Anomalies in September ice conditions

There has been a general downward trend in Arctic sea ice extent (the region with at least 15% ice concentrations) during the passive microwave era (1979 onwards)

Plate 5 Mean September ice extent (all colored regions) and anomalies in ice concentration (see color bar) for the years 1990, 1993, 1995, 1998, 2002 and 2003 based on SSM/I data. The panels for 2002 and 2003 show the median September ice extent. Concentration anomalies and median extent are both referenced to the period 1988-2000 (courtesy of N. Geiger-Wooten, NSIDC, Boulder, CO). See color plates section.

(see Chapter 11). These ice reductions have been greatest in late summer and early autumn. Plate 5 shows fields of September sea ice extent and concentration anomalies for 1990, 1993, 1995, 1998 and 2003. These years had the lowest sea ice extents observed in the passive microwave record over the period 1979-2003. September 2004 (not shown) also had very mild ice conditions. For each year, we show the ice extent

(the total colored area) and the ice concentration anomalies with respect to 1988-2000 means. The median September ice extent over this period appears in the panels for 2002 and 2003. By September, the ice is beginning to refreeze such that one can get reasonable estimates of concentration from passive microwave imagery. Several of these anomalous years have been examined as case studies. They are reviewed here to illustrate how ice dynamics and thermodynamics interact to produce large sea ice anomalies. In Chapter 11,we will revisit recent trends in sea ice extent (and thickness) in the broader contexts of the Arctic Oscillation and recent warming (Rigor et al., 2002; Rigor and Wallace, 2004; Rothrock and Zhang, 2005, Lindsay and Zang, 2005).

Our first example is for 1990. The low extent of the Arctic sea ice cover for September of this year was mostly due to negative anomalies in the Chukchi, East Siberian and Laptev seas. The onset of the anomaly was observed in spring. As outlined by Serreze et al. (1995b), over the period May-June, positive temperature anomalies were found over all of the Arctic Ocean, but were largest along the Siberian coast. The sea level pressure field for May was characterized by a mean low of 996 hPa centered at about 80° N, 100° E (Figure 7.11). This corresponds to a local anomaly of about 20 hPa. This was a very significant atmospheric event by any Arctic standards and was associated with strong southerly winds along the Siberian coast. The location of the mean low is furthermore consistent with the temperature anomalies (strong positive anomalies were found east of the low) and the observation of early coastal melt. The strong winds also explain the development during this month of areas of open water and low concentration ice in the Laptev and East Siberian Seas. This resulted in an early reduction of the surface albedo, further enhancing the melt process through the ice-albedo feedback effect (see Chapter 5). Increased cloud cover in June may have also played a role (the radiative cloud forcing is positive in this month). While the ice anomaly was already large by July, it grew rapidly during August. This resulted from development of a strong anticyclone centered north of Alaska, associated with strong easterly winds in the Chukchi and East Siberian seas. As the ice motion tends to be to the right of the geostrophic wind, the result was a poleward motion of ice away from the coast.

We next examine the 1998 anomaly, which is detailed by Maslanik et al. (1999). It is sometimes referred to as the "SHEBA anomaly" as it occurred during the SHEBA field year. The ice loss for this year was most pronounced in the Beaufort and Chukchi seas. Open water formed earlier in this region than in prior years, and September ice extent was 25% less than the previous minimum over the period 1953-97. The times series of ice extent in the Beaufort Sea at the end of September clearly shows 1998 standing out as an extreme (Figure 7.12). Ice extent was quite low in the preceding summer of 1997. Over the period of record examined by Maslanik et al. (1999), 1997 and 1998 are the only two successive years when summer ice extent in the Beaufort Sea was at least one standard deviation below the mean.

Rogers (1978) diagnosed variability in ice conditions in the Beaufort Sea from composite differences in summer-averaged sea level pressure fields between heavy and light ice years. Light ice years are associated with positive pressure differences

Figure 7.11 Mean sea level pressure and anomalies for May 1990 from NCEP/NCAR

Figure 7.12 Sea ice extent at the end of September for the Beaufort Sea region over the period 1953-98 (from Maslanik etal., 1999,byper-mission of AGU).

(from Serreze et al., 1995b, by permission of AGU).

Figure 7.12 Sea ice extent at the end of September for the Beaufort Sea region over the period 1953-98 (from Maslanik etal., 1999,byper-mission of AGU).

over the northern Canadian Arctic Archipelago and north-central Siberia, and negative differences over the East Siberian Sea. This strengthens the mean easterly and southerly winds over the Beaufort Sea and favors positive temperature anomalies, promoting melt and the transport of ice to the west and north. The sea level pressure and temperature anomalies for the summer of 1997 exemplify the pattern outlined by Rogers (1978) for light ice years. The situation in summer 1998 was somewhat different. Stronger regional pressure gradients favored stronger ice motion to the west and north. There were also more positive temperature anomalies over the southern Beaufort Sea as compared to 1997 and other light ice years. In response to the general westward ice drift, ice concentration and extent were actually well above normal in the East Siberian Sea.

The region of open water that developed in the previous summer of 1997 was replaced the following autumn and winter by primarily thin, firstyear ice. Firstyear ice, being relatively thin, will tend to melt out preferentially the next summer. McPhee et al. (1998) predicted that because of this preconditioning, a negative ice anomaly would develop in summer 1998. While their prediction turned out to be correct, the atmospheric conditions favoring ice reductions were even better expressed in 1998 than

1997. The larger size of the 1998 anomaly hence appears consistent with the dual effects of preconditioning and summer atmospheric forcing. As pointed out by Maslanik et al. (1999), the extreme size of the 1998 anomaly also requires consideration of other aspects of preconditioning. The November 1997 through April 1998 circulation at sea level points to a decrease in northerly winds over the Canada Basin, reducing the transport of older, thicker ice into the Beaufort Sea that would normally be supplied by the Beaufort Gyre circulation. This allowed the large expanse of thin firstyear ice to remain in the Beaufort Sea, contributing to the large ice loss in the summer of 1998.

Finally, mention should be made of the 2002 ice anomaly. Based on passive microwave records through September 2002, this year had the lowest ice extent observed since 1979. It shares features common with those described for 1990 and

1998, specifically, a preconditioning mechanism, and subsequent "unusual" atmospheric forcing. In this case, a preconditioning mechanism appears to have been offshore winds along the Siberian coast from March through May, with strong positive temperature anomalies in the East Siberian Sea. This was manifested as an early development of coastal leads and polynyas. This was followed by a strongly cyclonic circulation regime in summer. A closed low in the mean sea level pressure field was observed over the central Arctic Ocean for each of the three summer months. Low pressure was also found in September centered over the Laptev Sea. Following from earlier discussion, this would promote ice divergence, consistent with the widespread development of anomalously low ice concentrations over the inner pack ice seen in the September field (Plate 5). Enhanced absorption of solar radiation in the dark open water areas would then foster rapid ice melt. A contributing factor was continued unusual warmth during the summer months. There was also an unusual loss of ice in the Greenland Sea. Serreze et al. (2003c) interpreted this as a response to the cyclonic atmospheric regime over the central Arctic Ocean that reduced the Fram Strait ice outflow. However, other work, reviewed in Chapter 11, indicates that the low ice extents in September of 2002, 2003 and (possibly) 2004 are partly a delayed response to winter atmospheric forcing as far back as the period 1989-95. This forcing, associated with the positive phase of the Arctic Oscillation, reduced the age and consequently the thickness of ice over much of the Arctic Ocean. This thin ice eventually drifted into the Beaufort Sea, setting the stage for extreme summer ice losses in 2002 and 2003.

0 0

Post a comment