| 英文摘要: | In 1974, newly available satellite observations unveiled the presence of a giant ice-free area, or polynya, within the Antarctic ice pack of the Weddell Sea, which persisted during the two following winters1. Subsequent research showed that deep convective overturning had opened a conduit between the surface and the abyssal ocean, and had maintained the polynya through the massive release of heat from the deep sea2, 3. Although the polynya has aroused continued interest1, 2, 3, 4, 5, 6, 7, 8, 9, the presence of a fresh surface layer has prevented the recurrence of deep convection there since 19768, and it is now largely viewed as a naturally rare event10. Here, we present a new analysis of historical observations and model simulations that suggest deep convection in the Weddell Sea was more active in the past, and has been weakened by anthropogenic forcing. The observations show that surface freshening of the southern polar ocean since the 1950s has considerably enhanced the salinity stratification. Meanwhile, among the present generation of global climate models, deep convection is common in the Southern Ocean under pre-industrial conditions, but weakens and ceases under a climate change scenario owing to surface freshening. A decline of open-ocean convection would reduce the production rate of Antarctic Bottom Waters, with important implications for ocean heat and carbon storage, and may have played a role in recent Antarctic climate change. Antarctic Bottom Water (AABW) is the coldest, densest and most voluminous11 water mass of the world ocean and its shrinking in recent decades12, 13 has been linked to deep ocean heat uptake12, 14. Produced at present on Antarctic continental shelves, AABW is exported northwards to fill the deepest layers of the three oceanic basins and feed the deep branch of the meridional overturning circulation11, 15. In 1928, on the basis of early hydrographic observations, it was suggested16 that open-ocean convection also contributes to the production of AABW, as it does to North Atlantic Deep Water in the Labrador Sea. It was argued that deep convection occurred within the Weddell Gyre, but because of difficulty monitoring the Weddell Sea during austral winter, this contention went unverified until the mid-1970s6. Microwave observing satellites were first launched in December 1972, providing global observations of sea ice, and soon thereafter revealed the presence of a 250,000-km2 ice-free area within the seasonally ice-covered Weddell Sea1 (Fig. 1a). The huge polynya, located near Maud Rise (65° S, 0°), reappeared during the winters of 1974 to 1976, slowly drifting westward with the background flow1. The polynya was maintained by vigorous convective mixing, whereby the upward flux of relatively warm deep waters supplied enough heat to prevent sea ice formation1, 2, 3. Heat loss at the surface drove cooling to depths of about 3,000 m, producing new deep waters3 that could have fed the observed surge in Southern Ocean AABW volume during the following decade13. Together with the inference from hydrographic data of a Weddell convective event circa 19603, these observations confirm that deep convection in the open Weddell Sea has been a significant mode of AABW ventilation5, 6. However, following 1976, no similar polynya has been observed. The continuing quiescence over the past 37 of 41 available years of satellite observation makes it tempting to assume that deep convection in the open Southern Ocean occurs rarely, with little global consequence. Here we propose, instead, that deep convection was more common in the pre-industrial state, but that the hydrological changes associated with global warming17, 18, 19 are now suppressing this convective activity.
|
