The Influence of Atmospheric Circulation on Temperature

It is not possible to investigate the reasons for recent air temperature variations without discussing atmospheric circulation changes. It is widely known that the importance of circulation in the formation of climate is much greater here than at lower latitudes (see Alekseev et al. 1991, their Table 1). Alekseev et al. (1991) also found that the advcction of warmth from lower latitudes by atmospheric and oceanic circulation provides more than half the energy annually available in the Arctic climate system. This makes such advection more important than solar irradiance flux. The share of advection is especially large in the cold season, when there is only a negligible inflow of solar irradiation. During the polar night it is equal to 100%. As mentioned in Chapter 1, atmospheric circulation provides as much as 95% of warmth advection to the Arctic, while occanic circulation provides only 5%. Vangengeim (1952, 1961) found that changes of synoptic processes in the Arctic arc about 1.5 times faster than in moderate latitudes. As a consequence, it is possible to conclude that the Arctic is significantly more sensitive and vulnerable to atmospheric circulation changes (such as those, for example, driven by the AO and the NAO phenomena) than any other area.

Przybylak (1996a, 2002a) determined the relations between atmospheric circulation and air temperature in the Arctic using daily data. He found that changes observed after 1975 in atmospheric circulation led to the Arctic cooling in the period 1976-1990 (as was mentioned earlier). For example, the intensification of zonal circulation in mid-latitudes, which is noted since 1975 (see Kozuchowski 1993; Jônsson and Barring 1994), accounts for Arctic cooling in all seasons except spring (Table 10.3). Cooling is significant mainly in the cold half-year, but only in autumn is it present in all climatic regions. The greatest influence of atmospheric changes represented by the zonal index on temperature is noted in the Baffin Bay and Canadian regions, where statistically significant negative correlations in winter and autumn were found. On the other hand, the observed intensification of zonal circulation leads to the Northern Hemisphere wanning in all seasons except summer. Statistically significant conelations were computed for winter and annual values (Table 10.3).

In the previous sub-chapters, some relations of the Arctic climate with the NAO were suggested. More recently a new tenn has been introduced into the literature: the Arctic Oscillation (AO) (Thompson and Wallace 1998). The NAO and the AO arc dominant patterns of atmospheric circulation variability over the North Atlantic and over the Northern Hemisphere poleward of 30°N, respectively (Himell 1995; Hurrell and van Loon 1997; Thompson and Wallace 1998, 2000; Houghton et al. 2001; Mysak 2001). The NAO/AO patterns can be obtained as the leading empirical orthogonal functions of the sea-level pressure (SLP) fields over the domains mentioned above. It is widely known that the largest north-south air mass exchanges associated with the AO occur over the Atlantic Ocean. That is why the NAO is often regarded as the regional representative of the AO (Delworth and Dikson 2000; Mysak 2001). Moreover, Deser (2000) found that the AO time series is nearly indistinguishable from the leading structure of variability in the Atlantic sector (the NAO). The correlation coefficient of monthly SLP anomalies during November-April 1947-97 is 0.95. According to Wang and Ikeda (2000), the main difference between the AO and the NAO is that the AO operates seasonwide and correlates to the surface air temperature, also seasonwide, while the NAO (index) correlates to surface air temperature anomalies in winter (strongest), spring, and autumn, but not in summer. The AO is strongly coupled with atmospheric fluctuations at the 50-hPa level on the intraseasonal, interannual, and interdecadal time scales and therefore, as Thompson and Wallace (1998) write, "can be interpreted as the surface signature of modulations in the strength of the polar vortex aloft".

Table 10.3, Correlation coefficients between areally averaged seasonal (DJF. MAM, JJA and SON) and annual air temperature for regions, the Arctic, the Northern Hemisphere and zonal index over the period 1951-1990 (after Przybylak 1996a)

Area

DJF

MAM

JJA

SON

ANNUAL

Atlantic reginn

0.11

0.13

-0.08

-0.01

0.10

Siberian reginn

-0.07

0.25

-0.32

-0.13

0.15

Pacific region

-0.07

0.18

-0.16

-0.06

0.21

Canadian region

-0.40*

-O.ll

-0.08

-0.40*

-0.28

Baffin Bay region

-0.48**

-0.18

-0.25

-0.55***

-0.46**

Arctic

-0.31

0.09

-0.14

-0.33*

-0.05

NH (land-t-ocean)

-0.43**

0.31

-0.06

-0.20

-0.39*

- Correlation coefficients statistically significant at the levels of 0.05; 0.01 and 0.001, respectively; Arctic - areally averaged temperature based on data from 33-35 Arclic stations; NH (land ^ocean) - areally averaged temperature for Northern Hemisphere (source: Jones 1994, updated)

- Correlation coefficients statistically significant at the levels of 0.05; 0.01 and 0.001, respectively; Arctic - areally averaged temperature based on data from 33-35 Arclic stations; NH (land ^ocean) - areally averaged temperature for Northern Hemisphere (source: Jones 1994, updated)

The above patterns of atmospheric circulation significantly influence the Arctic climate system. Researchers have investigated the relationships between the NAO and the AO indices, on the one hand, and, on the other hand, factors such as surface air temperatures (e.g. Hurrell 1995; Hurrell and van Loon 1997; Thompson and Wallace 1998, 2000; Przybylak 2000a; Rigor et al. 2000; Wang and Ikcda 2000; Broccoli et at. 2001; Slonosky and Yiou 2001), atmospheric precipitation (Hurrell and van Loon 1997; Przybylak 2002a, b), and sea-ice area (including its extent) and sea-ice motion (K.wok and Rothrock 1999; Yi et ctl. 1999; Dickson et al 2000; Hilmer and Jung 2000; Kwok 2000; Wang and Ikeda 2000; Vinje 2001). Here, the influence of these indices only on surface air temperature in the Arctic is briefly presented mainly using the results of investigations obtained by Przybylak (2000a) and Rigor et al. (2000).

Figure 10.22. Spatial distribution of the coefficients of correlation between mean annual air temperatures in the Arctic and the NAO (upper map) and NP (lower map) indices over the period 1951-1995 (after Przybylak 2000a). Statistically significant correlations are hatched. Other key as in Figure 10.16.

Figure 10.22. Spatial distribution of the coefficients of correlation between mean annual air temperatures in the Arctic and the NAO (upper map) and NP (lower map) indices over the period 1951-1995 (after Przybylak 2000a). Statistically significant correlations are hatched. Other key as in Figure 10.16.

Rigor el al. (2000) estimated the contribution of the AO to trends in winter (Dec-Feb) surface air temperature over the Arctic. They found that the AO explains more than 50% of the temperature trends in a large portion of the Arctic (60-70%) - see Figure 14d in Rigor et al. (2000). The relationship between these two elements is especially strong in the eastern part of the Arctic, where the AO accounts for as much as 74% of the wanning during winter. Thompson et al. (2000) have associated ca. 30% of the recent wintertime warming of the extratropical Northern Hemisphere with the multidecadal trend in the AO. Thus, this means that the influence of the AO is greater on the climate in the Arctic than in the moderate latitudes.

Figure 10.23. Spatial distribution of the coefficients of correlation between mean seasonal temperatures in the Arctic and the NAO index over the period 1951-1995 (after Przybylak 2000a). Statistically significant correlations are hatched. Other key as in Figure 10.16.

Przybylak (2000a) found that the relations between the NAO indices (after Hurrcll 1995 and after Jones et al. 1997) computed as the normalised SLP differences between series taken from Lisbon/Gibraltar (Iberian Peninsula) and Stykkisholmur/Akureyri (Iceland), respectively, and the Arctic air temperature (Figures 10.22-10.24) are roughly similar to those presented earlier for the zonal index. The strongest statistically significant relations exist with mean annual air temperatures in the Baffin Bay and the Canadian Arctic (negative correlations) and in the southern part of Atlantic regions (positive correlation) (Figure 10.22). Changes in the NAO index here explain about 10- 25% of the air temperature variance. As was mentioned above, the relationships between changes in atmospheric circulation in the North Atlantic and temperature, not only in the Arctic, are strongest in the winter months (Hurrell 1995, 1996; Jones et at. 1997). Correlation coefficients computed for particular seasons confirm this finding (Figure 10.23). In winter, as for annual values, negative correlations occur in the Baffin Bay and Canadian regions. However, they are significantly greater and, for example, in the central part of the Baffin Bay region, explain as much as 40-50% of winter air temperature variance. Statistically significant positive correlations are present mainly in the eastern parts of the Atlantic region and in the western part of the Siberian region (Figure 10.23).

Figure 10.24. Differences of air temperature {in °C> between the most extreme 7-year run of NAO+ winters (December February 1989 1995) and NAO- winters (December February 1963-1969) (after Przybylak 2000a). The NAO+ and NAO- winters were taken after Dickson et ai (1997). Negative differences are hatched. Other key as in Figure 10.16.

Changes in winter air temperatures between 7 years with positive modes and 7 years with negative modes of Ihe NAO index are in very high agreement with those presented above (compare Figure 10.24 with Figure 10.23). Strong cooling connected with the highest positive values of the NAO index occurred in the Baffin Bay region (4-7°C) and in the eastern part of the Canadian Arctic (1-5°C). On the other hand, warming was observed in the south-

Figure 10.24. Differences of air temperature {in °C> between the most extreme 7-year run of NAO+ winters (December February 1989 1995) and NAO- winters (December February 1963-1969) (after Przybylak 2000a). The NAO+ and NAO- winters were taken after Dickson et ai (1997). Negative differences are hatched. Other key as in Figure 10.16.

cm and eastern parts of the Atlantic region reaching l~4°C. These results are in agreement with those presented by Hurrell (1995, Figure 3) and Hurrcll (1996, Figure 3). In light of the results presented by Serreze et al. (1997, their Figure 6) concerning the positive minus negative NAO index difference field of cyclone events in the cold season, the causes of warming in the Barents and Kara seas regions are difficult to explain. Serreze et al. (1997) found a decrease in cyclone events. Recent results published by Dickson et al. (1997, 2000) may help to resolve this issue. Dickson et al. (1997) found that the "increasingly anomalous southerly airflow that accompanies such a change over Nordic seas is held responsible for a progressive warming in the two streams of Atlantic water that enter the Arctic Ocean across the Barents Sea shelf and along the Arctic Slope west of Spitsbergen". The temperatures of these two Atlantic-inflow streams were between 1°C and 2°C higher than normal in the late 1980s and early 1990s. Alekseev (1997) and Zhang et al. (1998) also presents similar results.

Figwv 10.25. Differences of mean winter (December-February) air temperature (in °C) between 10 years with the strongest El Niño phenomena and 10 years with the strongest La Niña phenomena (after Przybylak 2000a). Negative differences are hatched. Other key as in Figure 10.16.

Przybylak (2000a) has also used the North Pacific (NP) index to check for possible Pacific influences on Arctic air temperature. This index is calculated as the area-weighted mean SLP over the region 30°N to 65°N, 160°E to 140°W (Trenberth and Hurrell 1994). The NP index signal-strength dominates that of the NAO index only in some fragments of the Pacific region and in the south-western part of the Canadian region (Figure 10.22). Only here are the correlation coefficients statistically significant. A roughly similar situation also occurs in all seasons (Przybylak 2000a). Changes in atmospheric circulation in the North Pacific have the greatest influence on Arctic air temperature in winter and the lowest in summer (as with the NAO index).

The influence of ENSO (El Nino - Southern Oscillation) on Arctic air temperature is significantly lower than that associated with circulation changes in the North Atlantic (NAO index) - compare Figures 10.24 and 10.25. During the El Nino phenomena, decreases of winter temperature may be observed in the Kara Sea region (by about 2°C), in the Baffin Bay region, and in the eastern part of the Canadian Arctic (0.5°C l.5°C). On the other hand, significant warming is present only in Alaska. In other seasons, the general patterns of air temperature differences are similar to that for winter but these differences are significantly less. It is clear that the influence of ENSO on Arctic air temperature is indirect and occurs mainly through changes in atmospheric circulation in the North Pacific and North Atlantic. Hurrell (19%) found a statistically significant correlation (r = 0.51) between the NP index and the Southern Oscillation (SO) index (the index is calculated as the normalised SLP difference between series taken from Tahiti (French Polynesia) and Darwin (Australia)) but no correlation between the NAO and SO indices based on data for the period 1935-1994. The present author has repeated Hurreirs calculations and has found significantly lower correlations between the NP and SO indices (r = 0.19). Similar results (r = 0.23 and r = 0.11) have been obtained for other periods: 1899-1995 and 1951 -1995, respectively. In the 1899 1995 period, the correlation coefficient is statistically significant at the level of 0.05. However, after 1975 significant and more consistent changes for all these indices arc observed (see Figure 2 in Hurrell 1996). Computations of correlation coefficients between the analysed indices for the periods 1956 1975 and 1976-1995 confirm this conclusion. For example, correlation coefficients between SO and NAO indices for the winter (December March) are equal to r = 0.00 and r = -0.23, respectively. It follows that in the last two decades the ENSO could to a greater degree cause the changes in atmospheric circulation also noted in the North Atlantic.

Chapter 11

Was this article helpful?

0 0

Post a comment