Introduction

The sea ice cover in the Arctic Ocean has declined in the last 40 years and reaches the extreme condition as shown in Figs. 1 and 2, while its decadal variability has increased (e.g., Wang and Ikeda 2000). This fact makes us imagine that the ice cover will change into a seasonal one in near future. Actually, the observed ice decrease seems faster than that in the IPCC Report prediction, in which the ice cover in summer is predicted to become minimal near the end of the 21st century (Stroeve et al. 2007). In particular, the summer ice cover hit a record low in 2007 so that some specialists may believe disappearance by 2020. On the basis of significant year-to-year variability, however, such an early disappearance might be exaggeration. Therefore, we really need to examine the mechanisms crucial for the rapid ice decrease.

Fig. 1. Observed ice cover concentration in the arctic basin. Ice reduction is significant in last 30 years.

• Beaufort and Chukchi

• East Siberian and Laptev

• Barents and Kara

1950 1960 1970 1980 1990 2000 Year

Fig. 2. Ice cover trends and decadal variability in three regions, the Beaufort-Chukchi Sea, the East Siberian-Laptev Sea and the Barents-Kara Sea. A 3-year Fanning file is applied to the ice cover time series.

Winter summer

1950 1960 1970 1980 1990 2000 Year

Fig. 2. Ice cover trends and decadal variability in three regions, the Beaufort-Chukchi Sea, the East Siberian-Laptev Sea and the Barents-Kara Sea. A 3-year Fanning file is applied to the ice cover time series.

The Polar Vortex is a major cyclonic circulation in the Arctic atmosphere and has more significant decadal oscillations than the trend. The recently archived data extending from clouds, the atmospheric boundary layer to biogeochemical components in the Arctic Ocean have been analyzed for providing a close insight into Arctic environmental change, which may occur in response to global warming or as part of natural variability. The achieved results include the following signals: the cloud cover has increased and is estimated to contribute to the ice reduction through the radiation balance at the magnitude similar to the ice-albedo feedback (Ikeda et al. 2003). The atmospheric boundary layer thickness has reduced, and the stratosphere is cooling as a result of global warming. The biogeochemical data indicate vertical motion in the ocean interior responding to the variable Polar Vortex (Ikeda et al. 2005)

In the Arctic, the most pronounced atmospheric pattern is the Northern Annular Mode (NAM), which was first reported as the Arctic Oscillation by Thompson and Wallace (1998). A difference between this mode and the North Atlantic Oscillation (NAO) has been discussed from the viewpoints of statistics and dynamics: which mode is dynamically meaningful. The horizontal pattern is the intensified/weakened Polar Vortex with a significant vertical coherence from the surface to the stratosphere. The decadal signal had a clear peak around 1990 with strong Polar Vortex in Fig. 3. The less ice anomalies occurred around this peak,

■3

1950 1960 1970 1980 1990 2000 Year

2 year lag

winter

1950 1960 1970 1980 1990 2000 Year

2 year lag

iummer

1950 1960 1970 1980 1990 2000 Year

Fig. 3. Decadal variability in the arctic oscillation and sea ice cover produced by the AO. The AO is defined by a difference in the zonal mean sea level pressure between 70° N and 85° N.

1950 1960 1970 1980 1990 2000 Year iummer

Fig. 3. Decadal variability in the arctic oscillation and sea ice cover produced by the AO. The AO is defined by a difference in the zonal mean sea level pressure between 70° N and 85° N.

propagating from the Beaufort-Chukchi Sea, the East Siberian-Laptev Sea to the Barents-Kara Sea in several years. In contrary, a less ice anomaly occurred around 1998 in the Beaufort-Chukchi Sea but did not correlate with a positive AO. We attempt to explore the cause of this finding and expect to give insight into possible mechanisms responsible for the rapid ice decrease in last 10 years.

The decrease is enhanced in the Beaufort-Chukchi Sea, which leads us to examine inflow of the warmer Pacific Water through the Bering Strait. In section 4 Arctic pathway' ', sea surface height is examined to see if the higher sea level in the Bering Sea contributed to ice reduction. Then, the Arctic atmospheric circulation is further analyzed in section ' Wind-induced ice cover variability' ' to find correlation with ice anomalies in the Pacific sector. In section ' ' Discussion' ' , these results are discussed.

Arctic pathway

Aagaard and Carmack (1989) reported that the Pacific Water is an important source of fresh water in the Arctic Ocean. The Pacific Water is also a heat source. An analysis is now extended to the sea level in the Bering Sea vs. the Greenland Sea and compared with the current meter data through the Bering Strait. The interannual variability is examined to show whether a peak was induced as a consequence of atmospheric circulation and contributed to inflow of the Pacific Water into the Arctic Basin. This idea will be verified with the timing of a sea ice

Year

Fig. 4. Wind stress curl over the Bering Sea and the East Greenland Sea.

Year

Fig. 4. Wind stress curl over the Bering Sea and the East Greenland Sea.

reduction in the Pacific sector. Since a significant increase in heat flux may be associated with the inflow, this component should be included in the model prediction for answering the question when the Arctic sea ice becomes a seasonal ice cover.

It is well accepted that a basin-scale sea level varies responding to wind-driven circulation. As cyclonic (anticyclonic) wind stress curl drives the ocean, the first baroclinic mode develops in time associated with surface Ekman divergence (convergence) and propagates westward, and the sea level descends (rises) in the central part of the basin. Following this simple concept, the NCEP/NCAR data were processed over the Bering Sea and also the East Greenland Sea for 50 years. Both analysis domains have zonal lengths of 40° and meridional lengths of 20°: i.e., the Bering Sea domain covers (170° E-150° W, 40° N-60° N), and the East Greenland Sea domain covers (30° W-10° E, 60° N-80° N). As shown in Fig. 4, a multi-decadal variability is evident over the Bering Sea, while it is mainly decadal over the East Greenland Sea.

Fig. 5. Sea level in the central part of the Bering Sea estimated from the wind stress curl over the Bering Sea. It is compared with the steric heights in three zonal sections of the Bering Sea produced from World Ocean Atlas, the TOPEX/POSEIDON altimeter height in the Bering Sea during the ice-free season and the current meter data in the Bering Strait (Woodgate et al. 2006).

Fig. 5. Sea level in the central part of the Bering Sea estimated from the wind stress curl over the Bering Sea. It is compared with the steric heights in three zonal sections of the Bering Sea produced from World Ocean Atlas, the TOPEX/POSEIDON altimeter height in the Bering Sea during the ice-free season and the current meter data in the Bering Strait (Woodgate et al. 2006).

The wind-driven general circulation is established as a non-dispersive, first baroclinic Rossby wave, whose phase speed is determined to be about 0.01 m s-1 from the stratification in the Bering Sea. Within the analysis domain with a 2,500km zonal length, the sea level is essentially similar to the reversed wind stress curl smoothed in time and lagged by 2-3 years. Figure 5 shows the sea level in the Bering Sea with a distinct peak around 1998, following a generally low sea level for the period of 1975 through 1995. This peak is also shown in the altimeter data in the Bering Sea so that the wind-driven sea level may be meaningful and interpreted to cause the low ice anomaly in the Pacific sector. The hydrographic data in World Ocean Atlas were analyzed for the layer above 1,000-m depth and indicate a growing trend in the Bering Sea, whereas the recent data are still sparse. The current meter data are unfortunately missing for the peak, although the trough around 2002 is consistent with the trough in the altimeter data (Woodgate et al. 2006).

An additional interesting feature is an earlier high sea level all way through 1960s in a consistent manner with the steric height. This anomaly could be related to the Great Salinity Anomaly (GSA), during which freshwater was exhausted toward the Greenland Sea and the entire northern North Atlantic in late 1960s to early 1970s (Dickson et al. 1988). Although its sign has been well recognized, it is debatable whether the source existed in the Pacific.

The thermodynamic effect is estimated for the pathway transport. Once it increases the transport by 10% (105 m3 s-1) of the Pacific Water at 3°C, extra heat flux melts 2-m thick sea ice over 105 km2 in a year. In winter, sea ice decays from its bottom but does not enhance air-sea heat flux so much, and then, the albedo-ice feedback works in summer and melts more ice.

In this paper, a higher sea level in the Bering Sea was suggested to be one of the causes of a less ice anomaly in the Pacific sector. A wind-driven general circulation was shown to be related to the sea level variability. It is noted that this view is based on the assumption that pressure field is horizontally uniform in the lower ocean. However, there is a topographic barrier between the Bering Sea and the Arctic Ocean. A further study is needed for clarifying validity of the steric height concept. One of the possible methods is to use a global ocean model to simulate the sea surface heights in the Bering Sea and the Arctic Ocean, which are connected all way across the Pacific and Atlantic equatorial regions and the Antarctic Circumpolar Current region. Various processes in these regions are well imagined to produce a pressure difference in the lower parts of the Bering Sea and the Arctic Ocean.

In addition to sea level variability in the Bering Sea, a wind stress along the Bering Strait must play some roles on the pathway, as suggested by Woodgate et al. (2005). Actually, the pathway transport increased from 2002 to 2004 (Woodgate et al. 2006). It is an urgent task to evaluate which is dominant for the pathway transport, sea level in the Bering Sea or wind stress along the Bering Strait.

0 0

Post a comment