Ocean-atmosphere mechanisms involved in the equatorial heat buildup leading to El Niño events
Joan
1,2[*]
Ballester
,
Simona
1
Bordoni ,
Desislava
2
Petrova ,
Xavier
2,3
Rodó
1. California Institute of Technology (Caltech), Pasadena, California, United States
2. Institut Català de Ciències del Clima (IC3), Barcelona, Catalonia, Spain
3. Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Catalonia, Spain
[*] Contact: [email protected]
1. LA NIÑA, NEUTRAL AND EL NIÑO CONDITIONS
The equatorial Pacific is characterized by a sharp upper-ocean zonal salinity front near the warm pool edge around
160E-180 (Fig. 1b), which results from the convergence of zonally-advected low (high) salinity water masses from
the western (central) Pacific. The zonal contrast in upper-ocean salinity is explained by both the intense thermallydriven atmospheric convection and rainfall to the west and the strong trade winds and evaporation to the east.
The longitudinal migration of the warm pool explains the intimate link between ENSO variability and the zonal
thermohaline structure of the upper equatorial Pacific, with increased (decreased) zonal contrast of upper-ocean
temperature and salinity during La Niña (El Niño) events (Fig. 1a,c-e). These associations in turn determine the
zonal position and vertical tilt of local upper-ocean isopycnals, through simultaneous changes in both the warm
pool edge and the salinity front (green and grey straight lines in Fig. 1a-c).
2. SUBSURFACE HEAT BUILDUP FAVORED BY LA NIÑA CONDITIONS
On average, around two years before the peak
of El Niño, interannual La Niña-like conditions
are found in the upper equatorial ocean (cf. Fig.
2a-d,g-j with Fig. 1d), inducing anomalous
downward heat advection in the upper 150
meters just underneath the steeper-thannormal warm pool edge and salinity front.
When negative feedbacks start to weaken these
conditions (Schopf and Suarez 1988), the
coupled atmosphere and upper-ocean system
returns to its climatological state (Fig. 2e,f,k,l).
The accumulated subsurface warm waters are
allowed to travel eastwards through the
thermocline (Fig. 2m), characterizing a
transition period of increased heat content in
the equatorial Pacific (Jin 1997).
Some few months later, the Bjerknes feedback
(Bjerknes 1969) is finally activated when the
released heat reaches the surface layer in the
eastern equatorial Pacific, leading to the fast
growth of an El Niño event (Fig. 2n).
Figure 1. Interannual salinity (g/kg, shading) and temperature (°C, contours) values (a-c) and anomalies (d,e) for
La Niña years (a,d), all years (b) and El Niño years (c,e).
The thick contour depicts the 28°C isotherm in panels a-c, which encloses the warm pool, and the 0°C anomaly in
panels d,e. Solid (dashed) thin contours show isotherms greater (lower) than the thick contour. The slope of the
salinity front (warm pool edge) is depicted in green (grey) in panels a-c.
Figure 2. Interannual composite anomalies for El Niño events.
Salinity (g/kg, shading in a-f), temperature (°C, shading in g-n) and
ocean currents (m/s, arrows) for lags 32 (a,g), 28 (b,h), 24 (c,i), 20 (d,j),
16 (e,k), 12 (f,l), 08 (m), 04 (n) months before the major El Niño events.
Contours indicate significant salinity and temperature anomalies.
3. OSCILLATORY COMPONENT OF ENSO
4. IS ENSO CYCLIC OR IRREGULAR?
The transition between opposite phases of ENSO is connected through the strengthening or weakening of the heat
buildup mechanism in the western equatorial Pacific subsurface, where the memory of the system is stored during
an event (Fig. 3a,c), as well as through the eastern equatorial Pacific surface, where the memory is released
allowing the rapid growth of a new episode of opposite sign (Fig. 3b,d).
ENSO variability is characterized by an irregular (i.e. non-oscillatory) succession of El Niño, La Niña and neutral
years. The period in which the subsurface temperature anomaly propagates through the thermocline is especially
sensitive to external forcing, such as other variability modes or atmospheric noise. This is the phase of the
oscillation with minimum ocean-atmosphere coupling and ENSO signal, the so-called spring barrier in ENSO
predictability around lags -12 to -9 months, just before the shoaling of the subsurface temperature anomaly.
For example, the propagation of the heat
buildup through the thermocline was stopped
and reversed around winter 1974/75 due to a
sudden change in the tendency of the trade
winds in the central Pacific, aborting a
potential El Niño episode and leading to a
new La Niña event in 1975/76 (Fig. 4a,b).
Stronger El Niño episodes, activated by more
intense subsurface heat buildups, might be
less sensitive to external sources (e.g.
1957/58 event in Fig. 4c,d), explaining why
only prominent El Niño episodes are
predictable beyond the spring barrier at lead
times of up to two years (Chen et al. 2004).
EL NIÑO
LA NIÑA
LA NIÑA
EL NIÑO
LA NIÑA
Figure 4. Time-long interannual anomalies.
Temperature at the thermocline (°C, shading
in a,c), zonal wind stress tendency (Pa/month,
contours in a,c), sea surface temperature (°C,
shading in b,d) and zonal wind stress (Pa,
contours in b,d). Blue (purple) contours depict
positive (negative) anomalies.
Figure 3. A sketch illustrating the transition between opposite ENSO events (La Niña => El Niño => La Niña).
Bjerknes (1969) Mon Weather Rev 97, 163-172.
Chen et al. (2004) Nature 428, 733-736.
Jin (1997) J Atmos Sci 54, 811-829.
Schopf and Suarez (1988) J Atmos Sci 45, 549-566.