Evidence for 800 years of North Atlantic multi-decadal variability from a Puerto Rican speleothem Amos Winter a, ⁎, Thomas Miller b , Yochanan Kushnir c , Ashish Sinha d , Axel Timmermann e , Mark R. Jury f , Christina Gallup g , Hai Cheng h , R. Lawrence Edwards h a Department of Marine Sciences, University of Puerto Rico at Mayagüez, PR 00681, USA b Department of Geology, University of Puerto Rico at Mayagüez, PR 00681, USA c Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY 10964, USA d Department of Earth Science, California State University Dominguez Hills, Carson, CA 90747, USA e International Pacific Research Center, SOEST, University of Hawaii at Manoa, Honolulu, HI 96822, USA f Department of Physics, University of Puerto Rico at Mayagüez, PR 00681, USA g Department of Geological Sciences, University of Minnesota at Duluth, MN 55812, USA h Department of Geology and Geophysics, University of Minnesota at Minneapolis, MN 55455, USA abstract article info Article history: Received 3 March 2010 Received in revised form 13 May 2011 Accepted 14 May 2011 Available online 15 June 2011 Editor: P. DeMenocal Keywords: climate change Atlantic multidecadal oscillation rainfall variability speleothems Puerto Rico The long-term behavior of the tropical Atlantic ocean/atmospheric system prior to the 20th century is not well characterized due to a lack of high-resolution proxy records to extend the short instrumental record. Here we present the first reconstruction of rainfall variability for the western tropical Atlantic that spans the past 8 centuries and is derived from the δ 18 O of speleothem calcite. The δ 18 O of speleothem calcite at this Puerto Rican location varies primarily in response to changes in the amount of summer-time precipitation. The speleothem documents multi-decadal to centennial length oscillations in δ 18 O that point to large variations in rainfall that have not been manifest in the short instrumental period. Since AD 1850, variations in δ 18 O have tracked shifts in the Atlantic Multidecadal Oscillation (AMO). We tentatively suggest that the speleothem δ 18 O-based rainfall record from Puerto Rico extends the history of the AMO to the 12th century. © 2011 Elsevier B.V. All rights reserved. 1. Introduction The tropical North Atlantic exhibits climatically complex ocean/ atmosphere interactions that include local processes and localized responses to climate variability centered outside of the Caribbean itself (Chiang et al., 2002; Czaja et al., 2002; Xie and Carton, 2004). The trade winds modulate the seasonal cross-equatorial sea surface temperature (SST) gradient and influence the location and the intensity of the Atlantic Intertropical Convergence Zone (ITCZ). The tropical Atlantic SSTs vary in response to forcing centered in the tropical Pacific (El Niño/Southern Oscillation (ENSO) phenomenon), the North Atlantic (North Atlantic Oscillation), NAO, as well as from, the Atlantic Multi-decadal Oscillation (AMO, Knight et al., 2006). The AMO is a term coined by Kerr (2000) to refer to the spatially coherent, slow (multidecadal) variation of North Atlantic SSTs. An early description of this pattern of SST variability appears in Folland et al. (1986) in connection with the publication of their comprehensive atlas of global SST variability and their interest in the causes of the 1970s drought in the Sahel region. Kushnir (1994) examined the relationship between the SST pattern and atmospheric circulation variability and showed that it is different than the ocean–atmosphere relationship associated with interannual variability in the North Atlantic Basin. Schlesinger and Ramankutty (1994) argued for an influence of the AMO on Northern Hemisphere temperatures. The latter study also showed that as far as can be discerned from the relatively short observational record, the AMO “oscillates” with a period of ~ 70 years and amplitude of 0.4 °C. Later, Enfield and Mestas-Nuñez (1999) showed that the AMO emerges as the first rotated EOF of non-ENSO global SST. Coupled ocean–atmosphere model experiments reproduce multidecadal climate modes with patterns similar to that of the AMO (Delworth et al., 1993; Delworth and Mann, 2000; Knight et al., 2006; Timmermann et al., 1998), which suggests that it can arise from internal ocean dynamics. The AMO related ocean–atmosphere pattern suggested that ocean dynamics are primarily responsible for the SST changes, a hypothesis that was supported at that time by the modeling study of Delworth et al. (1993) in which oscillations with an ~ 50 year Earth and Planetary Science Letters 308 (2011) 23–28 ⁎ Corresponding author. E-mail address: amoswinter@gmail.com (A. Winter). 0012-821X/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2011.05.028 Contents lists available at ScienceDirect Earth and Planetary Science Letters journal homepage: www.elsevier.com/locate/epsl