Amino acid racemization of planktonic foraminifera: pretreatment effects and temperature reconstructions

Billups, KatharinaAmino acid racemization (AAR) is a geochronological method that uses the ratio of D- to L- configurations in optically active amino acids from carbonate-based fossils to determine the time elapsed since the death of an organism. In well-dated fossil samples, the extent of racemization can be used to calculate post-depositional temperatures (also known as the effective diagenetic temperature). Calculated post-depositional temperatures of bracketed time intervals have uncertainties ranging from ±2 to 4°C with the dominant source of error in the D/L ratios (Kaufman 2003). Here, I aim to reduce these uncertainties using a bleach pretreatment that isolates the intra-crystalline fraction of amino acids in order to reduce the variability in foraminiferal D/L values to improve the precision of environmental paleotemperature estimates. I investigate the effect of this pretreatment method on the D/L ratios in three species of planktic foraminifera (Globorotalia tumida, Pulleniatina obliquiloculata, and Globorotalia truncatulinoides) from Holocene (~4-5 ka) deep sea sediments of similar environmental settings (Ocean Drilling Program Sites 1056, 1059, and 1062) and early Holocene to Pleistocene (~10.5-410 ka) sediments down-core (Ocean Drilling Program Site 1056 and KNR140 JPC-37). Results are reported for aspartic acid (Asp) and glutamic acid (Glu) because they are among the most abundant amino acids in foraminiferal protein and are the best resolved chromatographically. I analyzed 42 samples, each with an average of 9 replicates per sample and 5-10 individual tests per replicate, depending on the pretreatment method. Comparing D/L ratios from bleached versus unbleached samples indicates that bleaching only slightly reduces the variability in D/L values within the same sample (i.e. same species from the same core interval) by, on average, 1.1% and 3.0% for Asp and Glu, respectively. Furthermore, comparison of D/L ratios from the same species found at more than one site does not show statistical differences whether bleached or not. Bleaching does not appear to reduce the rejection rate or variability among replicates from the same sample from core intervals of similar age and environmental setting enough to warrant adding the additional time required to pick more tests for the bleaching procedure. Post-depositional temperatures calculated from unbleached D/L measurements of P. obliquiloculata and species corrected G. truncatulinoides give an account of the thermal history at ~2000 m (ODP Site 1056) and ~3000 m (ODP Site 1059 and KNR140 JPC-37) water depth at the Blake-Bahama Outer Ridge in the Western Atlantic. Paleotemperature estimates for the bracketed time interval of the Last Glacial Maximum (LGM; 15.4-30.8 ka) at ~3000 m water depth indicate >4°C cooling between the LGM and Holocene, consistent with other paleoclimate proxies (benthic foraminiferal δ18O and Mg/Ca). Currently, uncertainties of effective diagenetic temperatures averaged for the amino acids are ≤ ±0.6°C for T and ≤ ±1.6°C for T(t2-t1) due to an average inter-shell variation in D/L (±1σx) of ±1.6% and ±2.9% for Asp and Glu, respectively. Therefore, it is important to have minimal variability (standard deviation and standard error) in D/L values for temperature estimates that are reflective of less than 4°C uncertainties in bottom water temperature. For a more complete comparison of water temperatures at the ~2000 m and ~3000 m water depth, I would obtain a measurement on a 410 ka sample from JPC-37 (~3000 m water depth) in order to compare the effective diagenetic temperature of glacial-interglacial periods during the past 410 kyr in the Western Atlantic.University of Delaware, School of Marine Science and PolicyM.S