HIGH-RESOLUTION TEMPERATURE RECORD ACROSS THE K-PG BOUNDARY (67.1 TO 65.9 MA) USING CLUMPED ISOTOPES IN GAR SCALES FROM NORTH DAKOTA, USA

We report a stratigraphically well-constrained terrestrial climate record from the Late Maastrichtian to the earliest Danian with a duration of 1.2 Myr. We apply a new clumped isotope calibration for the carbonate in vertebrate bioapatite (Gray and Brandon, in review). The calibration is based on the bioapatite in modern gar scales collected from a latitudinal transect across northern North America. Gars are useful as isotope thermometers because their scales grow throughout their multiyear lifespan, they do not migrate, and they live in shallow freshwater settings (< 3 m depth). As a result, they provide a time-averaged record of the local water surface temperature. We account for seasonal temperature variations by calculating an effective temperature, T_e, which is the equivalent steady temperature that would result in the same Δ₄₇ value as that produced by a gar with a variable temperature history. Our T_e results more accurately represent the climate for continental interiors, particularly those at high latitudes with large seasonal changes in temperature. The prediction error for the calibration has a 95% confidence interval of ± 2.5 °C.

We report 12 temperatures for samples collected from the Hell Creek and Fort Union formations in North Dakota, USA, at a paleolatitude of ca. 52 °N. The stratigraphic sequence encompasses the recession of the Western Interior Seaway and mainly consists of fluvial deposits with minor lignite. T_e values are 16.8 °C at 67.1 Ma, reach a maximum of 35.6 °C at 66.3 Ma, and decrease to 22.6 °C at the K-Pg boundary. We find a T_e rise of 13 °C prior to the K-Pg boundary, coincident with the main eruption phase of the Deccan Traps, as determined by other studies (Hull et al., 2020). Overall trends in our data are also consistent with those from paleosols (Zhang et al., 2018), and freshwater mollusks (Tobin et al., 2014). These temperature estimates correlate well with the moderate atmospheric CO₂ GCM experiments (Niezgodski et al. 2017).