Little Ice Age climate in the Western Tropical Atlantic inferred from coral geochemical proxies

Abstract

Paleoclimate archives place the short instrumental record of climate variability in a longer temporal context and allow better understanding of the rate, nature and extent by which anthropogenic warming will impact natural and human systems. The ocean is a key component of the climate system and records of past ocean variability are thus essential for characterizing natural variability and quantifying climate sensitivity to radiative forcing. Coral skeletons are high-resolution archives of tropical sea surface temperatures (SSTs), but inconsistencies call the accuracy of existing coral proxy records into question.In this thesis, I first quantify the errors associated with the traditional coral thermometer, Sr/Ca, by comparing in situ logged SST with Sr/Ca-derived SST in four corals on the same reef. I show that intercolony disparities in mean Sr/Ca, amplitude of variability, and trend are not due to differences in water temperature, but rather to "vital effects" that result in a ± 2 C uncertainty on reconstructed SST.I then expand, refine, and test a new paleothermometer, Sr-U, across multiple coral species and through time. I show that Sr-U captures spatial SST variability with an uncertainty of ± 0.6 C. When applied to two corals outside of the calibration, Sr-U accurately captures the mean SST and the 20 th century trend in the Western Tropical Atlantic.Finally, I apply Sr-U to a coral from the Little Ice Age (LIA) to address uncertainties in the magnitude of western tropical Atlantic cooling during a 95-year period spanning 1465-1560. Results suggest the region was 1.1 C±0.6°C cooler than the 1958-1988 mean, but within error of early 20 th century SST at this site. Critically, several periods of warmth, equivalent to the 1958-1988 mean, occurred during a solar minimum that is widely believed to have been a cool period of the LIA. My results indicate that Sr/Ca exaggerates the actual cooling by almost 3 °C. My record demonstrates the value of Sr-U and highlights the need for continuous accurate SST records to better constrain the amplitude, drivers, and mechanisms of LIA tropical climate change.

Mid‐Holocene, Coral‐Based Sea Surface Temperatures in the Western Tropical Atlantic

“…A consistency standard accounted for instrumental drift. Reported error of the measurements on the Element2 is quantified by repeated measurements over 2 years of the consistency coral standard and indicates an external precision of ±0.04 mmol/mol for Sr/Ca (1σ, n = 274) and ±0.02 μmol/mol for U/Ca (1σ, n = 274) as reported by Alpert (). Error on the iCAP was calculated using the same consistency standard and indicates a precision of ±0.05 mmol/mol for Sr/Ca (1σ, n = 29) and ±0.01 μmol/mol for U/Ca (1σ, n = 29).…”

“…Sr/Ca and U/Ca data and Sr‐U estimates from two of the cores were published previously. They are rebinned and the Sr‐U values recalculated here (Alpert, ; Alpert et al, ). One core was extracted from a living Siderastrea siderea colony in the Puerto Morelos Reef Park, Yucatan Peninsula (core ID# Jardin C), and one from a living Orbicella faveolata colony in Mona Island, Puerto Rico (core ID# PR‐OF‐001).…”

“…colonies of mid‐Holocene age, collected in the Enriquillo Valley, western Dominican Republic, were analyzed for this study (Figure and Table ). In addition, we rebinned previously generated Sr/Ca and U/Ca data (A. Alpert, ) for the early LIA, recalculated the Sr‐U values, and applied the new temporal calibration to derive LIA SSTs. At the Enriquillo Valley Canada Honda fossil reef site, 11 corals were initially sampled from strata with an estimated age range of 7,000 to 5,000 years BP (Collazo, ; Cuevas et al, ; Hubbard et al, ; Figures b and c).…”

“…Sr‐U values were also calculated from Sr/Ca and U/Ca data previously generated from a fossil Orbicella faveolata coral from Pinnacles Reef, Puerto Rico (Alpert, ; Table ). U‐series dates of the coral E1P indicate the colony died around 1670 CE (Kilbourne et al, ).…”

“…We then applied the calibration to two U‐series‐dated, mid‐Holocene‐aged corals from the Enriquillo Valley, Dominican Republic, to reconstruct mid‐Holocene SSTs. In addition, we applied the new temporal calibration to existing Sr‐U data from a LIA coral spanning the time period 1445–1669 Common Era (CE; Alpert, ) and present an averaged LIA SST estimate from that coral for comparison with average mid‐Holocene and twentieth‐century SSTs.…”