3.2.3. Small pelagic regimes and tropical-extratropical interdecadal variability

In previous sections (see 3.1.3 to 3.1.5), the connection between ENSO variability and global climate regimes was explored. The results suggested some relationship between both is detectable but not very close. Therefore, we consider that other influences, besides the tropical variability, are also important when looking for the regime signals, both climatic and among small pelagic populations.

However, extratropical systems are not as well studied as ENSO variations, thus information on their relative importance for determining global variability is scarce in most cases. One exception is the Aleutian Low Pressure System, a dominant meteorological feature in the winter and spring North Pacific atmosphere that has strong links to North Pacific oceanography (Namias 1969). Figure 21 shows the smoothed Aleutian Low normalized sea level pressure (AL), together with the smoothed SOI series. Viewing their interdecadal variability, it seems evident both systems show parallel trends for most of the time, the exceptions being two periods during the late 1920s and early 1970s. These periods of unrelated behavior are more evident when looking at the absolute differences between both indices, in the bottom panel of Figure 21.

Figure 21. Upper panel: Transformed (standardize/I 1-year moving average smoothed) series of SOI and AL indices. Lower panel: Absolute differences between transformed SOI and AL series.

As mentioned at the introductory section, the origin of decadal-scale variability has been attributed to both the tropical and the extratropical systems by different authors, and to date no definitive conclusions have been reached. The controversy can be summarized as follows: whereas the identification of decadal signals in ENSO indices is considered by some (Graham 1994, Jacobs et al. 1994, Zhang et al. 1996) as strong evidence supporting the idea of a tropical forcing, some modeling results suggest that realistic simulations of extratropical variability of the North Pacific can be generated even without any input from the tropical Pacific system (Latif and Barnet 1994, 1996; White and Cayan 1996). If tropical forcing is to be considered the main source of extratropical variability, we would expect a parallel behavior of the series shown in Figure 21; these parallel trends do seem to occur generally but not all the time. Therefore, we believe the SOI-AL relationship is significant because it suggests the tropical forcing of the extratropics is not a constant mechanism.

Figure 22 shows the series of absolute differences between the SOI-AL indices together with the transformed sunspot numbers series. Results strongly suggest an inverse relationship between the two series, with the observed negative correlation coefficient (-0.5) being statistically significant. Therefore, large (small) SOI-AL absolute differences seem to occur during periods of low (high) solar activity. Assuming the degree of coupling of the tropical and extratropical systems could be reflected by a gross index such as the SOI-AL series, our results would suggest that those periods of relatively high solar activity tend to promote this coupling and therefore result in small SOI-AL differences. Somewhat independent tropical-extratropical variability (and therefore large SOI-AL differences) would tend to occur during periods of moderate to low solar activity. However, the effects of solar variability on the climate system of the earth are still a matter of controversy.

Figure 22. Upper panel: Absolute SOI-AL differences (lighter line) and detrended 11 year moving average anomalies of number of sun spots (area; data from the Sunspot Index Data Center-Royal Observatory of Belgium). Lower panel: Regression line and correlation coefficient between transformed Sunspots signal (x-axis) and absolute SOI-AL differences (y-axis).

Figure 23 shows the SOI-AL series as compared to the RIS regimes. These latter are identified by regression lines showing the general trends during selected periods. Periods when the SOI-AL differences are large tend to be of sardine (anchovy) population growth (decline), while anchovy (sardine) growth (decline) periods tend to occur when the SOI-AL differences are small. From the previously suggested relationships, it would follow that sardine growing periods (upward RIS trends) would tend to occur during years of diminished solar activity, when the tropical and the extratropical systems tend to behave independently. Anchovy growing periods (downward RIS trends) would take place during periods of increased solar activity and coupled tropical-extratropical interdecadal variability.

Figure 23. Absolute SOI-AL differences (shaded area) and RIS (open circles) with trends evident by the partial linear regressions (fine lines).