Mid-latitude source of the ENSO-spread in SINTEX-F ensemble
· Takeshi Doi
· Yushi Morioka
· Swadhin Behera
Received: 4 October 2017 / Accepted: 27 May 2018
© Springer-Verlag GmbH Germany, part of Springer Nature 2018
The ensemble spread of seasonal prediction is investigated in this study to understand its role in the predictability of El
Niño/Southern Oscillation (ENSO) based on the results of SINTEX-F2, a coupled ocean–atmosphere general circulation
model. In the SINTEX-F2 seasonal prediction system, the ﬁrst ENSO precursor appears as a cyclonic wind anomaly over
the central north Paciﬁc in boreal winter (January). This is followed by warm SST, positive rainfall and cross-equatorial
southerly wind anomalies in the northern hemisphere during spring (particularly in April). These anomalies in April are
accompanied by westerly wind anomaly in the western equatorial Paciﬁc. Finally, El Niño-like conditions with warm SST
and positive rainfall anomalies become dominant in the ensemble standard deviation after boreal summer. The 500 hPa
geopotential height suggests that stochastic atmospheric variability excites El Niño-like spread through air-sea interaction.
The oceanic response in the form of upper heat content (in the top 150 m) appears to result from the equatorial wind forcing
during boreal spring and summer. These model results suggest that air-sea interaction related to the seasonal footprinting
mechanism (SFM) is important for ENSO spread and the “spring predictability barrier”. The dependence of ENSO spread
on the background ensemble-mean state is also investigated.
Keywords ENSO · Seasonal prediction · Ensemble spread
El Niño/Southern Oscillation (ENSO) is the most domi-
nant air-sea coupled mode in the tropics. Coupled atmos-
phere–ocean feedbacks (Bjerknes 1966) generate a unique
warm sea surface temperature (SST) pattern over the eastern
equatorial Paciﬁc (Rasmusson and Carpenter 1983) during
El Niño together with a Matsuno-Gill like atmospheric
response (Matsuno 1966; Gill 1980; Philander et al. 1984).
Once developed, ENSO aﬀects global climate through tel-
econnection patterns that are set up by Rossby wave trains
(Horel and Wallace 1981; Hoskins and Karoly 1981).
Therefore, the improvement of ENSO prediction has been
an important goal. Since the pioneering work of Cane et al.
(1986), ENSO modeling for prediction has evolved from
intermediate complexity models to state of the art general
circulation models (GCMs).
A longstanding unresolved issue in ENSO prediction
has been the drop of prediction skill during boreal spring.
This problem is known as the “spring predictability barrier”
(McPhaden 2003; Webster and Hoyos 2010; Duan and Wei
2013) and has been attributed to the weak air-sea coupling
associated with the warm SST over the eastern equatorial
Paciﬁc and double inter-tropical convergence zone (ITCZ) in
that season (Harrison and Vecchi 1999; Vecchi and Harrison
2003; Chang et al. 2007; Ham and Kug 2014). Stochastic
forcing has also been considered as a potential reason for
the predictability barrier. Such forcing includes the sea-
sonal footprinting mechanism (SFM), which originates in
the mid-latitudes (Vimont et al. 2001, 2003a, b; Anderson
et al. 2003, 2007; Alexander et al. 2010; Di Lorenzo 2015),
and westerly wind bursts (WWB), which are intrinsic to the
tropics (McPhaden 1999; Tziperman and Yu 2007; Seiki and
Takayabu 2007; Lopez and Kirtman 2013; Fedorov et al.
Previous studies have shown that a cyclonic atmospheric
circulation anomaly associated with SFM generates a posi-
tive heat ﬂux (downward ﬂux) anomaly in boreal winter (e.g.
Vimont et al. 2001; Alexander et al. 2010). This, in turn,
* Tomomichi Ogata
Japan Agency for Marine-Earth Science and Technology,