A COMPARISON OF RESULTS FROM UK’37 AND MG/CA PALEOTHERMOMETERS IN THE NORTH ATLANTIC OCEAN DURING THE PLIOCENE (2.7-3.4 MA)

The Pliocene, the most recent period of sustained warmth in Earth’s history has been extensively studied as an analog for future warm climate conditions. A diversity of paleotemperature proxies have been employed to characterize the pattern and evolution of Pliocene surface temperatures. These climate reconstructions often use estimates from various paleotemperature proxies interchangeably. However, there have been very few efforts to explore the validity of this approach. Here, we present an alkenone (U^k’₃₇) derived sea surface temperature (SST) record from DSDP Site 609 (50ºN, 24ºW, 3884 m water depth) in the North Atlantic Ocean for the interval from 2.7-3.4 Ma, which we compare to previously published Mg/Ca SST estimates from the same site and time interval based on G. bulloides. The two time series are structurally similar and yield similar mean values (U^k’₃₇ = 18.0ºC; Mg/Ca = 17.7ºC) and standard deviations (U^k’₃₇ = 1.1ºC; Mg/Ca 1.4ºC). In contrast, the range of temperature variations is considerably larger for Mg/Ca record (7.4ºC) as compared to U^k’₃₇ record (5.2ºC). The records also have very different spectral signatures with the U^k’₃₇ record dominated by 100k beats whereas the Mg/Ca record is dominated by 41k beats. Correlation analysis reveals a modest (r=.48), but statistically significant relationship between temperature estimates from the two proxies. Our data suggest that whether or not these proxies can be used interchangeably depends on the research question being explored and the time scale of comparison. At this locality, studies interested in the average climate conditions during this time interval would be insensitive to which of these SST proxies was employed. In contrast, the inconsistent results for some metrics are potentially problematic for comparisons on orbital timescales and for specific time slices that may be of interest to climate modelers. It is imperative that we consider potential causes for the differences in estimates that arise from temperature proxy comparisons at this and other sites, including seasonality, ecology, diagenesis, and calibration differences.