|Empirical investigation on the relationship between climate and small pelagic global regimes and El Niño-southern oscillation (ENSO) (1997)|
|PREPARATION OF THIS DOCUMENT|
|1.1. Small pelagic regimes|
|1.2. Climate variability|
|2. DATA AND METHODS|
|3. RESULTS AND DISCUSSION|
|3.1. Global climate regimes|
|3.1.1. Identification of global climate regimes|
|3.1.2. Global climate regimes as reflected by wide-coverage air temperature series|
|3.1.3. Climate regimes and El Niño relative strength and frequency|
|3.1.4. Climate regimes as reflected by the SOI index|
|3.1.5. Decadal scale El Niño relative strength and frequency and the SOI|
|3.2. Small pelagic regimes|
|3.2.1. Definition of small pelagic regimes|
|3.2.2. Small pelagic regimes and global climate regimes|
|3.2.3. Small pelagic regimes and tropical-extratropical interdecadal variability|
|3.2.4. Small pelagic regimes and regional interdecadal climate variability|
Up to this point, we believe the analyses have yielded some preliminary but interesting results. While some findings do not seem to have a simple explanation at this point, others may allow us to draw some preliminary hypotheses of the possible relationships linking climate regimes, small pelagic regimes, and ENSO activity. Before attempting to do so, however, it would be useful to summarize the results and conclusions obtained so far:
· The GSAT seems to be a sensible index of the variability of the climate regimes; in other words, the regime signal seems evident when looking at this global climate index. Within this series, and for most of the present century, regimes are evident as sustained (cooling and warming) trends and (cold and warm) periods.
· Except for the most recent, this centurys climate regime trends and periods lasted for several decades (more than two and less than four). The 1870 to 1915 period seems to behave in a different way on the interdecadal time scale: trends and periods are shorter, and do not account for as much variability as the regimes of the present century.
· As reflected by the GSAT, global climate regimes grossly correspond to regimes of small pelagic abundance as previously described by Lluch-Belda et al. (1989). Thus, regime changes are to be considered as a global phenomena linking climate variability of distant areas, and having wide and similar effects on (at least some of) the biological components of oceanic systems.
· Despite being global phenomena, even a gross examination indicates that climate regimes are not homogeneous at different latitudes. As reflected by the GSAT series, regimes are strong and directly related to the air temperature over large areas at the northern and equatorial latitudes. There, climate regimes may account for as much as 80% of the regional variability on the interdecade time scale. However, at southern latitudes the regime signal becomes weak; and there are some indications of a weak and inverse relation between the GSAT and the southernmost areas.
· Although not a close one, there seems to be an inverse relationship between the frequency of El Niño events and the climate regimes. Warming (cooling) trends grossly correspond to periods of few (many) events. El Niño may be 2.5 times more frequent during cooling than during warming periods. Similarly, the so-called relative strength of the events (as stated by Quinn 1992) seems to be inversely related to the climate regimes: warming trends grossly correspond to periods of events weaker than those occurring during cooling periods.
· For the SOI index, climate regime shifts tend to occur when extreme SOI values are reached; highly positive (negative) values have been observed at the onset of warming (cooling) trends.
· Warming (cooling) trends of climate regimes may be characterized as periods of generally low (high) average values of the SOI.
· In time-scales longer than the interannual, extreme SOI values (El Niño and La Niña events) could mask the more common condition of the tropical Pacific as reflected by the positive-negative SOI values rate.
· At a decadal time-scale, positive (negative) SOI values tend to be more frequent than negative (positive) years in periods when ENSO events are more (less) frequent and intense.
· Despite its limitations, the RIS seems to be a good indicator of the small pelagic regimes; at least a better one than any individual catch record because this composite series tends to emphasize the common long-term variability while reducing the effects of the individual year-to-year variations.
· As reflected by the RIS, sardines were abundant worldwide from 1925 to 1950, peaking around the mid-1930s. Thereafter, anchovies dominated from the early 1950s up to the late 1970s, peaking around the late 1960s. The late 1970s and most of the 1980s were again years of sardine abundance, but recently (early 1990s) the sardine dominance has been declining. When considered globally, it seems evident that the alternation between sardine and anchovy stocks is basically an interdecadal signal.
· If the RIS-GSAT lagged-relationship holds true, recent trends of small pelagics suggest that both a period of worldwide anchovy abundance and a global cooling trend may be on their way, starting around the mid 1990s.
· For the relationship between the ENSO, as a tropical mechanism, and the Aleutian Low, as a component of the extratropical climate system, our results suggest that those periods of relatively high solar activity tend to promote tropical-extra tropical coupling. Somewhat independent tropical-extratropical variability tends to occur during periods of moderate to low solar activity.
· For the global regimes of small pelagics, sardine growing periods (upward RIS trends) seem 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.
· At the regional level, the SST of most systems where large populations of sardines and anchovies grow suggest that periods of (1) diminished solar activity, (2) large SOI-AL differences, and (3) sardine population growth tend to be warmer than those periods when the opposite conditions are observed. However, no such a globally coherent picture was observed for thermocline depth or pressure gradient.
Some of these observations and preliminary results can be included in an hypothetical mechanism linking ENSO and regimes, which is schematically presented in Figure 30. In this figure, we propose two basic modes or stages for the tropical Pacific ocean-atmosphere system: a weak circulation stage characterized by relatively low values of the SOI index, and a strong circulation stage characterized by relatively high values. Each of these modes would correspond to high and low levels of interdecadal solar activity; increased (diminished) activity would promote an intensification (relaxation) of the atmospheric circulation, probably because of stronger (weaker) pressure gradients. From the observed relationship between solar activity and the SOI-AL series, we suggest the coupling of the tropical-extratropical systems is probably forced by high solar activity and the resulting strong circulation mode, whereas weak circulation resulting from comparatively low solar input would promote a decoupling of the tropics and the extratropics.
In the weak circulation mode (upper panel), trade winds over the tropical Pacific would tend to be generally weak, in response to a comparatively small pressure difference between eastern and western areas. Thus, SOI index is to remain relatively low. If trade winds are weak, the sea level difference between western and eastern Tropical Pacific can not become too large.
Though trade winds tend to pile surface water towards the west, some of this water is redistributed by surface poleward currents (not shown) and the equatorial undercurrent systems (indicated as U). Thus, the sea level difference (P) caused by the weak trade winds may be partially overcome. This difference in sea level, we propose, may be regarded as the potential for an El Niño events, in the sense that it must be large for a strong El Niño to develop. The idea is that, at low values of P, any sudden shift of the Southerns Oscillation (SO) will probably not result in the onset of an El Niño event, or at most, the event will rank from weak to moderate.
The hypothesis contemplates another feature that may inhibit the onset of strong events during the weak circulation phase. Because the results show that during these periods the GSAT undergoes a warming trend, some heat transfer from the tropical ocean to the atmosphere may be involved. Some authors (e.g. Jones and Wigley 1990) mentioned that, regarding the GSAT, El Niño events are to be considered as cold events because a net heat transfer occurs from the atmosphere to the tropical ocean, thus cooling the air. La Nina episodes are considered by Jones and Wigley (1990) as leading to years of high GSAT, because these events imply a heat transfer in the opposite direction.
Therefore, during a weak circulation phase El Niño events are probably weak and relatively infrequent; and their cooling effect on the atmosphere may also be weak and infrequent. Let us suppose that in average conditions (i.e. when no tropical event occurs, neither Niño or Niña) there is a net heat transfer from the tropical ocean to the atmosphere. If so, the cooling of the air produced by only a few and weak El Niños may not overcome the air average heating. Thus, in the long run, the air may gain heat (DHa = +h) as it is released from the tropical ocean (DHw = -h), therefore explaining the upward trend of the GSAT. Moreover, this warming of the air may result in a weak ocean-atmosphere heat gradient (Ha » Hw), which in turn may reinforce the weak circulation phase. In other words, the proposed mechanism implies a positive feedback, not to be overcome unless a major shift of the SO forces the system to the alternative strong circulation phase.
The rationale for the other phase is similar to that previously stated. Now trade winds are generally strong, as indicated by the comparatively high values of the SOI index. More water is now piled into the western Pacific, and the influx is not overcome by the surface poleward currents and the equatorial undercurrent (U < W). This results in an increased slope of the sea level, thus in a larger potential (P) for El Niño events. Under these conditions, any sudden shift of the SO will probably result in the onset of an El Niño event, ranking from moderate to strong. Since more and stronger El Niños are supposed to occur during this phase, frequent and intense cooling of the atmosphere may result. This effect may totally overcome the normal direction of the heat flux (from the tropical ocean to the atmosphere) and result in a net heat transfer from the air (DHa = -h) to the tropical ocean (DHw = +h) and therefore in a downward trend of the GSAT. As the global air temperature drops, the ocean-atmosphere heat gradient increases (Ha < Hw); a positive feedback of the strong circulation phase. Again, this feedback is not overcome until a major shift of the SO forces the system towards the alternative weak circulation phase.
Since each phase is characterized by a sustained trend within the GSAT, the transition from one phase to the other implies a shift in the climate regime. As previously observed, shifts in the climate regime have occurred at extreme values of the SOI index. The onset of a warming trend follows extreme positive values, while strongly negative SOI have preceded the onset of a cooling trend (see Fig. 15). Extending the proposed scheme, we suggest that the changes in the tropical ocean-atmosphere heat transfer that may occur at both extremes of the SO may explain the change from one phase to the other.
According to this idea, the weak circulation feed back, reinforced by the small heat gradient which results from the tropical ocean to atmosphere net heat transfer, may end if a strong El Niño develops. Such an event would imply the heating of the water as well as the cooling of the air, therefore, the heat gradient would increase. When the gradient becomes large enough, strong circulation may result, and this in turn would tend to maintain the high gradient. A strong circulation phase may end if a strong La Nina event develops, since such an event would result in a weakening of the ocean atmosphere heat gradient caused by both the warming of the air and the cooling of the water. A significantly smaller gradient may then result in weak circulation, and therefore in a tendency towards keeping the gradient weak.
At this point, the connection of climate regimes and small pelagic regimes still remains to be explored. Some earlier results should be remembered. First, upward (downward) trends of the RIS, indicative of periods of sardine (anchovy) predominance, tend to occur when solar activity is at relatively low (high) levels. The possible effects of high and low levels of solar activity were already discussed in connection to global climate regimes and ENSO activity; therefore, the connection of these to small pelagic regimes seems straightforward. Sardine (anchovy) populations would tend to high (low) abundance during the weak circulation phase, while strong circulation conditions would favor the expansion (contraction) of anchovy stocks worldwide (exception made for the Benguela stocks, as explained earlier).
Small pelagic regimes and global climate regimes seem to be behave in parallel but with a lag of about a decade; small pelagic variations leading global climate signals. This suggests that the regime signals are first evident at the boundary systems where small pelagics develop, and that some time is required before the signal can be detected on a truly global scale. We propose, that while variations of solar activity and resulting circulation phases may have short-term effects at the boundary systems, there would be some delay before the signal is detected within the GSAT because of the feedback mechanism linking the ENSO activity to the GSAT variations.
One consequence of the onset of a strong circulation phase would be El Niño events becoming stronger and more frequent as the slope of the sea level over the Tropical Pacific grows larger. More frequent and strong El Niño events would then result in a cooling of the atmosphere, assuming that, during an El Niño, DHa = -h as the heat is gained by the tropical ocean (DHw = +h). This in turn would result in a downward trend within the GSAT. Thus, a global climate cooling trend may not be evident until enough strong El Niño events develop and the tropical ocean heat gain is large enough. A similar argument could be developed for explaining the delay between upward trends of the RIS and warming periods within the GSAT.
This hypothesis, if correct, would be an outstanding result because it would imply that variations of small pelagic stocks are an early indicator of global decade-scale changes in climate. Although a great deal of research and understanding is required before many of the conclusions and results of this work can be considered definitive, it is our feeling that the recent changes of small pelagic regimes, the late signals of global temperature trends, and the very atypical ENSO variability observed during 1990 to 1995 are all related, and suggest the following years will have both anchovy dominance and a downward global temperature on the interdecade time scale.