Moisture budget

As part of their summary of high-latitude NAO impacts, Dickson etal.(2000) compared patterns of the vertically integrated moisture flux across 70° N for winters associated with NAO extremes. The data set spanned the years 1974-91. Their analysis (Figure 11.15) shows that the poleward meridional moisture flux for longitudes along the Nordic seas is much higher under the positive NAO phase. The positive NAO is also associated with a stronger equatorward moisture flux between about 70° and 140° W over Canada.

In turn, there are strong NAO precipitation signals in the Atlantic gateway to the Arctic. Dickson et al. (2000) compiled composite differences of Northern Hemisphere precipitation with respect to NAO extremes using a blended land/ocean data set for 1979-95. The largest precipitation increases (up to 150 mm per winter) during positive NAO conditions are found in the Norwegian-Greenland seas and Scandinavia. However, increases extend across Siberia to the Lena watershed. Thompson et al. (2000) used the same approach as in Figure 11.13 to examine the component of terrestrial precipitation trends, over the period 1968-96, linearly related to the AO (Figure 11.16). The most pronounced high-latitude trends, including increases over Eurasia extending to about 60° E, and over parts of central Eurasia and Alaska, are reasonably well captured by the AO. These signals support the idea that the observed aggregate increases in Siberian river discharge noted by Peterson et al. (2002) (Figure 11.7) relate in part to changes in the NAO/AO.

Rogers et al. (2001) performed a comprehensive assessment of relationships between the NAO and AO with P —ET as averaged over the Arctic basin (the region

35.00

35.00

Longitude

Figure 11.15 Vertically integrated meridional moisture flux crossing 70° N in winter expressed as a function of longitude for composites of winter (DJF) months representing high-index and low-index NAO extremes. The 1974-91 winter mean is also shown (from Dickson et al., 2000, by permission of AMS).

Longitude

Figure 11.15 Vertically integrated meridional moisture flux crossing 70° N in winter expressed as a function of longitude for composites of winter (DJF) months representing high-index and low-index NAO extremes. The 1974-91 winter mean is also shown (from Dickson et al., 2000, by permission of AMS).

north of 70° N). P—ET was calculated from the aerological method using NCEP/NCAR fields (Chapter 6). For winter, spring, summer and autumn, basin-averaged P—ET correlates with the NAO (AO) index at 0.49(0.56), 0.42(0.57), 0.29(0.56) and 0.49(0.36), respectively. The strongest relationship is for annual P—ET, which correlates with the NAO(AO) at 0.69(0.49). In accord with the analyses just described, the strongest P—ET signals associated with the NAO are found in the Atlantic sector and extending north into the Arctic Ocean during winter (see also Hurrell et al., 2003).

11.4.4 Sea ice

Several investigators have documented an upward trend in either ice volume or ice area fluxes through Fram Strait and into the North Atlantic from the late 1970s through much of the 1990s (e.g., Kwok and Rothrock, 1999). This has been related to the upward trend in the NAO over this period acting to strengthen the cross-strait SLP gradient (meaning a more northerly flow). Sea ice extent east of Greenland nevertheless tends to be below average under the positive NAO state (e.g., Deser et al., 2000; Dickson et al., 2000). As offered by Dickson et al. (2000), part of the explanation is that the positive NAO favors a stronger poleward flux of relatively warm ocean waters that inhibits ice formation. The relationship between the NAO index and the Fram Strait outflow is furthermore inconsistent. Even with a low NAO index, it is possible to achieve a strong pressure gradient over Fram Strait and a large ice outflow. This appears to

Figure 11.16 The 29-year (1968-96) linear trends (total % change) in winter precipitation and the component of the trends linearly congruent with the AO index. The precipitation database is for land areas only. Contour intervals are 10% (-10, 10, 20 . . .). Dark shading indicates increased precipitation of at least +10%, with light shading indicating reduced precipitation of at least 10%. The zero contour is omitted and negative contours are dashed (from Thompson et al., 2000, by permission of AMS).

have been the case during the period of the Great Salinity Anomaly (Chapter 7). The observed relationship between ice extent east of Greenland and the NAO assessed by Dickson et al. (2000) is summarized in Figure 11.17. The figure plots the median ice border at the end of April for the periods 1963-9 (low index) and 1989-95 (high index).

The downward trend in sea ice extent for the Arctic as a whole is actually dominated by late summer and early autumn reductions along the Siberian and Alaskan coasts. Recall from Section 11.4.1 that these reductions have been viewed in a general AO framework that involves "preconditioning". We review the issue in more detail with the aid of Figure 11.18, adapted from the study of Rigor et al. (2002). The top panel shows observed trends in summer sea ice concentration and winter sea ice motion based on the period 1979-98. The bottom panel shows the components of summer sea ice concentration and winter sea ice motion for each year regressed on the winter AO index. The observed trends in winter sea ice motion indicate a change to a more

Figure 11.17 The median ice border at the end of April for the periods 1963-9 and 1989-95, corresponding, respectively, to minimum and maximum phases of the NAO index. The more northerly position of the ice edge for the period 1989-95 corresponds to a reduction in ice extent of about 587 000 km2 from the period 1963-9 (from Dickson et al., 2000, by permission of AMS).

Figure 11.17 The median ice border at the end of April for the periods 1963-9 and 1989-95, corresponding, respectively, to minimum and maximum phases of the NAO index. The more northerly position of the ice edge for the period 1989-95 corresponds to a reduction in ice extent of about 587 000 km2 from the period 1963-9 (from Dickson et al., 2000, by permission of AMS).

cyclonic pattern. Care should be exercised in the interpretation - while the actual winter ice motion remained generally anticyclonic, the trend was toward more cyclonic conditions. The main point is that the patterns in the two panels are very similar - most of the change in summer sea ice concentration and winter ice motion over the study period relates to the trend in the winter AO.

The main argument is that changes in the wind field associated with the general upward tendency in the winter AO led to changes in the sea ice motion tending to both advect ice away from the Siberian and Alaskan coasts and (because of the cyclonic tendency) foster more ice divergence (or less convergence). This led to an anomalous coverage of thin ice. With thinner ice in spring, and higher spring SATs due to stronger heat fluxes from the ocean to the atmosphere, the stage was set for large summer ice losses. Part of the argument is that earlier melt onset was enhanced as thin, firstyear ice has relatively high salinity. It will hence melt at a lower temperature than thick, multiyear ice.

The Rigor etal. (2002) framework does not explain all aspects of the recent ice trend. Looking back at Figure 11.9, the winter NAO/AO has regressed back to a more neutral state in recent years, yet the sea ice cover continues to decline, as is well illustrated by the extreme September ice minima of 2002-4. Regarding the 2002 anomaly, Serreze et al. (2003c) highlight the role of the strongly cyclonic circulation of the atmosphere and sea ice during summer (see Chapter 7). The subsequent analysis of Rigor and Wallace (2004) gives a different perspective, which again implicates the AO. Using a simple sea ice model, they find that circulation patterns associated with high AO conditions, as observed from 1989 to 1995, decreased the areal extent of old thick ice from 80% of the Arctic to 30%. Most of the old ice exited through Fram Strait. The Arctic was then left with an anomalous coverage of younger and thinner ice. During

Figure 11.18 Large-scale trends in observed winter sea ice motion and summer sea ice concentration (a) and regressions on the prior winter AO index (b). Results are based on the period 1979-98. Contours are for every 5%. Dark contours indicate negative trends, while light contours indicate positive trends. The zero contour is eliminated (adapted from Rigor et al., 2002, by permission of AMS).

Scale 2 cm/s

Figure 11.18 Large-scale trends in observed winter sea ice motion and summer sea ice concentration (a) and regressions on the prior winter AO index (b). Results are based on the period 1979-98. Contours are for every 5%. Dark contours indicate negative trends, while light contours indicate positive trends. The zero contour is eliminated (adapted from Rigor et al., 2002, by permission of AMS).

the summers of 2002 and 2003, this younger, thinner ice circulated back into Alaskan coastal waters via the Beaufort Gyre circulation, where extensive ice melt occurred. The recent minima in ice extent are therefore viewed as a delayed response to the 1989-95 high AO index phase. The age of sea ice explains more than 50% of the modeled variance in summer sea ice extent. The extreme ice conditions in September 2004 may also manifest this lagged effect.

Understanding sea ice changes must also consider ocean circulation and vertical stability. As summarized by Dickson et al. (2000), results from a number of oceano-graphic cruises indicate that, in comparison with earlier climatologies, the Arctic Ocean in the 1990s was characterized by a more intense and widespread influence of Atlantic water (see Chapter 3 for a general discussion of Arctic oceanography). The Atlantic-derived sub-layer warmed 1-2 °C compared with Russian climatologies of the 1940s-1970s and the level of the subsurface temperature maximum extended upward (to about 200 m from some observations). The Atlantic water influence also spread west. Evidence also arose of a weakening of the cold halocline in the Eurasian Basin (1991-8), followed by partial recovery (Steele and Boyd, 1998; Boyd et al., 2002).

The temporary weakening of the cold halocline is largely attributed to eastward diversion of Russian river inflow in response to changes in the atmospheric circulation. The Atlantic layer changes are attributed to increases in the Atlantic inflow through Fram Strait and the Barents Sea region in the early 1990s, and some warming of this inflow. These processes are viewed as a response to altered surface winds associated with increasing dominance of the positive phase of the NAO/AO (Dickson et al., 2000). The modeling study of Maslowski et al. (2000) suggests that the Atlantic layer changes may have promoted a stronger upward heat flux in the Eurasian basin. This process, in conjunction with temporary weakening of the cold halocline (hence weakening of the vertical density stratification), may have contributed to sea ice losses. Maslowski et al. (2001) argue for an additional influence of a warmer inflow of Pacific waters through the Bering Strait. A full accounting of these different factors remains elusive.

As this book came to press, new studies were emerging using state-of-the art coupled ice-ocean models. Using a model driven by NCEP/NCAR winds and temperature, Rothrock and Zhang (2005) simulated sea ice thickness and volume changes over the period 1948-99. They argue that although wind forcing was dominant in promoting a rapid decline in thickness from the late 1980s through the mid 1990s, the sea ice response to rising temperatures is more steadily downward over the study period. Lindsay and Zhang (2005) make similar conclusions. Climate warming reduced equilibrium ice thickness. Changes in the atmospheric circulation also flushed some of this thicker ice out of the Arctic (cf. Rigor and Wallace, 2004). This led to more open water in summer, meaning greater absorption of solar radiation in the upper ocean. With more heat in the ocean, only thinner ice can grow in autumn and winter, which will then more easily melt the following summer.

1940 1960 1980 20QC

Year

Figure 11.19 Gradients of sea level in the Arctic Basin. Positive gradients indicate anticyclonic circulation, and negative gradients indicate cyclonic circulation. The solid line is the 5-year running mean of the sea level gradient (from Proshutinsky and Johnson, 1997, by permission of AGU).

1940 1960 1980 20QC

Year

Figure 11.19 Gradients of sea level in the Arctic Basin. Positive gradients indicate anticyclonic circulation, and negative gradients indicate cyclonic circulation. The solid line is the 5-year running mean of the sea level gradient (from Proshutinsky and Johnson, 1997, by permission of AGU).

0 0

Post a comment