1. Introduction
The Indo-China Peninsula (ICP) is composed of several agriculture-based countries with dense populations, including Thailand, Vietnam, Cambodia, Laos, and Myanmar. Surface air temperature (SAT) is vital to the countries in the ICP since adverse SAT can exert serious impacts on the local crop yields, the vegetation greenness, the physical health of citizens and eventually hindering the national economic growth (Gasparrini et al. 2015; Lamchin et al. 2018; Luong et al. 2019; Marjuki et al. 2016; Peng et al. 2004; Phosri et al. 2020; Thirumalai et al. 2017). As one of the rapid SAT rising regions against the background of global warming, the ICP is considered to be highly vulnerable to climate change (Choi et al. 2009; Ge et al. 2018; Marjuki et al. 2016; Mie Sein et al. 2021; Nguyen et al. 2014). It has been demonstrated that the pronounced increase in ICP SAT features great seasonality with the warming rate in winter much larger than that in summer (Choi et al. 2009; Ge et al. 2018; Nguyen et al. 2014). Both Ge et al. (2018) and Nguyen et al. (2014) noticed that the SAT fluctuations in winter are relatively amplified compared to those in other seasons. Thus, researches concerning the wintertime ICP SAT variability are of great significance.
The existing studies related to ICP SAT have targeted at different regions: some covered a larger scope of application including the ICP (Caesar et al. 2011; Chen and Song 2018, 2019; Choi et al. 2009; Li 2020; Manton et al. 2001; Nicholls 2005; Zhu et al. 2020a), some precisely focused on the ICP (Gao et al. 2019; Ge et al. 2018; Lin et al. 2018; Luo and Lau 2017; Thirumalai et al. 2017; Zhu et al. 2020b) and others concentrated only on individual nations in the ICP (Mie Sein et al. 2021; Nguyen et al. 2014). Generally, SAT may change in response to both thermodynamic and dynamic forcing (Clark and Feldstein 2020; Lee et al. 2011). For instance, SAT over the ICP can be modulated by surface heat fluxes and temperature advection in situ (Chen and Song 2018). Specifically, changes in local precipitation, soil moisture and cloud cover correlate to the variation of ICP SAT via altering the net surface heat flux (Gao et al. 2019; Halpert and Ropelewski 1992; Thirumalai et al. 2017; Trenberth 2002). Given that the lower-tropospheric northeasterly linked to the East Asian winter monsoon (EAWM) prevails over the ICP during winter (Buckley et al. 2014; Sengupta and Nigam 2019; Sooktawee et al. 2014), the temperature advection dominantly induced by anomalous horizontal winds could significantly facilitate the warming or cooling of the ICP (Chen and Song 2018, 2019; Luo and Lau 2017; Nguyen et al. 2014).
It has been found that the EAWM circulation, the intensity of Siberian High and the Arctic Oscillation (AO) are some potential predictors of ICP SAT (Buckley et al. 2014; Buntoung et al. 2020; Chen and Song 2018, 2019; Sooktawee et al. 2014). Meanwhile, the interannual variations of both temperature extreme events (Caesar et al. 2011) and seasonal mean temperature (Chen and Song 2018) over the ICP positively correlate to the sea surface temperature (SST) over the tropical central and eastern Pacific (TCEP), the Indian Ocean (IO), the Bay of Bengal (BOB), and the South China Sea (SCS). Above all, El Niño–Southern Oscillation (ENSO) is foremost since it is the strongest and most well-known mode of tropical climate variability on interannual time scale. The ENSO-induced extreme diabatic forcing in the tropical Pacific could significantly affect the large-scale atmospheric anomalies, further influencing SAT in both the tropics and extratropics (Chowdary et al. 2014; Halpert and Ropelewski 1992; Kiladis and Diaz 1989; Trenberth 2002).
Typically, an anomalous lower-tropospheric anticyclone over the western North Pacific (WNPAC) would emerge and sustain during El Niño cycle. Note that though the WNPAC anomaly is originally induced by ENSO, the diversity of its ultimate behavior also tends to be modulated by SST anomalies beyond the TCEP. Many preceding studies have concluded that the atmospheric response in Northern Hemisphere, especially over the WNP, to IO SST anomalies is of great concern even when ENSO peaking, emphasizing the potential dependence of WNPAC upon the ENSO–IO SST linkage (Annamalai et al. 2007; Kim et al. 2016; Watanabe and Jin 2002). For the past few decades, the WNPAC anomaly is recognized as the pivotal system that bridges ENSO and the climate of countries like China, Japan, or those in the ICP (Chen et al. 2013; Wang et al. 2000; Wu et al. 2010a; Xie et al. 2016, 2009; Zhang et al. 1996, 2014). Recently, Thirumalai et al. (2017) estimated that 49% of the record-breaking 2016 April SAT extreme event over the ICP (2°C warmer than normal) was caused by the 2015/16 El Niño. So, how ENSO connects to wintertime ICP SAT, hereafter referred to as the ENSO–ICP SAT relationship for brief, is one of our concerns.
Being the paramount source for seasonal predictions, ENSO teleconnections are not persistently stable but exhibit substantial interdecadal modulations, which presents great challenges for climate forecast. For instance, the out-of-phase ENSO–EAWM relationship experienced a decline around the mid-1970s and subsequently recovered around the late-1990s, which may be attributed to the regulations of the Pacific decadal oscillation (PDO), the Atlantic multidecadal oscillation (AMO) or other factors (Geng et al. 2016; Gong et al. 2019; Hao and He 2017; He and Wang 2013; He et al. 2013; Kim et al. 2013, 2016; Wang and He 2012; Wang et al. 2008). Given the close linkage of ICP SAT to the EAWM, one would question whether the ENSO–ICP SAT relationship is stable in time. Therefore, this study intends to investigate whether and why the ENSO–ICP SAT relationship changes on the interdecadal time scale. Furthermore, what are the main factors adjusting the atmospheric responses to ENSO and eventually regulating the oscillating ENSO–ICP SAT relationship is an interesting issue that remains unclear. All these questions would be addressed in the present study.
The rest of the paper is organized as follows. Datasets, methods and model experiment design are introduced in section 2. Section 3 describes the climatological condition of winter ICP SAT and its interannual linkage with ENSO for the whole studying period. The interdecadal change in the ENSO–ICP SAT relationship as well as the crucial role of the WNPAC anomaly are documented in section 4. Section 5 includes a general survey of tropical Indo-Pacific SST anomalies and the detailed investigation of the role played by SST anomalies in the northwest IO (NWIO) using both observational analyses and numerical experiments, followed by summary and discussion in section 6.
2. Datasets, methods, and model experiments
a. Datasets
Monthly mean SST from Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST) available on 1° grid from 1870 to the present (Rayner 2003) was used in this study. Monthly mean SAT since 1901 onward was adopted from the high-resolution (0.5° grid) CRU TS v4 dataset provided by the Climatic Research Unit in University of East Anglia (Harris et al. 2020). Other variables including atmospheric fields and heat fluxes were provided by twentieth Century Reanalysis v3 dataset (20CRv3) from 1836 to 2015 with a horizontal resolution of 1° (Slivinski et al. 2019), which is accessible via the website at https://psl.noaa.gov/ of the NOAA/OAR/ESRL PSL, Boulder, Colorado.
b. Methods
This study investigated the wintertime average [December–February (DJF)] of variability, and anomalies presented were defined as departures from climatological values for each month in the whole analysis period (1958–2015). On account of the dominant interannual variance contribution of winter ICP SAT (more than 65%) and ENSO, interannual variations on time scales shorter than 9 years were extracted using a harmonic analysis throughout the study.
The area-averaged SAT anomaly over the ICP (8°–22°N, 98°–110°E; denoted by green boxes in Fig. 1) was defined as an ICP SAT index (ICPTI) during winter, which is shown in Fig. 1c. The Niño-3.4 (5°S–5°N, 170°–120°W) SST in DJF was used to represent the variability of ENSO. According to the criterion of 0.8 standard deviations, 12 El Niño (1957/58, 1963/64, 1965/66, 1972/73, 1976/77, 1982/83, 1986/87, 1987/88, 1991/92, 1997/98, 2002/03, 2009/10) and 13 La Niña (1967/68, 1970/71, 1973/74, 1975/76, 1983/84, 1984/85, 1988/89, 1995/96, 1998/99, 1999/00, 2005/06, 2007/08, 2010/11) episodes were identified during 1958–2015.
(a) DJF-mean climatology of SAT (colors, °C) and 1000-hPa horizontal wind (arrows, m s−1) during 1958–2015. (b) Regression patterns of DJF-mean anomalous SAT (colors, °C) and 1000-hPa horizontal wind (arrows, m s−1) against the standardized PC1 associated with the first mode of MV-EOF. (c) Standardized time series of DJF Niño-3.4 index (bars), ICP SAT index (ICPTI, DJF-mean SAT averaged over the ICP: 8°–22°N, 98°–110°E, denoted by green boxes) based on CRU land temperature (red line) and 20CRv3 1000-hPa air temperature (green line) as well as PC1 (blue line) associated with (b). (d) Regression pattern of DJF-mean anomalous SST (colors, °C) against the standardized ICPTI. In (b) and (d), only regression coefficients exceeding 95% of confidence level are shown.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
Several common statistical techniques were employed, including regression, linear correlation and composite analysis, and the significance of these analyses was estimated by the two-tail Student’s t test. Besides, multivariate empirical orthogonal function (MV-EOF) analysis was applied to DJF-mean SAT and 1000-hPa horizontal wind anomalies over the ICP region.
1) Decomposition of horizontal surface temperature advection
2) Decomposition of running correlation
c. Model description and experiment design
The Community Atmospheric Model, version 4 (CAM4) is the seventh generation atmospheric general circulation model (AGCM) developed by the National Center for Atmospheric Research (NCAR) (Neale et al. 2013). In the current study, the horizontal resolution of CAM4 simulation was set to 1.9° latitude × 2.5° longitude. For purpose of identifying to what extent the IO SST anomalies could modulate the atmospheric circulation, four 40-yr numerical simulations were conducted and only the results of the last 35 years were analyzed, with the first 5 years truncated as model spinup. Note that since individual years are largely independent in the process of integration, the individual years are equivalent to ensemble members and thus the 35-yr averages are equivalent to ensemble means. The control run (CTRL run) was forced by the climatological mean annual cycles of worldwide SST and sea ice. The other three sensitive experiments (SEN run) were designed with different tropical Indo-Pacific SST anomalies superimposed onto the climatological SST during boreal winter, specifically from November to March (see details below in section 5). The significance of differences between any two simulations is assessed by the two-tail Student’s t test.
3. Winter SAT over the ICP and its interannual linkage to ENSO
During boreal winter, the lower-level northeasterly wind prevails from the SCS blowing through the ICP on the way to the BOB. Accompanied by the steady northeasterly flow, the SAT over the ICP is characterized by a large southwest–northeast gradient, with the higher SAT (above 26°C) in the southern part of Thailand and Cambodia as well as the lower SAT (below 20°C) in the northern part of Vietnam and Laos (Fig. 1a). As shown in Table 1, the seasonal average of SAT in winter is the lowest (about 22°C) throughout the year, while the climatic states in other three seasons are all above 25°C. Moreover, the fluctuation in winter is the most significant with the standard deviation (0.77°C) much higher than that in other seasons. These results are similar to the findings of Ge et al. (2018), Thirumalai et al. (2017), and Nguyen et al. (2014). In addition, this paper supplied another index, spatial variance which represents the uniformity of spatial SAT distribution. Small values of spatial variance in spring, summer and autumn indicate that SAT is relatively well distributed under a warmer mean state. In contrast, the high value in winter (11.24°C2), resulting from the large southwest–northeast SAT gradient, implies that the horizontal surface temperature advection may be prone to change the ICP SAT in situ readily.
Three indices related to ICP SAT in winter, spring, summer, and autumn during 1958–2015. Values in boldface font refer to those in winter, the season of concern in the present study.
It seems in Fig. 1c that the wintertime ICP SAT varies in phase with Niño-3.4 index and the correlation coefficient between them reaches 0.48 (p < 0.01) for the whole studying period. Figure 1d further presents the regression pattern of winter SST anomaly in regard to ICPTI, displaying large positive SST anomalies over the TCEP flanked by negative SST anomalies over the northwestern and southern Pacific. This pattern of SST anomalies is strongly reminiscent of El Niño, indicating that an abnormally warmer (cooler) winter in the ICP would occur during the mature phase of El Niño (La Niña). Moreover, the wintertime ICP SAT also positively correlates to SST in adjacent sea areas like the SCS or the BOB and to that over the northern IO.
The perception of the in-phase ENSO–ICP SAT relationship could date back to 1980s (Halpert and Ropelewski 1992; Kiladis and Diaz 1989), but the physical mechanisms bonding them two have been rarely investigated due to limitations of low station density and national conditions (Ge et al. 2018; Mie Sein et al. 2021). Figure 2 demonstrates the large-scale atmospheric circulation responses at the upper and lower troposphere to ENSO-related SST anomalies. Furthermore, the localized responses related to the ICP SAT variation are displayed in Fig. 3. The net surface heat flux in Fig. 3b is equivalent to the summation of surface sensible, latent, net shortwave radiative and net longwave radiative heat flux, and positive (negative) values indicate downward (upward) heat fluxes. Thus, Figs. 2 and 3 together passably establish the connection between ICP SAT and ENSO during winter with both thermodynamic and dynamic processes concerned.
Regression patterns of DJF-mean anomalous (a) velocity potential (colors, 106 m2 s−1) and divergent wind (arrows, m s−1) at 200 hPa and (b) streamfunction (colors, 106 m2 s−1) and horizontal wind (arrows, m s−1) at 850 hPa against the standardized DJF Niño-3.4 index. The stippling indicates the significance of the variables represented by colors exceeding 95% of confidence level. Only arrows at the 95% confidence level are shown. The green boxes here and hereafter denote the target domain of the ICP: 8°–22°N, 98°–110°E.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
Regression patterns of DJF-mean anomalous (a) total cloud cover (%), (b) net surface heat flux (W m−2), (c) horizontal surface temperature advection (°C month−1) and its components: (d) A1, (e) A2, and (f) A3 against the standardized DJF Niño-3.4 index. In (b), positive (negative) values indicate that net surface heat fluxes are downward (upward), acting to heat (cool) the surface. The shading indicates the significance of the variables represented by colors exceeding 95% of confidence level.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
Take the circumstance of El Niño events as an example, the warmer SST anomalies over the TCEP accompanied by enhanced convection in situ can induce large-scale zonal atmospheric overturning with the ascending branch over the tropical central Pacific and the descending one over the tropical western Pacific covering the ICP. The zonal atmospheric overturning anomaly, also known as the abnormal Walker circulation (Bjerknes 1969), is characterized by upper-tropospheric velocity potential anomalies (Fig. 2a). Controlled by the downward motion, the ICP features the cloudless condition of suppressed convection (Fig. 3a), which is conducive to more local incident solar radiation and hence the positive net surface heat flux (Fig. 3b).
Besides, the El Niño–related SST anomalies could generate an anomalous WNPAC as Matsuno–Gill response (Gill 1980; Matsuno 1966). Situated on the northwestern flank of the WNPAC anomaly, the ICP is strongly influenced by the anomalous southwesterly that opposes the climatic northeasterly (Fig. 2b). Further decomposition of the surface horizontal temperature advection reveals that the El Niño–induced above-normal warm temperature is primarily attributed to the horizontal temperature advection, among which A1—caused by anomalous horizontal wind—is dominated (Figs. 3c,d). Also, the El Niño–related SAT anomalies with higher values in the central part of the ICP together with the climatic northeasterly, that is, A2, could further warm the ICP to some degree (Fig. 3e). Unsurprisingly, the effect of the higher-order term (A3) is negligible (Fig. 3f). The circumstances of La Niña are concerned to be nearly opposite to those of El Niño.
To attach significance to the local temperature advection associated with the lower-level horizontal wind, the MV-EOF analysis of 1000-hPa air temperature and horizontal wind within the ICP was employed. The first mode of the MV-EOF explains 74.85% of the total variance, and the corresponding principal component (PC1) is shown in Fig. 1c (blue line). Figure 1b displays the regression patterns of SAT and 1000-hPa horizontal wind against the normalized PC1, indicating that the abnormal ICP warming is tightly coupled with the anomalous in situ southwesterly which transports amounts of warm air into the ICP. The strong correlations of PC1 to ICPTI (0.99 with p < 0.01) and to Niño-3.4 index (0.48 with p < 0.01) once again show that the lower-tropospheric horizontal wind anomalies play a dominant role in modulating winter ICP SAT through the temperature advection.
4. Oscillating ENSO–ICP SAT relationship modulated by the anomalous western North Pacific anticyclone
Although the positive correlation of winter ICP SAT to ENSO is significant for the whole analysis period, rescanning Fig. 1c unravels some clues of interdecadal variations. It seems that the ICPTI and Niño-3.4 index covary more evidently before 1976 and after 1993 than for the period in between. To verify the supposition of such interdecadal changes, Fig. 4a presents the 21-yr running correlation between winter ICP SAT and Niño-3.4 index. The sliding correlations using two different SAT datasets show high consistency, characterized by large values of correlation coefficient exceeding 95% of confidence level outside the period of 1976–93, during which the correlation is trivial (almost close to zero). The interdecadal modulation of the ENSO–ICP SAT relationship is also detected by a variety of moving windows (15, 17, and 19 yr), and the results are similar to that of Fig. 4a (not shown). Thus, based on the oscillating ENSO–ICP SAT relationship revealed by Fig. 4a, we further divided the whole study period (1957–2014) into three subperiods: 1957–75 (P1), 1976–93 (P2), and 1994–2014 (P3), marked by red, blue, and green, respectively.
(a) The 21-yr window correlation between the DJF Niño-3.4 index and ICPTI based on CRU land temperature (black line) and 20CRv3 1000-hPa air temperature (pink line), respectively. (b) The anomalies of ICPTI in El Niño years (orange bars) and La Niña years (purple bars). Scatterplot of the standardized DJF Niño-3.4 index vs the standardized ICPTI during (c) the whole period (1957–2014), (d) P1 (1957–75), (e) P2 (1976–93), and (f) P3 (1994–2014). The correlation coefficient and corresponding p value are shown at the bottom-right corner of each scatterplot. In (a) and (b) and hereafter, the red, blue, and green bars at the bottom represent the years covered by P1, P2, and P3, respectively.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
To confirm whether the interdecadal changes of the ENSO–ICP SAT relationship are generated by some outliers in years during P2, Fig. 4b illustrates the anomalies of ICPTI in the selected 12 El Niño and 13 La Niña years. For P1 and P3, nearly all El Niño (La Niña) years coincide with the emergence of an abnormally warm (cool) winter in the ICP. But that is not the same story for P2, during which the in-phase ENSO–ICP SAT relationship is broken. Only one out of five El Niño events is accompanied by the above-normal ICPTI, and even a slightly cooler winter instead of a warmer one appeared over the ICP during December–February of 1982/83, one of the few super El Niño events so far. Besides, the correlation coefficient equals 0.48 (p < 0.01) for the whole period as previously mentioned, while that for P1 and P3 reaches a higher value of 0.73 (p < 0.01) and 0.68 (p < 0.01), respectively. However, the correlation coefficient for P2 is a close-to-zero negative number (−0.22), failing to pass the test of significance. In addition, the ENSO-induced SAT anomalies exceeding 95% of confidence level overspread the entire ICP region for the whole period and for the two subperiods: P1 and P3 (Figs. 5a,b,d), while ENSO exerts little influence on the winter ICP SAT for P2 owing to insignificant values over the ICP (Fig. 5c).
Regression pattern of DJF-mean anomalous SAT (°C) against the standardized DJF Niño-3.4 index during (a) the whole period, (b) P1, (c) P2, and (d) P3. The shading indicates the significance of the regressed SAT exceeding 95% of confidence level.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
In short, there exist authentic interdecadal changes in the ENSO–ICP SAT relationship with strong positive correlation in P1 or P3 and insignificant negative one in P2. We also replicate Fig. 4 with extending the analysis back to 1901, and find that interdecadal changes in the ENSO–ICP SAT can still be seen before 1958 but with low consistency between the two SAT datasets (not shown). The dramatic drop of correlation in P2 is unprecedented since 1901, urging extra attention along with physical explanations.
Given the importance of surface temperature advection anomalies that are induced by anomalous horizontal wind [A1 in Eq. (1)], Figs. 6a–c show that ENSO-related horizontal wind anomalies over the ICP are able to induce substantial anomalies of temperature advection in situ for P1 and P3 but those for P2 being to no avail. Meanwhile, whether the winter ICP SAT per se remains interrelated with the in situ temperature advection in different subperiods is also examined in Figs. 6d–f. Understandably, whenever ICP SAT anomalies emerge, profound anomalies of local surface temperature advection tend to co-occur, even for P2. Hence, it is the malfunction of ENSO in generating effective surface temperature advection anomalies that results in the weak ENSO–ICP SAT relationship during P2.
Regression pattern of DJF-mean anomalous horizontal temperature advection induced by anomalous winds (A1; °C month−1) against the standardized DJF Niño-3.4 index during (a) P1, (b) P2, (c) P3, and against the standardized ICPTI during (d) P1, (e) P2, and (f) P3. The shading indicates the significance of the regressed advection exceeding 95% of confidence level. (g) The 21-yr window correlation of DJF-mean 1000-hPa zonal wind (black) or meridional wind (pink) averaged over the ICP with respect to the DJF Niño-3.4 index (solid lines) and ICPTI (dashed lines), respectively.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
Figure 6g exhibits the 21-yr sliding correlations both of the ICP-averaged zonal wind (black) and meridional wind (pink) with Niño-3.4 index (solid) and ICPTI (dashed), respectively. Both the moving correlations of the zonal and meridional wind averaged over the ICP with respect to ICPTI are significantly strong (exceeding 95% of confidence level) all the time as shown in Figs. 6d–f. However, as for Niño-3.4 index, both the moving correlations drop below the standard line of 90% of confidence level in P2. The correlation coefficient between the 21-yr ENSO-regarded sliding correlations to the winter ICP SAT (black solid line in Fig. 4a) and that to the ICP-averaged zonal (meridional) wind shown as the black (pink) solid line in Fig. 6g reaches 0.93 (0.96) with p < 0.01, indicating that the ENSO-induced horizontal wind anomalies over the ICP play a leading part in modulating the ENSO–ICP SAT relationship via changing temperature advection with zonal and meridional wind anomalies contributing equally.
We assume that the anomalous southwesterly wind over the ICP may depend upon the pressure gradient between East Asian continent and WNP, with the latter is our primary concern. Thus, Figs. 7a–c depict the regression distributions of 850-hPa streamfunction and horizontal wind anomalies over the areas surrounding the ICP. The characterizations of the abnormal lower-level circulations in P1 and P3 share something in common, which are visibly distinct from those in P2. The ENSO-driven atmospheric responses in P1 and P3 consist of two anticyclonic flows: the strong one centered over the Philippines with southwesterly over the ICP and southeastern China, and the relatively weak one over the BOB associated with powerful easterly in the deep tropics as well as mild southwesterly in the Indian subcontinent (Figs. 7a,c). This is consistent with the findings of Sengupta and Nigam (2019) in their Fig. 7a. Obviously, the abnormal anticyclonic circulation over the Philippines drastically weakens in P2, in line with the disappearance of anomalous southwesterly flows over the ICP, which directly leads to the deficit of in situ temperature advection. By this token, the intensity of the anomalous anticyclone over the Philippines or western North Pacific (also known as WNPAC) is essential to the lower-tropospheric atmospheric circulations over the ICP, straightly modulating the ENSO–ICP SAT relationship.
Regression patterns of DJF-mean anomalous streamfunction (colors, 106 m2 s−1) and horizontal wind (arrows, m s−1) at 850 hPa against the standardized DJF Niño-3.4 index during (a) P1, (b) P2, and (c) P3. The stippling indicates the significance of the regressed streamfunction exceeding 95% of confidence level. Only arrows at the 95% confidence level are shown. The letter A represents an anomalous anticyclone. (d) The intensity of the anomalous western North Pacific anticyclone (WNPAC) defined by area-averaged 850-hPa streamfunction over 0°–20°N, 110°–135°E (denoted by white boxes) in P1, P2, and P3, respectively.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
In fact, plenty of previous studies have attached great importance to the intensity of the anomalous WNPAC for its crucial role in linking ENSO and the climate of Southeast and East Asia (Chang et al. 2000; Chen et al. 2015; Feng et al. 2014; He et al. 2019; Jiang et al. 2019; Kim et al. 2016; Kim and An 2019; Wang et al. 2000; Xie et al. 2016, 2009; Yuan et al. 2012; Zhang et al. 1996). In the present study, the regressed 850-hPa streamfunction anomalies averaged over 0°–20°N, 110°–135°E (denoted by white boxes in Figs. 7a–c) are defined as the strengths of the anomalous WNPAC, which are shown in Fig. 7d. The anomalous WNPAC intensity in P1 or P3 is large approximating 0.8 × 106 m2 s−1 and that in P2 reduced to about 0.6 × 106 m2 s−1, corresponding with the interdecadal changes in the ENSO–ICP SAT relationship.
5. Differences in the ENSO-related tropical Indo-Pacific SST anomalies: Key role of anomalous SST over the northwest Indian Ocean
Since the unstable ENSO–ICP SAT relationship highly hinges on the ENSO-induced WNPAC anomaly, it needs to be settled what factors modulate the anomalous WNPAC making its performance in P2 differ from that in P1 or P3. As reviewed by Li et al. (2018), besides the effect of SST anomalies over the tropical Pacific, the IO SST anomalies can exert great influences on the generation and development of the anomalous WNPAC as well. Thus, necessarily required is the broader distribution of SST anomalies during the ENSO mature winter in three different subperiods, which is exhibited in Fig. 8. In this section, the discrepancy and potential role of the ENSO-related SST anomalies over each basin in regulating the abnormal WNPAC will be examined individually.
Regression pattern of DJF-mean anomalous SST (°C) against the standardized Niño-3.4 index during (a) P1, (b) P2, and (c) P3. The shading indicates the significance of the regressed SST exceeding 95% of confidence level. Four domains are displayed by black boxes: the northwest Indian Ocean (NWIO; 0°–20°N, 40°–75°E), the South China Sea (SCS; 0°–30°N, 100°–120°E), the western North Pacific (WNP; 0°–20°N, 130°–160°E), and the tropical central and eastern Pacific (TCEP; 5°S–5°N, 160°E–90°W), from left to right. The gray bars beneath each picture refer to the corresponding regressed SST averaged over each domain, with the blue dots signifying the area-mean values of P2.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
a. Tropical Pacific
Once the ENSO–ICP SAT relationship, recognized as a kind of ENSO teleconnections, changes over time, one will naturally deem that interdecadal changes of ENSO per se ought to be responsible for this relationship modulation. However, our analyses indicate that variation of the ENSO-related SST anomalies over the tropical Pacific makes limited contribution, though there indeed exist some discrepancies of SST anomalies in three subperiods. On the one hand, the patterns of the regressed SST in P1, P2 and P3 all resemble a traditional El Niño event, with positive SST anomalies over the TCEP (Figs. 8a–c). To alternate Niño-3.4 index with Niño-3 (5°S–5°N, 150°–90°W) or Niño-4 (5°S–5°N, 160°E–150°W) index when calculating the running ENSO–ICP SAT correlation yields similar results (not shown), implying that the spatial pattern of the tropical Pacific warming may do nothing to the interdecadal correlation changes. On the other hand, we quantified the warming magnitude using the area-averaged SST anomalies over the TCEP (5°S–5°N, 160°E–90°W), and found it in three decades all over 0.7°C with slight discrepancy.
In Fig. 9a, we further decomposed the 21-yr running correlation according to Eq. (2) following the method proposed by Geng et al. (2018) by substituting X and Y with ICPTI and Niño-3.4 index, respectively. The sum of the four terms (T1 + T2 + T3 + T4, orange line) nearly replicates the original running correlation (black line), whereas only T2 (red line) shows pronounced interdecadal changes similar to the original running correlation (R = 0.99) and thus the running covariance (T2) can account for most of the interdecadal correlation variation. Note that although there exist interdecadal changes in the variance of ICPTI (Ge et al. 2018) and Niño-3.4 index (Hu et al. 2020), which is also revealed in Fig. 9b, contributions of them (T3 and T4) are negligible compared to T2. In short, this study is no denial of the fact that ENSO properties have undergone interdecadal shifts, but the impacts of them on the ENSO–ICP SAT relationship are insignificant stemming from our analyses.
(a) The 21-yr window correlation between the DJF Niño-3.4 index and ICPTI as well as its four decomposed terms based on Eq. (2). (b) The 21-yr window standard deviation of ICPTI (blue) and Niño-3.4 index (green).
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
Apart from the TCEP warming, the WNP cooling is also considered as an important factor that would modulate the WNPAC anomaly (Li et al. 2018; Shi and Qian 2018; Wang et al. 2000; Wu et al. 2010a). The value of the negative SST anomalies averaged over the WNP (0°–20°N, 130°–160°E) in P2 is −0.2°C, stronger than that in P1 or P3 (−0.14°C). Since the WNPAC intensity maintains significantly anticorrelated with the WNP SST during 1958–2015 (not shown), cooler WNP SST anomalies should be theoretically accompanied by a stronger WNPAC anomaly. Nevertheless, it seems to be a paradox that the anomalous WNPAC of P2 is weaker with the relatively stronger WNP cooling. This implies that there may be other factors responsible for the reduction of the WNPAC anomaly in P2, further explorations about which are shown below.
b. South China Sea
The SCS-averaged (0°–30°N, 100°–120°E) SST anomalies in P1 (0.29°C) and P3 (0.22°C) are augmented with the warming extending more northward, compared to those in P2 (0.13°C) (Figs. 8a–c and 10a–c). As indicated by the study of Wang et al. (2006), variations of SCS SST are primarily governed by changes in shortwave radiation and latent heat flux. Pictures in the top and bottom panel of Fig. 10 illustrate these two processes driving anomalous SCS SST, respectively. Blue contours in Figs. 10a–c denote the net surface incident solar radiation with a positive (negative) value representative of more (less) shortwave radiation absorbed by the surface. Significant positive net surface solar radiation flux anomalies together with abnormal descending motions (yellow dots) cover the most part of the SCS, indicating that SCS SST anomalies are the responses to the increased solar radiation (attributed to the in situ suppressed convection). Moreover, the negative anomalies of surface upward sensible and latent heat flux induced by the anomalous southerly largely account for the northward extension of the SCS SST anomalies in P1 and P3 (Figs. 10d–f). The wind–evaporation–SST (WES) mechanism, responsible for interpretation of many atmosphere–ocean coupled phenomena, functions as well in the SCS as previously stated by Klein et al. (1999) and Wang et al. (2000, 2003).
Regression patterns of DJF-mean anomalous SST (colors, °C), 500-hPa vertical velocity (dots, green for upward motions and yellow for downward motions) and surface net incident solar radiation (blue contours, interval of 4 W m−2) against the standardized Niño-3.4 index during (a) P1, (b) P2, and (c) P3. The shading indicates the significance of the regressed SST exceeding 95% of confidence level. Only dots at the 95% confidence level are shown. The solid and dashed contours denote the positive and negative values, respectively, and the zero contours are omitted. (d)–(f) As in (a)–(c), but for sum of surface upward sensible and latent heat flux (colors, W m−2) and 1000-hPa horizontal wind (arrows, m s−1). Only regression coefficients exceeding 95% of confidence level are shown.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
According to the above analyses, the difference of ENSO-related SCS SST anomalies verified as the response to atmospheric forcing—stronger in P1 and P3 but weaker in P2—is in tandem with that of ENSO-related ICP SAT anomalies, again highlighting the role of the anomalous WNPAC in modulating the relationship between SCS SST as well as ICP SAT and ENSO during winter.
c. Northwest Indian Ocean
Finally, it remains to be investigated whether the discrepancy of the IO SST anomalies can actively induce different behaviors of the anomalous WNPAC. During the ENSO peaking winter, the IO is manifested with a uniform warming pattern (also known as the IO basin mode, IOBM) in different decades (Figs. 8a–c). Particularly, positive SST anomalies over the northwest IO (NWIO; 0°–20°N, 40°–75°E) exhibit most pronounced interdecadal changes with the magnitude in P2 (0.09°C) smaller than that in P1 (0.18°C) or P3 (0.16°C). To elucidate the interdecadal variation of the ENSO–NWIO SST linkage, black solid line in Fig. 11a shows that the correlation coefficient in P2 reduces to 0.5, albeit still significant at 95% confidence level, standing in stark contrast to that in P1 and P3. Seemingly, a stronger (weaker) WNPAC anomaly corresponds to the strong (weak) ENSO–NWIO SST linkage, which is similar to the finding of Kim et al. (2016).
(a) The 21-yr window correlation (left ordinate) between the DJF Niño-3.4 index and NWIO SST (black solid line) as well as the south IO SST (black dashed line), and 21-yr window standard deviation (right ordinate) of NWIO SST (pink line). P1 or P3 (P2) with a high (relatively low) correlation coefficient is referred to as the era of strong (weak) ENSO–NWIO SST linkage. (b) Simultaneous correlation between DJF-mean SST and 500-hPa vertical velocity (colors) as well as precipitation amount (contours, interval of 0.1) over the Indian Ocean. (c),(d) As in (b), but for correlation between DJF Niño-3.4 index and 500-hPa vertical velocity as well as precipitation amount in strong and weak ENSO–NWIO SST linkage era, respectively. In (b)–(d), the shading indicates the correlation coefficients related to 500-hPa vertical velocity exceeding 95% of confidence level. The green and yellow contours denote the positive and negative coefficients related to precipitation amount, respectively, and only contours at the 95% confidence level are shown. The solid and dashed black boxes denote the domain of NWIO (0°–20°N, 40°–75°E) and the south IO (10°–25°S, 55°–100°E), respectively.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
As for the IO SST anomalies, it remains debatable that whether they can exert effective impacts on the anomalous WNPAC during the ENSO mature winter, though the IO capacitor effect has been proved to be robust during the ENSO decaying summer (Chen et al. 2015; Wu et al. 2010b; Xie et al. 2016, 2009; Yang et al. 2007). Li et al. (2018) and Wu et al. (2009) proposed a minor impact of the IOBM on the WNP circulation in winter as a result of the counteraction between the active role over the western IO and the passive one over the eastern IO. Another school of researchers persisted that the active role of the IO SST anomalies in driving the atmosphere could not be neglected (Annamalai et al. 2007; Kim et al. 2016; Watanabe and Jin 2002). In the current study, the IO SST anomalies, especially over the NWIO, are proved to effectively modulate the ENSO–ICP SAT relationship via influencing the WNPAC anomaly by means of both observational analyses and numerical experiments.
1) Observational analyses
In Fig. 11b, we preliminarily probe into the degree of the role of IO SST in driving the atmosphere. Given the IO warming mode, a negative (positive) coefficient at any mesh corresponds to upward (downward) motion in situ, indicating the SST-forcing-atmosphere (atmosphere-forcing-SST) situation could dominate there during winter. It is obvious that significant negative values mainly cover two regions, the NWIO and the south IO, with the profound impact of the latter on Northern Hemisphere circulation has been unraveled by Annamalai et al. (2007). Situations are different elsewhere particularly with large positive values spreading the eastern IO, hence, the warming there is attributed to atmospheric forcing such as more downward solar radiation (not shown). Therefore, small discrepancy of SST anomalies over the eastern IO (Figs. 8a–c) is of no account in modulating the anomalous WNPAC. Note that when concerning the IO basin as a whole, the active role of SST is trivial because the net value averaged over the IO is small just as said by Wu et al. (2009) and Li et al. (2018). But with a careful examination on regional scales, evidences for the significant SST-ascending association over the NWIO and the south IO are revealed.
In the strong linkage era (P1&P3) when the strong NWIO warming coincide with ENSO in winter, significant upward motion anomalies correspondingly co-occur in situ (Fig. 11c). However, no significant vertical motion anomalies exist over the NWIO in the weak linkage era (P2) since the NWIO warming is weakened (Figs. 8b and 11d). Given that the ENSO-related anomalies of both SST and vertical velocity over the south IO are nearly the same in different decades, the ENSO–NWIO SST linkage may be the key factor modulating the anomalous WNPAC as well as the ENSO–ICP SAT relationship. The trivial interdecadal change in the impact of the south IO SST is also hinted by the black dashed line in Fig. 11a with a clear manifestation of high correlation (exceeding 99% of confidence level) throughout the whole analysis duration. Similar results using precipitation amount (contours in Figs. 11b–d) support these findings.
Furthermore, a conditional composite analysis was performed in Fig. 12. Because the NWIO SST anomalies are highly correlated to the ENSO amplitude among the chosen ENSO episodes (R = 0.8), the abnormal NWIO SST index is scaled by the concurrent Niño-3.4 index to rule out the potential impact of ENSO amplitude (Fig. 12a). Note that a large value of the scaled NWIO SST index represents the El Niño or La Niña winters with strong NWIO SST anomalies (strong ENSO–NWIO SST linkage), and a smaller value denotes those with weak NWIO SST anomalies (weak ENSO–NWIO SST linkage). Divided by the average value of the scaled NWIO SST index separately in El Niño and La Niña years, all events were classified into four quadrants. Be aware that the anomalous NWIO SST reaches 0.27° and 0.02°C in the strong and weak ENSO–NWIO SST linkage case, respectively, but the difference of SST anomalies over the TCEP (1.08° and 0.9°C) and south IO (0.29° and 0.27°C) is small. Then, differences of variables between Q1 and Q2 (Q4 and Q3) concretely depict the circumstances in the presence of strong (weak) NWIO SST anomalies.
(a) Scatterplot of the anomalous DJF Niño-3.4 index vs the NWIO SST anomalies scaled by the simultaneous Niño-3.4 index. The diamonds (triangles) stand for El Niño (La Niña) years. Years in P1, P2, and P3 are marked by red, blue, and green, respectively. Quadrants 1–4 are partitioned by dashed lines in (a). Composite patterns of DJF-mean anomalous SAT (colors, °C) and 850-hPa horizontal wind (arrows, m s−1) for (b) strong ENSO–NWIO SST linkage events (Q1 minus Q2 and further divided by 2) and for (c) weak ENSO–NWIO SST linkage events (Q4 minus Q3 and further divided by 2). (d),(e) As in (b) and (c), but for anomalous SST (colors, °C), 850-hPa streamfunction (black contours, interval of 0.3 × 106 m2 s−1) and 500-hPa vertical velocity (dots, green for upward motions and yellow for downward motions). The shading indicates the anomalous SAT or SST exceeding 95% of confidence level. The solid and dashed contours denote the positive and negative values, respectively, and the zero contours are omitted. Only arrows or dots at the 95% confidence level are shown. The values at the top right of (b) and (c) stand for the ICPTI anomalies, and those of (d) and (e) for the anomalous WNPAC intensity, which is defined as the area-averaged 850-hPa streamfunction over the blue boxes (0°–20°N, 110°–135°E), identical to Fig. 7d.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
Excitingly, after eliminating the influence possibly caused by super strong ENSO events, the anomalous WNPAC is much more intensified accompanied by strong (1.04 × 106 m2 s−1, Fig. 12d) than weak (0.56 × 106 m2 s−1, Fig. 12e) NWIO warming, and the stronger WNPAC anomaly is in tandem with significant anomalies of southwesterly wind and SAT (0.63°C) over the ICP (Fig. 12b). Consistent with Figs. 11b and 11c, upward motion anomalies are distributed over the NWIO and south IO where the underlying SST is abnormally warm in the strong linkage composite (Fig. 12d). Indicated by Watanabe and Jin (2002, in their Fig. 4), the NWIO warming with enhanced convection (corresponding to upper-level divergence in situ, not shown) may drive an anomalous zonal large-scale circulation with ascending branch over the NWIO (denoted by green dots) and descending one over the WNP (denoted by yellow dots). This WNP descending branch associated with the NWIO strengthened convection further intensifies the original downward motion anomalies over the WNP and hence a stronger WNPAC anomaly as a Rossby response (Gill 1980; Matsuno 1966). However, in the case of weak linkage composite, no evident convection anomalies are found over the NWIO because of weak SST anomalies, albeit the significant warming-ascending association exists in the south IO. At this point, the descending motion anomalies over the WNP are still significant under the impact of ENSO but with a smaller magnitude, resulting in a weaker WNPAC anomaly (Fig. 12e) and no significant anomalous wind and SAT (−0.03°C) over the ICP (Fig. 12c). Therefore, the ENSO–NWIO SST linkage could account for the modulation of the WNPAC anomaly to a large extent via inducing an anomalous zonal atmospheric circulation.
2) Numerical experiments
To verify the aforementioned assumption, we conducted four AGCM simulations (CTRL, SEN1, SEN2, and SEN3 run) by means of CAM4. Prescribed SST anomalies (colors in Fig. 13) for three sensitive runs to be added onto the climatological SST (contours in Fig. 13) from November to March are derived from the previous composite analysis in Figs. 12d and 12e, while the SST setting in the rest of the year is kept equivalent to the control run (viz., utilizing the climatological SST with no additionally superimposed SST anomalies). According to Fig. 13, differences between three SEN runs and the CTRL run could examine the effect of no, weak and strong IO warming, respectively. Particularly, differences between SEN3 and SEN2 run allow us to assess the specific impact of positive NWIO SST anomalies.
Prescribed SST anomalies (colors, °C) and the climatological SST (contours, interval of 2°C) upon which these SST anomalies are added in winter (from November to March), the sum of the two was used as the boundary condition to force CAM4 simulation for (a) SEN1, (b) SEN2, and (c) SEN3 run. Note that during the other times of the year, the three sensitive runs were designed with the global climatological SST with annual cycles, same as the control run setting introduced in section 2.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
When only SST anomalies in the tropical Pacific are prescribed, the pair of cyclones in the Pacific and the pair of anticyclones in the IO are captured but with a westward displacement compared to the observations (Figs. 13a and 14a). In the SEN1 run, anomalous northerly wind instead of southwesterly wind dominates in the ICP (Fig. 14e), implying the in-phase ENSO–ICP SAT relationship could not be established in this situation. In the presence of the weak IO warming forcing (primarily distributed in the southern IO; Fig. 13b), the anticyclonic flow slightly shifts eastward (Fig. 14b), but still resulting in unfavorable circulation to the ICP SAT variation (weak horizontal wind anomalies in Fig. 14f).
Difference of simulated DJF-mean 850-hPa streamfunction (colors, 106 m2 s−1) between (a) SEN1 and CTRL run, (2) SEN2 and CTRL run, (3) SEN3 and CTRL run, and (d) SEN3 and SEN2 run. The stippling indicates the significance of the discrepancy exceeding 90% of confidence level. (e)–(h) As in (a)–(d), but for 850-hPa horizontal wind (arrows, m s−1). Arrows exceeding 90% of confidence level are marked by red. The gray shading indicates southwesterly wind anomalies.
Citation: Journal of Climate 35, 3; 10.1175/JCLI-D-21-0477.1
Given IO SST anomalies obtained from the case of the strong ENSO–NWIO SST linkage (significant NWIO warming in the NWIO in Fig. 13c), great changes in the simulated circulation take place compared with the first two runs. The anticyclone gets intensified and further displaces eastward, implying the WNPAC anomaly is relatively stronger. Located at the northwestern flank of the anomalous WNPAC, the ICP is controlled by profound southwesterly wind anomalies (Figs. 14c,g).
Since the largest difference of SST anomalies between Figs. 13b and 13c is in the NWIO, results of SEN3 run subtracted from those of SEN2 run could directly yield effects of the NWIO warming. As shown in Figs. 14d and 14h, the NWIO warming could indeed induce the enhancement of the WNPAC anomaly and resultant southwesterly wind anomalies in the ICP, with the latter being the key to the ENSO–ICP SAT relationship. Therefore, numerical experiments also agree that atmospheric responses over the ICP are heavily sensitive to SST anomalies not only in the tropical Pacific but also in the NWIO, supporting the findings above from observational analyses.
6. Summary and discussion
Situated in the tropics, the ICP relies heavily on the local SAT which can exert great impacts on the safety of people’s life and national economic growth. The distinctive characteristics of the ICP SAT in winter compared with in other seasons—the lowest seasonal mean temperature, the largest standard deviation and warming trend as well as the highest spatial variance—entail researches specially focusing on the wintertime variability of the ICP SAT.
Among various predictive factors, ENSO plays a dominant role in regulating the winter ICP SAT with a high correlation reaching 0.48 (p < 0.01) between the two during 1958–2015. Thermodynamically, the ENSO-driven large-scale atmospheric overturning can alter the ICP cloud cover and the resulting surface heat flux to induce change of the local SAT. Dynamically, an anomalous anticyclone (cyclone) over the WNP emerges as a Rossby response to the ENSO-related SST anomalies during the mature phase of El Niño (La Niña). The prevailing southwesterly (northeasterly) wind anomalies in the northwestern flank of the anomalous anticyclone (cyclone) are conducive to abundant warm (cold) temperature advection, acting to heat (cool) the ICP. In consideration of a large southwest–northeast SAT gradient in winter, the horizontal temperature advection at lower-level primarily induced by the ENSO-related anomalous wind governs the variation of the ICP SAT.
This study reveals an unstable relationship between the wintertime ICP SAT and ENSO with high correlation during P1 (1957–75, 0.73 with p < 0.01) as well as P3 (1994–2014, 0.68 with p < 0.01) and low correlation during P2 (1976–93, −0.22 with p = 0.38). Whether ENSO is able to effectively influence the ICP SAT heavily depends on whether or not ENSO can induce significant temperature advection anomalies related to the anomalous horizontal wind. Take the El Niño event as an example, more significant southwesterly wind anomalies over the ICP are concurrent with a stronger WNPAC anomaly and vice versa.
The behavior of the anomalous WNPAC is supposed to be the combined effects of the tropical Indo-Pacific SST anomalies. While the contribution of the SST variability over the tropical Pacific to the running ENSO–ICP SAT correlation coefficient is negligible by using a decomposition method, there are great differences of SST anomalies over the NWIO and SCS in P1–P3: the NWIO and SCS SST anomalies are larger in P1 and P3 than in P2. Note that the SCS SST anomalies are the atmospheric responses associated with the incident solar radiation heating and the WES mechanism, thus weaker SCS SST anomalies in P2 primarily result from weaker atmospheric forcing. By contrast, the anomalous NWIO SST, also reflective of the ENSO–NWIO SST linkage, is proposed as the key factor for modulating the WNPAC anomaly and further altering the ICP SAT. Generally during the strong ENSO–NWIO SST linkage period, NWIO SST anomalies are larger and play an active role in enhancing convection in situ and driving atmospheric circulation anomalies, which results in more suppressed convection over the tropical western Pacific and hence the stronger WNPAC anomaly. But during the weak ENSO–NWIO SST linkage period, a weaker WNPAC anomaly response without the impact of NWIO warming is insufficient to maintain the southwesterly wind anomalies over the ICP. The effects of IO SST anomalies on the atmospheric circulation over the WNP are verified by both observational analyses and several CAM4 sensitive experiments in the current study. The AGCM simulation outputs also indicate that the anomalous SST strength over the IO, especially over the NWIO, is of great significance to induce anomalous atmospheric circulation over the ICP.
Precisely, we attempt to propose a following chain of processes involved in the interdecadal modulation of the ENSO–ICP SAT relationship: ENSO–NWIO SST linkage (NWIO SST anomalies in the presence of ENSO) → intensity of the WNPAC anomaly → surface temperature advection induced by anomalous horizontal wind over the ICP → wintertime ICP SAT, with the right-pointing arrows representing the causality. Fully understanding this chain of processes will greatly help with improving the prediction capacity of the ICP SAT in winter.
In the end, brief discussions about what determines changes in the ENSO–NWIO SST linkage, the beginning of the proposed chain, are presented. In fact, the non-stationarity of the ENSO–NWIO SST linkage in winter has been noticed before and its possible interpretations were preliminarily explored. Chowdary et al. (2012) has owned this interdecadal variation to the changes in atmospheric forcing (primarily the latent heat flux) which mainly controls the NWIO SST, while Ghasemi (2020) thought that the number of ENSO episode in a different epoch may likely determine the correlation coefficient between ENSO and NWIO SST. Additionally, the wintertime basinwide IO warming could also hinge on whether the IO dipole (IOD) pattern occurs in the ENSO developing autumn, which would induce processes of oceanic dynamics to change the western IO SST (Chowdary and Gnanaseelan 2007). It provides a perspective of understanding whether the preceding IO SST variation may expect to contribute the interdecadal oscillation of ENSO–IO SST linkage. Stemming from our analyses, we offer a new angle that the interdecadal correlation variation may be attributed to the interdecadal change in amplitude of the NWIO SST per se. As shown by the pink curve in Fig. 11a, the standard deviation of NWIO SST substantially decreases in P2 compared to P1 and P3, in parallel with the ENSO–NWIO SST correlation. We assume that when the NWIO SST variability is suppressed, covariance between ENSO and NWIO SST becomes minimized, resulting in weak NWIO SST anomalies in the presence of ENSO. Albeit various possible explanations mentioned above for the instable ENSO–NWIO SST linkage, deeper explorations are urgently required in the future.
Acknowledgments.
Our special thanks go to Dr. Yanke Tan, Dr. Ruifen Zhan, and Dr. Jiacan Yuan for inspiring discussions and valuable suggestions in the group seminar. The authors also appreciate the constructive advice from three anonymous reviewers, which greatly help with improving the manuscript. This research is jointly supported by National Natural Science Foundation of China (42030601, 41875087) and Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies (2020B1212060025).
Data availability statement.
Datasets analyzed during the current study were derived from the Hadley Centre Sea Ice and Sea Surface Temperature (HadISST) dataset, the Climatic Research Unit TS v4 (CRU TS v4) dataset and the twentieth Century Reanalysis v3 (20CRv3) dataset, which are all openly available at locations cited in the reference section. To be more specific, we downloaded these datasets from the following public domain resources: https://www.metoffice.gov.uk/hadobs/ and https://www.uea.ac.uk/groups-and-centres/climatic-research-unit/ and https://psl.noaa.gov/, respectively.
REFERENCES
Annamalai, H., H. Okajima, and M. Watanabe, 2007: Possible impact of the Indian Ocean SST on the Northern Hemisphere circulation during El Niño. J. Climate, 20, 3164–3189, https://doi.org/10.1175/JCLI4156.1.
Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97, 163–172, https://doi.org/10.1175/1520-0493(1969)097<0163:ATFTEP>2.3.CO;2.
Buckley, B. M., R. Fletcher, S.-Y. S. Wang, B. Zottoli, and C. Pottier, 2014: Monsoon extremes and society over the past millennium on mainland Southeast Asia. Quat. Sci. Rev., 95, 1–19, https://doi.org/10.1016/j.quascirev.2014.04.022.
Buntoung, S., S. Janjai, M. Nunez, S. Pattarapanitchai, P. Nimnuan, and J. Pariyothon, 2020: Spatial and temporal changes of precipitable water vapour in Thailand. Phys. Geogr., 41, 467–488, https://doi.org/10.1080/02723646.2019.1710433.
Caesar, J., and Coauthors, 2011: Changes in temperature and precipitation extremes over the Indo-Pacific region from 1971 to 2005. Int. J. Climatol., 31, 791–801, https://doi.org/10.1002/joc.2118.
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, S., and L. Song, 2018: The leading interannual variability modes of winter surface air temperature over Southeast Asia. Climate Dyn., 52, 4715–4734, https://doi.org/10.1007/s00382-018-4406-x.
Chen, S., and L. Song, 2019: Recent strengthened impact of the winter Arctic Oscillation on the Southeast Asian surface air temperature variation. Atmosphere, 10, 164, https://doi.org/10.3390/atmos10040164.
Chen, Z., Z. Wen, R. Wu, P. Zhao, and J. Cao, 2013: Influence of two types of El Niños 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.
Chen, Z., Z. Wen, R. Wu, X. Lin, and J. Wang, 2015: Relative importance of tropical SST anomalies in maintaining the Western North Pacific anomalous anticyclone during El Niño to La Niña transition years. Climate Dyn., 46, 1027–1041, https://doi.org/10.1007/s00382-015-2630-1.
Choi, G., and Coauthors, 2009: Changes in means and extreme events of temperature and precipitation in the Asia-Pacific Network region, 1955–2007. Int. J. Climatol., 29, 1906–1925, https://doi.org/10.1002/joc.1979.
Chowdary, J. S., and C. Gnanaseelan, 2007: Basin-wide warming of the Indian Ocean during El Niño and Indian Ocean dipole years. Int. J. Climatol., 27, 1421–1438, https://doi.org/10.1002/joc.1482.
Chowdary, J. S., S.-P. Xie, H. Tokinaga, Y. M. Okumura, H. Kubota, N. Johnson, and X.-T. Zheng, 2012: Interdecadal variations in ENSO teleconnection to the Indo–Western Pacific for 1870–2007. J. Climate, 25, 1722–1744, https://doi.org/10.1175/JCLI-D-11-00070.1.
Chowdary, J. S., N. John, and C. Gnanaseelan, 2014: Interannual variability of surface air-temperature over India: Impact of ENSO and Indian Ocean sea surface temperature. Int. J. Climatol., 34, 416–429, https://doi.org/10.1002/joc.3695.
Clark, J. P., and S. B. Feldstein, 2020: What drives the North Atlantic Oscillation’s temperature anomaly pattern? Part I: The growth and decay of the surface air temperature anomalies. J. Atmos. Sci., 77, 185–198, https://doi.org/10.1175/JAS-D-19-0027.1.
Feng, J., L. Wang, and W. Chen, 2014: How does the East Asian summer monsoon behave in the decaying phase of El Niño during different PDO phases? J. Climate, 27, 2682–2698, https://doi.org/10.1175/JCLI-D-13-00015.1.
Gao, C., and Coauthors, 2019: Land–atmosphere interaction over the Indo-China Peninsula during spring and its effect on the following summer climate over the Yangtze River basin. Climate Dyn., 53, 6181–6198, https://doi.org/10.1007/s00382-019-04922-x.
Gasparrini, A., and Coauthors, 2015: Mortality risk attributable to high and low ambient temperature: A multicountry observational study. Lancet, 386, 369–375, https://doi.org/10.1016/S0140-6736(14)62114-0.
Ge, F., and Coauthors, 2018: Assessment of trends and variability in surface air temperature on multiple high-resolution datasets over the Indochina Peninsula. Theor. Appl. Climatol., 135, 1609–1627, https://doi.org/10.1007/s00704-018-2457-x.
Geng, X., W. Zhang, M. F. Stuecker, P. Liu, F.-F. Jin, and G. Tan, 2016: Decadal modulation of the ENSO–East Asian winter monsoon relationship by the Atlantic Multidecadal Oscillation. Climate Dyn., 49, 2531–2544, https://doi.org/10.1007/s00382-016-3465-0.
Geng, X., W. Zhang, F.-F. Jin, and M. F. Stuecker, 2018: A new method for interpreting nonstationary running correlations and its application to the ENSO–EAWM relationship. Geophys. Res. Lett., 45, 327–334, https://doi.org/10.1002/2017GL076564.
Ghasemi, A. R., 2020: Influence of northwest Indian Ocean sea surface temperature and El Niño–Southern Oscillation on the winter precipitation in Iran. J. Water Climate Change, 11, 1481–1494, https://doi.org/10.2166/wcc.2019.274.
Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447–462, https://doi.org/10.1002/qj.49710644905.
Gong, H., L. Wang, and W. Chen, 2019: Recently strengthened influence of ENSO on the wintertime East Asian surface air temperature. Atmosphere, 10, 720, https://doi.org/10.3390/atmos10110720.
Halpert, M. S., and C. F. Ropelewski, 1992: Surface temperature patterns associated with the Southern Oscillation. J. Climate, 5, 577–593, https://doi.org/10.1175/1520-0442(1992)005<0577:STPAWT>2.0.CO;2.
Hao, X., and S. He, 2017: Combined effect of ENSO-like and Atlantic multidecadal oscillation SSTAs on the interannual variability of the East Asian winter monsoon. J. Climate, 30, 2697–2716, https://doi.org/10.1175/JCLI-D-16-0118.1.
Harris, I., T. J. Osborn, P. Jones, and D. Lister, 2020: Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Sci. Data, 7, 109, https://doi.org/10.1038/s41597-020-0453-3.
He, C., T. Zhou, and T. Li, 2019: Weakened anomalous western North Pacific anticyclone during an El Niño–decaying summer under a warmer climate: Dominant role of the weakened impact of the tropical Indian Ocean on the atmosphere. J. Climate, 32, 213–230, https://doi.org/10.1175/JCLI-D-18-0033.1.
He, S., and H. Wang, 2013: Oscillating relationship between the East Asian winter monsoon and ENSO. J. Climate, 26, 9819–9838, https://doi.org/10.1175/JCLI-D-13-00174.1.
He, S., H. Wang, and J. Liu, 2013: Changes in the relationship between ENSO and Asia–Pacific midlatitude winter atmospheric circulation. J. Climate, 26, 3377–3393, https://doi.org/10.1175/JCLI-D-12-00355.1.
Hu, Z.-Z., A. Kumar, B. Huang, J. Zhu, M. L’Heureux, M. J. McPhaden, and J.-Y. Yu, 2020: The interdecadal shift of ENSO properties in 1999/2000: A review. J. Climate, 33, 4441–4462, https://doi.org/10.1175/JCLI-D-19-0316.1.
Jiang, W., G. Huang, P. Huang, R. Wu, K. Hu, and W. Chen, 2019: Northwest Pacific anticyclonic anomalies during post–El Niño summers determined by the pace of El Niño decay. J. Climate, 32, 3487–3503, https://doi.org/10.1175/JCLI-D-18-0793.1.
Kiladis, G. N., and H. F. Diaz, 1989: Global climatic anomalies associated with extremes in the Southern Oscillation. J. Climate, 2, 1069–1090, https://doi.org/10.1175/1520-0442(1989)002<1069:GCAAWE>2.0.CO;2.
Kim, J.-W., and S. I. An, 2019: Western North Pacific anticyclone change associated with the El Niño–Indian Ocean Dipole coupling. Int. J. Climatol., 39, 2505–2521, https://doi.org/10.1002/joc.5967.
Kim, J.-W., S.-W. Yeh, and E.-C. Chang, 2013: Combined effect of El Niño–Southern Oscillation and Pacific decadal oscillation on the East Asian winter monsoon. Climate Dyn., 42, 957–971, https://doi.org/10.1007/s00382-013-1730-z.
Kim, J.-W., S.-I. An, S.-Y. Jun, H.-J. Park, and S.-W. Yeh, 2016: ENSO and East Asian winter monsoon relationship modulation associated with the anomalous northwest Pacific anticyclone. Climate Dyn., 49, 1157–1179, https://doi.org/10.1007/s00382-016-3371-5.
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.
Lamchin, M., W. K. Lee, S. W. Jeon, S. W. Wang, C. H. Lim, C. Song, and M. Sung, 2018: Long-term trend and correlation between vegetation greenness and climate variables in Asia based on satellite data. Sci. Total Environ., 618, 1089–1095, https://doi.org/10.1016/j.scitotenv.2017.09.145.
Lee, S., T. Gong, N. Johnson, S. B. Feldstein, and D. Pollard, 2011: On the possible link between tropical convection and the Northern Hemisphere Arctic surface air temperature change between 1958 and 2001. J. Climate, 24, 4350–4367, https://doi.org/10.1175/2011JCLI4003.1.
Li, T., B. Wang, B. Wu, T. Zhou, C.-P. Chang, and R. Zhang, 2018: Theories on formation of an anomalous anticyclone in western North Pacific during El Niño: A review. J. Meteor. Res., 31, 987–1006, https://doi.org/10.1007/s13351-017-7147-6.
Li, X. X., 2020: Heat wave trends in Southeast Asia during 1979–2018: The impact of humidity. Sci. Total Environ., 721, 137664, https://doi.org/10.1016/j.scitotenv.2020.137664.
Lin, L., C. Chen, and M. Luo, 2018: Impacts of El Niño–Southern Oscillation on heat waves in the Indochina peninsula. Atmos. Sci. Lett., 19, e856, https://doi.org/10.1002/asl.856.
Luo, M., and N.-C. Lau, 2017: Synoptic characteristics, atmospheric controls, and long-term changes of heat waves over the Indochina Peninsula. Climate Dyn., 51, 2707–2723, https://doi.org/10.1007/s00382-017-4038-6.
Luong, L. M. T., D. Phung, P. D. Sly, T. N. Dang, L. Morawska, and P. K. Thai, 2019: Effects of temperature on hospitalisation among pre-school children in Hanoi, Vietnam. Environ. Sci. Pollut. Res. Int., 26, 2603–2612, https://doi.org/10.1007/s11356-018-3737-9.
Manton, M. J., and Coauthors, 2001: Trends in extreme daily rainfall and temperature in Southeast Asia and the South Pacific: 1961–1998. Int. J. Climatol., 21, 269–284, https://doi.org/10.1002/joc.610.
Marjuki, G. van der Schrier, A. M. G. Klein Tank, E. J. M. van den Besselaar, Nurhayati, and Y. S. Swarinoto, 2016: Observed trends and variability in climate indices relevant for crop yields in Southeast Asia. J. Climate, 29, 2651–2669, https://doi.org/10.1175/JCLI-D-14-00574.1.
Matsuno, T., 1966: Quasi-geostropic motions in the equatorial area. J. Meteor. Soc. Japan, 44, 25–43, https://doi.org/10.2151/jmsj1965.44.1_25.
Mie Sein, Z. M., I. Ullah, S. Syed, X. Zhi, K. Azam, and G. Rasool, 2021: Interannual variability of air temperature over Myanmar: The influence of ENSO and IOD. Climate, 9, 35, https://doi.org/10.3390/cli9020035.
Neale, R. B., J. Richter, S. Park, P. H. Lauritzen, S. J. Vavrus, P. J. Rasch, and M. Zhang, 2013: The mean climate of the Community Atmosphere Model (CAM4) in forced SST and fully coupled experiments. J. Climate, 26, 5150–5168, https://doi.org/10.1175/JCLI-D-12-00236.1.
Nguyen, D.-Q., J. Renwick, and J. McGregor, 2014: Variations of surface temperature and rainfall in Vietnam from 1971 to 2010. Int. J. Climatol., 34, 249–264, https://doi.org/10.1002/joc.3684.
Nicholls, N., 2005: The El Niño–Southern Oscillation and daily temperature extremes in east Asia and the west Pacific. Geophys. Res. Lett., 32, L16714, https://doi.org/10.1029/2005GL022621.
Peng, S., and Coauthors, 2004: Rice yields decline with higher night temperature from global warming. Proc. Natl. Acad. Sci. USA, 101, 9971–9975, https://doi.org/10.1073/pnas.0403720101.
Phosri, A., T. Sihabut, and C. Jaikanlaya, 2020: Short-term effects of diurnal temperature range on hospital admission in Bangkok, Thailand. Sci. Total Environ., 717, 137202, https://doi.org/10.1016/j.scitotenv.2020.137202.
Rayner, N. A., 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.
Sengupta, A., and S. Nigam, 2019: The northeast winter monsoon over the Indian subcontinent and Southeast Asia: Evolution, interannual variability, and model simulations. J. Climate, 32, 231–249, https://doi.org/10.1175/JCLI-D-18-0034.1.
Shi, J., and W. Qian, 2018: Asymmetry of two types of ENSO in the transition between the East Asian winter monsoon and the ensuing summer monsoon. Climate Dyn., 51, 3907–3926, https://doi.org/10.1007/s00382-018-4119-1.
Slivinski, L. C., and Coauthors, 2019: Towards a more reliable historical reanalysis: Improvements for version 3 of the Twentieth Century Reanalysis system. Quart. J. Roy. Meteor. Soc., 145, 2876–2908, https://doi.org/10.1002/qj.3598.
Sooktawee, S., U. Humphries, A. Limsakul, and P. Wongwises, 2014: Spatio-temporal variability of winter monsoon over the Indochina Peninsula. Atmosphere, 5, 101–121, https://doi.org/10.3390/atmos5010101.
Thirumalai, K., P. N. DiNezio, Y. Okumura, and C. Deser, 2017: Extreme temperatures in Southeast Asia caused by El Niño and worsened by global warming. Nat. Commun., 8, 15 531, https://doi.org/10.1038/ncomms15531.
Trenberth, K. E., 2002: Evolution of El Niño–Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107, 4065, https://doi.org/10.1029/2000JD000298.
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.
Wang, B., R. Wu, and T. Li, 2003: Atmosphere–warm ocean interaction and its impacts on Asian–Australian monsoon variation. J. Climate, 16, 1195–1211, https://doi.org/10.1175/1520-0442(2003)16<1195:AOIAII>2.0.CO;2.
Wang, C., W. Wang, D. Wang, and Q. Wang, 2006: Interannual variability of the South China Sea associated with El Niño. J. Geophys. Res., 111, C03023, https://doi.org/10.1029/2005JC003333.
Wang, H., and S. He, 2012: Weakening relationship between East Asian winter monsoon and ENSO after mid-1970s. Chin. Sci. Bull., 57, 3535–3540, https://doi.org/10.1007/s11434-012-5285-x.
Wang, L., W. Chen, and R. Huang, 2008: Interdecadal modulation of PDO on the impact of ENSO on the East Asian winter monsoon. Geophys. Res. Lett., 35, L20702, https://doi.org/10.1029/2008GL035287.
Watanabe, M., and F.-f. Jin, 2002: Role of Indian Ocean warming in the development of Philippine Sea anticyclone during ENSO. Geophys. Res. Lett., 29, 1478, https://doi.org/10.1029/2001GL014318.
Wu, B., T. Zhou, and T. Li, 2009: Seasonally evolving dominant interannual variability modes of East Asian climate. J. Climate, 22, 2992–3005, https://doi.org/10.1175/2008JCLI2710.1.
Wu, B., T. Li, and T. Zhou, 2010a: Asymmetry of atmospheric circulation anomalies over the western North Pacific between El Niño and La Niña. J. Climate, 23, 4807–4822, https://doi.org/10.1175/2010JCLI3222.1.
Wu, B., T. Li, and T. Zhou, 2010b: Relative contributions of the Indian Ocean and local SST anomalies to the maintenance of the western North Pacific anomalous anticyclone during the El Niño decaying summer. J. Climate, 23, 2974–2986, https://doi.org/10.1175/2010JCLI3300.1.
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. Kosaka, Y. Du, K. Hu, J. S. Chowdary, and G. Huang, 2016: Indo-western Pacific Ocean capacitor and coherent climate anomalies in post-ENSO summer: A review. Adv. Atmos. Sci., 33, 411–432, https://doi.org/10.1007/s00376-015-5192-6.
Yang, J., Q. Liu, S.-P. Xie, Z. Liu, and L. Wu, 2007: Impact of the Indian Ocean SST basin mode on the Asian summer monsoon. Geophys. Res. Lett., 34, L02708, https://doi.org/10.1029/2006GL028571.
Yuan, Y., S. Yang, and Z. Zhang, 2012: Different evolutions of the Philippine Sea anticyclone between the eastern and central Pacific El Niño: Possible effects of Indian Ocean SST. J. Climate, 25, 7867–7883, https://doi.org/10.1175/JCLI-D-12-00004.1.
Zhang, R., A. Sumi, and M. Kimoto, 1996: Impact of El Niño on the East Asian monsoon: A diagnostic study of the ’86/87 and ’91/92 events. J. Meteor. Soc. Japan, 74, 49–62, https://doi.org/10.2151/jmsj1965.74.1_49.
Zhang, R., T. Li, M. Wen, and L. Liu, 2014: Role of intraseasonal oscillation in asymmetric impacts of El Niño and La Niña on the rainfall over southern China in boreal winter. Climate Dyn., 45, 559–567, https://doi.org/10.1007/s00382-014-2207-4.
Zhu, S., F. Ge, Y. Fan, L. Zhang, F. Sielmann, K. Fraedrich, and X. Zhi, 2020a: Conspicuous temperature extremes over Southeast Asia: Seasonal variations under 1.5°C and 2°C global warming. Climatic Change, 160, 343–360, https://doi.org/10.1007/s10584-019-02640-1.