Modelyear

Figure 6. NINO3 SST anomalies calculated from (a) the Pacific Run and (b) the Indo-Pacific Run. These monthly values are low-pass filtered to remove variations shorter than 12 months. (From Yu et al., 2002.)

interannual variability in the Indian Ocean and its interactions with the ENSO cycle. The Indian Ocean SST variability resulting from these interactions, such as the basinwide warming/cooling and the Indian Ocean zonal/dipole mode, may feed back to affect the ENSO evolutions. In addition, since the interactive part of the tropical warm pool covers both the western Pacific Ocean and the eastern Indian Ocean in the Indo-Pacific Run, the eastern Indian Ocean part of the warm pool can respond interactively to Pacific ENSO events. This can amplify the overall feedbacks from the atmosphere during ENSO events. It was also suggested by Kug et al. (2006) that an interactive Indian Ocean can affect surface winds in the western Pacific/maritime continent, which can further affect the ENSO evolution.

Yu (2005) noticed that the phase-locking of ENSO to the annual cycle is enhanced and becomes more realistic in the Indo-Pacific Run compared to the Indian Ocean Run (not shown).

Figure 7. Lagged autocorrelation coefficients of NINO3 SST anomalies calculated from the (a) Pacific Run, (b) Indo-Pacific Run, and (c) GISST data. The curves are shifted to line up their one-month lag with the calendar month on the abscissa (From Yu, 2005.)

As a result of the phase-locking, the spring persistence barrier becomes more obvious in the Indo-Pacific than in the Pacific Run. The spring barrier is a well-known feature of the observed ENSO cycle. Lagged autocorrelation analyses with various ENSO indices, such as the NINO3 SST anomalies, Southern Oscillation pressure differences, and central Pacific rainfall anomalies, show sharp declines in the correlation coefficients in boreal spring (Troup, 1965; Wright, 1979; Webster and Yang, 1992; Torrence and Webster, 1998; Clarke and Gorder, 1999). Figure 7 shows the lagged correlation coefficients of the NINO3 calculated from the Pacific and Indo-Pacific runs and the observations. For the Pacific Run, Fig. 7(a) shows a more gradual decrease in the correlations and a weaker dependence of the decline on calendar months. This experiment produces a weaker spring persistence barrier than the observed [Fig. 7(c)]. Correlation coefficients in Fig. 7(c) are calculated using observed SST's from 1901 to 2000, based on the Global Sea-Ice and Sea Surface Temperature Data Set (GISST) (Rayner et al., 1996). In the Indo-Pacific Run [Fig. 7(b)], the spring barrier is stronger and more realistic, with a rapid decline in the correlations occurring in March-May for most of the 12 curves.

It is well recognized that the ENSO has a low-frequency (3-7 years) and a biennial

(~2 years) component (Rasmusson and Carpenter, 1982; Rasmusson et al., 1990; Barnett, 1991; Gu and Philander, 1995; Jiang et al., 1995; Wang and Wang, 1996). Yu (2005) found that the biennial ENSO component is very weak in the Pacific Run, but is significantly enhanced in the Indo-Pacific Run. This is clearly shown in the power spectra of NINO3 index of Fig. 8. By analyzing the persistence barrier in the decades of strong and weak biennial and low-frequency ENSO in the Indo-Pacific Run, Yu (2005) found that the overall amplitude of ENSO is not a primary factor in determining the strength of the persistence barrier. It is the amplitude of the biennial component of ENSO that affects the barrier the most. The persistence barrier is consistently strong (weak) when biennial ENSO variability is large (small). No such clear relationship is found between the strength of the barrier and the amplitude of the low-frequency ENSO component.

Results obtained from these two basin-coupling CGCM experiments (i.e. the Pacific Run and the Indo-Pacific Run) support the hypotheses that the spring persistence barrier is a result of the phase locking of ENSO (Torrence and Webster, 1998; Clarke and Gorder, 1999) and that the biennial ENSO component is crucial to the phase locking (Clarke and Gorder, 1999). Yu (2005) further suggests

Figure 8. Power spectra of NINO3 SST anomalies calculated from the (a) Pacific Run and (b) Indo-Pacific Run. Dashed lines indicate the 95% significance level. Thin-solid lines are red-noise spectra. (From Yu, 2005.)

(a) Pacific Run (b) Indo-Pacific Run

(a) Pacific Run (b) Indo-Pacific Run

Figure 8. Power spectra of NINO3 SST anomalies calculated from the (a) Pacific Run and (b) Indo-Pacific Run. Dashed lines indicate the 95% significance level. Thin-solid lines are red-noise spectra. (From Yu, 2005.)

that the Indian Ocean coupling plays a key role in producing the biennial component of ENSO. The mechanisms for this are not yet understood, but are being studied. It is believed that the TBO in the Indian and Australian monsoons may be involved in the enhancement of the biennial ESNO component.

6. ENSO's Interactions with the Tropospheric Biennial Oscillation (TBO)

The TBO is a major climate variation feature of the Indian-Australian monsoon system. Years with above-normal summer rainfall tend to be followed by ones with below-normal rainfall and vice versa. The dynamics behind this phenomenon has not yet been fully understood. In work that suggests that the TBO has its own dynamics, the interactions between the monsoon and the tropical Indian and/or Pacific Oceans are emphasized to play a central role in the TBO (e.g. Nicholls, 1978; Meehl, 1987, 1993; Clarke et al., 1998; Chang and Li, 2000; Webster et al., 2002; Yu et al., 2003; Li et al., 2006). However, different theories emphasized different parts of the Indo-Pacific Oceans for the importance. Webster et al. (2002) argued that the TBO is resulted from the monsoon-ocean interaction in the Indian Ocean. The wind-driven Ekman transport provides the needed phase reversal mechanism for the biennial oscillation. Meehl (1993) believed that the TBO involves the interactions between the monsoon and the Indian Ocean and both the western and the eastern Pacific Ocean. In contrast to this view, the TBO theory of Chang and Li (2000) assigned a passive role to the eastern Pacific Ocean. Instead, they emphasized monsoon-ocean interactions in the Indian and the western Pacific Ocean for the TBO.

The basin-coupling CGCM experiments are capable of isolating the monsoon-ocean interactions in the Pacific or the Indian Ocean and are, therefore, a useful tool for examining these TBO theories. Yu et al. (2003) contrasted the Indian monsoon variability produced in all three basin-coupling CGCM experiments (i.e. Pacific, Indo-Pacific, and Indian-Ocean runs) and noticed interesting differences among them. Figure 9 shows the power spectra of the Indian summer monsoon rainfall index (IMRI; rainfall averaged over an area between 10° N between 30°N, and between 65°E and 100°E) calculated from the experiments. The figure shows that there is virtually no biennial monsoon variation in the simulation including only the Pacific Ocean coupling (i.e. the Pacific Run).

Figure 9. Power spectra of the Indian summer monsoon rainfall index calculated from the (a) Pacific Run, (b) Indian-Ocean Run, and (c) Indo-Pacific Run. The 95% significance levels are indicated by the dashed curves. The shaded area is the period for the quasi-biennial oscillation.

With only the Indian Ocean coupling, the biennial peak is enhanced but is not strong enough to pass the statistical significance level (i.e. the Indian Ocean Run). A statistically significant biennial peak shows up only in the CGCM simulation that includes both the Pacific and Indian Ocean couplings (Indo-Pacific Run). These results suggest that the monsoon-ocean interaction in the Indian Ocean is able to produce weak biennial monsoon variability, and the biennial variability is further enhanced when the interactions between the Indian and the Pacific Ocean are included. The interactions and feedbacks between TBO and the biennial ENSO component are probably responsible for this enhancement.

An important aspect of the TBO is that the biennial tendency in the Indian summer monsoon is related to the biennial tendency in the Australian summer monsoon (Meehl, 1987, 1993). A strong (weak) Indian summer monsoon is often followed by a strong (weak) Australian summer monsoon. The anomalies then reverse sign as they return to the northern hemisphere and lead to a weak (strong) Indian monsoon during the northern summer of the following year. The in-phase transition from Indian summer monsoon to Australian summer monsoon and the out-of- phase transition from Australian summer monsoon to Indian summer monsoon of the next year are two key features of the TBO.

Yu et al. (2003) examined the role of the Indian and Pacific Oceans in these two monsoon transitions of the TBO. They noticed that the in-phase monsoon transition was produced more often in the CGCM experiments that included the Pacific Ocean coupling (the Pacific and Indo-Pacific Runs), while the out-of-phase transition was produced more often in the experiments that included the Indian Ocean coupling (the Indian Ocean and Indo-Pacific Runs). These results are demonstrated in Fig. 10, which displays the lagged correlation coefficients between the simulated monthly IMRI and Australian monsoon rainfall index (AMRI; rainfall averaged over an area between 100°E-150°E and 20S-5°N) anomalies. The figure shows that both the observations and the Indo-Pacific CGCM Run produce two large correlation coefficients: a positive coefficient with the IMRI leads the AMRI by about two seasons, and a negative coefficient with the AMRI leads the IMRI by about two seasons. The large positive correlation represents the in-phase transition from Indian to Australian summer monsoons.

(a) CMAP Observation

■ +0.4

i 1 i 1 i 1 i 1 i

i 1 i 1 i 1 i 1 i

§ +0.2

I IMRI lags AMRI

:

o

Was this article helpful?

0 0

Post a comment