Kristina A. Dahl Æ Anthony J. Broccoli Ronald J. Stouffer Assessing the role of North Atlantic freshwater forcing in millennial scale climate variability: a tropical Atlantic perspective Received: 15 March 2004 / Accepted: 5 November 2004 / Published online: 22 February 2005 Ó Springer-Verlag 2005 Abstract This study analyzes a three-member ensemble of experiments, in which 0.1 Sv of freshwater was ap- plied to the North Atlantic for 100 years in order to address the potential for large freshwater inputs in the North Atlantic to drive abrupt climate change. The model used is the GFDL R30 coupled ocean–atmo- sphere general circulation model. We focus in particular on the effects of this forcing on the tropical Atlantic region, which has been studied extensively by paleocli- matologists. In response to the freshwater forcing, North Atlantic meridional overturning circulation is reduced to roughly 40% by the end of the 100 year freshwater pulse. Consequently, the North Atlantic re- gion cools by up to 8°C. The extreme cooling of the North Atlantic increases the pole-to-equator tempera- ture gradient and requires more heat be provided to the high latitude Atlantic from the tropical Atlantic. To accommodate the increased heat requirement, the ITCZ shifts southward to allow for greater heat transport across the equator. Accompanying this southward ITCZ shift, the Northeast trade winds strengthen and precip- itation patterns throughout the tropical Atlantic are al- tered. Specifically, precipitation in Northeast Brazil increases, and precipitation in Africa decreases slightly. In addition, we find that surface air temperatures warm over the tropical Atlantic and over Africa, but cool over northern South America. Sea-surface temperatures in the tropical Atlantic warm slightly with larger warm anomalies developing in the thermocline. These re- sponses are robust for each member of the ensemble, and have now been identified by a number of freshwater forcing studies using coupled OAGCMs. The model responses to freshwater forcing are generally smaller in magnitude, but have the same direction, as paleoclimate data from the Younger Dryas suggest. In certain cases, however, the model responses and the paleoclimate data directly contradict one another. Discrepancies between the model simulations and the paleoclimate data could be due to a number of factors, including inaccuracies in the freshwater forcing, inappropriate boundary condi- tions, and uncertainties in the interpretation of the paleoclimate data. Despite these discrepancies, it is clear from our results that abrupt climate changes in the high latitude North Atlantic have the potential to signifi- cantly impact tropical climate. This warrants further model experimentation into the role of freshwater forc- ing in driving climate change. 1 Introduction This study examines the response of the tropical Atlantic to a freshwater input in the North Atlantic using a coupled ocean–atmosphere general circulation model. The motivation for performing freshwater forcing experiments such as this one stems from the hypothesis that millennial-scale climate changes are driven by rapid increases in freshwater input to the North Atlantic from the melting of continental ice sheets (Broecker et al. 1985, 1988). Such freshwater inputs may cause a density stratification of the water column sufficient enough to weaken the formation of North Atlantic Deep Water (NADW) and northward surface flow in the Atlantic, K. A. Dahl (&) Massachusetts Institute of Technology/ Woods Hole Oceanographic Institution Joint Program, WHOI MS #22, Woods Hole, MA 02543, USA E-mail: kdahl@whoi.edu Tel.: +1-508-2892853 Fax: +1-508-4572187 A. J. Broccoli Department of Environmental Sciences, Rutgers University, 14 College Farm Road, New Brunswick, NJ 08901, USA R. J. Stouffer Geophysical Fluid Dynamics Laboratory, P.O. 308, Princeton, NJ 08542, USA Climate Dynamics (2005) 24: 325–346 DOI 10.1007/s00382-004-0499-5