The direct effects of atmospheric forcing are crucial to the dynamics of the south-eastern Bering Sea. Because the mean flow over the shelf tends to be sluggish, and any direct oceanographic connection between the shelf and the North Pacific is obstructed by the Alaska Peninsula, the linkages between the south-eastern Bering Sea and the climate system are largely mediated by the atmosphere. Given the basin-wide scale of the atmospheric anomalies that dominate the seasonal and longer-term variability, the conditions that occurred in the south-east Bering Sea in 1997-1998 are related to the state of the entire North Pacific climate system. Previous work on this climate system concentrated on the wintertime atmospheric forcing and oceanic response (Miller et al., 1994; Trenberth and Hurrel, 1994), but interest in spring and summer conditions is increasing (e.g. Overland et al., 2001). Here we take a somewhat broader and less detailed view and consider both the winter and summer conditions during 1997 and 1998, and compare them to their climatological norms.
The atmospheric forcing of the south-east Bering Sea during winter is substantial, and through the long-lasting effects of sea ice, its influence can persist through the summer. The atmospheric circulation over the North Pacific and Bering Sea features large interannual fluctuations. It is important to recognize that this circulation is inherently variable on time scales longer than a few days, and much of this variability defies simple explanation. Nevertheless, frameworks do exist for accounting for some aspects of the variability. Most notable are the correlations between the North Pacific circulation and the El Niño-Southern Oscillation (ENSO) (Horel and Wallace, 1981) which varies on 2-7 year time scales, and the Pacific Decadal Oscillation (PDO) of North Pacific sea surface temperature (Mantua et al., 1997), which varies on decadal time scales. Figure 2 shows time series for ENSO (in terms of the NINO3 index) and the PDO, and their relationship to a more direct parameter for the Bering Sea, namely, the strength of the seasonal mean Aleutian Low. The strength (and position) of the Aleutian Low is important to the Bering Sea through its impact on the winds and surface heat fluxes, which in turn affect the formation and advection of sea ice. In general, deep, strong Aleutian Lows are associated with warmer-than-normal winters in the south-eastern Bering Sea, because the individual storm systems preferentially pump warm air poleward. On the other hand, weaker Aleutian Lows tend to be associated with an abundance of migrating anticyclones, which usually transport cold air equatorward. The effects of these transients can be modulated by the mean meridional wind anomalies that can also accompany changes in the position and strength of the Aleutian Low. Both the NINO3 and PDO indices underwent a marked change in the mid-late 1970s; this is the 'regime shift' noted by Trenberth (1990) and others. These time series (Fig. 2) reveal that the Aleutian Low is generally stronger when the NINO3 and especially the PDO indices are positive, and vice versa. The Aleutian Low during 1996, 1997 and 1998 was slightly stronger than normal, while during 1995 it was weaker (Fig. 2). This variability contributed to substantial differences in the timing and duration in the sea ice over the Bering Sea shelf, as will be shown in the following section.
Figure 2. Time series for ENSO (the NINO3 index), the PDO8 (after Mantua et al., 1997), and the Aleutian Low.
Mooring measurements have been collected on the shelf from 1995 to 1998 (see following sections) and can be used to examine the response of the south-east Bering Sea to a variety of wintertime conditions. For summertime, we have examined atmospheric forcing during 1997 and 1998 in the context of the historical record. This analysis complements the results of Overland et al. (2001), which focused on the anomalous heating during the summer of 1997 and its links to the 1997-1998 El Niño. As mentioned above, the bulk of previous research on the atmospheric circulation over the North Pacific has concentrated on the winter season. Its variability has been characterized in terms of preferred modes, such as the Pacific-North American pattern (PNA). The atmospheric variability is weaker in the summer, but it can also be characterized in terms of a small number of modes, which can differ from their winter counterparts. (For a general summary of these modes and their calculation, see Barnston and Livezey, 1987.)
The primary mode for the south-east Bering Sea from the spring to early summer is generally the North Pacific (NP) pattern. This mode consists of a dipole in tropospheric pressure anomalies between (roughly) northern Alaska and the west central Pacific along 40°N. Two other modes that flank the NP, the West Pacific (WP) and East Pacific (EP) patterns, are also prominent during the early and middle stages of the warm season, respectively. As shown by Overland et al. (2001), the NP and EP indices have been systematically positive and negative, respectively, since about 1990. Both modes have contributed to anomalously high pressures that have occurred over Alaska during the warm season for the last decade. During the spring and summer the WP was systematically negative (i.e. contributing towards higher pressure over the central Bering Sea) from the early 1980s to about 1990.
The large-scale pressure anomalies that comprise these modes are important to the south-east Bering Sea through their modulation of the processes that determine air-sea interaction. During the warm season, the two most important aspects of atmospheric forcing are the magnitude of solar radiation and wind speed at the sea surface. The incident solar radiation, together with mixed layer depth, controls the heating of the upper ocean during the summer; the wind speed regulates the depth of the mixed layer and the turbulent exchange at the base of the upper mixed layer, and hence the distribution of this heat in the water column.
Solar heating anomalies were strongly positive in 1997, especially from late May to mid-July (Overland et al., 2001), but were near zero in 1998 (not shown). These anomalies have been systematically positive since the early 1980s, and were systematically negative from the late 1950s to the early 1970s. Winds during the warm season, as indicated by observations at St. Paul Island in the Pribilof Islands (Fig. 3), were weak in 1997, with the exception of one prominent event in May. In 1998, winds were strong into June and after mid-August. By way of comparison, the winds in 1995 (1996) tended to be weaker (stronger) than typical. In general, the warm-season winds have tended to be weaker than their climatological norms since the early 1980s.
Figure 3. Times series of wind speed cubed (an indicator of the strength of wind mixing) measured at St Paul Island, 1995-1998. The bold line in each panel is the smoothed (3 day running average) daily mean value established from the 48 year long data set.
Return to previous section or go to the next section