2

lP

Blake, R.E., O'Neil, J.R., Garcia, G.A., 1998. Effects of microbial activity on the 18O of

1516-1531.

na

dissolved inorganic phosphate and textural features of synthetic apatites. Am. Mineral 83,

ur

Blois, J.L., Hadly, E.A., 2009. Mammalian Response to Cenozoic Climatic Change. Annu. Rev.

Jo

Earth Planet. Sci. 37, 181–208. doi:10.1146/annurev.earth.031208.100055 Böhme, M., 2003. The Miocene Climatic Optimum: evidence from ectothermic vertebrates of Central Europe. Palaeogeogr. Palaeoclimatol. Palaeoecol. 195, 389–401. doi:10.1016/S0031-0182(03)00367-5 Bryant, J.D., Froelich, P.N., Showers, W.J., Genna, B.J., 1996. A tale of two quarries: biologic and taphonomic signatures in the oxygen isotope composition of tooth enamel phosphate from modern and Miocene equids. Palaios 11, 397–408. Cerling, T.E., Harris, J.M., 1999. Carbon isotope fractionation between diet and bioapatite in

34

Journal Pre-proof ungulate mammals and implications for ecological and paleoecological studies. Oecologia 120, 347–363. Cerling, T.E., Harris, J.M., Macfadden, B.J., Leakey, M.G., Quade, J., Eisenmann, V., Ehleringer, J.R., 1997. Global vegetation change through the Miocene/Pliocene boundary. Nature 389, 153–158. Chamberlain, C.P., Mix, H.T., Mulch, A., Hren, M.T., Kent-Corson, M.L., Davis, S.J., Horton,

of

T.W., Graham, S.A., 2012. The Cenozoic climatic and topographic evolution of the western

ro

North American Cordillera. Am. J. Sci. 312, 213–262. doi:10.2475/02.2012.05

-p

Chaney, R.W., Axelrod, D.I., 1959. Miocene floras of the Columbia Plateau, 617th ed. Carnegie

re

Institute Washington Publication.

Chen, S.T., Smith, S.Y., Sheldon, N.D., Strömberg, C.A.E., 2015. Regional-scale variability in

lP

the spread of grasslands in the late Miocene. Palaeogeogr. Palaeoclimatol. Palaeoecol. 437,

na

42–52. doi:10.1016/j.palaeo.2015.07.020

Coplen, T.B., Kendall, C., 2000. Stable isotope and oxygen isotope ratios for selected sites of the

ur

U.S. Geological Survey’s NASQAN and Benchmark surface-water networks.

Jo

Cotton, J.M., Sheldon, N.D., Strömberg, C.A.E., 2012. High-resolution isotopic record of C4 photosynthesis in a Miocene grassland. Palaeogeogr. Palaeoclimatol. Palaeoecol. 337–338, 88–98. doi:10.1016/j.palaeo.2012.03.035 D’Angela, D., Longinelli, A., 1990. Oxygen isotopes in living mammal’s bone phosphate: Further results. Chem. Geol. 86, 75–82. Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus 16, 436–468. Delgado Huertas, A., Lacumin, P., Stenni, B., Snachez Chillon, B., Longinelli, A., 1995. Oxygen isotope variations of phosphate in mammalian bone and tooth enamel. Geochim.

35

Journal Pre-proof Cosmochinica Acta 59, 4299–4305. Dettman, D.L., Kohn, M.J., Quade, J., Ryerson, F.J., Ojha, T.P., Hamidullah, S., 2001. Seasonal stable isotope evidence for a strong Asian monsoon throughout the past 10.7 m.y. Geology 29, 31–34. doi:10.1130/0091-7613(2001)029<0031:SSIEFA>2.0.CO;2 Diefendorf, A.F., Mueller, K.E., Wing, S.L., Koch, P.L., Freeman, K.H., 2010. Global patterns in leaf 13C discrimination and implications for studies of past and future climate. Proc. Natl.

of

Acad. Sci. 107, 5738–5743. doi:10.1073/pnas.0910513107

ro

Drewicz, A.E., Kohn, M.J., 2018. Stable isotopes in large herbivore tooth enamel capture a mid-

-p

Miocene precipitation spike in the interior Pacific Northwest. Palaeogeogr. Palaeoclimatol.

re

Palaeoecol. 495, 1–12.

Dunn, R.E., Strömberg, C.A.E., Madden, R.H., Kohn, M.J., Carlini, A.A., 2015. Linked canopy,

na

doi:10.1126/science.1260947

lP

climate, and faunal change in the Cenozoic of Patagonia. Science. 347, 258–261.

Eagle, R.A., Schauble, E.A., Tripati, A.K., Tutken, T., Hulbert, R.C., Eiler, J.M., 2010. Body

ur

temperatures of modern and extinct vertebrates from 13C-18O bond abundances in bioapatite.

Jo

Proc. Natl. Acad. Sci. U. S. A. 107, 10377-82. Edwards, E.J., Osborne, C.P., Strömberg, C.A.E., Smith, S.A., Bond, W.J., Christin, P.-A., Cousins, A.B., Duvall, M.R., Fox, D.L., Freckleton, R.P., Ghannoum, O., Hartwell, J., Huang, Y., Janis, C.M., Keeley, J.E., Kellogg, E.A., Knapp, A.K., Leakey, A.D.B., Nelson, D.M., Saarela, J.M., Sage, R.F., Sala, O.E., Salamin, N., Still, C.J., Tipple, B., 2010. The origins of C4 grasslands: integrating evolutionary and ecosystem science. Science. 328, 587–91. doi:10.1126/science.1177216 Ehleringer, J.R., Sage, R.F., Flanagan, L.B., Pearcy, R.W., 1991. Climate change and the

36

Journal Pre-proof evolution of C4 photosynthesis. Trends Evol. Ecol. 6, 95–99. Farquhar, G.D., Ehleringer, J.R., Hubick, K.T., 1989. Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 40, 503–537. Feranec, R.S., Hadly, E.A., Paytan, A., 2009. Stable isotopes reveal seasonal competition for resources between late Pleistocene bison (Bison) and horse (Equus) from Rancho La Brea, southern California. Palaeogeogr. Palaeoclimatol. Palaeoecol. 271, 153–160.

of

doi:10.1016/j.palaeo.2008.10.005

ro

Fields, R.W., Rasmussen, D.L., Tabrum, A.R., Nichols, R., 1985. Cenozoic rocks of the

-p

intermontane basins of western Montana and eastern Idaho: A summary. Cenozoic

re

Paleogeography of the West-Central United States 9–36. Figueirido, B., Janis, C.M., Perez-Claros, J.A., De Renzi, M., Palmqvist, P., 2012. Cenozoic

lP

climate change influences mammalian evolutionary dynamics. Proc. Natl. Acad. Sci. U. S.

na

A. 109, 722–727. doi:10.1073/pnas.1110246108 Flower, B.P., Kennett, J.P., 1994. The Middle Miocene climatic transition: East Antarctic ice

ur

sheet development, deep ocean circulation and global carbon cycling. Palaeogeogr.

Jo

Palaeoclimatol. Palaeoecol. 108, 537–555. Fox, D.L., Koch, P.L., 2004. Carbon and oxygen isotopic variability in Neogene paleosol carbonates: constraints on the evolution of the C4-grasslands of the Great Plains, USA. Palaeogeogr. Palaeoclimatol. Palaeoecol. 207, 305–329. Fricke, H.C., Clyde, W.C., O’Neil, J.R., Gingerich, P.D., 1998. Evidence for rapid climate change in North America during the latest Paleocene thermal maximum: oxygen isotope compositions of biogenic phosphate from the Bighorn Basin (Wyoming). Earth Planet. Sci. Lett. 160, 193–208.

37

Journal Pre-proof Fricke, H.C., O’Neil, J.R., 1996. Inter- and intra-tooth variation in the oxygen isotope composition of mammalian tooth enamel phosphate: implications for palaeoclimatological and palaeobiological research. Palaeogeogr. Palaeoclimatol. Palaeoecol. 126, 91–99. doi:10.1016/S0031-0182(96)00072-7 Graham, A., 1999. Late Cretaceous and Cenozoic history of North American vegetation. Oxford University Press, New York.

ro

gradients. Geology 23, 1044–1048. doi:10.1130/0091-

of

Greenwood, D.R., Wing, S.L., 1995. Eocene continental climates and latitudinal temperature

-p

7613(1995)023<1044:ECCALT>2.3.CO;2

re

Harris, E.B., 2016. Effects of the mid-Miocene Climatic Optimum on ecosystem structure and plant-animal interactions: a phytolith and stable isotope perspective. Ph.D. dissertation.

lP

University of Washington, Seattle, WA.

na

Harris, E.B., Strömberg, C.A.E., Sheldon, N.D., Smith, S.Y., Ibañez-Mejia, M., 2017a. Revised chronostratigraphy and biostratigraphy of the early-middle Miocene Railroad Canyon

ur

section of central-eastern Idaho, USA. Bull. Geol. Soc. Am. 129, 1241–1251.

Jo

doi:10.1130/B31655.1

Harris, E.B., Strömberg, C.A.E., Sheldon, N.D., Smith, S.Y., Vilhena, D.A., 2017b. Vegetation response during the lead-up to the middle Miocene warming event in the Northern Rocky Mountains, USA. Palaeogeogr. Palaeoclimatol. Palaeoecol. 485, 401–415. doi:10.1016/j.palaeo.2017.06.029 Hinojosa, L.F., Villagrán, C., 2005. Did South American mixed paleofloras evolve under thermal equability or in the absence of an effective Andean barrier during the Cenozoic? Palaeogeogr. Palaeoclimatol. Palaeoecol. 217, 1–23. doi:10.1016/j.palaeo.2004.11.013

38

Journal Pre-proof Hinz, E., Kohn, M.J., 2010. The effect of tissue structure and soil chemistry on trace element uptake in fossils. Geochim. Cosmochim. Acta 74, 3213–3231. Hobbs, K.M., Parrish, J.T., 2016. Miocene global change recorded in Columbia River basalthosted paleosols. Geol. Soc. Am. Bull. 128, 1543–1554. Holbourn, A., Kuhnt, W., Kochhann, K.G.D., Andersen, N., Sebastian Meier, K.J., 2015. Global perturbation of the carbon cycle at the onset of the Miocene Climatic Optimum. Geology

of

43, 123–126. doi:10.1130/G36317.1

ro

Holbourn, A., Kuhnt, W., Schulz, M., Flores, J.A., Andersen, N., 2007. Orbitally-paced climate

-p

evolution during the middle Miocene “Monterey” carbon-isotope excursion. Earth Planet.

re

Sci. Lett. 261, 534–550. doi:10.1016/j.epsl.2007.07.026 Hoppe, K.A., Amundson, R., Vavra, M., McClaran, M.P., Anderson, D.L., 2004. Isotopic

lP

analysis of tooth enamel carbonate from modern North American feral horses: implications

na

for paleoenvironmental reconstructions. Palaeogeogr. Palaeoclimatol. Palaeoecol. 203, 299– 311.

ur

Hoppe, K.A., Koch, P.L., Carlson, R.W., Webb, S.D., 1999. Tracking mammoths and

Jo

mastodons: reconstruction of migratory behavior using strontium isotope ratios. Geology 27, 439–442.

Hyland, E.G., Sheldon, N.D., Smith, S.Y., Strömberg, C.A.E., 2018. Late Miocene rise and fall of C4 grasses in the western United States linked to aridification and uplift. Geol. Soc. Am. Bull. 131, 224–234. Iacumin, P., Bocherens, H., Mariotti, A., Longinelli, A., 1996. Oxygen isotope analyses of coexisting carbonate and phosphate in biogenic apatite: a way to monitor diagenetic alteration of bone phosphate? Earth Planet. Sci. Lett. 142, 1–6.

39

Journal Pre-proof Ivanov, D., Utescher, T., Mosbrugger, V., Syabryaj, S., Djordjević-Milutinović, D., Molchanoff, S., 2011. Miocene vegetation and climate dynamics in Eastern and Central Paratethys (Southeastern Europe). Palaeogeogr. Palaeoclimatol. Palaeoecol. 304, 262–275. doi:10.1016/j.palaeo.2010.07.006 Jacobs, B.F., Kingston, J.D., Jacobs, L.L., 1999. The origin of grass-dominated ecosystems. Ann. Missouri Bot. Gard. 86, 590–643. doi:10.2307/2666186

of

Janis, C.M., Damuth, J., Theodor, J.M., 2004. The species richness of Miocene browsers, and

ro

implications for habitat type and primary productivity in the North American grassland

-p

biome. America (NY). 207, 371–398. doi:10.1016/j.palaeo.2003.09.032

re

Janis, C.M., Scott, K.M., Jacobs, L.L., 1998. Evolution of Tertiary Mammals of North America, 1st ed. Cambridge University Press, Cambridge.

lP

Jiménez-Moreno, G., Fauquette, S., Suc, J.-P., 2008. Vegetation, climate and palaeoaltitude

na

reconstructions of the Eastern Alps during the Miocene based on pollen records from

2699.2008.01911.x

ur

Austria, Central Europe. J. Biogeogr. 35, 1638–1649. doi:10.1111/j.1365-

Jo

Julien, M.A., Bocherens, H., Burke, A., Drucker, D.G., Patou-Mathis, M., Krotova, O., Péan, S., 2012. Were European steppe bison migratory? 18O, 13C and Sr intra-tooth isotopic variations applied to a palaeoethological reconstruction. Quat. Intl. 271, 106–119. doi:10.1016/j.quaint.2012.06.011 Kayseri-Özer, M.S., 2017. Cenozoic vegetation and climate change in Anatolia—A study based on the IPR-vegetation analysis. Palaeogeogr. Palaeoclimatol. Palaeoecol. 467, 37–68. Kent-Corson, M.L., Barnosky, A.D., Mulch, A., Carrasco, M. a., Chamberlain, C.P., 2013. Possible regional tectonic controls on mammalian evolution in western North America.

40

Journal Pre-proof Palaeogeogr. Palaeoclimatol. Palaeoecol. 387, 17–26. doi:10.1016/j.palaeo.2013.07.014 Koch, P., Tuross, N., Fogel, M., 1997. The effects of sample treatment and diagenesis on the isotopic integrity of carbonate in biogenic hydroxylapatite. J. Archaeol. Sci. 24, 417–429. Koch, P.L., 1998. Isotopic Reconstruction of Past Continental Environments. Annu. Rev. Earth Planet. Sci. 26, 573–613. doi:10.1146/annurev.earth.26.1.573 Kohn, M.J., 2010. Carbon isotope compositions of terrestrial C3 plants as indicators of

of

(paleo)ecology and (paleo)climate. Proc. Natl. Acad. Sci. U. S. A. 107, 19691–5.

ro

doi:10.1073/pnas.1004933107

-p

Kohn, M.J., 1996. Predicting animal δ18O: accounting for diet and physiological adaptation.

re

Geochim. Cosmochinica Acta 60, 4811–4829.

Kohn, M.J., Cerling, T.E., 2002. Stable Isotope Compositions of Biological Apatite. Rev.

lP

Mineral. Geochem. 48, 455–488.

na

Kohn, M.J., Dettman, D.L., 2007. Paleoaltimetry from stable isotope compositions of fossils. Rev. Mineral. Geochem. 66, 119–154.

ur

Kohn, M.J., Fremd, T.J., 2008. Miocene tectonics and climate forcing of biodiversity, western

Jo

United States. Geology 36, 783–786. doi:10.1130/g24928a.1 Kohn, M.J., Fremd, T.J., 2007. Tectonic controls on isotope composition and species diversification, John Day Basin, central Oregon. PaleoBios 27, 48–61. Kohn, M.J., McKay, M.P., 2012. Paleoecology of late Pleistocene–Holocene faunas of eastern and central Wyoming, USA, with implications for LGM climate models. Palaeogeogr. Palaeoclimatol. Palaeoecol. 326–328, 42–53. doi:10.1016/j.palaeo.2012.01.037 Kohn, M.J., McKay, M.P., Knight, J.L., 2005. Dining in the Pleistocene - Who’s on the menu? Geology 33, 649–652. doi:10.1130/G21476.1

41

Journal Pre-proof Kohn, M.J., Schoeninger, M.J., Barker, W.W., 1999. Altered states: effects of diagenesis on fossil tooth chemistry. Geochim. Cosmochinica Acta 63, 2737–2747. Kohn, M.J., Schoeninger, M.J., Valley, J.W., 1998. Variability in oxygen isotope compositions of herbivore teeth: reflections of seasonality or developmental physiology? Chem. Geol. 152, 97–112. doi:10.1016/S0009-2541(98)00099-0 Kohn, M.J., Schoeninger, M.J., Valley, J.W., 1996. Herbivore tooth oxygen isotope

of

compositions: effects of diet and physiology. Geochim. Cosmochinica Acta 60, 3889–3896.

ro

Kohn, M.J., Strömberg, C.A.E., Madden, R.H., Dunn, R.E., Evans, S., Palacios, A., Carlini,

-p

A.A., 2015. Quasi-static Eocene–Oligocene climate in Patagonia promotes slow faunal

re

evolution and mid-Cenozoic global cooling. Palaeogeogr. Palaeoclimatol. Palaeoecol. 435, 24–37. doi:10.1016/j.palaeo.2015.05.028

lP

Larcher, W., Winter, A., 1981. Frost susceptibility of palms: experimental data and their

na

interpretation. Principes 25, 143–152.

Lee-Thorp, J., van der Merwe, N.J., 1987. Carbon isotope analysis of fossil bone apatite. S. Afr.

ur

J. Sci. 83, 712–715.

Jo

Leopold, E.B., Denton, M.F., 1987. Comparative age of grassland and steppe east and west of the Northern Rocky Mountain. Ann. Missouri Bot. Gard. 74, 841. doi:10.2307/2399452 Levin, N.E., Cerling, T.E., Passey, B.H., Harris, J.M., Ehleringer, J.R., 2006. A stable isotope aridity index for terrestrial environments. Proc. Natl. Acad. Sci. U. S. A. 103, 11201. Liebrand, D., Beddow, H.M., Lourens, L.J., Pälike, H., Raffi, I., Bohaty, S.M., Hilgen, F.J., Saes, M.J., Wilson, P.A., van Dijk, A.E., 2016. Cyclostratigraphy and eccentricity tuning of the early Oligocene through early Miocene (30.1–17.1 Ma): Cibicides mundulus stable oxygen and carbon isotope records from Walvis Ridge Site 1264. Earth Planet. Sci. Lett. 450, 392-

42

Journal Pre-proof 405. Liu, L., Stegman, D.R., 2012. Origin of Columbia River flood basalt controlled by propagating rupture of the Farallon slab. Nature 482, 386–389. doi:10.1038/nature10749 Longinelli, A., 1984. Oxygen isotopes in mammal bone phospate: a new tool for paleohydrological and paleoclimatological research? Geochim. Cosmochinica Acta 48, 385–390.

of

Luz, B., Cormie, A.B., Schwarcz, H.P., 1990. Oxygen isotope variations in phosphate of deer

ro

bones. Geochim. Cosmochim. Acta 54, 1723–1728. doi:10.1016/0016-7037(90)90403-8

-p

Luz, B., Kolodny, Y., 1985. Oxygen isotope variations in phosphate of biogenic apatites, IV.

re

Earth Planet. Sci. Lett. 75, 29–36.

MacFadden, B.J., 2000. Cenozoic mammalian herbivores from the Americas: reconstructing

lP

ancient diets and terrestrial communities. Annu. Rev. Ecol. Syst. 31, 33–59.

na

doi:10.1146/annurev.ecolsys.31.1.33

Maguire, K.C., Samuels, J.X., Schmitz, M.D., 2018. The fauna and chronostratigraphy of the

ur

middle Miocene Mascall type area, John Day Basin, Oregon, USA. PaleoBios 35, 1–51.

Jo

McMillan, M.E., Angevine, C.L., Heller, P.L., 2002. Postdepositional tilt of the MiocenePliocene Ogallala Group on the western Great Plains: Evidence of late Cenozoic uplift of the Rocky Mountains. Geology 30, 63–66. doi:10.1130/00917613(2002)030<0063:PTOTMP>2.0.CO;2 McNab, B.K., 2002. The physiological ecology of vertebrates. A view from energetics. Cornell University Press, Ithica, NY. Miller, K.G., Fairbanks, R.G., Mountain, G.S., 1987. Tertiary oxygen isotope synthesis, sea level history, and continental margin erosion. Paleoceanography 2, 1–19.

43

Journal Pre-proof Mix, H.T., Chamberlain, C.P., 2014. Stable isotope records of hydrologic change and paleotemperature from smectite in Cenozoic western North America. Geochim. Cosmochim. Acta 141, 532–546. doi:10.1016/j.gca.2014.07.008. Mosbrugger, V., Utescher, T., Dilcher, D.L., 2005. Cenozoic continental climatic evolution of Central Europe. Proc. Natl. Acad. Sci. U. S. A. 102, 14964–14969. doi:10.1073/pnas.0505267102

of

Nelson, S. V., 2005. Paleoseasonality inferred from equid teeth and intra-tooth isotopic

ro

variability. Palaeogeogr. Palaeoclimatol. Palaeoecol. 222, 122–144.

-p

doi:10.1016/j.palaeo.2005.03.012

re

Nichols, R., 1979. Additional early Miocene mammals from the Lemhi Valley of Idaho. Tebiwa 17, 1–12.

lP

Nichols, R., 1976. Early Miocene mammals from the Lemhi Valley of Idaho. Tebiwa 18, 9–37.

na

O’Leary, M.H., 1988. Carbon isotopes in photosynthesis. Biosci. 38, 328–336. Palmqvist, P., Gröcke, D.R., Arribas, A., Fariña, R.A., 2003. Paleoecological reconstruction of a

ur

lower Pleistocene larege mammal community using biogeochemical (δ13C, δ15N, δ18O,

Jo

Sr:Zn) and ecomorphological approaches. Paleobiology 29, 205–229. Passey, B.H., Cerling, T.E., 2002. Tooth enamel mineralization in ungulates: implications for recovering a primary isotopic time-series. Geochim. Cosmochim. Acta 66, 3225–3234. Passey, B.H., Robinson, T.F., Ayliffe, L.K., Cerling, T.E., Sphonheimer, M., Dearing, M.D., Roeder, B.L., Ehleringer, J.R., 2005. Carbon isotope fractionation between diet, breath CO2, and bioapatite in different mammals. J. Archaeol. Sci. 32, 1459–1470. doi:10.1016/j.jas.2005.03.015 Pekar, S.F., DeConto, R.M., Harwood, D.M., 2006. Resolving a late Oligocene conundrum:

44

Journal Pre-proof Deep-sea warming and Antarctic glaciation. Palaeogeogr. Palaeoclimatol. Palaeoecol. 231, 29–40. Pierce, K.L., Morgan, L.A., 1992. The track of the Yellowstone hot spot: volcanism, faulting and uplift. Geol. Soc. Am. Mem. 179, 1–53. Pound, M.J., Haywood, A.M., Salzmann, U., Riding, J.B., 2012. Global vegetation dynamics and latitudinal temperature gradients during the Mid to Late Miocene (15.97-5.33Ma). Earth-

of

Sci. Rev. 112, 1–22. doi:10.1016/j.earscirev.2012.02.005

ro

Reichgelt, T., West, C.K., Greenwood, D.R., 2018. The relation between global palm distribution

-p

and climate. Sci. Rep. 8, 4721.

re

Retallack, G.J., 2009. Refining a pedogenic-carbonate CO2 paleobarometer to quantify a middle Miocene greenhouse spike. Palaeogeogr. Palaeoclimatol. Palaeoecol. 281, 57–65.

lP

doi:10.1016/j.palaeo.2009.07.011

doi:10.1086/512753

na

Retallack, G.J., 2007. Cenozoic paleoclimate on land in North America. J. Geol. 115, 271–294.

ur

Retallack, G.J., 2004. Late Oligocene bunch grassland and early Miocene sod grassland

Jo

paleosols from central Oregon. Palaeogeogr. Palaeoclimatol. Palaeoecol. 207, 203–237. Royer, D.L., Osborne, C.P., Beerling, D.J., 2002. High CO2 increases the freezing sensitivity of plants: Implications for paleoclimatic reconstructions from fossil floras. Geology 30, 963– 966. doi:10.1130/0091-7613(2002)030<0963:HCITFS>2.0.CO;2 Rozanski, K., Araguás-Araguás, L., 1992. Relation between long-term trends of oxygen-18 isotope composition of precipitation and climate. Science 258, 981–985. Rozanski, K., Araguás-Araguás, L., Gonfiantini, R., 1993. Isotopic patterns in modern global precipitation, in: Swart, P.K., Lohmann, K.C., McKenzie, J.A., Savin, S. (Eds.), Climate

45

Journal Pre-proof Change in Continental Isotopic Records. American Geophysical Union, Washington, pp. 1– 36. Samuels, J.X., Hopkins, S.S.B., 2017. The impacts of Cenozoic climate and habitat changes on small mammal diversity of North America. Glob. Planet. Change 149, 36–52. Shackleton, N.J., Kennett, J.P., 1975. Paleotemperature history of the Cenozoic and initiation of Antarctic glaciation: oxygen and carbon isotopic analyses in DSDP Sites 277, 279, and 281,

ro

Drilling Program, College Station, TX, United States.

of

Initial Reports of the Deep Sea Drilling Project, vol. 29. Texas A & M University, Ocean

-p

Sheldon, N.D., Retallack, G.J., Tanaka, S., 2002. Geochemical climofunctions from North

re

American soils and application to paleosols across the Eocene-Oligocene boundary in Oregon. J. Geol. 110, 687–696. doi:10.1086/342865

lP

Stewart, G.R., Turnbull, M.H., Schmidt, S., Erskine, P.D., 1995. 13C natural abundance in plant

na

communities along a rainfall gradient: A biological integrator of water availability. Aust. J. Plant Physiol. 22, 51–55.

ur

Stinchcomb, G.E., Nordt, L.C., Driese, S.G., Lukens, W.E., Williamson, F.C., Tubbs, J.D., 2016.

Jo

A data-driven spline model designed to predict paleoclimate using paleosol geochemistry. Am. J. Sci. 316, 746–777. Strömberg, C.A.E., 2011. Evolution of grasses and grassland ecosystems. Annu. Rev. Earth Planet. Sci. 39, 517–544. doi:10.1146/annurev-earth-040809-152402 Strömberg, C.A.E., 2005. Decoupled taxonomic radiation and ecological expansion of openhabitat grasses in the Cenozoic of North America. Proc. Natl. Acad. Sci. U. S. A. 102, 11980–4. doi:10.1073/pnas.0505700102 Strömberg, C.A.E., McInerney, F.A., 2011. The Neogene transition from C3 to C4 grasslands in

46

Journal Pre-proof North America : assemblage analysis of fossil phytoliths. Paleobiology 37, 50–71. Strömberg, C.A.E., Werdelin, L., Friis, E.M., Saraç, G., 2007. The spread of grass-dominated habitats in Turkey and surrounding areas during the Cenozoic: Phytolith evidence. Palaeogeogr. Palaeoclimatol. Palaeoecol. 250, 18–49. doi:10.1016/j.palaeo.2007.02.012 Swanson, D.A., Wright, T.L., Hooper, P.R., Bentley, R.D., 1979. Revisions in stratigraphic nomenclature of the Columbia River Basalt Group. U.S. Geol. Surv. Bull. 1457-G.

of

Syabryaj, S., Utescher, T., Molchanoff, S., Bruch, A., 2007. Vegetation and palaeoclimate in the

ro

Miocene of Ukraine. Palaeogeogr. Palaeoclimatol. Palaeoecol. 253, 153–168.

-p

Tedford, R.H., Albright III, L.B., Barnosky, A.D., Ferrusquia-Villafranca, I., Hunt Jr., R.M.,

re

Storer, J.E., Swisher III, C.C., Voorhies, M.R., Webb, S.D., Whistler, D.P., 2004. Mammalian biochronology of the Arikareean through Hemphillian interval (late Oligocene

lP

through early Pliocene epochs), in: Woodburne, M.O. (Ed.), Late Cretaceous and Cenozoic

na

Mammals of North America. Columbia University Press, pp. 169–231. Tipple, B.J., Meyers, S.R., Pagani, M., 2010. Carbon isotope ratio of Cenozoic CO2: A

ur

comparative evaluation of available geochemical proxies. Paleoceanography 25, 1–11.

Jo

doi:10.1029/2009PA001851

Trayler, R.B., Kohn, M.J., 2017. Tooth enamel maturation reequilibrates oxygen isotope compositions and supports simple sampling methods. Geochim. Cosmochim. Acta 198, 32– 47. doi:10.1016/j.gca.2016.10.023 U.S. climate data [WWW Document], 2016. Your Weather Serv. URL http://www.usclimatedata.com/ (accessed 10.20.15). Utescher, T., Bruch, A.A., Micheels, A., Mosbrugger, V., Popova, S., 2011. Cenozoic climate gradients in Eurasia — a palaeo-perspective on future climate change? Palaeogeogr.

47

Journal Pre-proof Palaeoclimatol. Palaeoecol. 304, 351–358. doi:10.1016/j.palaeo.2010.09.031 Utescher, T., Djordjević-Milutinović, D., Bruch, A., Mosbrugger, V., 2007. Palaeoclimate and vegetation change in Serbia during the last 30 Ma. Palaeogeogr. Palaeoclimatol. Palaeoecol. 253, 141–152. van Dam, J.A., Reichart, G.J., 2009. Oxygen and carbon isotope signatures in late Neogene horse teeth from Spain and application as temperature and seasonality proxies. Palaeogeogr.

of

Palaeoclimatol. Palaeoecol. 274, 64–81. doi:10.1016/j.palaeo.2008.12.022

ro

Vogel, J.C., 1978. Isotopic assessment of the dietary habits of ungulates. S. Afr. J. Sci. 74, 298–

-p

301.

re

Wolfe, J., 1994. Tertiary climatic changes at middle latitudes of western North America. Palaeogeogr. Palaeoclimatol. Palaeoecol. 108, 195–205. doi:10.1016/0031-0182(94)90233-

lP

X

na

Woodburne, M.O., Swisher, C.C., 1995. Land mammal high-resolution geochronology, intercontinental overland dispersals, sea level, climate, and vicariance, in: Berggren, W.A.,

ur

Kent, D.V., Aubry, M.-P., Hardenbol, J. (Eds.), Geochronology, Time Scales and Global

Jo

Stratigraphic Correlation. Soc. Sedim. Geol. Spec. Publ. 54, 336–364. Yakir, D., 1992. Variations in the natural abundance of oxygen-18 and deuterium in plant carbohydrates. Plant Cell Environ. 15, 1005–1020. Yakir, D., DeNiro, M.J., Gat, J.R., 1990. Natural deuterium and oxygen-18 enrichment in leaf water of cotton plants grown under wet and dry conditions: evidence for water compartmentation and its dynamics. Plant Cell Environ. 13, 49–56. You, Y., Huber, M., Müller, R.D., Poulsen, C.J., Ribbe, J., 2009. Simulation of the Middle Miocene Climate Optimum. Geophys. Res. Lett. 36. doi:10.1029/2008GL036571

48

Journal Pre-proof Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686–93. doi:10.1126/science.1059412 Zachos, J.C., Dickens, G.R., Zeebe, R.E., 2008. An early Cenozoic perspective on greenhouse warming and carbon-cycle dynamics. Nature 451, 279–83. doi:10.1038/nature06588 Zanazzi, A., Kohn, M.J., MacFadden, B.J., Terry, D.O., 2007. Large temperature drop across the

of

Eocene-Oligocene transition in central North America. Nature 445, 639–42.

ro

doi:10.1038/nature05551

-p

Zazzo, A., Lecuyer, C., Sheppard, S.M.F., Grandjean, P., Mariotti, A., 2004. Diagenesis and the

re

reconstruction of paleoenvironments: a method to restore original 18O values of carbonate

Jo

Figure captions:

ur

na

lP

and phosphate from fossil tooth enamel. Geochim. Cosmochim. Acta 68, 2245-2258.

Figure 1. Location of the Railroad Canyon Section (RCS) in central-eastern Idaho, U.S.A. For detailed locality information, see Barnosky et al. (2007) and Harris et al. (2017b).

Figure 2. Geologic and paleontological context for the time bins used in this study. A) Composite stratigraphic section for the Railroad Canyon section (RCS) showing formational boundaries, North American Land Mammal Ages (NALMAs), location of dated ashes within the

49

Journal Pre-proof section (RCS1-RCS4), magnetostratigraphy, lithostratigraphy, location of paleosols sampled for δ13Corg and phytoliths (Harris et al., 2017b), location of vertebrate fossil sites (Barnosky et al., 2007), with solid black bars representing portions of each site where fossil teeth were sampled from, and time bins used in this study (TB1-5). The gap between time bins 1 and 2 is due to a lack of fossils, and therefore data, in that part of the section. Note that these time bins capture pre- and post-MMCO warming time (i.e., time bins 1-4 and 5, respectively). B) Benthic marine

of

δ13C and δ18O record for the period of interest, with dark/light orange lines based on data from

ro

Holbourn et al. (2015) and purple/blue lines based on data from Liebrand et al. (2016); δ13C and

-p

δ18O values of Holbourn et al. (2015) have been increased by 0.5 and 0.3‰ respectively to align

re

with data of Liebrand et al. (2016). Red shaded box and “Peak warmth” signifies maximum inferred MMCO temperatures. Vertebrate fossil locality names are abbreviated as follows: WH4,

lP

Whiskey Springs 4; MG, Mollie Gulch; WRC, West Railroad Cut; ERC, East Railroad Cut; SF

na

e&w, Snowfence east and west; ST1, Snowfence Turtle 1; ST2/3, Snowfence Turtle 2 and 3; DS 3/4, Dead Squirrel 3 and 4; T1, Turtle 1; HDS3, High Dead Squirrel 3; HDS2, High Dead

Jo

ur

Squirrel 2; DP, Deadman Pass. EMU = early Miocene unconformity.

Figure 3. Estimates of Mean Annual Precipitation (MAP in mm·yr-1) through time in the Railroad Canyon section (RCS). Bold horizontal bars represent MAP estimates calculated from average enamel δ13C values for each RCS time bin (TB1-5) and hollow circles are MAP estimates calculated from the CIA-K index of Bt horizons in RCS paleosols (Retallack 2007). Vertical lines for the enamel-based MAP estimates are error bars propagated from 2s.e. of the average δ13C value. Dashed line represents modern MAP for Leadore, Idaho located c. 12 miles south of the RCS.

50

Journal Pre-proof

Figure 4. Composition of meteoric water (δ18O) for all taxa during each time bin (TB1-5) in the Railroad Canyon section. For each boxplot, the dark line indicates the median, the box represents the interquartile range, the dashed lines indicate the range of measured values, circles indicate outliers (greater than 1.5 times the interquartile range from the median), and n = the number of

of

samples. The width of each boxplot represents the range of time captured by each time bin.

ro

Figure 5. Bulk δ18O enamel results from 23 serially sampled equid and rhino teeth (some of

-p

which show quasi-sinusoidal seasonal variation) placed into the stratigraphic framework of the RCS. Data within TB1, TB3 and TB4 have been grouped according to similarity in seasonal

re

variation curves and have been aligned by the lowest δ18O value in each curve. Hollow symbols

lP

are data from equid teeth, and filled symbols are data from rhino teeth. Different symbol and line

na

types signify different individual teeth, which are labeled as follows: A) UMVP 4400; B) UWBM 100084; D) UMVP 5724; E) UMVP 5723; F) UMVP 5721; G) UWBM 100086; H)

ur

UMVP 4154F; I) UMVP 4159; J) UMVP 4154E; K) UWBM 100062; L) UMVP 4452A; M)

Jo

UWBM 100066; N) UWBM 100055; P) UMVP 4453E; Q) UWBM 100081; R) UWBM 100038; S) UMVP 9305; T) UMVP 6424; U) UMVP 9426; V) UWBM 100040; W) UWBM 100082; X) UMVP 4482D; Y) UMVP 4482E; Z) UMVP 4482C. o = occlusal surface; c = cervical margin.

Figure 6. Survey of mid-latitude (c. 35-50 °N) sites points to variability in regional climate change across the MMCO based on studies using climate-proxy data (e.g., paleobotany, stable isotope data, sedimentology) for Mean Annual Temperature (MAT) and Mean Annual

51

Journal Pre-proof Precipitation (MAP). Time bins, represented as stacks: pre-MMCO (bottom of stack), MMCO (middle) and post-MMCO (top of stack). A. Relative + absolute changes in MAT. MAT bins: orange = warm (MAT >15 °C); pale orange = rel. warm (but cooler than peak MAT >15 °C); blue = cool (MAT <15 °C). B. Relative + absolute changes in MAP. MAP bins: green = wet (MAP >800 mm/year); pale green = rel. wet (but drier than peak MAP >1,000 mm/year); yellow = dry (MAP <800 mm/year). C. Relative changes in MAT, only showing the trend. MAT bins:

of

orange = warmer; blue = cooler; both regardless of absolute MAT values. D. Relative changes in

ro

MAP, only showing the trend. MAP bins: green = wetter; yellow = drier. Climate zones from

-p

Scotese (2002). See Supplementary Materials for a more detailed description of binning

lP

re

methods.

Tables:

na

Table 1. Railroad Canyon Mean Annual Precipitation (MAP) estimates calculated from

Jo

ur

herbivore δ13C enamel compositions.

52

Journal Pre-proof

Table 1. Railroad Canyon Mean Annual Precipitation (MAP, mm/yr) estim herbivore bulk δ13 C enamel compositions.

MAP 2σ

All time bins Mean

194 +317

511

26

All time bins Median

162

452

12

TB1 Mean

128 +320

ro

Min MAP

448

0

-p

(mm/yr)

Max MAP

275

0

+362 -193

568

13

100 +252

353

0

101

350

0

166 +332

498

0

106

365

0

56 206

na

lP

TB2 Mean

Jo

TB3 Median

ur

TB3 Mean

TB4 Mean

-162 +219 -103

TB4 Median

of

re

TB1 Median

-169 +290 -150

-124 +248 -122

-173 +259

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☒The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

53

Jo

ur

na

lP

re

-p

ro

of

Journal Pre-proof

54

Journal Pre-proof Highlights * Regional faunal change in N. Rocky Mountains has been tied to Miocene global warming * We used stable isotopes to infer central Idaho climates and test climate-fauna link * Isotope data from teeth suggest stable, warm, arid climate during early-mid Miocene

Jo

ur

na

lP

re

-p

ro

of

* The contrast with warm-wet Miocene climate elsewhere points to key regional variation

55