Abstract
The El Niño-Southern Oscillation (ENSO) is known to have two major flavors, the traditional cold tongue type (CT) and the recently recognized warm pool type (WP). Both flavors of ENSO bring widespread climate impacts. This study focuses on the variability of WP ENSO: first, the variability of its position, and second, its dominance in overall ENSO activity. The study period begins in 1950 when robust data become available.A pattern-correlation based detection technique for WP ENSO is designed. Pre-existing detection methods assume WP ENSO to occur in a fixed location. In contrast, the newly designed detection method allows position variability and is hence more suitable for this study.
With the sea surface temperature anomaly (SSTA) defined as deviation from a long-term mean (from 1910 to 2015, or from 1979 to 2008), WP El Niño (La Niña) exhibits a clear long-term westward (eastward) shift. Slight interdecadal variability and prominent residual event-to-event variation are also found for both WP El Niño and La Niña. However, if the SSTA is defined as deviation from a low-frequency basic state (12 years), the low-frequency variability of WP ENSO vanishes and only the event-to-event variation remains. Both the long-term shift and interdecadal variability are mostly artifacts due to the definition of the SSTA. Only the event-to-event variation occurs regardless of how the SSTA is defined. A simple advection mechanism can explain this variability.
An atmospheric general circulation model is used to study the effects of the event-to-event position variability on the boreal winter climate impacts of WP ENSO. The position of WP El Niño is found to influence the orientation of an anomalous temperature dipole in North America and a temperature anomaly in southern Europe. The position of WP La Niña modulates the strength and extension of an anomalous temperature tripole in North America, and a temperature anomaly in continental Europe and North Africa. A preliminary analysis suggests that the position of WP ENSO is more influential than its magnitude in some parts of the world, e.g., North America and continental Europe.
The second part of the study focuses on ENSO activity across the equatorial Pacific. ENSO activity has shifted between the Eastern Pacific and Central Pacific on a decadal timescale since 1970. The cycle was less regular before. While the shift from the Eastern Pacific to Central Pacific happens rapidly, the shift from the Central Pacific to Eastern Pacific is separated by a period of low ENSO activity. When the Bjerknes feedback is strong (weak), ENSO activity is strongest in the Eastern Pacific (Central Pacific).
The Eastern Pacific to Central Pacific transition is hypothesized to be caused by a persistent Eastern Pacific downdraft after CT El Niño. Such a downdraft weakens the Bjerknes process for cold ENSO, leading to WP La Niña. The persistently cold ENSO leads to a Northern Pacific response that in turn enhances the equatorial Pacific trade wind. This persistently weakens the Bjerknes process in general, and a long period of Central Pacific activity follows. As the Bjerknes process is weakened, overall ENSO activity is damped significantly. ENSO activity weakens over time, and a period of low ENSO activity takes place. By chance, CT El Niño is triggered at the end of this period of low ENSO activity. Eastern Pacific ENSO activity becomes strong again and a new cycle begins.
An idealized ENSO model is designed and used to examine the responses of the cycle to global warming. The transition from the Eastern Pacific active period to the Central Pacific active period is less likely when the world is cold. Possibly because of this, the cycle appears to have been less regular in the 1950s. The transition has more readily taken place as the world has warmed, and the cycle has become more regular as a result.
| Date of Award | 25 Mar 2019 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Wen ZHOU (Supervisor) |
Cite this
- Standard