Climate and ecology in the Rocky Mountain interior after the early Eocene Climatic Optimum

As atmospheric carbon dioxide (CO2) and temperatures increase with modern climate change, ancient hothouse periods become a focal point for understanding ecosystem function under similar conditions. The early Eocene exhibited high temperatures, high CO2 levels, and similar tectonic plate configuration as today, so it has been invoked as an analog to modern climate change. During the early Eocene, the greater Green River Basin (GGRB) of southwestern Wyoming was covered by an ancient hypersaline lake (Lake Gosiute; Green River Formation) and associated fluvial and floodplain systems (Wasatch and Bridger formations). The volcaniclastic Bridger Formation was deposited by an inland delta that drained from the northwest into freshwater Lake Gosiute and is known for its vast paleontological assemblages. Using this well-preserved basin deposited during a period of tectonic and paleoclimatic interest, we employ multiple proxies to study trends in provenance, parent material, weathering, and climate throughout 1 million years. The Blue Rim escarpment exposes approximately 100 m of the lower Bridger Formation, which includes plant and mammal fossils, solitary paleosol profiles, and organic remains suitable for geochemical analyses, as well as ash beds and volcaniclastic sandstone beds suitable for radioisotopic dating. New 40Ar / 39Ar ages from the middle and top of the Blue Rim escarpment constrain the age of its strata to∼ 49.5–48.5 Myr ago during the “falling limb” of the early Eocene Climatic Optimum. We used several geochemical tools to study provenance and parent material in both the paleosols and the associated sediments and found no change in sediment input source despite significant variation in sedimentary facies and organic carbon burial. We also reconstructed environmental conditions, including temperature, precipitation (both from paleosols), and the isotopic composition of atmospheric CO2 from plants found in the floral assemblages. Results from paleosol-based reconstructions were compared to semi-co-temporal reconstructions made using leaf physiognomic techniques and marine proxies. The paleosol-based reconstructions (near the base of the section) of precipitation (608–1167 mm yr−1) and temperature (10.4 to 12.0 C) were within error of, although lower than, those based on floral assemblages, which were stratigraphically higher in the section and represented a highly preserved event later in time. Geochemistry and detrital feldspar geochronology indicate a consistent provenance for Blue Rim sediments, sourcing predominantly from the Idaho paleoriver, which drained the active Challis volcanic field. Thus, because there was neither significant climatic change nor significant provenance change, variation in sedimentary facies and organic carbon burial likely reflected localized geomorphic controls and the relative height of the water table. The ecosystem can be characterized as a wet, subtropical-like forest (i.e., paratropical) throughout the interval based upon the floral humidity province and Holdridge life zone schemes. Given the mid-paleolatitude position of the Blue Rim escarpment, those results are consistent with marine proxies that indicate that globally warm climatic conPublished by Copernicus Publications on behalf of the European Geosciences Union. 2516 R. A. Stein et al.: Climate and ecology in the Rocky Mountain interior ditions continued beyond the peak warm conditions of the early Eocene Climatic Optimum. The reconstructed atmospheric δ13C value (−5.3 ‰ to −5.8 ‰) closely matches the independently reconstructed value from marine microfossils (−5.4 ‰), which provides confidence in this reconstruction. Likewise, the isotopic composition reconstructed matches the mantle most closely (−5.4 ‰), agreeing with other postulations that warming was maintained by volcanic outgassing rather than a much more isotopically depleted source, such as methane hydrates.


Figure S6
Comparisons of paleoclimate reconstructions using multiple proxies at Blue Rim, southwestern Wyoming. Paleosol-based proxies are shown in blue (y-axis), with opaque blue representing the range of reconstructed precipitation (a-b) and temperature (c-d) and transparent blue shows error for each proxy (1σ for Chemical Index of Alteration and Paleosol Weathering Index). Two plant-based proxies are shown in red (x-axis), with opaque red representing the range of reconstructed precipitation using leaf margin analysis and leaf area analysis based off multiple regional and global equations (e.g., Wolfe 1979;Wing & Greenwood 1993;Wilf 1997;Wilf et al., 1998;Gregory-Wodzicki 2000;Jacobs 2004;Kowalksi & Dilcher 2003;Miller et al., 2006;Peppe et al., 2011), and error (standard error) shown in opaque red. Climate Leaf Analysis Multivariate Program (CLAMP, a & c;e.g., Spicer et al. 2009) is shown in purple, with opaque purple to show the range of reconstructed values based on regional meteorological stations and global reconstructions, and transparent purple showing standard deviation (1σ). CLAMP was not done on the upper horizon. The precipitations and temperatures for which both proxies overlap (within error) are outlined in a dashed box, and grey boxes show the precipitations and temperatures that overlap for reconstructed ranges (excluding error). The Lower (plant macrofossil) Horizon is shown in panels a and c, the Upper (plant macrofossil) Horizon is shown in panels b and d (Allen, 2017b).

Figure S7
Current boundaries of Wyoming, USA with localities plotted with same symbols as seen in Figure 10. Blue Rim escarpment is plotted in stars (blue for this study, red for Allen 2017b). The Paleocene-Eocene Thermal Maximum record from Bighorn Basin is plotted in a yellow circle, while the EECO record from the Bighorn Basin is plotted in a red circle. The middle Eocene climatic optimum (MECO) is plotted in a blue circle.

Floral humidity province
To contextualize climate variables (temperature, precipitation) to ecoregion and humidity, Gulbranson et al. (2011) developed a life-zone proxy based on Rasmussen et al. (2005) and Rasmussen and Tabor's (2007) pedogenic energy model. This energy quantifies energy influxes due to solar radiation (and subsequent net primary productivity: NPP) and precipitation.
The total energy input into a soil (Ein) is related to energy supplied by NPP (ENPP) and precipitation (EPPT). EPPT and ENPP are calculated using weathering indices (CIA). EPPT is plotted against evapotranspiration (ET) and divided into humidity zones using Eq. 9: Where ΔT is the temperature difference between 273.16 °K and mean annual temperature.
Mean annual precipitation calculated using the relationship between CIA-K and precipitation was used in this relationship (Eq. 5).
In modern environments, effective precipitation (Peff) is a linear function of MAP and ET (Eq.

Holdridge Life Zones
The Holdridge Life Zone classification system (Holdridge 1947) matches climate with vegetation. Higher precipitation adds energy to a soil system, which mobilizes elements and weathers the soil. Evapotranspiration, represented on the other axis of the Holdridge biome diagram (Fig. 7b) represents energy loss from the soil profile. These two plotted together can then be divided into biome space, allowing for estimation of climate and vegetation through the same diagram. Using the ratio of evapotranspiration (calculated by Equation 10) to mean annual precipitation, which represents potential evapotranspiration ratio, as compared to precipitation, we plotted each paleosol in life zones.

Global context of climate
Chronologically, the region of interest (specifically, the Rocky Mountain Wyoming , and hot, dry summers (Fig. 8). The locations of each of these sites is plotted on Figure S7, to demonstrate their proximity.
The Parachute Creek, Laney, and Fossil Butte Members of the Green River Formation, all slightly older than the Blue Rim escarpment and deposited during the EECO, have also been interpreted as semi-deciduous with seasonally dry subtropical taxa (Wing 1987;Allen 2015).
Further away in the Okanagan Highlands of the North American Pacific Northwest, climate reconstructions yield temperature ranges of 10-13.5 °C (Wolfe et al. 1994;Wolfe et al. 1998;Greenwood et al. 2005), similar to those values reconstructed at Blue Rim, demonstrating the equability of North America during the early Eocene.

Blue Rim climate based on floral vs paleosol reconstructions
Mean annual precipitation (MAP) reconstructions from this study based on inorganic proxies in paleosols ranged from 608-1167 mm yr -1 (average: 845 mm yr -1 ± 255 mm yr -1 , standard deviation). A high influence of carbonate in the parent material resulted in a lower CIA- ). This could be due to the location in the section (stratigraphically older, see Figure 5) of the studied paleosols. Temperature results based off inorganic geochemistry in paleosols in this study were slightly lower than those reconstructed using leaf physiognomy (CLAMP: 14 to 15 °C, LMA: 14 to 20 °C; Fig. 7; Allen 2017b, originally calculated using Wolfe 1979; Wing and Greenwood 1993;Wilf 1997;Kowalski & Dilcher 2003;Miller et al., 2006;Spicer et al., 2009;and Peppe et al., 2011), with PWI-based temperatures ranging from 10 to 12 °C (average 11.0 °C ± 0.7 °C standard deviation; Fig. 7).