Oxygen isotope and paleobotanical estimates of temperature and δ18 O-latitude gradients over North America during the early Eocene

Empirical estimates of climate parameters such as mean annual temperature (MAT) are essential to describe both ancient climate and to ground truth climate model simulations of past climates. In terrestrial settings, no single proxy record is able to provide detailed temporal yet geographically widespread information, and it thus becomes essential that paleoclimatic information obtained from different proxies be comparable. As an example of how to compare estimates, and to illustrate what can be learned by doing so, we measured oxygen isotope ratios of phosphate in the tooth enamel of fossil mammals and the body scales of freshwater fish in order to estimate the mean annual temperature (MAT) in which they lived. The samples were collected from Paleogene rocks of the Bighorn, Green River, and Powder River basins of Wyoming. Earliest Eocene samples provide MAT estimates that are similar to MAT estimates for the same times and areas obtained from leaf margin analysis. Furthermore, isotopic data from the Paleocene Eocene Thermal Maximum (PETM), which has not yet yielded fossil floras, indicate a significantly higher MAT of 26degreesC. We also used temperature estimates obtained by both methods to investigate paleoelevation and to reconstruct geographic patterns of MAT over North America for the early Eocene. Larger Laramide basins have higher temperatures and lower inferred elevations compared to associated volcanic highlands where cooler temperatures are consistent with substantial elevations. In turn, latitudinal gradients in temperature indicate that warming centered on polar regions played an important role during the Eocene hothouse. Lastly, we show that the reconstructed Eocene relation between oxygen isotope ratios of precipitation and MAT is significantly different from the modem relation between these two variables. Because the relation has changed over time, the modern correlation should not be used to estimate paleo-MAT from measurements of oxygen isotope ratios in fossil material.