This study attempts to understand contributions of ENSO and the boreal summer sea surface temperature anomaly (SSTA) in the East Indian Ocean (EIO) to the interannual variability of tropical cyclone (TC) frequency over the western North Pacific (WNP) and the involved physical mechanisms. The results show that both ENSO and EIO SSTA have a large control on the WNP TC genesis frequency, but their effects are significantly different. ENSO remarkably affects the east–west shift of the mean genesis location and accordingly contributes to the intense TC activity. The EIO SSTA affects the TC genesis in the entire genesis region over the WNP and largely determines the numbers of both the total and weak TCs. ENSO modulates the large-scale atmospheric circulation and barotropic energy conversion over the WNP, contributing to changes in both the TC genesis location and the frequency of intense TCs. The EIO SSTA significantly affects both the western Pacific summer monsoon and the equatorial Kelvin wave activity over the western Pacific, two major large-scale dynamical controls of TC genesis over the WNP. In general the warm (cold) EIO SSTA suppresses (promotes) the TC genesis over the WNP. Therefore, a better understanding of the combined contributions of ENSO and EIO SSTA could help improve the seasonal prediction of the WNP TC activity.
The interannual variability of tropical cyclone (TC) activity, including genesis location, frequency, track, intensity, lifetime, and landfall, over the western North Pacific (WNP) has received increasing attention in recent years (e.g., Chan 1985; Dong 1988; Wu and Lau 1992; Wu et al. 2004; Camargo and Sobel 2005; Chen et al. 2006; and many others). Although other factors such as the quasi-biennial oscillation (Chan 1995) and Madden–Julian oscillation (e.g., Sobel and Maloney 2000) may also affect the interannual variability of the WNP TC activity, most studies have mainly focused on the possible effect of the El Niño–Southern Oscillation (ENSO), a signal that dominates tropical interannual variability. A well-documented effect of ENSO is an eastward displacement of the mean location of TC genesis over the WNP accompanying the El Niño event (e.g., Chan 2000; Wang and Chan 2002; Chia and Ropelewski 2002). Accordingly, there is a tendency toward more intense and long-lived typhoons in El Niño years than in La Niña years (Chia and Ropelewski 2002; Camargo and Sobel 2005; Chen et al. 2006). However, the total annual TC genesis frequency over the WNP has no significant correlation with ENSO index (Ramage and Hori 1981; Lander 1994; Wang and Chan 2002; Chen et al. 2006). Frank and Young (2007) suggested that factors that control TC formation differ in important ways from those that ultimately determine storm intensity. Furthermore, the relationship between ENSO and WNP TCs is likely complicated by the difference in the timing between the mature phase of ENSO (winter) and the peak of the typhoon season (June–October). In other words, other factors may be dominant in determining the annual TC genesis totals. It is thus still a puzzle what controls the interannual variability of TC genesis frequency over the WNP.
In recent years, the Indian Ocean (IO) has been found to have a vital impact on climate variability over the WNP and East Asia in boreal summer (e.g., Wu et al. 1995; Yoo et al. 2006). More recently, a capacitor mechanism was elaborated on to explain how the El Niño affects the IO sea surface temperature (SST) and how the latter causes the atmospheric anomalies over the Indo-western Pacific and East Asia (Yang et al. 2007; Xie et al. 2009). By this mechanism, an El Niño charges (warms) the tropical IO as a battery charges a capacitor, and the persistent tropical IO warming exerts its climatic effect as a discharging capacitor after the El Niño itself has dissipated. Therefore, in this sense, the IO SST is not just a passive response to ENSO but an important agent modulating climate variability over the WNP and East Asia.
Xie et al. (2009) found that the persistent IO warm (cold) SST anomaly (SSTA) could excite a warm (cold) equatorial Kelvin wave to the east in the troposphere, significantly affecting the atmospheric circulation and resulting in atmospheric anomalies over the WNP and East Asia. Since the large-scale atmospheric setting is critical to TC genesis over the WNP, questions arise as to whether the IO SSTA could have any significant effect on the WNP TC activity and, if so, which physical mechanisms are involved.
In this study, we will focus on the effect of east IO (EIO) SSTA on the interannual variability of the WNP TC genesis frequency. To provide a more complete picture of such variability, we will also reexamine the relationship between ENSO and WNP TC activity since previous studies have shown that ENSO may exert a strong modulation to the activity of intense TCs over the WNP. We will show that both the EIO SSTA and ENSO contribute to the interannual variability of the WNP TC genesis frequency but in different ways. EIO SSTA can have a significant effect on the total TC genesis frequency, especially on the frequency of weak TCs, while ENSO determines largely the number of intense TCs. We will also discuss the dynamical processes behind the statistical relationships and provide insights into the involved physical mechanisms. Details of data and methodology used in this study are described in section 2. Statistical relationships of ENSO and EIO SSTAs with the WNP TC activity are presented in section 3. Effects of ENSO and EIO SSTAs on WNP TC genesis and possible mechanisms are discussed in sections 4 and 5, respectively. The main conclusions are presented in the last section.
2. Data and methodology
The best-track TC data (6-hourly position and intensity) for the period 1980–2007 were obtained from the Shanghai Typhoon Institute of China Meteorological Administration (CMA). To complement our analyses of the relationship between the TC activity and SSTAs, the best-track data of Joint Typhoon Warning Center (JTWC) and Regional Specialized Meteorological Center of Japan Meteorological Agency (JMA) were also used. As for TC intensity definition, different agencies use the maximum wind speed near the TC center in different averaging time intervals. The CMA best-track dataset records 2-min mean sustained wind speed, the JMA records 10-min mean sustained wind speed following a recommendation by the World Meteorological Organization (WMO), and the JTWC uses 1-min mean sustained wind speed. The 2-min and 10-min mean sustained winds are statistically around 87.1% (Knapp and Kruk 2010) and 88% (Simiu and Scanlon 1978) of the 1-min mean sustained wind, respectively. Camargo and Sobel (2005) and Chen et al. (2006) have addressed the close relationship between ENSO and TC intensity using JTWC and JMA datasets, respectively. Here, CMA intensity scale will be mainly considered. Following Wang and Chan (2002), to minimize subjectivity in the identification of weak systems, we only considered TCs that reached at least tropical storm (TS) intensity (with the maximum sustained wind speed Vmax ≥ 17 m s−1) in the data. Based on CMA scale, TC intensity was categorized into five groups: TS (17 m s−1 ≤ Vmax < 24.5 m s−1), severe tropical storms (STS; 24.5 m s−1 ≤ Vmax < 32.7 m s−1), typhoons (TY; 32.7 m s−1 ≤ Vmax < 41.5 m s−1), severe typhoons (STY; 41.5 m s−1 ≤ Vmax < 51.0 m s−1), and super typhoons (SuperTY; Vmax ≥ 51.0 m s−1). Furthermore, in some of the analyses, we also regrouped the five TC categories into weak TCs (TS, STS, and TY) and intense TCs (STY and SuperTY). Note that TCs formed over the South China Sea (SCS) and the WNP in the typhoon season (June–October) were considered together and simply called WNP TCs (TS-SuperTY) in this study. The typhoon season here was defined as the time period from June to October, rather than June to November as used in other studies (e.g., Wang and Chan 2002; Camargo and Sobel 2005). The reason is that the genesis frequency is generally low in November and the atmospheric circulation is close to that in winter, and thus the TC genesis mechanism in November may differ significantly from that during June–October during which TC are geneses largely controlled by the position and strength of the SCS and WNP summer monsoon trough.
The monthly Niño-3.4 index averaged over the region (5°S–5°N, 120°–170°W) was obtained from the Climate Prediction Center (CPC), and the EIO (10°S–22.5°N, 75°–100°E) SST index was extracted from the extended reconstructed SST (ERSST) analyses from the National Oceanic and Atmospheric Administration (NOAA; Smith and Reynolds 2003). Outgoing longwave radiation (OLR) data were also obtained from NOAA. Daily and monthly mean wind, temperature, relative humidity, and sea level pressure fields were from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalyses (Kalnay et al. 1996).
3. Statistical analysis
Table 1 shows the correlation matrix among the WNP TC genesis frequency during the typhoon season, ENSO, and EIO SSTA. It can be seen that the TC genesis frequencies from the three best-track datasets are highly correlated, indicating that they show consistent interannual variability in TC genesis frequency. In particular, the agreement between the JMA and CMA best-track datasets is more evident, with a correlation coefficient of 0.94. Consistent with previous studies (Ramage and Hori 1981; Lander 1994), the total number of TC genesis events over the WNP is not significantly correlated with the Niño-3.4 index no matter which dataset was used for TCs. It is interesting, however, that all three TC datasets show high negative correlation of the WNP TC genesis frequency with the EIO SSTA from spring to summer, well above 99% confidence level. Especially, for the TC data from CMA, the correlation coefficients reach −0.49, −0.65, and −0.69 for the previous winter, spring, and summer, respectively. In other words, the interannual variability of summer EIO SSTA can explain as much as 50% of the total variance of the variability in TC genesis frequency over the WNP. Since the TC datasets from three agencies are highly consistent with each other in the interannual variability, we will only use the CMA best-track data in our following analyses.
We extended the CMA best-track data to the period 1960–2007 and found that the high correlations remained between the WNP TC genesis frequency and the EIO SSTA (with correlation coefficients of −0.43, −0.56, and −0.62, respectively, for the EIO SSTA in the previous winter, spring, and summer). We also examined the correlations of the annual (the whole year) TC genesis frequency with the EIO SSTA during 1980–2007 and got the corresponding correlation coefficients of −0.53, −0.66 and −0.65, respectively, at the 99% confidence level. The consistent high correlation suggests that the EIO SSTA plays a vital role in controlling the interannual variability of TC genesis over the WNP.
High positive correlation between the ENSO index and the EIO SSTA index in the following season (Table 1) suggests that ENSO could exert its climatic effect on the WNP TC genesis frequency indirectly through the tropical IO by the capacitor mechanism proposed by Xie et al. (2009) though there is no direct significant correlation between them. Meanwhile, the low correlations of the Niño-3.4 index in summer with those in previous seasons and EIO SSTA in all seasons also imply that the EIO SSTA is likely to be independent of summer Niño-3.4 SSTA.
We calculated the correlation coefficients of WNP TC frequencies to the east–west of various longitudes with EIO SSTA and ENSO indices in the summer (Fig. 1). For example, the correlation at 140°E indicates that between the number of WNP TCs to the east–west of 140°E and the SSTA index (viz., either Niño-3.4 SSTA or EIO SSTA index). Although the EIO SSTA is highly correlated with the WNP TC frequency for the whole basin, the correlation obviously changes with longitude (Fig. 1b). The correlation between the EIO SSTA and the TC frequency to the east of 150°E is not statistically significant. This suggests that the effect of the EIO SSTA could be more important in the west WNP. It is interesting to note that the effect of ENSO is evident in some regions although the correlation between the total WNP TC genesis frequency and ENSO index is not statistically significant. In particular, the ENSO index has the maximum positive (negative) correlation with TC genesis frequency to the east (west) of 145°E, which confirms the previous results (Wang and Chan 2002). The reverse relationships of ENSO index with TC frequency in the west and east WNP greatly weaken its correlation with the total TC genesis frequency in the basin. Therefore, ENSO is indeed a factor contributing to the east–west shift of TC genesis location, but not the total basin-scale TC genesis frequency.
We also examined the correlations of ENSO and EIO SSTA with the WNP TC genesis frequencies in three categories: all TCs, weak TCs (including TS, STS, and TY), and intense TCs (including STY and SuperTY). As we can see from Table 2, ENSO tends to be correlated strongly with the intense TC genesis frequency, with a positive correlation coefficient of 0.75, but weakly with the weak TCs. In contrast, the correlation between EIO SSTA and weak TCs is high with a negative correlation coefficient of −0.78. The results in Table 2 thus strongly suggest that weak and intense TCs could respond differently to changes in external forcing as will be discussed in the next sections.
4. Effect of ENSO on WNP TC genesis frequency
According to the normalized time series of the summer ENSO index during 1980–2007 (not shown), we can define five El Niño years (1982, 1987, 1991, 1997, and 2002 with SSTA ≥ 0.8 standard deviation) and five La Niña years (1984, 1985, 1988, 1998, and 1999 with SSTA ≤ −0.8 standard deviation). Table 3 lists the numbers of all TCs and intense TCs over the WNP in each of the El Niño and La Niña years. For TCs including all categories, there is no obvious difference in the genesis frequency between El Niño and La Niña years, consistent with the statistical analysis discussed in section 3. The total number of WNP TCs in both El Niño years and La Niña years fluctuates around 20 except in 1984 and 1998, which are likely to be dominantly controlled by the EIO SSTA (see section 5). However, the number of intense TCs in El Niño years is much higher than that in La Niña years. The former is significantly above normal (10.2 vs 7), while the later is far below normal (4.2 vs 7). These differences are significant at the 99% confidence level based on the Student’s t test.
Figure 2 shows the formation frequencies and positions of intense and weak TCs during the typhoon season in El Niño years and La Niña years. Since the correlation coefficient between ENSO index and the WNP TC genesis frequency reaches the maximum around 145°E (Fig. 1) and shows different trends westward and eastward, we divided the WNP into two subregions: the west region west of 145°N and the east region east of 145°E (Fig. 2). A prominent feature in Fig. 2 is the east–west shift of the mean location of TC genesis in El Niño years and La Niña years. More TCs form in the east WNP in El Niño years than those in La Niña years. The chi-square test indicates that the differences in the east WNP and west WNP are significant at the 99% confidence level. Another distinct feature is the sharp contrast between the numbers of weak and intense TCs in El Niño years and La Niña years. The number of weak TCs in the east WNP in El Niño years is almost as much as that in La Niña years. Therefore, the difference in TC genesis totals is a result of much more intense TCs formed in the east WNP. The ratio of intense TCs formed in the east region in El Niño years to those in La Niña years is 3.6:1 (Fig. 2). The differences for intense TCs are significant at the 90% confidence level based on the chi-square test. On the contrary, fewer TCs in the west region in El Niño years are greatly attributed to the activity of weak TCs. The ratio of weak TC genesis frequency in the west WNP in El Niño years to those in La Niña years is 1:1.7. The effect of ENSO on the WNP TC activity found here is generally consistent with previous studies (Chia and Ropelewski 2002; Wang and Chan 2002; Camargo and Sobel 2005; Chen et al. 2006).
The effect of ENSO on the WNP TC genesis was explained in previous studies by the zonal and meridional shifts of the WNP summer monsoon trough and changes in the summer teleconnection wave train over the WNP associated with ENSO (Chen et al. 1998; Chia and Ropelewski 2002; Wang and Chan 2002; Chen et al. 2006). Wang and Chan (2002) suggested that large meridional shear associated with the equatorial westerly anomalies increases low-level cyclonic vorticity in the southeast quadrant of the WNP, which is favorable for TC genesis and development in the east WNP. Here we provide further insights in terms of barotropic eddy–mean flow interaction.
Following Seiki and Takayabu (2007), an environmental variable was decomposed into a basic state (with an overbar) defined as an 11-day running mean and an eddy or synoptic component (with a prime) defined as the residual. The eddy kinetic energy tendency equation linearized about the basic state can be written as
is the eddy kinetic energy (EKE), u and υ are the zonal and meridional winds, V is the three-dimensional wind vector, Vh is the horizontal wind vector, R is the gas constant for dry air, p is pressure, ω is the vertical p velocity, T is temperature, Φ is geopotential, and D is the EKE dissipation rate. The first term on the right-hand side in Eq. (1) represents the barotropic conversion (KmKe). The second and third terms represent the advection of EKE by the mean flow and by the eddy flow, respectively. The fourth term represents the conversion from eddy available potential energy. The fifth term corresponds to the divergence of the eddy geopotential flux or the work done by the pressure gradient force associated with the activity of eddies.
Although latent heat release and baroclinic processes clearly play important roles in TC intensification, previous observational studies have found that the barotropic energy conversion in the lower troposphere is an essential energy source for the developing tropical depressions (e.g., Shapiro 1978; Maloney and Hartmann 2001). We assume that the barotropic energy conversion from the circulation anomalies provides a significant extra energy source for eddies during the El Niño years and that these growing eddies contribute to the TC genesis and development when other environmental conditions are favorable. We thus focus on the contribution of barotropic energy conversion term (KmKe) to the EKE tendency in Eq. (1). The term KmKe can be written in the Cartesian coordinates as
where the four terms on the right-hand side are barotropic energy conversions due to the convergence and the meridional shear of zonal wind and the zonal shear and convergence of meridional wind. These terms are largely determined by both the structure of the disturbances and the large-scale flow pattern.
Figure 3 shows the differences in the composite 850-hPa environmental wind, EKE, barotropic energy conversion, and OLR during the typhoon season between the five El Niño and the five La Niña years listed in Table 3. The SSTA associated with El Niño induces convective anomalies over the central equatorial Pacific as inferred from the difference in OLR between the El Niño and La Niña years from Fig. 3b. Convective heating triggers strong equatorial westerly anomalies, enhancing the low-level cyclonic shear in the deep tropics off the equator in the central Pacific (Figs. 3a,d). The cyclonic shear in the lower troposphere in the TC genesis region between 130°E and 180° is also enhanced by the Rossby wave response to the convective heating in the central equatorial Pacific (Gill 1980). Strong cyclonic shear is the key to the eastward shift of the TC genesis region in the WNP in El Niño years (Wang and Chan 2002). Furthermore, in the large-scale zonal wind, there is a northwest–southeast-oriented area (5°–20°N, 125°E–180°) with anomalous zonal convergence (marginally significant at the 95% confidence level) overlapped by cyclonic shear anomalies (Figs. 3c,d), both of which are favorable for the genesis and intensification of TCs (Holland 1995). In contrast, an anomalous zonal divergence and an anticyclonic shear in the large-scale zonal wind appear over the Philippine Sea in the western WNP, suppressing convective activity and TC genesis in the region (Figs. 3c,d).
Consistent with the convergence and cyclonic shear in the large-scale zonal wind in the northwest–southeast-oriented band in Figs. 3c,d is a positive anomaly in EKE (Fig. 3e). This implies that the cyclonic shear and zonal convergence of the large-scale zonal wind anomalies associated with the El Niño events provide a more favorable environment for the activity of mesoscale and synoptic disturbances, including TCs, in the east WNP. Note that the negative anomalies in EKE in the SCS and the Philippine Sea are generally small compared with the positive anomalies in the northwest–southeast-oriented band in the east WNP. This indicates that the difference in EKE between El Niño and La Niña years is not simply a result of eastward shift of the convective center but could be contributed to stronger disturbances in El Niño years. In other words, the El Niño does contribute to the genesis of intense TCs in the east WNP, as we have already seen in Table 3.
From the difference in barotropic energy conversion (KmKe) at 850 hPa between the El Niño and La Niña years shown in Fig. 3f, we can see that the barotropic energy conversion could be a considerable energy source to the positive anomalies in EKE in El Niño years, as seen in Fig. 3e. Again, the anomalies in barotropic energy conversion are biased toward positive anomalies, further indicating that the large-scale environmental flow associated with El Niño transfers barotropic energy to synoptic disturbances and favors the genesis of intense TCs.
To further examine the contributions of different dynamical processes to the barotropic energy conversion, we calculated all four terms on the right-hand side in Eq. (3) and showed the results in Fig. 4. Changes in the KmKe are dominated by the −u′υ′∂u/∂y term, which is related to the meridional shear of the mean zonal flow (Fig. 4b). This suggests that changes in meridional shear of the large-scale zonal wind induced by ENSO play a dominant role in the interannual variability of the barotropic energy conversion from the basic flow to the EKE in the east WNP. The term −u′2∂u/∂x also contributes to the total barotropic energy conversion and the anomalies in EKE over the WNP TC genesis region (Fig. 4a). Similar to the energy conversion related to the meridional shear of the large-scale zonal flow, the term related to the zonal convergence of the large-scale zonal flow also shows the positive anomalies in the band with high positive EKE anomalies (shown in Fig. 3e) and is biased toward positive anomalies as well. Note that this term shows negative anomalies in the Philippine Sea, indicating a response to the anomalous divergence induced by the equatorial westerly anomalies associated with convection in the central Pacific in El Niño years. In contrast, the contributions from the other two terms related to the large-scale meridional flow are considerably smaller (Figs. 4c,d). The term related to the convergence of the large-scale meridional flow shows positive anomalies over the Philippine Sea (Fig. 4d) and seems to partially offset the negative contribution by the divergence of the large-scale zonal wind there (Fig. 4a). The results from the difference in barotropic energy conversion between the El Niño and La Niña years suggest that the zonal convergence of the large-scale zonal wind, which is also often termed the large-scale confluence zone (Holland 1995), could contribute to the eastward shift of the TC genesis location and the meridional cyclonic shear of the large-scale zonal wind may contribute to more intense TCs in El Niño years.
5. Effect of EIO SSTA on WNP TC genesis frequency
Figure 5 shows the normalized time series of the WNP TC genesis frequency in the typhoon season and the EIO SSTA in summer during 1980–2007. A prominent feature is a clear negative correlation between the WNP TC genesis frequency and the EIO summer SSTA with a correlation coefficient of −0.69, indicating that fewer (more) WNP TCs form in response to the positive (negative) EIO summer SSTA.
To better understand the effect of the EIO SSTA, we also defined five strong warm EIO years (1987, 1998, 2002, 2003, and 2005 with SSTA ≥ 1 standard deviation) and five strong cold EIO years (1982, 1984, 1985, 1989, and 1994 with SSTA ≤ −1 standard deviation) from Fig. 5. Table 4 shows the numbers of all TCs and intense TCs over the WNP in each of the warm and cold EIO years. The TC frequency including all categories in each of the cold years is significantly above normal, while in each of the warm years it is generally below normal. The total number of WNP TC genesis in the five cold years is 126 (25.2 yr−1) whereas in the five warm years it is 79 (15.8 yr−1); that is, the TC genesis frequency in the cold years is about 1.6 times that in the warm years. The differences of all TC frequencies and the weak TC frequencies between the cold and warm years are significant at the 99% confidence level based on the Student’s t test. In contrast, there is no obvious difference in the intense TC frequencies between the cold and warm years.
Figure 6 shows frequencies and positions of TC formation over the WNP in the five cold and five warm EIO years during the typhoon season and the differences between the warm and cold years in frequencies of TC genesis and occurrence in each 5° longitude by 5° latitude grid box. Different from the effect of ENSO, the EIO SSTA shows no significantly different effects on TC genesis frequency in the west and east WNP. The increase (decrease) in the WNP TC activity in the cold (warm) EIO years seems to occur in the whole basin. However, the effects of the EIO SSTA on weak and intense TCs are different. As shown in the top-right corners in Figs. 6a,b, the weak TC numbers in the west and east regions in the cold years show a uniform increase relative to those in the warm years. The number of weak TCs in the west (east) region in the cold years is almost 2 (3) times that in the warm years. However, there is no significant difference in the number of intense TCs over the two subregions between the cold and warm EIO years. It is thus suggested that the EIO SSTAs dominantly affect the basin-scale TC genesis frequency while its signal is not strong enough to affect the activity of intense TCs over the WNP. The latter is mainly controlled by the ENSO cycle as discussed in section 4. The frequencies of both TC genesis and occurrence are much lower in the warm years than in the cold years and show three local minima: one over the SCS, one over the area southeast of Taiwan, and one between 140° and 160°E (Figs. 6c,d). The basinwide decrease of both frequencies in the warm years, especially for the northwestward and westward paths as inferred from the frequency of occurrence (Fig. 6d), is consistent with the decrease in TC genesis frequency over the WNP as shown in Figs. 6a,b.
Chen et al. (2004) pointed out that the majority (about 70%) of tropical cyclogenesis in the WNP is linked to the SCS and WNP summer monsoon trough. We thus calculated the correlations between the EIO SSTAs in previous winter, spring, and summer, respectively, and the WNP summer monsoon index defined by Wang and Fan (1999). The high correlation coefficients of −0.42, −0.50, and −0.48, respectively, indicate that the persistent EIO cold (warm) SSTA would lead to strong (weak) summer monsoon over the SCS and the WNP. This can also be seen in the composite 850-hPa wind and OLR fields for the warm and cold years, respectively, during the typhoon season shown in Fig. 7. As we can see from the low OLR (<218 W m−2) and wind fields in Figs. 7a,b, the monsoon trough over the WNP in the EIO warm years is located over the SCS and east of the Philippines, whereas in the cold years the trough deepens and extends eastward and northward, a situation more favorable for TC genesis. This is because the cold EIO SSTA implies an increased land–sea thermal contrast, leading to stronger SCS and WNP summer monsoon, more favorable for the WNP TC genesis.
In addition to the effect on the WNP TC activity through its forced anomaly in the WNP summer monsoon trough, the summer EIO SSTA also affects the western Pacific tropical circulation through the equatorial Kelvin wave dynamics as proposed by Xie et al. (2009) and further enhances its effect on the WNP TC activity. Figure 8 shows the differences between the warm and cold EIO years in 850-hPa winds (Fig. 8a), deep layer mean tropospheric temperature (Fig. 8b), 850-hPa vorticity (Fig. 8c), 925-hPa divergence (Fig. 8d), sea level pressure (Fig. 8e), and 500-hPa vertical velocity (Fig. 8f). The positive summer SSTA in the warm EIO years enhances convective activity over the EIO. The anomalous convective heating over the EIO drives anomalous equatorial easterlies between 15°S and 15°N over the SCS and the WNP (Fig. 8a). This largely reduces the cyclonic shear of zonal winds over the WNP TC genesis region. As a result, the lower troposphere over the main WNP TC genesis region is dominated by an anomalous anticyclonic circulation (Fig. 8a) with low-level divergence and anticyclonic vorticity anomalies (Figs. 8c,d), positive surface pressure (Fig. 8e), anomalous subsidence (Fig. 8f), and reduced midtropospheric relative humidity (not shown). These anomalies can be easily understood by the wave dynamics, as elaborated by Xie et al. (2009). Anomalous convective heating associated with the warm SSTA in the EIO excites a warm equatorial Kevin wave in the troposphere to the east over the western Pacific, as seen from the warm temperature anomalies averaged in the deep troposphere along the equator (Fig. 8b). The warm temperature anomaly lowers the surface pressure (Fig. 8e) and produces anomalous low-level convergence toward the equatorial region (Fig. 8d). An immediate response to the above equatorial anomalies is the divergence and subsidence in the tropics off the equator, leading to an increase in surface pressure and an anomalous anticyclonic circulation in a latitude band roughly between 5° and 30°N over the WNP, namely in the main WNP TC genesis region. These dynamical conditions greatly inhibit TC genesis over the WNP in the warm EIO years. Therefore, the EIO SSTA highly modulates TC formation over the WNP through affecting both the equatorial atmospheric wave dynamics and the East Asian and WNP summer monsoon circulation. Since both processes coexist and interact with each other, it is hard to evaluate their relative importance. Nevertheless, we believe that the two processes work cooperatively and not exclusively to lead to the observed response of the TC genesis frequency over the WNP in typhoon season to the EIO SSTA in summer. As a result, the EIO SSTA largely determines the interannual variability of the WNP TC genesis frequency.
In this study, we have investigated contributions of ENSO and EIO SSTA in boreal summer to the interannual variability of TC genesis frequency over the WNP and the involved physical mechanisms. Both ENSO and EIO SSTA are found to play important roles in modulating the WNP TC genesis frequency, but their effects are significantly different. ENSO remarkably affects the east–west shift of the mean TC genesis location and accordingly contributes little to the basin-scale genesis frequency but predominantly controls the activity of intense TCs. In contrast, the EIO SSTA affects the TC genesis in the entire TC genesis region over the WNP. As a result, the numbers of the total WNP TCs and weak TCs are predominantly determined by the EIO SSTA.
The results show that the main TC genesis region shifted eastward in El Niño years compared to that in La Niña years and that more frequent intense TCs formed in El Niño years. This ENSO–TC relationship is dynamically explained by changes in the large-scale environmental flow and the associated barotropic eddy–mean flow interaction. Anomalous convective heating induced by El Niño over the central-eastern Pacific triggers anomalous large-scale equatorial westerlies and enhances the activity of equatorial Rossby waves in the tropical WNP. This large-scale atmospheric setting results in a reverse pattern of the anomalous zonal convergence and meridional shear of the large-scale zonal wind in the west and east WNP. Both contribute to the barotropic energy conversion from the large-scale flow to the synoptic disturbances. It is found that the eastward shift of the convergence (confluence) zone of the zonal wind is responsible for the eastward shift of the main TC genesis region while the cyclonic shear of the large-scale zonal wind contributes to the more intense TCs in the east WNP in El Niño years.
The interannual variability of the basin-scale TC genesis frequency over the WNP is largely determined by the SSTA over the EIO in summer, which can explain as much as 50% of the variance in the former. The WNP TC genesis frequency is considerably higher in the years when the EIO is cold than it is warm in summer. Two mechanisms by which the EIO SSTA affects the WNP TC genesis frequency have been highlighted. The first mechanism is related to the effect of the EIO summer SSTA on the East Asian and WNP summer monsoon. Warm (cold) EIO SSTA implies the reduced (enhanced) land–sea thermal contrast, leading to weaker (stronger) than normal East Asian and WNP summer monsoon and its associated monsoon trough, thus suppressing (promoting) TC genesis over the WNP. The second mechanism is associated with the equatorial Kelvin wave dynamics as identified by Xie et al. (2009). The EIO warm (cold) SSTA can excite a warm (cold) equatorial Kelvin wave to the east, lowering (increasing) the surface pressure in the equatorial region and leading to anomalous anticyclonic (cyclonic) vorticity and divergence (convergence) in the tropics off the equator in the WNP TC genesis region. These would result in anomalous descending (ascending) motion and dry (wet) midtroposphere, thus suppressing (promoting) convective activities and TC genesis over the WNP.
Although earlier studies revealed that ENSO plays a key role in triggering the IO SSTA, recent studies all suggest that the IO SSTA is not just a passive response to ENSO but an important agent of climate variability over the WNP and East Asia. In a recent study, Du et al. (2011) find that the vertical shear increases in the summer following strong El Niño events due to the development of a warm Kevin wave from the tropical IO, suppressing TC genesis over the WNP. In this study, we have extended recent studies and documented the significant contribution of the EIO SSTA to the interannual variability of TC genesis frequency over the WNP. Our results suggest that the EIO SSTA could be a good predictor for the WNP TC genesis frequency while the ENSO could be a good predictor for the frequency of intense TCs in the typhoon season. Thus, these results may help improve the statistical seasonal prediction scheme for the WNP TC activity since the EIO SSTA in spring also shows significant correlations with the TC activity in the typhoon season.
This study has been supported in part by NSF Grant ATM-0754029 awarded to the University of Hawaii and NSFC Grants 40805040 and GYHY200806009. Additional support has been provided by the Japan Agency for Marine-Earth Science and Technology (JAMSTEC), NASA, and NOAA through their sponsorship of the International Pacific Research Center at the University of Hawaii at Manoa.
Corresponding author address: Dr. Yuqing Wang, IPRC/SOEST, University of Hawaii at Manoa, Rm. Post 409G, 1680 East-West Rd., Honolulu, HI 96822. Email: firstname.lastname@example.org
* School of Ocean and Earth Science and Technology Publication Number 8021 and International Pacific Research Center Publication Number 721.