|
|
Further
discussion on the possible development of a La Ni
a in 2003
Issued on 16 June 2003
Further to our earlier
discussion on the possibility of a La Ni
a in 2003 (issued on 24 April
2003), it looks like that more and more signs are indicating that it is likely
to happen. The following summarizes
our analyses.
A.
Actual observations
1.
Variations in actual SSTA
Similar
to our earlier study (Chan and Xu 2000), the
variations of SSTA in the Nio3.4 area can be divided into two groups: a warm
event either maintaining itself to the second year or becoming neutral (the
non-transition cases, or E0 to E1), and a warm event becoming a cold event (the
transition cases, or E0 to L0).
It can be seen that for the E0 to E1 cases (Fig. 1a), the decrease in
SSTA is slower than that in 2002/03 except perhaps for the 1976/77 case.
Because of this, the average SSTA of all the previous cases is still
quite high by MAM, way above the value in 2002/03.
On the other hand, the SSTA evolution in 2002/03 is very similar to those
of the transition cases (Fig. 1b).
If we examine the time-longitude cross-section along the equatorial region, we
see that the evolution in 2002/03 (Fig. 2c) is similar to that of the summer
(SU)-type warm-to-cold transition (Fig. 2a) than to the case of no transition
(Fig. 2b). In
the case of no transition, the SSTA never becomes negative in the central
equatorial Pacific but in 2003, negative SSTA is quite obvious from April 2003
on.
Thus,
a simple “analog” forecast based on the results of Figs. 1 and 2 would
suggest the occurrence of a La Nia in the summer of 2003.
  |
|
|
Fig.
1. |
Ni |
  |
|
|
Fig.
2. |
Longitude-time section of SSTA over the equatorial area (5oS - 5oN) for the (a) summer (SU) type warm-to-cold transition composite, (b) non-transition composite, and (c) 2002/03. Red (blue) shaded areas indicate SSTA > 0.5oC (< -0.5oC). (a) and (b) are from Chan and Xu (2000). |
2. Low-level zonal wind anomalies
As pointed out in the April discussion, the easterly wind anomalies appear to be
propagating eastward in 2003 (Fig. 3c), which is very similar to that composite
for the SU-type composite (Fig. 3a).
By early summer, the easterlies would have reached near the dateline,
which is happening now in 2003.
For cases that did not have the warm-to-cold transition (Fig.
3b),
easterlies are confined in the western Pacific west of ~160oE.
This again suggests the possibility of a La Nia in 2003.
This conclusion can be further corroborated from the 925-hPa zonal wind
anomalies. For
a warm year that does not transition into a cold year, the zonal flow anomalies
in April and May in the central equatorial Pacific are likely to remain westerly
whereas the trades will be stronger for the transition cases, which is indeed
observed (Fig. 4).
For 2002/03, the zonal wind anomaly averaged between April and May is
negative (-1.1 m s-1), thus suggesting the likelihood of a cold
event.
  |
|
|
Fig. 3. |
As in Fig. 2 except for the 1000-hPa zonal wind anomalies. |
 |
|
|
Fig.
4. |
Scatterplot
of Ni |
3. Sea-level pressure anomalies
The sea-level pressure (SLP) anomalies for the SU-type warm-to-cold transition cases become positive over much of the Pacific by April (Fig. 5a). However, those for the non-transition cases remain negative throughout the year after the warm event (Fig. 5b). In 2002/03, while the SLP anomalies did remain negative during spring of 2003, they have mostly become positive after April except near the equatorial regions (Fig. 5c). However, according to Chan and Xu (2000), it is the subtropical high in both hemispheres that determine whether a transition would occur. That is, the SLP anomalies outside the equatorial region are more important. Therefore, with these anomalies becoming positive, especially in the Southern Hemisphere, it is likely that a cold event would occur in the summer of 2003.
  |
|
|
Fig.
5. |
Latitude-time section of SLP anomalies over the eastern Pacific (180o - 120oW) for the (a) summer (SU) type warm-to-cold transition composite, (b) non-transition composite, and (c) 2002/03. The shaded areas indicate SLPA < 0. Unit: 0.01 hPa. (a) and (b) are from Chan and Xu (2000). |
B. Model forecasts
4.
Nino plumes
As mentioned in our earlier discussion, the Nino plumes from the ECMWF model are showing a downward trend, more so for the more recent forecasts. Comparing the April and May plumes (Fig. 6), it is obvious that more ensemble members are predicting the trend towards a cold event.
  |
|
Fig. 6. |
Ni |
5.
Other dynamical and statistical forecasts
According to the IRI summary: http://iri.columbia.edu/climate/ENSO/currentinfo/SST_table.html,
the average dynamical model forecasts give a NINO3.4 anomaly of - 0.5oC
in JJA, dropping to - 0.6oC by OND.
The IRI itself gives a 55% (44%) probability of the occurrence of a cold
(neutral) event: http://iri.columbia.edu/climate/ENSO/currentinfo/QuickLook.html.
However, the statistical model forecasts vary quite significantly, with some
predicting the continuation of a warm event and some a severe cold event.
As a result, the average statistical forecast gives a neutral prediction.
C.
Summary
It
appears that based on the examination of actual observations, past history and
model forecasts, it is now quite likely that 2003 will be a La Nia
year, which could occur as early as June but most likely before August.
Chan, J. C. L. and J. J. Xu, 2000: Physical mechanisms responsible for the transition from a warm to a cold state of the El Nino Southern Oscillation. J. Climate, 13, 2056-2071. Abstract
