Reassessment of the lake-based Pliocene hydroclimate proxy network in western North America

Hydroclimate proxy data from lakes, vegetation, and stable isotopes suggest western North America was wetter in the Pliocene than today. This is unexpected because CO2 was higher than preindustrial levels and models predict future drying in western North America. Pliocene proxy data also disagree with most Pliocene climate models, which simulate extensive drying. These discrepancies raise questions: 1) Are proxy data misinterpreted or biased? 2) Do models accurately simulate moisture delivery to western North America in warm climates? 3) Is the Pliocene indeed analogous to the future? Here we address 1) by reassessing and expanding past syntheses using lake-based Pliocene hydroclimate data. Past efforts were hindered by poor age control; data were often diachronous and reflected differing orbital forcing across sites. We reassess age control at all sites and update paleomagnetic dates, North American Land Mammal Ages, and correlated tephras to the most recent published ages. We then compare Pliocene and modern environments and assign each site wetter, drier, or unclear using lake presence/absence, type (perennial/ephemeral), lake extent, and salinity. To illustrate uncertainty, we create wet and dry maps. The dry map displays only perennial lakes with unambiguous ages. The wet map adds perennial lakes with large age uncertainty, ephemeral lakes, and potential lakes, for which geophysical data indicate basin fill of probable Pliocene age. Preliminary results imply that mid-Pliocene glacial/interglacial cycles may not have caused much change in western North American hydroclimate. Specifically, the KM5c interglacial at 3.2 Ma (targeted by the PRISM and PlioMIP proxy and modeling efforts) and M2 glacial at 3.3 Ma have similar wet and dry maps, though we note that small changes in lake extent are not captured by our approach. The early Pliocene (4.0-5.3 Ma) appears regionally drier than the mid-Pliocene. For all timeslices, the wet and dry maps differ substantially, highlighting the utility of ongoing multiproxy syntheses. Overall, results show reconstructed Pliocene wetness (relative to Pliocene and future simulations) is not an artifact of age uncertainty. Our data suggest the early Pliocene was not much wetter than today, but the mid-Pliocene was wetter, especially in the southern Great Basin.