Info

^JAN \

0

JAN

JUL

JUL

-

-5 -I

JL

JN

,ii

JN

>_JUL

%

%

1 -

%

%

\ 10-

■10

—-Í----

■10

JUL ___

10i

r*"""jAN

10-

■10

\ JAN,.' v.__

-»iH

ANNUAL

18 15 12

FIGURE 12.21 Departures of (top to bottom): surface temperature (°C) sea-level pressure (mb), precipitation (mm day"'), and precipitation minus evaporation (P-E, mm day"') for January and July, for northern hemisphere (left) and southern hemisphere (right) land areas for the period 18 ka B.P. to the present at 3000-yr intervals. Departures are expressed as the paleoclimatic experiment minus the control (0 years B.P.). Large dots indicate the departure is statistically significant (2-sided t-test) at or above the 95% confidence level, based on the model's natural variability. For precipitation, the ordinate on the right indicates ± 10% departures for January (JN) and July (JL) (Kutzbach and Guetter, 1986).

and coupled ocean-atmosphere GCMs have been used to examine this issue and several studies show that the oceanic circulation in the Atlantic can switch between these modes and the transitions can be rapid (Manabe and Stouffer, 1988; Weaver and Hughes, 1994; Rahmstorf, 1994, 1995). The system is strongly non-linear so that at some critical threshold an abrupt switch from one mode to another can occur, with dramatic consequences for climate in the northern hemisphere, especially in western Europe. However, several authors have argued that the notion of a conveyor system either on or off is not consistent with geological evidence and it is more likely that there were many "modes" or situations in which some regions of deepwater formation ceased to operate, and the main centers shifted geographically. Indeed, there may have been times when deepwater was replaced by intermediate water (see discussion in Sections 6.9 and 6.10.2) (Boyle and Keigwin, 1987; Lehman and Keigwin, 1992a). Experiments with the GFDL ocean circulation model, coupled to an atmospheric GCM show that there can be multiple equilibria in which the main convection sites in the North Atlantic shift in response to varying inputs of meltwater or atmospheric circulation changes, or both (Rahmstorf, 1994). Sea-surface temperatures can change by up to 5 °C during these shifts with Intermediate Water replacing NADW.

Manabe and Stouffer (1997) used a coupled ocean-atmosphere model to examine the response of the thermohaline circulation (and global climate in general) to the effect of a sustained freshwater input across the North Atlantic (50-70° N) compared to a similar input across the western subtropical Atlantic (20-29° N, 52-90° W). This experiment was designed to distinguish between the effects of meltwater drainage from the Laurentide ice sheet southward into the Gulf of Mexico, and thence to the Atlantic, vs drainage through the Gulf of St. Lawrence. Broecker et al. (1989) argued that this switch may have been responsible for the Younger Dryas cold episode, which has been noted in many parts of the world, but especially in western Europe (see Section 6.10.2). In Manabe and Stouffer's two experiments, they added 0.1 Sv (106 m3 s"1) of freshwater for 500 yr to each region, then ran the model for another 750 yr to determine the long-term response of the overall climate system. Such long simulations are very computer-intensive and so few studies have been carried out to examine questions of this sort. Figure 12.22 shows changes resulting from the northern freshwater input, in sea-surface salinity (SSS), sea-surface temperature (SST), and sea-ice thickness in the Denmark Strait (30° N, 65.3° W) off the coast of East Greenland, where the response to changes in thermohaline circulation is large. Also shown is a measure of the overall thermohaline circulation, and air temperatures over Summit, Greenland (site of the GISP2/GRIP ice cores). Salinity fell by -3% over the 500-yr period of freshwater input, accompanied by a reduction in thermohaline circulation strength, from -18 Sv to -4 Sv. The SSTs also declined rapidly, accompanied by more extensive sea-ice growth in the Denmark Strait. Superimposed on this overall decline are large amplitude decade-to-century scale oscillations of SSS, SSTs, and air temperatures, as shown here for Greenland. These are reminiscent of the high amplitude changes of 8lsO seen in the late glacial section of the GISP2/GRIP ice cores (see Fig. 5.22). In a similar experiment, Manabe and Stouffer (1995) added large amounts of freshwater to the North Atlantic much more rapidly (over 10 yr) and found similar oscillations, but larger in amplitude, associated with a rapid cessation of the thermohaline circulation. Such abrupt changes therefore seem to be characteristic of the change in ocean circulation mode. Once the freshwater input ceased, SSS and the thermohaline circulation recovered their original values within 250 yr, with SSTs returning to control period values more slowly, delayed by the sea-ice cover, which constrained the warming. Figure 12.23 shows the geographical pattern of surface air temperature anomalies (from the control period) for four 100-yr intervals, over the course of the experiment. The largest anomalies at the end of the freshwater influx (years 401-500) are in the North Atlantic (as much as -7 °C over southeast Greenland), with cooler conditions extending throughout the high latitudes of the northern hemisphere. However, low- and mid-latitude experienced little change in temperature, and, indeed, the overall global mean temperature change in years 401-500 was close to zero. However, a notable anomaly is observed in the Southern Circumpolar Ocean, resulting from the reduction in thermohaline circulation and its consequences in the Southern Ocean. This anomaly continued long after the

1250

FIGURE 12.22 A coupled ocean-atmosphere GCM experiment was carried out, in which freshwater was introduced into the North Atlantic for a period of 500 yr, followed by a 750-yr period with conditions like those at the start of the experiment.The resulting mean annual time series are shown of: (a) sea-surface salinity; (b) sea-surface temperature; (c) sea-ice thickness (all in Denmark Strait, 30° W, 65.3° N, off the coast of east Greenland); (d) shows the strength of the thermohaline circulation (in Sverdrups: units = I06 m3 s '); and (e) gives the air temperature over Summit Greenland (degrees Kelvin). Note the decline in thermohaline circulation as the freshwater input continues, accompanied by large amplitude shifts in SSTs and air temperatures around Greenland. Recovery to pre-experiment (control) conditions after the freshwater input ceased (in year 500) was completed within 250 years (Manabe and Stouffer, 1997).

FIGURE 12.23 Surface air temperature anomalies (from the control experiment) for selected periods (years 201-300,401-500, etc.) during and after the North Atlantic freshwater influx experiment. Note the persistent anomaly in the Southern Ocean, and the minimal impact of North Atlantic changes on low- to mid-latitudes (Manabe and Stouffer, 1997).

FIGURE 12.23 Surface air temperature anomalies (from the control experiment) for selected periods (years 201-300,401-500, etc.) during and after the North Atlantic freshwater influx experiment. Note the persistent anomaly in the Southern Ocean, and the minimal impact of North Atlantic changes on low- to mid-latitudes (Manabe and Stouffer, 1997).

601 St-700th

120E 150E

601 St-700th

120E 150E

freshwater input to the North Atlantic ceased — a persistent cool episode that was especially pronounced west of the Antarctic Peninsula. A comparison with the observed paleoclimatic data shows that the pattern of anomalies is similar to that of the Younger Dryas, though displaced somewhat poleward. This may be because the model did not include an ice cover representative of late glacial conditions; had it done so, the temperature anomaly field would probably have looked even more like the Younger Dryas oscillation, as mapped, for example, by Peteet (1995). It is also of interest that the response to North Atlantic freshwater input is delayed cooling (by up to several centuries in some regions) and this may help to account for the apparent diachronous evidence for late glacial cooling that has often confounded attempts to define a "Younger Dryas" chron.

How do these changes differ from the experiment in which freshwater was injected farther south? Figure 12.24 shows SST anomalies in years 401-500 of the two experiments. The two patterns are remarkably similar, although the magnitude of anomalies associated with the southern freshwater injection is considerably smaller. Determining which forcing was more important will therefore be difficult, as there may have been a wide spectrum of conditions, spanning the range of these two exper-

(b)FWN

FIGURE I 2.24 Sea-surface temperature anomalies in years 401-500 resulting from a 500-yr influx of freshwater (a) into the subtropical Atlantic and (b) into the North Atlantic.The patterns are similar, but the magnitude of anomalies is larger with the North Atlantic freshwater influx (Manabe and Stouffer, 1997).

FIGURE I 2.24 Sea-surface temperature anomalies in years 401-500 resulting from a 500-yr influx of freshwater (a) into the subtropical Atlantic and (b) into the North Atlantic.The patterns are similar, but the magnitude of anomalies is larger with the North Atlantic freshwater influx (Manabe and Stouffer, 1997).

iments, at various times in the past. Fanning and Weaver (1997), for example, found in their ocean-atmosphere GCM experiments that North Atlantic freshwater discharge alone (via the St Lawrence) could not shut down the thermohaline circulation, but it could when the Atlantic was first "preconditioned" by freshwater discharge from the Mississippi drainage. These experiments help to explore further the dramatic changes in glacial and late glacial time recorded by paleoclimatic records. They must now be reconciled with arguments over the magnitude and timing of Laurentide (and Scandinavian) discharge as well as the role of ice discharged from the Arctic Ocean. Data from marine sediments, interpreted in terms of deep and intermediate water formation, and the shifting sources of these water masses must also be fitted into the puzzle. Although these lines of evidence do not all fit together perfectly at this time, it is clear that changes in thermohaline circulation and freshwater influx to the North Atlantic are critical factors in resolving these questions, with important implications for understanding the sensitivity of the thermohaline circulation to greenhouse gas-induced changes in the hydrological balance of the North Atlantic basin.

0 0

Post a comment