Atmospheric Associations

Superposed on these SST patterns, and interacting with them, are the atmospheric counterparts (and consequences) of the CT and GR SST variations. Regression coefficients relating 500-mbar height anomalies (deviations of the heights of 500 mbar pressure surfaces in the atmosphere from their long-term seasonal averages) to the CT and GR series are shown in Figs. 4a and 4c. Note that regression coefficients are related to correlation coefficients by p = m500 / CTclim, where p is a regression coefficient, r is the corresponding correlation coefficient, ct500 is the local standard deviation of 500-mbar height anomalies, and CTclim is the standard deviation of the climate index. Hence, these maps can be interpreted as indicating the typical atmospheric anomaly that accompanies a modest excursion (1 ct) of the climate index. Negative coefficients relate the 500-mbar height anomalies and CT over vast


60'E 120'E


120'W 60'W

60°E 12CTE 180° 120'W 60'W

FIGURE 4 Regression coefficients relating (a) 500-mbar height anomalies to the Cool Tongue (CT) index, (b) 500-mbar storminess to the CT index, (c) 500-mbar height anomalies to the Global Residual (GR) index, and (d) 500-mbar storminess to the GR index. The contour intervals in (a) and (c) are 5 m per climate-index standard deviation, and in (b) and (d) they are 0.5 m per climate-index standard deviation. Storminess in panels (b) and (d) is measured by the monthly standard deviation of band-pass-filtered (3-10 days) 500-mbar height anomalies.

Siberian and North Pacific regions, whereas a northwest-southeast slanted dipole of positive and negative coefficients is indicated over the South Pacific (Fig. 4a). The negative coefficients correspond to lower than normal 500-mbar heights when CT is positive (El Niños). The broad region of negative coefficients over the North Pacific corresponds to an average 500-mbar height decline of as much as 20 m for a 1 ct rise in CT. Positive CTs (El Niños) are associated with deeper than normal Aleutian Lows, and the resulting steep north-south coefficient gradients near 35°N, 160°W in Fig. 4a correspond to anomalously strong 500-mbar height gradients, and anomalously strong westerly winds, when CT is positive. Thus, anomalies in atmospheric circulation patterns associated with positive CT values (El Niños) bring storms and precipitation to the southwestern and southeastern parts of North America. The circulation pattern associated with positive CTs also results in enhanced midlatitude blocking and the diversion of storms away from northwestern North America. The effects of these circulation changes on North American climate and hydrology will be discussed in Section 1.5.

The dipole of positive and negative coefficients over the southeast Pacific in Fig. 4b corresponds to diversions of low-pressure systems and attendant storms toward the subtropical belt of South America when CT is positive (El Niños), in a rough parallel to the changes in circulation in the Northern Hemisphere. Observations of 500-mbar heights in the far south are rare enough, however, that the coefficients may reflect dynamic consistency requirements of the reanalysis process (Kalnay et al., 1996) more than an observed feature of the atmosphere. Anomalously strong westerly winds are associated with positive CT along a belt extending from the Southern Ocean near the international date line to the southern Archipiélago de los Chonos of Chile. These patterns have been described as the Pacific South American pattern, a primary mode of Southern Hemisphere atmospheric variability (Mo and Ghil, 1987; Kidson, 1988; Mo and Higgins, 1998; Garreaud and Battisti, 1999).

Together, the low-pressure bands (of negative coefficients) associated with positive CT in the Northern and Southern Hemispheres reflect the extratropical extensions of the SOI seesaw of pressures. The Southern Oscillation pattern is not well represented at 500 mbars in the tropics (Fig. 4a), but along 30°N and 30°S it appears as the equatorward increase in coefficients (toward positive values in the tropics). In the tropical eastern Pacific, the increase in height at 500 mbars is mirrored as a decrease in pressure in the lower troposphere and a slackening of easterly trade winds during El Niños.

The relations between CT and midlevel winds and low-pressure systems are inferred from regression coefficients mapped in Fig. 4a. Corresponding relations between CT and storminess can be inferred from regression coefficients relating monthly standard deviations of high-frequency 500-mbar height anomalies to CT, as in Fig. 4b. Positive values in Fig. 4b correspond to regions in which 500-mbar heights are more variable, on 3- to 10-day timescales, than normal when CT is positive; regions with negative values are less variable and less stormy. Positive CTs (El Niños) are associated with less than normal storminess across Siberia, Alaska, northwestern Canada, and the northwestern United States. By this measure, the southwestern United States is more stormy than normal during El Niños, but the eastern Pacific storm track, near 35°N, 160°W, is clearly enhanced, is extended farther eastward over the southern United States, and is displaced farther south than normal. Storminess may also be enhanced during El Niños over northeastern Canada. A similar analysis (not shown here) of monthly standard deviations of high-frequency SLP variability in the Northern Hemisphere (where daily SLP records span the entire twentieth century) indicates that the associations of SLP-based storminess with CT closely parallel those in the northern half of Fig. 4b both before and after the 1950s (the period encompassing the bulk of available 500-mbar height observations).

In the Southern Hemisphere, positive CTs (El Niños) are associated with greater storminess over the central Pacific along the band from 30°-50°S. Between South America and the date line, storminess appears to be diverted northward from the midlatitudes (40°-60°S) toward about 30°S. Less storminess is indicated over southern Argentina.

The atmospheric expressions of GR are broadly similar to CT's influences, but—as with SSTs—with different spatial emphases. In Fig. 4c, notice that regression coefficients relating 500-mbar height anomalies to GR are dominated by negative coefficients over the North Pacific and a weaker north-south dipole of negative and positive coefficients over the South Pacific. The negative coefficients over the North Pacific are very similar to coefficients associated with CT. However, the less similar circulation anomalies over North America indicate that the circulations associated with CT are more zonal than those associated with GR. In the Southern Hemisphere, the dipole of coefficients is much less intense, implying relatively less expression of GR than CT in Southern Hemisphere 500-mbar heights and winds.

Storminess patterns associated with GR loosely parallel those associated with CT, but variations of GR (Fig. 4d) seem to yield storminess anomalies that are more focused over the North Pacific and the Pacific Northwest, with a much weaker expression over the southern United States than is associated with CT. In the mid-latitude Pacific of both hemispheres, the relationship between storminess anomalies and the average 500-mbar height anomalies during ENSO variations is similar to the relationship between storminess anomalies and the average 500-mbar heights during the decadal ENSO-like variability: a decrease (increase) in stormi-ness is found on the poleward (equatorward) flank of the negative midlatitude height anomaly in the central Pacific (e.g., on the flanks of the Aleutian Low) when the indices are positive. The Northern Hemisphere storminess patterns associated with GR (Fig. 4d) also are reflected in correlations of pre-1950s GR values with SLP-based storminess indices from that earlier period (not shown).

Taken together, the panels of Fig. 4 suggest qualitative similarities in the atmospheric expression of inter-annual ENSO variability (CT) and decadal ENSO-like variability (GR). These atmospheric similarities, and the SST similarities discussed previously, will be related to their similar hydroclimatic effects in the Americas in Section 1.5. Overall, the atmospheric expressions of both CT and GR (as well as tropical and North Pacific PC indices—not shown) are remarkably symmetric about the equator, especially over the Pacific Ocean: positive CT and positive GR variations are associated with equatorward diversions of westerlies, low-pressure systems, and storms from the midlatitude Pacific basin toward subtropical latitudes. These cross-equatorial symmetries may reflect tropical roots for both CT and GR (and for the tropical and North Pacific PCs). These strong interhemispheric symmetries of ENSO-like variations of Pacific climate are likely to produce strong interhemispheric climate variations along the entire length of the North and South American cordillera on both interannual and decadal timescales.

0 0

Post a comment