Recent deuterium excess icecore studies

The Vostok ice-core has provided a great deal of information about environmental changes in the past 420,000yr (Petit et al., 1999). In a SD-S18O diagram, the ice samples are beautifully aligned on a regression straight line with a slope of 7.94. Vimeux et al. (1999, 2001) have interpreted Vostok deuterium excess variations as depending on fluctuations of the temperature of the oceanic moisture source only, by using a relationship between relative humidity of the air and sea-surface temperature. The variation of the ocean isotopic composition has been removed from the deuterium excess profile by using the marine isotope record. This correction does not modify the long-term variations but significantly affects the amplitudes. Central to the analysis ofVimeux et al. (1999,2001) is the physically based assumption that the temperature difference between the vapour source and the precipitation site principally controls the SD of Antarctic snow, because net isotopic distillation during atmospheric transport from the source to the precipitation site is driven by fractional reduction of the air-mass water content. New information on source-region temperatures can be obtained by measuring the deuterium excess in precipitation.

Deuterium excess in the most recent 250,000 yr of the Vostok ice-core is dominantly modulated by the obliquity periodicity. During periods of low obliquity, the annual mean insolation at high latitudes is low and the latitudinal insolation gradient between 20°S and 60°S is maximized. This increases evaporation at low latitudes and latitudinal atmospheric moisture transport. The latter enhances the contribution of remote oceanic moisture sources and reduces the contribution of local cooler sources. All this together acts to increase the deuterium excess. An anticorre-lation has in fact been observed between the obliquity and the deuterium excess. It is interpreted in terms of the relative contribution of low and high latitudes to the precipitation at Vostok. The Vostok data allow all the interglacial periods and glacial inceptions in the past 420,000 yr to be examined. A constant relationship can be observed between deuterium and deuterium excess during these periods. The deuterium excess starts to increase during the warmest period and reaches a maximum value at the beginning of the next cold stage. Then the deuterium excess decreases through the glacial period. In all cases, the glacial inceptions occur when the obliquity is low and the relative contribution of low latitudes to Vostok precipitation is maximized. This suggests that, at glacial inceptions, the temperature of the oceanic surface at low latitudes remains at its interglacial level for some time after the high latitudes have abruptly cooled. Prior to 250,000yr, there is a lack of correlation with obliquity, probably due to a remote origin for the deep ice at Vostok station because of the ice flow.

The Vostok core also shows a correlation of CO2 and SD for the past 150,000yr, the strength of which is r2 = 0.64. This strength is limited primarily by the rapid decrease of SD during and immediately following the last interglacial. Such a large temperature drop is puzzling. Cuffey & Vimeux (2001), by using measurements of deuterium excess for the temperature reconstruction, were able to show that the relative mismatch of the CO2 and deuterium records is an artefact caused by variations of climate in the water vapour source regions. Using a model correcting for this effect, the co-variation of CO2 and temperature is clearly improved for the past 150,000yr (r2 = 0.89). This excellent correlation strongly supports the role of carbon dioxide as a forcing factor of climate.

Below a depth of 3310 m, where the age of the ice is about 420,000 yr and above accreted ice from subglacial Lake Vostok at 3538 m depth, there is evidence of complex ice deformation resulting in folding and intermixing of ice at a submetric scale and, for the upper part of this sequence (3310-3405 m), in interbedding of ice layers from distinct origins at a larger scale. Souchez et al. (2002) have used deuterium and deuterium excess properties of the ice to document the build up of these highly deformed basal ice layers. First, there is a damping of SD variations with depth from top to bottom of this core section. Second, the plotting of ice samples on a d-SD diagram (Fig. 35.1) shows striking features. The two oldest glacial-interglacial climatic cycles are displayed. They represent ice between 2755 and 3108 m depth (black circles) and between 3109 and 3309 m depth (open circles) respectively. The dot distribution takes the shape of a ring. Now, samples of ice below 3405 m depth (crosses) are also plotted in this figure. The crosses representing these ice samples cluster inside the ring. Moreover, if a vertical line (dotted) is drawn from the centre of the ring, splitting into two equal parts the maximum SD range for the climatic cycles displayed, it can be seen that most of the sample points from the two oldest climatic cycles are within the part containing the more negative values. As one sample was measured per metre of core, this indicates that the colder periods are more developed in terms of ice thickness than the warmer ones. Now, the crosses representing ice samples below 3405 m depth are all included within this part containing the more negative values. Such characteristics point to the occurrence of a folding/mixing process for the basal ice. Indeed folding/mixing

" r ? •

" J ■ ;




X " O

. o r

O 9° #

• MQQtiet i'x.. vS**


• o

S1** W



o ' • • • X

x 0


° !

X ».

. . •


1 1

Figure 35.1 d-SD diagram in meteoric ice from the Vostok ice-core. (Reproduced by permission of American Geophysical Union from Souchez et al. (2002). Copyright American Geophysical Union.)

produces ice having isotopic compositions intermediate between those of the ice layers involved, before they were deformed, hence the clustering of the sample points in the hole of the ring. There is a higher probability that the intermediate isotopic compositions produced result from folding/mixing of ice from colder periods, more frequently present at depth. The distribution within the part containing the more negative SD values thus can be understood. Within this context, the reduction in amplitude in the SD variations with depth can be viewed as implying more complex ice deformation at depth, a very likely situation if one considers the uneven bedrock topography where the ice is grounded upstream from Vostok station. Deformation more complex than simple shear is most pronounced close to the bed, where ice viscosity is reduced by higher temperatures.

In the EPICA ice-core at Dome C in East Antarctica, deuterium excess study of the ice from the last deglaciation period reveals the timing and strength of the sea-surface temperature changes at the source regions for Dome C precipitation (Stenni et al., 2001). It can be shown that an Oceanic Cold Reversal took place in the southern Indian Ocean 800 yr after the Antarctic Cold Reversal, a cold period of the last deglaciation. During this deglaciation period, the temperature gradient between the oceanic moisture source and the Antarctic continent shows a temporal trend similar to the Dome C sodium profile, illustrating the strong link between this temperature gradient and the strength of the atmospheric circulation.

Was this article helpful?

0 0

Post a comment