摘要

Proxy reconstructions of the mid-Piacenzian warm period (mPWP, between 3.264 and 3.025 Ma) suggest terrestrial temperatures were much warmer in the northern high latitudes (55 degrees-90 degrees N, referred to as NHL) than present-day. Climate models participating in the Pliocene Model Intercomparison Project Phase 1 (PlioMIP1) tend to underestimate this warmth. For instance, the underestimate is similar to 10 degrees C on average across NHL and up to 17 degrees C in the Canadian Arctic region in the Community Climate System Model version 4 (CCSM4). Here, we explore potential mPWP climate forcings that might contribute to this mPWP mismatch. We carry out seven experiments to assess terrestrial temperature responses to Pliocene Arctic gateway closure, variations in CO2 level, and orbital forcing at millennial time scale. To better compare the full range of simulated terrestrial temperatures with sparse proxy data, we introduce a pattern recognition technique that simplifies the model surface temperatures to a few representative patterns that can be validate with the limited terrestrial proxy data. The pattern recognition technique reveals two prominent features of simulated Pliocene surface temperature responses. First, distinctive patterns of amplified warming occur in the NHL, which can be explained by lowered surface elevation of Greenland, pattern and amount of Arctic sea ice loss, and changing strength of Atlantic meridional overturning circulation. Second, patterns of surface temperature response are similar among experiments with different forcing mechanisms. This similarity is due to strong feedbacks from responses in surface albedo and troposphere water vapor content to sea ice changes, which overwhelm distinctions in forcings from changes in insolation, CO2 forcing, and Arctic gateway closure. By comparing CCSM4 simulations with proxy records, we demonstrate that both model and proxy records show similar patterns of mPWP NHL terrestrial warmth, but the model underestimates the magnitude. High insolation, greater CO2 forcing, and Arctic gateways closure each contributes to reduce the underestimate by enhancing the Arctic warmth of 1-2 degrees C. These results highlight the importance of considering proxy NHL warmth in the context of Pliocene Arctic gateway changes, and variations in insolation and CO2 forcing.

  • 出版日期2017-5-15