Rapid climate change and Arctic Ocean freshening W.R. Peltier Department of Physics, University of Toronto, Toronto, Ontario M5S 1A7, Canada

The idea that global climate can change rap- in character. This is the so-called Younger- idly has been clearly demonstrated on the basis Dryas cold interval that began at ca.12.8 ka, of oxygen isotopic stratigraphies from the deep during which the general climate warming ice cores drilled at Summit, Greenland (Green- that occurred during the transition from Last land Ice-core Project [GRIP] Members, 1993; Glacial Maximum to Holocene conditions was Taylor et al., 1993). Because δ18O measured in interrupted by a return to the cold conditions of the ice is a proxy for the temperature of the air the glacial that lasted ~1000 yr. Until recently from which Greenland precipitation is derived, it was generally accepted that this event was the ice sheet itself acts as a “paleothermometer,” linked to a freshening of surface conditions over one that may be accurately calibrated using gas the Greenland-Iceland-Norwegian seas, due to isotope data and air thermal diffusion constants a massive release of fresh water from glacial (e.g., Gatchev and Severinghaus, 2005). These Lake Agassiz into the North Atlantic Ocean that records make it clear that extremely rapid excur- occurred when the fl ow of fresh water from the sions in the air temperature over Greenland have waning Laurentide Ice Sheet was diverted from been characteristic of both Northern Hemi- the south, through the Mississippi River System, sphere glacial conditions during oxygen isotope to the east, through the Great Lakes and the St. stage 3 (OIS3, 65– 25 ka) and of the last glacial- Lawrence river system (Broecker et al., 1989). Figure 1. Freshwater “hosing” areas over the to-interglacial transition (ca. 21–10 ka). The Very recently, this explanation for the Younger- North Atlantic and Arctic Oceans employed for the purpose of comparing the impact variability during these intervals has been char- Dryas event has been called into question by of surface ocean freshening on Atlantic acterized by millennial-timescale excursions the diffi culty encountered in identifying the thermo haline circulation. from warm to cold and cold to warm conditions spillway(s) through which the required massive in which the local temperature variations asso- fl ood of fresh water is suggested to have passed ciated with the individual events have been on (Lowell et al., 2005). This led to a recent detailed application of the same intensities of freshwater the order of 10 °C. Individual fl uctuations have analysis of the time-dependent deglacial routing forcing over the two regions. been characterized by a “sawtooth” form, rather of meltwater to the sea, based upon the applica- The study by Knies et al. (2007, p. 1075 of this like the 100 k.y. Late Quaternary ice-age cycle tion of a carefully calibrated continental-scale issue) considers the question of Arctic Ocean itself, with the transition from cold to warm ice dynamics model which was coupled to a freshening from the perspective of the entire atmospheric conditions occurring on a time detailed model of the glacial isostatic adjustment Late Pleistocene interval of time. They employ scale that is short compared to the time scale of process and superimposed upon a hydrologically calcium carbonate records from deep sea cores cooling. This two-timescale characterization of self-consistent digital elevation model (Tarasov raised from the vicinity of Fram Strait, through the individual pulses suggests that they are the and Peltier, 2005, 2006). This analysis strongly which Arctic Ocean surface water fl ows south- consequence of a physical process that involves suggested that the meltwater fl ood that occurred wards onto the surface of the Greenland-Iceland- what is referred to in the language of dynamical at the Younger-Dryas onset did not occur as a Norwegian seas. They discuss an 800 ka record systems as a “relaxation oscillation.” consequence of a switch in fl ow direction from that covers the time from the mid-Pleistocene The most compelling explanation of these the south to the east, but rather from the south climate transition, following which the 100 k.y. events is in terms of variability in the strength of to the north through the McKenzie River outlet glacial cycle dominated climate system variabil- the thermohaline overturning circulation (THC) onto the Arctic Ocean surface over the Beau- ity. The quality of this record has enabled them of the Atlantic Ocean, which is primarily driven fort Gyre. Peltier et al. (2006) showed that such to identify a strong imprint of Arctic-derived by the formation of deep water in the Greenland- freshening of the surface of the Arctic Ocean freshwater pulses that have passed through the Iceland-Norwegian seas (e.g., Broecker et al., would have been as effi cient a means of shutting Fram Straight. Such events appear to have been 1989). The atmosphere is warmed over the down the Atlantic THC as would direct Atlan- an enduring sub-Milankovitch–timescale char- North Atlantic adjacent to Greenland when tic freshening. Figure 1 illustrates the different acteristic of the glaciation-deglaciation process, the THC is strong, but is cooler when the THC is geographical regions over which freshwater and are shown to have been strongly linked to weak. The fast time scale of the relaxation oscil- forcing was applied in order to demonstrate this covariation in the strength of the THC. Knies lation is determined by the convective destabili- equivalence. Direct paleoceanographic evidence et al.’s summary comment, that “most spectac- zation of the water column that occurs when also shows an Arctic source of the freshwater ular are freshwater pulses during glacial termi- the THC fl ips from an “off” to an “on” mode. loading event that triggered the Younger-Dryas nations” (p. 1077 of this issue), suggests that The long time scale is determined by a slow, cold reversal. Figure 2 compares the results of a Younger-Dryas–type cooling events may have diffusion-dominated process that determines the sequence of Arctic and Atlantic “water hosing” been a rather common feature of the variability time needed to re-establish convective instability experiments, performed using the National Cen- that occurs when the climate system undergoes of the water column in the Greenland-Iceland- ter for Atmospheric Research (NCAR) Climate a glacial-to-interglacial transition. Norwegian seas after the THC has collapsed System Model (CSM 1.4), that demonstrate the Knies et al.’s work is especially interesting (e.g., see Peltier and Sakai, 2001). almost identical response of the strength of both because it raises the question of why Arctic The millennial-timescale event that has been the Atlantic overturning circulation and North freshening events may be especially effi cient most studied is actually the most anomalous Atlantic surface air temperature (SAT) to the in their ability to impact the THC and thus

© 2007 The Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. GEOLOGY,Geology, December December 2007; 2007 v. 35; no. 12; p. 1147–1148; doi: 10.1130/focus122007.1; 2 fi gures. 1147 Figure 2. A–B: Response of Atlantic thermohaline circulation and North- ern Hemisphere sea ice ex tent to 100 yr of fresh- en ing at the rate of either 1 Sv (1 Sv = 106 m3/s) or 0.3 Sv applied to the Atlantic and Arctic Oceans, respectively, over the areas illustrated in Figure 1. C–D: Variations in surface air tempera- ture predicted at Sum- mit, Greenland, due to the alternative freshen- ing scenarios. Note that the response to Atlantic and Arctic freshening in both the strength of the thermohaline overturning circulation (THC) and surface air temperature is essentially identical.

Northern Hemisphere climate. It is now clear, Fram Strait, most probably as pack ice, it Reviews of Geophysics, v. 42, p. RG1005, doi: for example, that even massive fl oods of fresh is then delivered to the Greenland-Iceland- 10.1029/2003RG000128. water into the Gulf of Mexico, such as occurred Norwegian seas where it melts and is as effec- Lowell, T., Waterson, N., Fisher, T., Loope, H., Glover, K., Comer, G., Hajdas, I., Denton, G., during meltwater pulse 1a at ca. 14.2 ka, syn- tive in arresting the THC as are the Heinrich Schaefer, I., Rinterknecht, V., Broecker, W., chronous with the onset of the Bölling warm event-related armadas. The Knies et al. study and Teller, J., 2005, Testing the Lake Agassiz interval, failed to impact the strength of the is extremely useful in establishing this process meltwater trigger for the Younger Dryas: EOS, THC. I have previously argued, citing the lab- as having operated ubiquitously under glacial Transactions, American Geophysical Union, v. 86, p. 365, doi: 10.1029/2005EO400001. oratory experiments of Parsons et al. (2001), conditions. However, more detailed analysis Parsons, J.D., Bush, J.W., and Syvitski, J.P., 2001, that this was most probably due to the heavily is warranted to establish that Younger-Dryas– Hyperpycnal plume formation from river- sediment-laden nature of the riverine outfl ow, type events were characteristic of the glacial- ine outfl ows with small sediment concentra- which caused the meltwater plume to enter interglacial transitions that preceded the most tions: Sedimentology, v. 48, p. 465–478, doi: the ocean hyperpycnally (i.e., the fresh water recent such event. 10.1046/j.1365-3091.2001.00384.x. Peltier, W.R., and Sakai, K., 2001, Dansgaard- fl ows under the salt water rather than over REFERENCES CITED Oeschger oscillations: A hydrodynamic theory, the surface). This hypothesis has since been Aharon, P., 2006, Entrainment of meltwaters in in Schafer, P., Ritzrau, W., Schuter, M., and verifi ed by Aharon (2006) based upon time- hyperpycnal fl ows during deglaciation super Thiedi, J., eds., The Northern North Atlantic: synchronous analyses of δ18O in both planktic fl oods in the Gulf of Mexico: Earth and Plan- A Changing Environment: Berlin, Springer- etary Science Letters, v. 241, p. 260–270, doi: Verlag, p. 423–440. and benthic forams from Gulf of Mexico sedi- Peltier, W.R., Vettoretti, G., and Stastna, M., 2006, ment cores. The effi ciency with which a given 10.1016/j.epsl.2005.10.034. Broecker, W.S., Kennet, J.P., Flower, B.P., Teller, Atlantic meridional overturning and climate freshwater pulse may impact the strength of J.T., Trumbore, S., Bonani, B., and Wolfl i, W., response to Arctic Ocean freshening: Geo- the THC therefore depends upon the mecha- 1989, Routing of meltwater from the Lauren- physi cal Research Letters, v. 33, p. L06713, nism by which the fresh water is delivered to tide Ice Sheet during the Younger-Dryas cold doi: 10.1029/2005GL025251. Tarasov, L., and Peltier, W.R., 2005, Arctic fresh- the Greenland-Iceland-Norwegian seas. Dur- episode: Nature, v. 341, p. 318–321, doi: 10.1038/341318a0. water forcing of the Younger-Dryas cold rever- ing Heinrich events, the mechanism is effi cient Gatchev, A.M., and Severinghaus, J.P., 2005, A revised sal: Nature, v. 435, p. 662–665, doi: 10.1038/ because this delivery occurs through the inter- +10 ± 4 °C magnitude of the abrupt change in nature03617. mediary of “armadas of icebergs” whose tracks Greenland temperature at the Younger-Dryas Tarasov, L., and Peltier, W.R., 2006, A calibrated termination using published GISP2 gas iso- deglacial chronology for the North American are preserved on the ocean fl oor in the form of continent: Evidence of an Arctic trigger for the North Atlantic Ice Rifted Debris belt (e.g., tope data and air thermal diffusion constants: Quaternary Science Reviews, v. 24, p. 513–519, the Younger-Dryas event: Quaternary Science Hemming, 2004). Arctic freshwater outfl ow doi: 10.1016/j.quascirev.2004.10.016. Reviews, v. 25, p. 659–688. may be especially effective because the stabil- Greenland Ice-core Project (GRIP) Members, 1993, Taylor, K.C., Lamorey, G.W., Doyle, G.A., Alley, R.B., ity of the water column in the Arctic Ocean is Climate instability during the last interglacial Grootes, P.M., Mayewski, P.A., White, J.W.C., and Barlow, L.K., 1993, The “fl icking switch” especially strong. Freshwater outfl ows, even if period recorded in the GRIP ice core: Nature, v. 364, p. 203–207, doi: 10.1038/364203a0. of Late Pleistocene climate change: Nature, heavily sediment laden, are therefore unlikely Hemming, S.R., 2004, Heinrich events: Massive v. 361, p. 432–436, doi: 10.1038/361432a0. to enter the ocean hyperpycnally. When the Late Pleistocene detritus layers of the North surface water thus produced is forced through Atlantic and their global climate imprint: Printed in USA

1148 GEOLOGY, December 2007