1. Introduction
As a huge energy reservoir, the western Pacific Ocean (0°–30°N, 100°–180°E) provides abundant moisture and heat for East Asian climate variability (Tao and Chen 1987; Wu 2002; Zhou and Yu 2005; He and Wu 2014). Changes in the western Pacific convective activities may lead to different regional climate over Japan, the Korean peninsula, eastern China, and even North America (Chang et al. 2000; Wu and Wang 2002; Chen et al. 2014; Tan et al. 2016). Understanding the western Pacific climate variability is of profound significance for regional climate prediction and disaster prevention.
This is the second part of a series, with the earlier related paper (He and Wu 2018) studying the recent climate shift over the western Pacific around the early 2000s. In that earlier paper, He and Wu (2018) detected a weakened coherence of interannual summer rainfall variations between the western Pacific and the South China Sea accompanied by an eastward shift of the large-scale rainfall anomaly pattern since the early 2000s. This is related to an adjustment in the simultaneous Indo-Pacific sea surface temperature (SST) influence. Before the early 2000s the combined effect of the northern Indian Ocean cooling and the tropical Pacific warming resulted in an elongated anomalous lower-level cyclonic circulation and a high coherence of interannual rainfall variations over the western Pacific, whereas after the early 2000s the Indian Ocean SST contribution was largely weakened, and the tropical Pacific SST anomaly led to an eastward retreated lower-level wind and rainfall anomalies. Here, the question still remains: Why were the Indo-Pacific SST anomalies altered significantly in the recent decade and what are the possible causes?
Numerous studies have demonstrated that El Niño–Southern Oscillation (ENSO) can exert pronounced impacts on the Indian Ocean (Wu et al. 2008; Du et al. 2009; Xie et al. 2009; He and Wu 2014). An El Niño event induces the Indian Ocean basinwide warming by modulating the atmospheric circulation and surface heat flux (Klein et al. 1999; Lau and Nath 2003; Xie et al. 2009). The western Indian Ocean warming is established in winter via downwelling Rossby wave propagation (Webster et al. 1999; Xie et al. 2002) and wind-induced surface latent heat flux (Wu et al. 2008), followed by a warming over the southern Indian Ocean in spring (Wu et al. 2008). The north–south SST gradient drives a south–north asymmetric mode, and the followed changes in winds and surface heat fluxes lead to enhanced rainfall over the northern Indian Ocean (Wu et al. 2008), and the northern Indian Ocean SST is subsequently cooled in summer (Du et al. 2009; He and Wu 2014). ENSO induces diverse atmospheric and oceanic responses in both the Indian and Pacific Oceans (Wang et al. 2000; Xie et al. 2002; Wu et al. 2008). Changes in the ENSO characteristics may induce different atmospheric teleconnections and associated changes in the tropical oceans.
Is the change in the Indo-Pacific SST anomalies in summer in the recent decade related to changes in ENSO properties? If so, how? The primary objective of this study is to clarify the physical mechanism for ENSO’s influence on the recent decadal modulation of tropical interannual variability since the early 2000s. The paper is organized as follows. Data and methods are introduced in section 2. Section 3 presents changes in the atmospheric and oceanic variability over the tropical Indo-Pacific region since the early 2000s. The role of ENSO in the shift of the Indo-Pacific summer climate variability is investigated in section 4. CGCM experiments are conducted to validate the proposed physical mechanism in section 5, followed by a summary and discussion in section 6.
2. Data and methods
The following datasets are used in this study: 1) the monthly-mean SST (1° × 1° grid) from the National Oceanic and Atmospheric Administration optimum interpolation, version 2, dataset (Reynolds et al. 2002; http://www.esrl.noaa.gov/psd/data/gridded/data.noaa.oisst.v2.html), 2) the monthly-mean precipitation (2.5° × 2.5° grid) from the Global Precipitation Climatology Project, version 2.2, dataset (Adler et al. 2003; http://www.esrl.noaa.gov/psd/data/gridded/data.gpcp.html), 3) the monthly-mean wind (2.5° × 2.5° grid) from the National Centers for Environmental Prediction (NCEP)–Department of Energy Reanalysis 2 (Kanamitsu et al. 2002; http://www.esrl.noaa.gov/psd/data/gridded/data.ncep.reanalysis2.html), 4) the monthly-mean net surface shortwave radiation and net surface latent heat flux (1° × 1° grid) from the Woods Hole Oceanographic Institution Objectively Analyzed Air–Sea Fluxes project (Zhang et al. 2004; Yu et al. 2008; http://oaflux.whoi.edu), and 5) the monthly-mean sea surface height (SSH) relative to geoid (⅓° latitude × 1.0° longitude with 40 vertical levels) from the NCEP Global Ocean Data Assimilation System (GODAS) reanalysis data (Ji et al. 1995; http://www.esrl.noaa.gov/psd/). Here, all of the variables cover the period from January 1982 to December 2016 except for the shortwave radiation data, which are available from January 1984 to December 2009.
The Community Earth System Model (CESM), version 1.0.4, is used in this study. CESM is a fully coupled climate model with five component models. The atmosphere component, the Community Atmosphere Model, version 4.0 (CAM4), is configured with 26 hybrid sigma levels on a 1.9° × 2.5° finite-volume grid. The ocean component—Parallel Ocean Program, version 2 (POP2), handles dipole and tripole grids for the horizontal grid and a fixed Eulerian grid for the vertical grid with depth as the vertical coordinate. A detailed introduction to CESM can be found in Smith et al. (2010), Vertenstein et al. (2011), and Eaton (2012). In section 5, CAM4 is coupled with POP2 for regional coupling experiments to validate ENSO’s role in altering the tropical Indo-Pacific variability.
The anomalies are obtained by using individual monthly means minus climatological monthly means. Empirical orthogonal function (EOF) analysis, correlation analysis, regression analysis, and composite analysis are adopted for observational statistical diagnosis. The Student’s t test is employed in estimating the significance of correlation, regression, and composite anomalies.
3. Shifted rainfall and SST anomaly pattern over the Indo-Pacific region
Figures 1a and 1b show the simultaneous correlations with respect to the first and second principal components of EOF of the western Pacific (5°–25°N, 105°–160°E) summer rainfall anomaly for the period 1982–2016. The circulation patterns are apparently different between these two modes. In Fig. 1a, the anomalous cyclonic circulation reaches as far west as the Indo-China Peninsula, and the rainfall anomaly pattern features a north–south tripole pattern over the western Pacific. In Fig. 1b, the anomalous cyclonic circulation retreats eastward away from the South China Sea, and the western Pacific rainfall anomaly changes to a west–east dipole pattern.
Simultaneous correlation of precipitation anomalies (mm day−1; shading) and 850-hPa wind anomalies (m s−1; vector) with respect to the (a) first and (b) second EOF principal component of the western Pacific (5°–25°N, 105°–160°E) summer rainfall anomaly for the period 1982–2016. (c) The time series of detrended JJA-mean precipitation anomalies over the WNP (black solid), the SCS (red dashed), and JK (green dashed). The boxes in (a) highlight the key regions used to define rainfall indices over the WNP (5°–20°N, 130°–155°E), the SCS (10°–25°N, 108°–118°E), and JK (25°–40°N, 130°–150°E).
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
Accordingly, rainfall responses are distinct over the western North Pacific (WNP; 5°–20°N, 130°–155°E), the South China Sea (SCS; 10°–25°N, 108°–118°E), and Japan and Korea (JK; 25°–40°N, 130°–150°E) between these two modes (Figs. 1a,b). Figure 1c presents the rainfall indices over the above regions. The WNP or SCS rainfall index is calculated by the rainfall anomaly averaged over the WNP or SCS, respectively, and the JK rainfall index is defined as the area-mean rainfall anomaly over JK multiplied by −1. According to Fig. 1c, the rainfall relationships among the three regions experienced a prominent change around 2003–04. Before 2003–04, the rainfall indices display high coherence over the WNP, the SCS, and JK (Fig. 1c) with a correlation coefficient between the WNP rainfall index with the SCS and JK rainfall indices to be 0.63 and 0.55, respectively. The rainfall relationships are consistent with the circulation anomaly pattern in the first EOF mode (Fig. 1a). In contrast, after 2003–04, such coherence in rainfall variations is largely weakened. The WNP rainfall index tends to show opposite sign to the SCS and JK rainfall indices, with a correlation coefficient decreasing to −0.47 and −0.41, respectively (Fig. 1c). This result matches with the large-scale rainfall and circulation anomaly pattern in the second EOF mode (Fig. 1b). The coherence among the rainfall indices signifies that the circulation and rainfall variability over the western Pacific seem to experience a decadal change around the early 2000s. For further examination, Table 1 lists the abnormal rainfall years by setting ±0.7 mm day−1 as the criterion for an abnormal rainfall. Here, we define a same-sign case as when all the WNP, SCS, and JK rainfall indices are abnormal and of the same sign and an opposite-sign case as when the WNP rainfall index is abnormal and has an opposite sign to that the SCS and JK rainfall indices. It is clear that there are five same-sign years (1983, 1984, 1988, 1998, and 2001) and only one opposite-sign year (1995) before 2003–04. In contrast, after 2003–04, the number of opposite-sign cases increases to four (2006, 2010, 2011, and 2013) and there are no significant same-sign cases. The result coincides with those of He and Wu (2018), showing a weakened covariability between the WNP and the SCS summer rainfall anomalies after 2003–04 although the abnormal years chosen in these two studies are not same because of the use of somewhat different criteria. Hence, in the following study, we separate two periods by 2003–04 and investigate the possible causes for the decadal change of summer rainfall variability over the western Pacific around the early 2000s.
Relationships of the western Pacific rainfall anomaly with the JJA NIO SST anomalies and the DJF–MAM EEP SST anomalies for the period 1982–2016 (the leading 19 or 20 is omitted). Plus and minus signs represent significant positive and negative rainfall (SST) anomalies, respectively, and zero represents insignificant anomalies as based on the criterion of ±0.7 mm day−1 for rainfall anomaly, ±0.5 K for EEP SST anomaly, and ±0.3 K for NIO SST anomaly. For example, “−0+” means the EEP SST anomalies develop from negative in DJF to normal in MAM and positive in JJA; “I√” denotes the years with an opposite-sign relationship of the JJA WNP rainfall with the NIO SST anomalies. Years with same-sign relationships of the JJA WNP rainfall with the SCS and JK rainfall anomalies are highlighted in boldface type, and years with opposite sign relationships of the JJA WNP rainfall with the SCS and JK rainfall anomalies are highlighted in italics.
Previous studies have demonstrated that the Indo-Pacific SST anomalies play a dominant role in the summer rainfall variability over the western Pacific. The northern Indian Ocean SST anomalies contribute to the western Pacific summer rainfall variability via modulating an east–west overturning circulation, and the tropical central Pacific SST anomalies affect local rainfall via a Rossby wave response and moisture convergence (Wu et al. 2014; He and Wu 2014, 2018). Different configurations of these SST anomalies may lead to different rainfall anomaly distributions over the western Pacific (Hu et al. 2014; Wu et al. 2014). To examine whether change in the rainfall anomaly pattern is related to change in the tropical SST anomalies or not, Fig. 2 presents simultaneous correlations with respect to the WNP summer rainfall index for the periods 1982–2003 and 2004–16, respectively. In the prior-to-2003 period, the north–south tripole rainfall pattern over the western Pacific (Fig. 2a) is highly correlated with SSTs in the northern Indian Ocean (NIO; 10°–20°N, 50°–100°E), the central North Pacific (7°–17.5°N, 150°E–175°W), and the central equatorial Pacific (CEP; 5°S–5°N, 160°E–130°W) (Fig. 2c). However, in the post-2004 period, the rainfall anomaly distribution turns to a zonal dipole pattern (Fig. 2b). In accordance, the SST signal over the Indian Ocean is largely weakened and the western Pacific summer rainfall anomaly is more correlated with the Pacific SST anomalies (Fig. 2d). These rainfall–SST relationships provide evidence that the weakened teleconnection with the Indian Ocean SST anomalies seems to be responsible for changes in circulation anomaly pattern over the western Pacific. This echoes the study in He and Wu (2018) that pointed out that an eastward-shifted large-scale circulation anomaly pattern over the western Pacific after the early 2000s is attributed to a weakened teleconnection with the Indian Ocean SST anomalies and an enhanced effect of the Pacific Ocean SST anomalies. More detailed evidence and numerical experiments are presented by He and Wu (2018).
Simultaneous correlations of the WNP summer rainfall anomalies with (top) large-scale precipitation and (bottom) SST anomalies for the periods (a),(c) 1982–2003 and (b),(d) 2004–16. Thick contours indicate that the correlations are statistically significant at the 90% confidence level.
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
This finding brings up another important question: What are the possible causes for the weakened Indian Ocean SST forcing of the western Pacific summer rainfall anomalies after the early 2000s? Section 4 tries to unravel the plausible mechanism by considering changes in ENSO’s properties and its teleconnection with the Indian Ocean variability.
4. ENSO influence
a. Changes in ENSO period and amplitude
Figure 3 shows the lead–lag correlations of large-scale SST anomalies with respect to the WNP summer rainfall anomaly for the period 1982–2003 and 2004–16, respectively. Signals in both the Indian and Pacific Ocean are apparently different between these two periods. In the prior-to 2003 period, the tropical Pacific features a La Niña pattern. The negative SST correlations reach up to 0.8 and last from the preceding autumn to spring (Figs. 3a–c). Accordingly, robust SST anomalies develop in the tropical Indian Ocean, evolving from a zonal asymmetric pattern in autumn to a meridional asymmetric pattern in spring (Figs. 3a–c). In contrast, in the post-2004 period, SST correlations are relatively weak in the equatorial central-eastern Pacific, with a phase transition from negative in the preceding autumn to positive in summer (Figs. 3d–f). In this case, the SST signal in the Indian Ocean is largely attenuated and becomes insignificant (Figs. 3d–f). Previous studies have pointed out that El Niño (La Niña) events can invigorate surface warming (cooling) in the NIO in the following summer through a series of air–sea interaction processes (Wu et al. 2008; Du et al. 2009; Xie et al. 2010; He and Wu 2014). Hence, the distinct evolution of the equatorial Pacific SST anomalies between the two periods implies that ENSO properties may experience changes in the recent decade and this may consequently lead to different responses in the tropical Indian Ocean.
Lead–lag correlations of the WNP summer rainfall anomalies with SST anomalies in (a),(d) JJA, (b),(e) the preceding MAM, and (c),(f) the preceding DJF for the periods (left) 1982–2003 and (right) 2004–16. Thick contours indicate that the SST correlations are statistically significant at the 90% confidence level.
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
For further verification, Fig. 4 presents an ENSO index obtained by averaging SST anomalies over the equatorial eastern Pacific (EEP; 5°S–5°N, 140°–90°W). Significant differences in the frequency and amplitude of EEP SST anomalies can be found before and after the early 2000s (Figs. 4a,c). According to the Morlet wavelet analysis, the dominant power is concentrated in the near-4-yr band from 1982 to 2000, yet there is an apparent shift to a period of 1.5–3 yr after 2004, and the wavelet power is significantly reduced in the post-2004 period (Fig. 4c). Thus, the EEP SST anomalies have shorter period with lower amplitude in the recent decade. Relevant changes in ENSO properties are also noticed by Hu et al. (2013, 2017). A shortened-period ENSO event features a fast developing phase and a rapid decaying phase, and this may likely affect seasonal phase locking of ENSO.
(a) Detrended 3-month-running-mean SST anomalies (K) averaged over the EEP (5°S–5°N, 140°–90°W). (b) Detrended JJA SST anomalies (K) averaged over the NIO (10°–20°N, 50°–100°E). The shading in (a) and (b) stands for ±1.0 standard deviation for the periods 1982–99 and 2000–16. (c) The normalized local wavelet power spectrum of the EEP SST time series using the Morlet wavelet function. The y axis is the Fourier period (yr), and the x axis is time (yr). The black contours enclose regions of greater than 90% confidence for a red-noise process. The black curve indicate the “cone of influence” where edge effects become important. More details about the Morlet wavelet analysis are in Torrence and Compo (1998).
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
The evolution of ENSO phase is examined in Table 1. Here, we define a lasting case as when the EEP SST anomalies are significant in winter and keep the same sign during spring, a transition case as when the EEP SST anomalies are significant in winter and turn to normal in spring and to opposite sign in summer, and a normal case as when the EEP SST anomalies are insignificant in more than two seasons. Before 2003–04, among all of the 11 abnormal years, there are 5 lasting cases (1983, 1984, 1985, 1996, and 1998), whereas after 2003–04, only 1 lasting case (2010) occurs, and the transition and normal cases increase up to 7 of 8 abnormal years. It is obvious that after the early 2000s, when ENSO period becomes shorter and ENSO amplitude becomes weaker, there are more ENSO events that feature a phase transition from winter to summer.
ENSO properties—including phase, intensity, and position—are crucial factors for determining remote responses in the Indian Ocean. A weak and rapidly decaying El Niño event probably cannot support the development of the second warming in the NIO in the following summer. It is found that the standard deviation of the June–August (JJA) NIO SST anomalies drops from 0.31 K for the period 1982–99 to 0.23 K for 2000–16 (Fig. 4b), following a decrease of the EEP SST variance from 1.11 to 0.81 K (Fig. 4a). We further compare the amplitude of NIO SST anomalies with the phase evolution of EEP SST anomalies in Table 1. It is apparent that the NIO SST anomalies show high coherence with the persistence of the EEP SST anomalies. Among all of the six years (1983, 1984, 1985, 1996, 1998, and 2010) in which the EEP SST anomalies last from the preceding winter to spring, there are five years in which the NIO SST anomalies are significant and affect the western Pacific summer rainfall anomaly. In contrast, the NIO SST anomalies are nearly insignificant when the EEP SST anomalies undergo a phase transition (except for 1988).
The above evidence shows that changes in the ENSO period and amplitude in the recent decade are accompanied by a rapid decay of the equatorial central-eastern Pacific SST anomalies from winter to summer. This is followed by a decrease in the NIO SST variance and a weakened NIO SST influence on the western Pacific climate variability.
b. Changes in ENSO teleconnection
To gain insight into how ENSO affects the way that the Indo-Pacific SST forcing works for the western Pacific summer climate variability, we perform composite analyses corresponding to the lasting ENSO cases (1983, 1984, 1985, 1996, 1998, and 2010) and the transition ENSO cases (1988, 1995, 2002, 2006, 2009, and 2012). The composite anomalies are calculated on the basis of the negative minus positive phase with respect to the December–February (DJF) EEP SST anomalies.
Figure 5 presents the composite anomalies for the lasting ENSO events from the preceding winter to summer. Corresponding to positive WNP summer rainfall anomalies, pronounced negative SST anomalies are detected in the EEP during the preceding winter with a mean magnitude larger than 1.2 K (Fig. 5a). These negative EEP SST anomalies can promote a basinwide cooling in the tropical Indian Ocean (Klein et al. 1999). The eastern Indian Ocean suffers enhanced rainfall and cloud (Fig. 5b), reducing the shortwave radiation reaching the ocean surface (Wu and Kirtman 2007; Wu and Yeh 2010), which subsequently lowers local SSTs and forms negative SST anomalies over the southeastern tropical Indian Ocean (Fig. 5c) (He et al. 2016). In the ensuing spring, the southern Indian Ocean cooling further drives an asymmetric mode of rainfall and lower-level wind anomalies in the tropical Indian Ocean (Figs. 5c,d), and this leads to a decrease of the NIO SST in the following summer through cloud-radiation and wind-evaporation effects (Wu et al. 2008; Wu and Yeh 2010; He et al. 2016) (Fig. 5e). The JJA NIO SST anomaly decreases up to −0.5 K and thereafter enhances the western Pacific precipitation by about 2 mm day−1 by inducing anomalous east–west vertical circulation (Fig. 5f; He and Wu 2014).
Negative minus positive composite anomalies for the lasting ENSO events during the period 1982–2016: the preceding (a) DJF, (c) MAM, and (e) MJJ SST (K); and the preceding (b) DJF, (d) MAM, and (f) MJJ precipitation (mm day−1; shading) and 850-hPa wind (m s−1; vectors). Thick contours and vectors indicate that the composite anomalies are statistically significant at the 90% confidence level. The scale for vectors is given at the bottom right of the figure.
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
In contrast, in the transition ENSO cases, the tropical Pacific SST anomaly distribution features an ENSO Modoki pattern with the largest SST anomalies in the central Pacific in winter (Fig. 6a). Compared with the lasting ENSO cases, ENSO Modoki has shorter period under a faster basinwide thermocline adjustment and zonal advection feedback (An and Wang 2000). Thus, the EEP SST anomalies decay faster from winter to spring with relatively weaker SST anomalies less than −1.0 K on average (Figs. 6a,c). In this case, ENSO is no longer strong enough to maintain its influence on the Indian Ocean, and SST responses in the tropical Indian Ocean are weak (Figs. 6a,c,e). Correspondingly, the relationship between the NIO SST and the WNP summer rainfall anomalies is insignificant in summer (Figs. 6e,f). Instead, following an early phase transition of ENSO, SST increases in the tropical Pacific by more than 1 K along with remarkable westerly anomalies near the equator (Figs. 6e,f). The Pacific warming simultaneously induces anomalous convection in the WNP via a Rossby-wave-type response (Gill 1980) and destabilization of the lower troposphere (He and Wu 2014) (Fig. 6f).
As in Fig. 5, but for the transition ENSO events.
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
The differences in coupled variations are further illustrated in Fig. 7. Following the persistent negative EEP SST anomalies with the maximal magnitude reaching 1.5 K, the upwelling Rossby wave propagation is pronounced across the southern Indian Ocean, with SSH declining by 20 mm from the east to the west since October (Fig. 7a). Accordingly, the tropical Indian Ocean experiences cooling lasting to winter. Soon afterward, a north–south asymmetric mode begins to form from January until May, with a north–south SST gradient reaching 0.2 K. This asymmetric mode prevents shortwave radiation and latent heat flux warming the NIO by −6 W m−2 from February to April. In turn, the asymmetric mode is reversed in summer, with strong SST cooling in the NIO owing to the negative contribution of heat flux. In this case, ENSO events mature in winter and decay during spring to summer, and the NIO serves as a medium for an indirect ENSO influence on the western Pacific summer rainfall variability. For the transition ENSO cases, accompanying the fast-decaying EEP SST anomalies, the Rossby wave response in the southern Indian Ocean is relatively weak, and SST cooling no longer forms in the tropical Indian Ocean. In spring, the meridional SST gradient is less than 0.1 K and the north–south asymmetric mode is not formed. Consequently, the negative NIO SST anomalies are insignificant in summer owing to the offsetting contribution of shortwave radiation and latent heat flux with a magnitude of ±0.3 W m−2, respectively. In this case, the equatorial Pacific SST anomalies feature a central Pacific–type La Niña or El Niño in winter, and it experiences a phase transition from winter to summer. In summer, the tropical Pacific warming or cooling, rather than the Indian Ocean forcing, contributes directly to the WNP summer rainfall anomalies (He and Wu 2018).
Evolution of composite anomalies for (a) lasting and (b) transition ENSO events during the period 1982–2016: the equatorial central Pacific (5°S–5°N, 180°E–120°W) SST (K; red columns), the northern (5°–20°N, 50°–90°E) minus southern (20°–5°S, 50°–90°E) Indian Ocean SST (K; red solid curve), the shortwave radiation over the NIO (W m−2, downward positive; green solid curve), the latent heat flux (W m−2, downward positive; green dashed curve), and SSH across 5°S (10−1 m; shading). Note that 1983, 1984, 2010, 2012, and 2013 are not included in composite of shortwave radiation anomaly because of a lack of data. The SST scale is at the left, and the heat flux scale is at the right.
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
The above evidence shows that ENSO events with a slow-decaying pace occur frequently before the early 2000s. The lasting EEP SST forcing allows robust SST anomalies to form in the NIO in summer, which thereby promotes the development of the zonally elongated lower-level cyclonic wind anomaly and the corresponding same-sign rainfall anomalies over the western Pacific. However, as the ENSO period becomes shorter with smaller amplitude after the early 2000s, the EEP SST anomalies decrease with a phase transition from winter to summer, and the influence of ENSO on the Indian Ocean SST anomalies is no longer as strong as that during the former period. Consequently, the NIO SST anomaly is weak in summer, and its contribution to the western Pacific summer rainfall variations becomes insignificant. Instead, the direct tropical Pacific SST forcing becomes relatively more important for the WNP summer climate variability in the recent period. This result partly echoes the recent studies that noticed that the central Pacific–type El Niño events occur more frequently than the conventional El Niño events (Yeh et al. 2009; Lee and McPhaden 2010; Yeh et al. 2011) and that the interannual variability in the tropical Pacific is significantly weaker around the late 1990s–early 2000s (Hu et al. 2013, 2017).
5. CGCM experiments
The above analysis underscores that a change in ENSO properties is responsible for the weakened Indian Ocean forcing of the western Pacific variability. In this section, CAM4 is coupled to POP2 to validate the proposed mechanism. The fully coupled model is run for 65 years, and the last 15 years are selected as a reference run to conduct experiments. Here, three integrations are performed: the control run, the long-period eastern Pacific forcing run (LEP run), and the short-period eastern Pacific forcing run (SEP run). In the control run, climatological mean SSTs (as obtained from the reference run) are kept fixed over the whole Pacific while elsewhere the atmospheric model is fully coupled to the oceanic model. In the LEP run, the atmosphere and ocean remain coupled except in the Pacific. In the Pacific, SSTs are relaxed to climatological means plus SST anomalies (Fig. 8a). The SST anomalies are imposed over 15°S–10°N, 170°E–70°W from September to August based on the composite anomalies in the lasting ENSO cases (Fig. 5). The SEP run is analogous to the LEP run except that SST anomalies are derived from the composite in the transition ENSO cases (Fig. 6). The three runs are integrated from October to the following August each year and the integrations are repeated for 15 years with different initial conditions to complete a 15-member ensemble. The initial condition each year is derived from that in the reference run. The difference between the LEP and SEP runs with the control run reveals different atmosphere and ocean responses to different types of ENSO. Figures 8 and 9 present the difference of the LEP run and the SEP run minus the control run, respectively.
Difference between the LEP run and the control run: (a) Evolution of equatorial Pacific SST anomalies (K). SST response (K) over the Indian Ocean in (b) December–January, (c) March–April, and (d) July–August. (e) Evolution of surface net heat flux (W m−2, downward positive; green columns) and SST (K; red solid curve) anomalies over the NIO (0°–15°N, 50°–90°E), along with the north (5°–20°N, 50°–90°E) minus south (20°–5°S, 50°–90°E) SST gradient (K; red dashed curve) over the Indian Ocean. (f) Precipitation (mm day−1) and 850-hPa wind anomalies (m s−1; vectors) in summer. The upper color bar is for SST and the lower one is for precipitation. In (e), the SST scale is at the left, and the heat flux scale is at the right. The scale for wind vectors is given at the bottom right of (f).
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
As in Fig. 8, but for the SEP run.
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
In the LEP run, the EEP SST anomalies last from the preceding autumn to spring with a magnitude larger than −1 K (Fig. 8a). Correspondingly, SST is cooled down by about −0.5 K in the eastern Indian Ocean during December–January (Fig. 8b), and a meridional asymmetric mode dominates the tropical Indian Ocean soon afterward in March–April (Fig. 8c). The NIO SST anomalies subsequently turn to be negative (Fig. 8c) under 6 W m−2 enhanced upward surface heat flux (Fig. 8e). In summer, the rainfall anomalies increase up to 2 mm day−1 and extend from the western Pacific to as far as the northeastern Indian Ocean when the NIO SST cooling coexists with the CEP SST warming (Fig. 8f). The LEP simulation confirms that a persistent EEP SST forcing favors for the development of a south–north asymmetric mode in the tropical Indian Ocean, and this allows the NIO SST anomalies to evolve from positive in March–May (MAM) to negative in JJA. In this case, the westward extended summer rainfall anomaly over the western Pacific is under the simultaneous influence of both the Indian and Pacific Ocean SST forcing.
In the SEP run, the EEP SST anomalies experience a faster decay with a relatively smaller amplitude of −1 K, and they turn to 1 K in the late spring (Fig. 9a). In accordance with this situation, SST responses in the Indian Ocean are weak (Figs. 9b–d). The north–south SST gradient in the tropical Indian Ocean is less than 0.2 K during MAM (Figs. 9c,e), and the NIO SST anomalies are insignificant in summer because of a reduced surface heat flux effect with a magnitude of less than −3 W m−2 (Figs. 9d,e). Relative to the LEP run (Fig. 8f), the western Pacific rainfall anomaly shifts more eastward and equatorward to around 120°E–140°W since the NIO SST anomalies decrease to less than −0.3 K and the positive CEP SST anomalies increase up to 1 K (Fig. 9f). The SEP simulation validates that ENSO with a fast decaying pace and lower variance is not strong enough to maintain its influence in the tropical Indian Ocean and this declines the SST signal in the NIO in the following summer. In this case, the summer rainfall anomaly over the western Pacific has a weakened teleconnection with the Indian Ocean SST anomalies, and subjects to an enhanced effect of the Pacific Ocean SST anomalies.
The numerical model results generally support the observation analysis although there is bias with too strong positive SST response over the southeastern Indian Ocean and the simulated atmospheric response cannot exactly match with the observations (Figs. 8 and 9). This may be attributed to the following reasons: 1) In the CGCM experiments, SST anomalies are only fixed in the tropical central-eastern Pacific. Lack of consideration of SST forcing and feedback processes in other regions (e.g., SSTs in the WNP and the “Maritime Continent”) may lead to bias in the model response. 2) In the decoupled regions, SSTs are relaxed to the reconstructed ones. The artificial change affects the ocean structure and wave propagation, and this may lead to a certain extent of errors in the decoupled boundary between the Pacific and the Indian Ocean. 3) Difference in the structure, amplitude, and position of the Indian Ocean SST anomalies between the simulation and the observation can affect model’s reproduction of the atmospheric changes.
The differences between the SEP and LEP runs include effects of both period and amplitude changes of ENSO. To get some hint of the relative contributions of period and amplitude changes, another regional decoupled experiment is conducted in which the SST anomalies imposed in the eastern Pacific are derived from those in the SEP run except that the amplitude of SST anomalies is enhanced to be as large as those in the LEP run. The differences in the response between the SEP and the extra run should be attributed to the effect of the difference in the amplitude of EEP forcing, and meanwhile the difference in the response between the LEP and the extra run should be attributed to the effect of the difference in the phase of EEP forcing. Analysis shows that the north–south SST gradient can be detected in the Indian Ocean in spring and the NIO SST is cooled down in summer in the extra run. The SST response in the Indian Ocean in the extra run is more significant than that in the SEP run, yet much less than that in the LEP run (figures not shown). The difference of the meridional Indian Ocean SST gradient between the extra and LEP runs appears a bit larger than that between the extra and SEP runs. Hence, both the shortening of the period and weakening of the amplitude of ENSO contribute to the weakening of SST signals in the Indian Ocean.
6. Summary and discussion
This study is the second part of a series of studies about the decadal modulation of summer climate variability over the western Pacific after the early 2000s. In the prior related paper, He and Wu (2018) argued that a reduced Indian Ocean forcing and a strengthened Pacific forcing is responsible for a weakened coherence of summer rainfall variations between the northwestern Pacific and the SCS since the early 2000s. Here we further investigate the potential mechanism for the shifted large-scale rainfall anomaly pattern over the western Pacific and the role of Indo-Pacific SST forcing by considering changes in ENSO properties.
Observational evidence suggests that ENSO has a shorter period and a weaker amplitude in the recent decade, and the eastern Pacific SST anomalies tend to undergo a phase transition from winter to summer. These changes in ENSO period and variance are responsible for the shifted western Pacific summer climate variability since the early 2000s. The potential physical mechanism is summarized in a schematic diagram shown in Fig. 10. Before the early 2000s, the slow-decaying ENSO with relatively high amplitude induces the NIO SST anomalies in JJA. In the Pacific Ocean, opposite-sign SST anomalies develop in JJA following the decay of ENSO. Under the combined effects of the Indian and Pacific Ocean SST forcing, the summer rainfall anomalies display a high coherence over the western Pacific. After the early 2000s, ENSO has a shorter period and a reduced amplitude and the EEP SST anomalies tend to experience a phase transition from winter to summer. Consequently, the influence of ENSO on the tropical Indian Ocean is weak and there are no notable JJA NIO SST anomalies. Meanwhile, positive SST anomalies develop in the tropical Pacific in summer following an early phase transition of ENSO. In this case, the direct Pacific SST forcing becomes relatively more important for the western Pacific summer climate variability. Correspondingly, summer rainfall anomalies feature an eastward-shifted pattern over the western Pacific in the recent period. The proposed mechanism is further supported by two sets of CGCM numerical experiments.
Idealized schematic diagram highlighting changes in the tropical atmospheric and oceanic conditions and their interactions in DJF, MAM, and JJA for the two periods 1982–2003 and 2004–16. (top) The long-lasting DJF La Niña or El Niño respectively promotes second cooling or warming in the NIO during summer through a series of air–sea interaction processes. These NIO SST anomalies, in conjunction with the Pacific SST anomalies, facilitate occurrence of zonally elongated lower-level cyclonic circulation and same-sign rainfall anomalies over the whole western Pacific during the 1982–2003. (bottom) The eastern Pacific SST anomalies tend to evolve with shorter period and smaller amplitude since the early 2000s. These weakened ENSO events undergo a phase transition from DJF to JJA. It lessens response in the tropical Indian Ocean and reduces SST variance over the NIO in summer. Meanwhile, the transition-phase ENSO events feature well-developed equatorial Pacific SST anomalies in the following JJA, making the Pacific SST heating become more significant. The declining Indian Ocean forcing and the strengthened Pacific forcing lead to an eastward-shifted western Pacific lower-level wind anomaly and rainfall anomaly pattern in summer. The shading background denotes SST anomalies, with contours highlighting the key changes of SST anomalies between the two periods. The figure is processed using GrADS and Adobe Illustrator.
Citation: Journal of Climate 33, 3; 10.1175/JCLI-D-19-0150.1
ENSO variability is controlled by a delicate balance of various amplifying and damping feedbacks. The causes for changes in the ENSO properties could be multiple. For example, the westward-shifted structure of the coupled mode in the recent decade (Yeh et al. 2009) may favor a high-frequency period and low-amplitude ENSO event by a quick damping oscillation and a shortened basinwide thermocline adjustment (Suarez and Schopf 1988; An and Wang 2000; Sun and Yu 2009). In addition, the external forcing may also play a role in modulating the ENSO properties. For instance, changes in the phase of the Pacific decadal oscillation, the strength of the Atlantic meridional overturning circulation, and the deep-ocean heat uptake may affect the Pacific mean state (Meehl et al. 2011; Katsman and van Oldenborgh 2011; Trenberth and Fasullo 2013) and adjust the oscillation by altering the coupling processes (Jin 1997; Fedorov and Philander 2000; Deser et al. 2010). It is not yet possible to say how ENSO will be in the future, and the causes for ENSO changes still need further investigation.
Acknowledgments
This study is jointly supported by National Key Research and Development Program of China (2016YFC1401401), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA20060500), Innovation Academy of South China Sea Ecology and Environmental Engineering (ISEE2018PY06), Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (GML2019ZD0306), Science and Technology Program of Guangzhou China (202002030540130002), Funds for Creative Research Groups of China (41521005), National Natural Science Foundation of China grants (41530425, 41775080, 41721004, 41676013, 41731173, 41506003, and 41776023), the Rising Star Foundation of the South China Sea Institute of Oceanology (NHXX2018WL0201), and the Independent Research Project Program of State Key Laboratory of Tropical Oceanography (LTOZZ1802). The authors gratefully acknowledge the use of the HPCC for all numeric simulations at the South China Sea Institute of Oceanology, Chinese Academy of Sciences.
REFERENCES
Adler, R. F., and Coauthors, 2003: The version 2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 1147–1167, https://doi.org/10.1175/1525-7541(2003)004<1147:TVGPCP>2.0.CO;2.
An, S. I., and B. Wang, 2000: Interdecadal change of the structure of the ENSO mode and its impact on the ENSO frequency. J. Climate, 13, 2044–2055, https://doi.org/10.1175/1520-0442(2000)013<2044:ICOTSO>2.0.CO;2.
Chang, C. P., Y. Zhang, and T. Li, 2000: Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part I: Roles of the subtropical ridge. J. Climate, 13, 4310–4325, https://doi.org/10.1175/1520-0442(2000)013<4310:IAIVOT>2.0.CO;2.
Chen, Z., Z. Wen, R. Wu, P. Zhao, and J. Cao, 2014: Influence of two types of El Niño on the East Asian climate during boreal summer: A numerical study. Climate Dyn., 43, 469–481, https://doi.org/10.1007/S00382-013-1943-1.
Deser, C., M. A. Alexander, S. P. Xie, and A. S. Phillips, 2010: Sea surface temperature variability: Patterns and mechanisms. Annu. Rev. Mar. Sci., 2, 115–143, https://doi.org/10.1146/annurev-marine-120408-151453.
Du, Y., S. P. Xie, G. Huang, and K. Hu, 2009: Role of air–sea interaction in the long persistence of El Niño–induced north Indian Ocean warming. J. Climate, 22, 2023–2038, https://doi.org/10.1175/2008JCLI2590.1.
Eaton, B., 2012: User’s guide to the Community Atmosphere Model CAM-5.1.1. NCAR Doc., 32 pp., http://www.cesm.ucar.edu/models/cesm1.0/cam/docs/ug5_1_1/ug.pdf.
Fedorov, A. V., and S. G. Philander, 2000: Is El Niño changing? Science, 288, 1997–2002, https://doi.org/10.1126/science.288.5473.1997.
Gill, A., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447–462, https://doi.org/10.1002/qj.49710644905.
He, Z., and R. Wu, 2014: Indo-Pacific remote forcing in summer rainfall variability over the South China Sea. Climate Dyn., 42, 2323–2337, https://doi.org/10.1007/s00382-014-2123-7.
He, Z., and R. Wu, 2018: Change in coherence of summer rainfall variability over the western Pacific around the early 2000s: Roles of Indo-Pacific forcing. J. Climate, 31, 3525–3538, https://doi.org/10.1175/JCLI-D-17-0687.1.
He, Z., R. Wu, and W. Wang, 2016: Signals of the South China Sea summer rainfall variability in the Indian Ocean. Climate Dyn., 46, 3181–3195, https://doi.org/10.1007/s00382-015-2760-5.
Hu, W., R. Wu, and Y. Liu, 2014: Relation of the South China Sea precipitation variability to tropical Indo-Pacific SST anomalies during spring-to-summer transition. J. Climate, 27, 5451–5467, https://doi.org/10.1175/JCLI-D-14-00089.1.
Hu, Z.-Z., A. Kumar, H. L. Ren, H. Wang, M. L’Heureux, and F.-F. Jin, 2013: Weakened interannual variability in the tropical Pacific Ocean since 2000. J. Climate, 26, 2601–2613, https://doi.org/10.1175/JCLI-D-12-00265.1.
Hu, Z.-Z., A. Kumar, J. Zhu, B. Huang, Y. Tseng, and X. Wang, 2017: On the shortening of the lead time of ocean warm water volume to ENSO SST since 2000. Sci. Rep., 7, 4294, https://doi.org/10.1038/S41598-017-04566-Z.
Ji, M., A. Leetmaa, and J. Derber, 1995: An ocean analysis system for seasonal to interannual climate studies. Mon. Wea. Rev., 123, 460–481, https://doi.org/10.1175/1520-0493(1995)123<0460:AOASFS>2.0.CO;2.
Jin, F.-F., 1997: An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model. J. Atmos. Sci., 54, 811–829, https://doi.org/10.1175/1520-0469(1997)054<0811:AEORPF>2.0.CO;2.
Kanamitsu, M., W. Ebisuzaki, J. Woollen, S. K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631–1643, https://doi.org/10.1175/BAMS-83-11-1631.
Katsman, C. A., and G. J. van Oldenborgh, 2011: Tracing the upper ocean’s “missing heat.” Geophys. Res. Lett., 38, L14610, https://doi.org/10.1029/2011GL048417.
Klein, S. A., B. J. Soden, and N.-C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917–932, https://doi.org/10.1175/1520-0442(1999)012<0917:RSSTVD>2.0.CO;2.
Lau, N. C., and M. J. Nath, 2003: Atmosphere–ocean variations in the Indo-Pacific sector during ENSO episodes. J. Climate, 16, 3–20, https://doi.org/10.1175/1520-0442(2003)016<0003:AOVITI>2.0.CO;2.
Lee, T., and M. J. McPhaden, 2010: Increasing intensity of El Niño in the central-equatorial Pacific. Geophys. Res. Lett., 37, L14603, https://doi.org/10.1029/2010GL044007.
Meehl, G. A., J. M. Arblaster, J. T. Fasullo, A. Hu, and K. E. Trenberth, 2011: Model-based evidence of deep-ocean heat uptake during surface-temperature hiatus periods. Nat. Climate Change, 1, 360–364, https://doi.org/10.1038/nclimate1229.
Reynolds, R. W., N. A. Rayner, T. M. Smith, D. C. Stokes, and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 1609–1625, https://doi.org/10.1175/1520-0442(2002)015<1609:AIISAS>2.0.CO;2.
Smith, R., and Coauthors 2010: The Parallel Ocean Program (POP) reference manual: Ocean component of the Community Climate System Model (CCSM). LANL Tech. Rep. LAUR-10-01853, 141 pp., http://www.cesm.ucar.edu/models/cesm1.0/pop2/doc/sci/POPRefManual.pdf.
Suarez, M. J., and P. S. Schopf, 1988: A delayed oscillator for ENSO. J. Atmos. Sci., 45, 3283–3287, https://doi.org/10.1175/1520-0469(1988)045<3283:ADAOFE>2.0.CO;2.
Sun, F., and J. Y. Yu, 2009: A 10–15-yr modulation cycle of ENSO intensity. J. Climate, 22, 1718–1735, https://doi.org/10.1175/2008JCLI2285.1.
Tan, W., X. Wang, W. Wang, C. Wang, and J. Zuo, 2016: Different responses of sea surface temperature in the South China Sea to various El Niño events during boreal autumn. J. Climate, 29, 1127–1142, https://doi.org/10.1175/JCLI-D-15-0338.1.
Tao, S. Y., and L. X. Chen, 1987: A review of recent research of the East Asian summer monsoon in China. Monsoon Meteorology, C. P. Chang and T. N. Krishnamurti, Eds., Oxford University, 60–92.
Torrence, C., and G. P. Compo, 1998: A practical guide to wavelet analysis. Bull. Amer. Meteor. Soc., 79, 61–78, https://doi.org/10.1175/1520-0477(1998)079<0061:APGTWA>2.0.CO;2.
Trenberth, K. E., and J. T. Fasullo, 2013: An apparent hiatus in global warming? Earth’s Future, 1, 19–32, https://doi.org/10.1002/2013EF000165.
Vertenstein, M., T. Craig, A. Middleton, D. Feddema, and C. Fischer, 2011: CESM1.0.4 user’s guide. NCAR Doc., 146 pp., http://www.cesm.ucar.edu/models/cesm1.0/cesm/cesm_doc_1_0_4/x42.html.
Wang, B., R. Wu, and X. Fu, 2000: Pacific–East Asian teleconnection: How does ENSO affect East Asian climate? J. Climate, 13, 1517–1536, https://doi.org/10.1175/1520-0442(2000)013<1517:PEATHD>2.0.CO;2.
Webster, P. J., A. M. Moore, J. P. Loschnigg, and R. R. Leben, 1999: Coupled ocean–atmosphere dynamics in the Indian Ocean during 1997–98. Nature, 401, 356–360, https://doi.org/10.1038/43848.
Wu, R., 2002: Processes for the northeastward advance of the summer monsoon over the western North Pacific. J. Meteor. Soc. Japan, 80, 67–83, https://doi.org/10.2151/jmsj.80.67.
Wu, R., and B. Wang, 2002: A contrast of the East Asian summer monsoon and ENSO relationship between 1962–77 and 1978–93. J. Climate, 15, 3266–3279, https://doi.org/10.1175/1520-0442(2002)015<3266:ACOTEA>2.0.CO;2.
Wu, R., and B. P. Kirtman, 2007: Regimes of local air–sea interactions and implications for performance of forced simulations. Climate Dyn., 29, 393–410, https://doi.org/10.1007/s00382-007-0246-9.
Wu, R., and S. W. Yeh, 2010: A further study of the tropical Indian Ocean asymmetric mode in boreal spring. J. Geophys. Res., 115, D08101, https://doi.org/10.1029/2009JD012999.
Wu, R., B. P. Kirtman, and V. Krishnamurthy, 2008: An asymmetric mode of tropical Indian Ocean rainfall variability in boreal spring. J. Geophys. Res., 113, D05104, https://doi.org/10.1029/2007JD009316.
Wu, R., G. Huang, Z. Du, and K. Hu, 2014: Cross-season relation of the South China Sea precipitation variability between winter and summer. Climate Dyn., 43, 193–207, https://doi.org/10.1007/s00382-013-1820-y.
Xie, S.-P., H. Annamalai, F. A. Schott, and J. P. McCreary, 2002: Structure and mechanisms of south Indian Ocean climate variability. J. Climate, 15, 864–878, https://doi.org/10.1175/1520-0442(2002)015<0864:SAMOSI>2.0.CO;2.
Xie, S.-P., K. Hu, J. Hafner, H. Tokinaga, Y. Du, G. Huang, and T. Sampe, 2009: Indian Ocean capacitor effect on Indo-western Pacific climate during the summer following El Niño. J. Climate, 22, 730–747, https://doi.org/10.1175/2008JCLI2544.1.
Xie, S.-P., Y. Du, G. Huang, X.-T. Zheng, H. Tokinaga, K. Hu, and Q. Liu, 2010: Decadal shift in El Niño influences on Indo-western Pacific and East Asian climate in the 1970s. J. Climate, 23, 3352–3368, https://doi.org/10.1175/2010JCLI3429.1.
Yeh, S. W., J. S. Kug, B. Dewitte, M. H. Kwon, B. P. Kirtman, and F. F. Jin, 2009: El Niño in a changing climate. Nature, 461, 511–514, https://doi.org/10.1038/nature08316.
Yeh, S. W., B. P. Kirtman, J. S. Kug, W. Park, and M. Latif, 2011: Natural variability of the central Pacific El Niño event on multi-centennial timescales. Geophys. Res. Lett., 38, L02704, https://doi.org/10.1029/2010GL045886.
Yu, L., X. Jin, and R. A. Weller, 2008: Multidecade global flux datasets from the Objectively Analyzed Air–Sea Fluxes (OAFlux) Project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. Woods Hole Oceanographic Institution OAFlux Project Tech. Rep. OA-2008-01, 64 pp., http://oaflux.whoi.edu/pdfs/OAFlux_TechReport_3rd_release.pdf.
Zhang, Y., W. B. Rossow, A. A. Lacis, V. Oinas, and M. I. Mishchenko, 2004: Calculation of radiative fluxes from the surface to top of atmosphere based on ISCCP and other global data sets: Refinement of the radiative transfer model and the input data. J. Geophys. Res., 109, D19105, https://doi.org/10.1029/2003JD004457.
Zhou, T. J., and R. C. Yu, 2005: Atmospheric water vapor transport associated with typical anomalous summer rainfall patterns in China. J. Geophys. Res., 110, D08104, https://doi.org/10.1029/2004JD005413.
