1. Introduction
The onset of the South China Sea summer monsoon (SCSSM) starts the abrupt transition from the dry season to the rainy season and the subsequent seasonal evolution over East Asia (Qian and Lee 2000; Ding and Chan 2005). Along with its slow-varying seasonal evolution, the monsoon onset process exhibits fast fluctuations on a subseasonal time scale (LinHo and Wang 2002). The onset of the SCSSM features the eastward retreat of the western North Pacific subtropical high (WNPSH), the generation of the monsoon trough and the cross-equatorial flow over the South China Sea (SCS) in the lower troposphere, and the eastward extension of the South Asian high (SAH) in the upper troposphere (Ding and Liu 2001; Wang et al. 2004; Wen et al. 2005; Liu and Zhu 2016). The onset time of the SCSSM could therefore indicate the evolution of summer monsoon rainfall over East Asia in the warm season (He and Zhu 2015; Jiang et al. 2018).
Many previous studies have reported the long-term variability of the onset time of the SCSSM. On an interdecadal time scale, the SCSSM onset dates have advanced since 1994 (Kajikawa et al. 2012). This phenomenon has been ascribed to the decadal changes in sea surface temperature (SST) in the equatorial western Pacific (Kajikawa and Wang 2012), the La Niña–like SST in the tropical Pacific (Xiang and Wang 2013), or the zonal wind over Kalimantan Island (Lin and Zhang 2020). The interannual variability of the SCSSM onset date depends on El Niño–Southern Oscillation (ENSO) and its resultant SST anomalies in the tropical Indo-Pacific warm pool, which provide the most crucial seasonal predictability (Wu and Wang 2000; Mao and Wu 2007; Zhou and Chan 2007; Yuan et al. 2008; Liu et al. 2016; He et al. 2017; Martin et al. 2019). The SCSSM tends to onset earlier after a La Niña but later following an El Niño event in the preceding winter. Another seasonal predictability of SCSSM onset time stems from air–land interactions over the Eurasian continent, including spring sensible heat over the Tibetan Plateau (TP; Li and Yanai 1996; Qian et al. 2001), land surface temperature in the mid–high latitudes (Liu et al. 2009), and surface thermal condition and soil moisture content over the Indochina Peninsula (Zhang and Qian 2002; Li et al. 2020). Based on these precursors of the underlying conditions, skillful physical-based empirical models have been built to predict the onset dates of the SCSSM (Zhu and Li 2017).
In addition, the subseasonal variability of atmospheric circulation and convection can trigger the SCSSM onset in the winter-to-summer transition episode (Wu and Zhang 1998; Straub et al. 2006). In the tropics, the Madden–Julian oscillation (MJO) and the northwestward-propagating 10–20-day intraseasonal oscillation (ISO) arrive at the SCS together to weaken the WNPSH (Mao and Chan 2005; Zhou and Chan 2005). In the extratropics, an equatorward-propagating 12–24-day ISO, which starts over Northeast China, could initiate the monsoon convection over the SCS via the southward intrusion of cold air in the lower troposphere (Chang and Chen 1995; Ding and Liu 2001; Tong et al. 2008). The SCSSM then builds up during the dry-to-wet phase transition of the ISO over the SCS (Wang and Xu 1997; Shao et al. 2015). Although most previous studies have examined the ISO of the SCSSM onset based on convection or 850-hPa zonal wind (U850) over the SCS, little attention has been paid to the ISO in the mid–upper troposphere. Following the evolution of meridional temperature gradient (MTG; ∂T/∂y) in the mid–upper troposphere, the SAH alters and leads the formation of cross-equatorial flow and 850-hPa westerly wind over the SCS, acting as an onset precursor of the SCSSM (Liang et al. 2013; Liu and Zhu 2016; Jiang et al. 2016). It implicates that the subseasonal variation in the mid–upper troposphere may provide additional predictability to the SCSSM onset.
The year-by-year behavior of the ISO over the SCS is distinct (Wang et al. 2018; Li et al. 2019). For instance, the onset of the SCSSM in 2018 was extremely delayed by the activity of the MJO and the 10–20-day ISO in the tropics, under the background of persisting weaker SAH and midlatitude Mongolian cyclone (Liu and Zhu 2019; Lu et al. 2020). In 2019, apart from the influence of the tropical 30–60-day ISO, the SCSSM onset was greatly advanced by a tropical cyclone over the Bay of Bengal and its upscaling effect on the diabatic heating of the TP, which resulted in the enhancement and northward shift of the SAH (Hu et al. 2020; Liu and Zhu 2020). The relationship between ENSO and onset time of the SCSSM was thus broken in these two years, increasing uncertainty in the seasonal prediction of the SCSSM onset. These cases confuse us whether the prediction of the SCSSM onset is a seasonal or subseasonal issue. Therefore, the present study aims to investigate the seasonal-to-subseasonal predictability of the SCSSM onset from a perspective of MTG in the mid–upper troposphere, in order to understand the relative role of ENSO and subseasonal variability in the onset of the SCSSM. The remainder is organized as follows. Section 2 describes the data and methods, especially regarding the SCSSM onset dates defined by MTG and the decomposition of the seasonal cycle and subseasonal process. The relative contribution of the seasonal cycle and subseasonal component to the onset of the SCSSM is shown in section 3. Section 4 demonstrates the slow-varying process of the monsoon circulation over the SCS and its association with ENSO. Section 5 elaborates on the features and maintain mechanism of the mid–upper tropospheric ISO, while the influences of this ISO on the SCSSM onset is presented in section 6. The conclusions and discussion are summarized in section 7.
2. Data and methods
a. Datasets
We used the daily mean JRA-55 reanalysis from 1980 to 2019, released by the Japanese Meteorological Agenda, with a horizontal resolution of 1.25° × 1.25° and 27 standard isobaric surfaces from 1000 to 100 hPa (Kobayashi et al. 2015; Harada et al. 2016), to describe the atmospheric circulation and thermal structure during the onset process of the SCSSM. The land rainfall records were verified by the CPC Global Unified Gauge-Based Analysis of Daily Precipitation dataset with a horizontal resolution of 0.5° × 0.5° (Xie et al. 2007; Chen et al. 2008). The monthly mean SST involving a resolution of 2° × 2° was obtained from the ERSST V5 dataset released by National Oceanic and Atmospheric Administration (NOAA; Huang et al. 2017). The monthly mean outgoing longwave radiation (OLR) provided by NOAA, with a horizontal resolution of 2.5° × 2.5°, was used to depict tropical convection (Liebmann and Smith 1996).
b. Definition of SCSSM onset dates
In each year, the onset dates of the SCSSM were identified by the negative-to-positive transition of the MTG averaged between 500 and 200 hPa, which directly reflected the thermal structure and indirectly represented the vertical shear of zonal wind over the SCS (Webster et al. 1998; Mao et al. 2004; Liu et al. 2016). We considered that the SCSSM started each year on the date satisfying the following criteria: 1) the daily 500–200 hPa averaged MTG over the SCS (10°–20°N, 110°–120°E) changes from negative to positive after 20 April and stays positive for the next five days; and 2) the MTG over the SCS remains positive in at least 15 of the subsequent 20 days (including the onset date). During the period of 1980–2019, the onset dates are consistent with the onset pentads of the SCSSM defined by Liu et al. (2015), showing a significant temporal correlation coefficient (TCC) of +0.91. The climatological MTG-based onset date of the SCSSM during 1980–2019 was 13 May with a standard deviation (STD) of 8.6 days. We termed the onset date of the SCSSM defined by the MTG criterion “MTGOD” for convenience herein.
c. Decomposition of seasonal cycle and subseasonal component
The daily MTG index over the SCS in every year was decomposed into the slow-varying seasonal cycle and the fast-fluctuating subseasonal part using Fourier spectral analysis. For each variable, the seasonal cycle part was the sum of the annual mean and the first three Fourier harmonics, while the subseasonal part was the sum of the residual 5th–36th harmonics, including an intraseasonal period from 10 to 70 days. When the seasonal cycle component of MTG switched from negative to positive and maintained positive for more than 15 days in April–June (AMJ), the switch date was defined as the seasonal cycle transition time of the MTG (MTGSCT) over the SCS. The climate-mean MTGSCT was 14 May from 1980 to 2019, along with a STD of 6.3 days.
Large difference exists between MTGOD and MTGSCT in some years. For instance, although the MTGSCT was on 15 May in 2018, the MTGOD was extremely delayed to 2 June in 2018 (Figs. 1a,c). In contrast, the MTGOD on 30 April advanced the MTGSCT on 10 May in 2019 (Figs. 1b,c). The climate-mean absolute difference between MTGOD and MTGSCT is 6.1 days with a STD of 4.1 days.
Time series of daily MTG and its components (10−7 K m−1) averaged over the SCS (10°–20°N, 110°–120°E). The case in (a) 2018 and (b) 2019 (black, blue, and red lines indicate the original MTG, its seasonal cycle, and the sum of seasonal cycle and 10–30-day component, respectively). (c) Time series of MTGOD (bars) and MTGSCT (black line) from 1980 to 2019 (Julian days). Red and blue bars in (c) indicate an El Niño and La Niña event in the previous winter, respectively.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
d. Statistical technology
3. Statistical relationship among subseasonal MTG, MTGOD, and MTGSCT
The subseasonal component of the MTG index exhibits a significant period of 10–30 days in spectral analysis during 1980–2019 (Fig. 2a). The climate-mean value of power spectral between 10 and 30 days is 15.00. In the years when the absolute difference between MTGOD and MTGSCT is larger than 5 days exceeding one STD, the 10–30-day period is statistically significant in the subseasonal component of the MTG index, showing the mean power spectrum of 16.59 (Fig. 2b). When the absolute difference is smaller than 5 days, the averaged power spectrum reduces to 13.42, even though the period remains statistically significant (Fig. 2c). The positive correlation between the 10–30-day power spectrum of MTG index and the absolute differences between MTGOD and MTGSCT is statistically significant during 1980–2019, regardless of the ENSO status in May (Fig. 3a). Thus, more active 10–30-day MTG ISO accompanies larger deviation of MTGOD from MTGSCT.
(a) Power spectrum of the subseasonal component of the April–May–June MTG index over the SCS averaged from 1980 to 2019. (b),(c) As in (a), but for the years with the absolute difference between MTGSCT and MTGOD greater than or equal to and less than 5 days, respectively. Numbers at the top-right corner are for the sample sizes in each category. Red, blue, and green dashed lines are for the Markov red noise spectrum, a priori 95% confidence bound, and a posterior 95% confidence level, respectively. Bold dashed lines indicate the mean value of the power spectrum between 10 and 30 days in each category.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
(a) Scatterplot between the standardized 10–30-day power spectrum of subseasonal component of the AMJ MTG index over the SCS and the absolute difference between MTGOD and MTGSCT. (b) Scatterplot between the standardized MTGOD and MTGSCT from 1980 to 2019. (c) As in (b), but the ordinate axis is about the onset dates defined by the sum of seasonal cycle and 10–30-day component of the MTG index. (d) As in (b), but the ordinate axis is about the onset dates defined by the sum of seasonal cycle and 10–70-day component of the MTG index. Black, red, and blue lines indicate the linear regression line in all years, the cases with May Niño-3.4 index > 0 and <0, respectively. The years with May Niño-3.4 index > +0.5 and <−0.5 are marked in red and blue dots, respectively. Numbers in the parentheses denote the sample sizes in each category. Values with superscript symbols *, **, and *** indicate the temporal correlation coefficients exceeding the 90%, 95%, and 99% confidence level, respectively.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
The MTGOD is significantly correlated with the MTGSCT from 1980 to 2019, presenting a remarkable TCC of +0.492, which exceeds the 99% confidence level (Fig. 3b). This relationship remains pronounced in the environment of either warmer or colder central-eastern equatorial Pacific in May. The 13th–36th and 5th–12th harmonics indicate the 10–30- and 31–70-day ISO component of the MTG, respectively. Based on the sum of seasonal cycle and 10–30-day component of the MTG index over the SCS (see red lines in Figs. 1a,b), we redefined the onset dates using the same criterion of MTGOD to examine the role of the 10–30-day MTG ISO. During the period of 1980–2019, the reconstructed onset date shows a climate-mean value of 14 May and a STD of 7.8 days. The TCC between MTGOD and reconstructed onset dates reaches +0.777, which is much higher than the TCC between MTGOD and MTGSCT (Fig. 3c). When excluding the MTGSCT, the partial TCC between MTGOD and reconstructed onset dates remains +0.371 exceeding the 95% confidence level. Furthermore, the positive TCC between MTGOD and reconstructed dates is significant under the different ENSO background. It suggests a separate effect of the 10–30-day MTG ISO on the onset of the SCSSM from the ENSO influences. In addition, we reconstructed the onset dates based on 10–30-day and 31–70-day ISO and seasonal cycle of MTG index over the SCS. This reconstructed onset dates are still significantly positively correlated with the MTGOD. Without the lower frequency ISO, the TCC decreases to +0.714 in general but increases to +0.663 (+0.843) in the El Niño (La Niña) years comparing with the previous results (Fig. 3d). This confirms the dominant role of the 10–30-day MTG ISO in the subseasonal variation of the SCSSM onset, consistent with dominant periods in power spectrum analysis. In particular, the 31–70-day ISO, whose period is not statistically significant in general, may develop and affect the onset process in the ENSO years, in conjunction with the 10–30-day ISO.
4. Interannual variability of the MTGSCT and its links with ENSO
The significant relationship between MTGOD and MTGSCT is stable during the last 40 years, as indicated by a 21-yr running correlation test (green line in Fig. 4). The MTGSCT tends to advance after a La Niña winter but delay following an El Niño winter (black line in Fig. 1c). This relationship is stable in the 21-yr running correlation test (blue line in Fig. 4). However, an unstable relationship exists between MTGOD and DJF-mean Niño-3.4 index (black line in Fig. 4). In particular, the 21-yr running TCC stably has exceeded the 99% confidence level since 2004, confirming the robust association between MTGOD and winter ENSO events after 1994 (Liu et al. 2016). However, the significant linkage between MTGOD and winter ENSO events would drop dramatically if the robust MTGSCT–ENSO relationship is removed in a partial correlation analysis (red line in Fig. 4).
21-yr running TCC of MTGOD and MTGSCT with the DJF-mean Niño-3.4 index without linear trend during 1980–2019. Green line: TCC between MTGSCT and MTGOD; blue line: TCC between MTGSCT and Niño-3.4 index; red line: partial TCC between MTGOD and Niño-3.4 index excluding MTGSCT; black line: TCC between MTGOD and Niño-3.4 index.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
The ENSO-related SST anomalies affect the MTGSCT via air–sea interaction in the tropics. After an El Niño event, warm SST anomalies emerge in the equatorial eastern Pacific (EEP) and the tropical Indian Ocean (TIO) from April to May, along with cooling in the western North Pacific (Figs. 5a,b). The tropical convection is then deepened over the tropical central Pacific but suppressed around the Indo-Pacific warm pool and the Maritime Continent (Figs. 5c,d), as a result of Walker circulation and WNPSH anomalies (Webster and Yang 1992; Ju and Slingo 1995; Wang et al. 2000). The anomalous air temperature in the mid–upper troposphere acts as a Matsuno–Gill response to the tropical SST anomalies (Matsuno 1966; Gill 1980). A horseshoe-like warming pattern is induced over the EEP, in contrast to a zonally elongating warming center over the TIO (Figs. 5e,f). Meanwhile, the extratropical anomalies of convection and thermal field are much weaker near the SCS. Because of the warming in the southern SCS, the in situ air temperature rises to prevent the negative-to-positive switch of the seasonal cycle component of MTG over the SCS, leading to later MTGSCT after El Niño event. Finally, the ENSO events affect the MTGOD by modulating the seasonal cycle of the MTG over the SCS.
Horizontal distribution of (a),(b) SST anomalies (K), (c),(d) OLR (W m−2), and (e),(f) average air temperature between 500 and 200 hPa (K) in (left) April and (right) May regressed against the MTGSCT index from 1980 to 2019. The linear trends have been removed, and the values exceeding the 95% confidence level are stippled. Bold black contours represent the 1500-m topography.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
5. The 10–30-day MTG ISO and its maintenance mechanism
A lead–lag autocorrelation of the 10–30-day MTG ISO index shows a life cycle with minimum TCC on day −7 and day +7 but maximum one on day 0 (Fig. 6). The sum of the 13th–36th harmonics of each variable is retained as the 10–30-day ISO component. We further divide the ISO lifespan into eight phases based on its autocorrelation coefficients and investigate the evolution of this ISO and its influences using phase-dependent composite analysis. Each phase is spaced by π/4 with an interval of about 3 days. Since the composite results based on the years with large difference between MTGOD and MTGSCT is similar (figure not shown), here we use all the 40-yr cases from 1981 to 2019 to obtain the general features of this 10–30-day MTG ISO.
Autocorrelation coefficient of the 10–30-day ISO component of the MTG index over the SCS in AMJ during 1980–2019. Red dashed lines indicate the 99% confidence level. Blue numbers denote the eight phases of the ISO.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
a. Evolution of the 10–30-day MTG ISO
The 10–30-day MTG ISO is closely associated with the ISO in the extratropics. In phase 1 when the ISO index is in the trough, the 500–200-hPa averaged temperature is colder to the north of the SCS, along with an evident local cyclonic anomaly in the upper troposphere (Fig. 7a). Based on the thermal wind relationship, this colder-in-north pattern decelerates the upper-level easterly and the low-level westerly over the SCS to enhance the WNPSH at 850 hPa and suppress the monsoon convection over the western North Pacific (Fig. 8a). In the meantime, a weak warm anomaly with an insignificant anticyclonic anomaly at 200 hPa emerges over the western TP (Fig. 7a). The negative phase of the MTG ISO starts to weaken in phases 2 and 3 (Fig. 6). The warmer air and anomalous anticyclone shift eastward and gets enhanced in phase 2, then becomes significant in phase 3, accompanied by the eastward retreat and damping of the cooling center and cyclonic anomaly over East Asia (Figs. 7b,c). In the lower troposphere, the anomalous WNPSH and suppressed convection over the SCS becomes attenuated, but the southwesterly wind and spring rainfall tends to strengthen over South China (Figs. 8b,c).
Phase-dependent composite of 10–30-day ISO of 200-hPa winds (vectors; m s−1; vectors exceeding the 95% confidence level are in black) and 500–200-hPa averaged air temperature (shading; K; values exceeding the 95% confidence level are stippled) referring to the eight phases of the 10–30-day ISO of the MTG index in AMJ from 1980 to 2019. (a)–(h) Phases 1–8, respectively. Bold purple contours represent the 1500-m topography.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
Phase-dependent composite of 10–30-day ISO of winds at 850 hPa (vectors; m s−1; vectors exceeding the 95% confidence level are in black) and OLR (shading; W m−2; values exceeding the 95% confidence level are stippled) referring to the eight phases of the 10–30-day ISO of the MTG index in AMJ from 1980 to 2019. (a)–(h) Phases 1–8, respectively. Bold purple contours represent the 1500-m topography.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
In phase 4 when the MTG ISO enters its positive phase, the remarkable anomaly of warm air and anticyclone in the mid–upper troposphere arrives at South China, along with development of the 850-hPa westerly wind and the monsoon convection over the SCS (Figs. 7d and 8d). Afterward, the MTG ISO peaks in phase 5. The maximum upper-tropospheric warming and anticyclonic anomaly take place to the north of the SCS (Fig. 7e). The pronounced anomaly of upper-level easterly and low-level westerly wind constitute a vertical shear of easterly wind, which facilitates monsoon onset convection over the SCS (Fig. 8e). In this phase, a cooling center in the mid–upper troposphere generates over the western TP, opposing to the situation in phase 1 (Figs. 7a,e). The positive phase of the MTG ISO subsequently weakens in phase 6, reaches the neutral phase in phase 7, and enters its negative phase in phase 8 (Fig. 6). During phases 6–8, the anomalous thermal field, circulation, and convection associated with the MTG ISO reverse their property in contrast to the situation from phase 2 to phase 4 (Figs. 7f–h and 8f–h).
This 10–30-day ISO in AMJ resembles the quasi-biweekly oscillation (QBWO) above the TP in the mei-yu season that termed as “AT pattern” by Fujinami and Yasunari (2009). But it is located more southward and shows much weaker signals upstream the TP. This is because the upper-tropospheric westerly basic flow is zonally distributed over East Asia in AMJ as that associated with the AT pattern of QBWO in early summer, along with the jet axis in AMJ settling to south of the TP. In this situation, the local influences of the TP thermal forcing in AMJ are more crucial for the 10–30-day ISO than the remote effects of the upstream wave train in the mid–high latitude over Eurasian continent. In addition, the 10–30-day ISO in AMJ is distinct from the northward propagation of QBWO over the TP in boreal summer (Wang and Duan 2015). The reason is that the vertical easterly wind shear, which guarantees the northward propagation of QBWO from the tropics, cannot be established until the full onset of the Asian summer monsoon after early June.
b. Maintenance mechanism
As presented by Fig. 9a, in the negative-to-positive spell of the MTG ISO, the extratropical warmer air first emerges over the western TP in phase 1. It then propagates eastward from phase 2 to phase 6. The eastward propagation of this ISO follows the positive zonal temperature advection (
(a) 20°–40°N averaged phase–longitude and (b) 110°–120°E averaged phase–latitude cross-section of the 10–30-day ISO of the 500–200-hPa mean air temperature (shading; K; values exceeding the 95% confidence level are stippled) referring to the eight phases of the 10–30-day ISO of the MTG index in AMJ from 1980 to 2019. Contours in (a) are for the horizontal advection of the 10–30-day ISO of the air temperature due to seasonal cycle component of zonal winds and 10–30-day ISO of air temperature (interval of 0.1 K day−1; dark green solid lines are for warmer but dark red dashed lines denote the colder advection, respectively). Contours in (b) denote the 10–30-day ISO of the diabatic heating (interval of 0.1 K day−1; dark green solid lines are for diabatic heating, but dark red dashed lines denote the diabatic cooling, respectively).
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
Since the extratropical ISO presents significant anomalies of air temperature and circulation, we further use the potential vorticity [
Phase-dependent composite of the pressure–longitude cross-section of the 10–30-day ISO of zonal circulation (vectors; m s−1; vectors exceeding the 95% confidence level are in black), diabatic heating (shading; K day−1; values exceeding the 95% confidence level are stippled), and potential vorticity (contours; interval of 0.1 PVU = 0.1 × 10−6 m2 s−1 K kg−1; dark green solid lines are for positive but dark red dashed lines denote negative anomalies of potential vorticity, respectively) averaged between 20° and 40°N, referring to the eight phases of the 10–30-day ISO of the MTG index in AMJ from 1980 to 2019. (a)–(h) Phases 1–8, respectively. Gray shading indicates the topography.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
Phase-dependent composite of 10–30-day ISO of 500-hPa winds (vectors; m s−1; vectors exceeding the 95% confidence level are in black), CPC rainfall amount (contours; interval of 0.5 mm day−1; dark green solid lines are for above-normal rainfall but dark red dashed lines denote below-normal rainfall amount, respectively), and atmospheric heat source over Tibetan Plateau above 1500-m topography (shading; W m−2; values exceeding the 95% confidence level are stippled), referring to the eight phases of the 10–30-day ISO of the MTG index in AMJ from 1980 to 2019. (a)–(h) Phases 1–8, respectively. Bold purple contours represent the 1500-m topography.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
Phase evolution chart verifies that the anomalous atmospheric heat source of TP is in-phase with the 10–30-day ISO of spring rainfall over South China, both of which leads the variation of mid–upper-tropospheric temperature to the north of the SCS by one phase (Fig. 12). Therefore, we ascribe the 10–30-day MTG ISO to the subseasonal variation of diabatic heating over TP and South China in AMJ. The coupling between atmospheric heat source of TP and spring rainfall over South China amplifies this extratropical ISO when it crosses the large topography over East Asia.
Time series of standardized 10–30-day ISO of atmospheric heat source over Tibetan Plateau (60°–105°E, 25°–45°N above 1500 m) (bars; values exceeding the 95% confidence level are in bold borders), CPC rainfall amount over South China (110°–120°E, 22.5°–32.5°N) (red line), and 500–200-hPa averaged air temperature over the northern SCS (110°–130°E, 15°–25°N) (black line). Solid dots represent the values exceeding the 95% confidence level in each phase.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
6. Influences of the 10–30-day MTG ISO on the SCSSM onset
The distinct feature between the phases of the 10–30-day MTG ISO indicates its phase-dependent influences on the onset of the SCSSM. In statistics, MTGOD tends to be later than MTGSCT when the MTG ISO is strong in phase 1; but when the ISO significantly enhanced in phase 5, the MTGOD becomes much earlier than MTGSCT (Fig. 13). The intensity of MTG ISO shows less difference in other phases. In physics, the stronger phase 1 tends to delay the MTGOD by cooling the tropospheric temperature to the north of the SCS, whereas the MTGOD is advanced by the warmer air to the north of the SCS when phase 5 is reinforced (Figs. 7a,e).
Boxplot of phase intensity of 10–30-day MTG ISO in the case of large difference between MTGOD and MTGSCT. Blue (red) bars indicate the situation when the MTGOD is more than 6 days earlier (later) than the MTGSCT. The phase whose intensity difference between the two categories exceeds the 99% confidence level is in boxes. The ISO intensity is defined as the average standardized 10–30-day MTG ISO within 31 days centered on the MTGSCT in each phase.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
In addition, the upper-level PV anomaly could predict the MTG influences on the SCSSM onset on the 10–30-day time scale. In phase 1 that inhibits the SCSSM onset, as a response of more stable atmosphere to upper-level positive PV anomaly over South China (Hoskins et al. 1985), the mid–upper troposphere is getting colder to the north of the SCS, along with anomalous sinking and suppressed spring rainfall over South China (Figs. 10a, 11a, and 14a). While the negative PV anomaly is propagating eastward in phases 2 and 3, the 10–30-day ISO of northerly wind (υ′ < 0) on its east gives rise to the positive PV advection (
Phase-dependent composite of pressure–latitude cross-section of 10–30-day ISO of meridional circulation (vectors; m s−1; vectors exceeding the 95% confidence level are in black), horizontal advection of seasonal cycle component of potential vorticity due to the 10–30-day meridional winds (shading; 0.1 PVU day−1; values exceeding the 95% confidence level are stippled), and air temperature (contours; K; dark green solid lines are for warmer but dark red dashed lines denote the colder air temperature, respectively) averaged between 110° and 120°E, referring to the eight phases of the 10–30-day ISO of the MTG index in AMJ from 1980 to 2019. (a)–(h) Phases 1–8, respectively. Gray shading indicates the topography.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0696.1
7. Conclusions and discussion
a. Conclusions
The onset time of the SCSSM indicates the beginning of the major rainy season over East Asia in boreal summer. The ENSO event has been treated as the primary source of its seasonal predictability. However, the ENSO-based prediction of the SCSSM onset time has failed in some cases. In a perspective of the MTG in the mid–upper troposphere, the present study decomposed the onset process of the SCSSM into seasonal cycle and subseasonal component and attributed the seasonal and subseasonal predictability of the SCSSM onset dates to the ENSO events and extratropical 10–30-day ISO in the mid–upper troposphere, respectively. The major conclusions are summarized as follows.
On the interannual time scale, the modulation of ENSO events on the onset time of the SCSSM is realized by changing the seasonal cycle component of the tropical thermal field. In April and May, the ENSO-related SST anomalies in the tropical Pacific and Indian Ocean modulate the MTGSCT by affecting the seasonal cycle component of air temperature in the south of the SCS via the Matsuno–Gill response. The negative-to-positive transition time of the MTG is therefore changed to determine the anomaly of the SCSSM onset date. The correspondence between ENSO and SCSSM onset vanishes when excluding the seasonal cycle component.
Apart from the seasonal cycle component, the subseasonal MTG index over the SCS features an ISO with a significant periodicity of 10–30 days. It is determined by the extratropical ISO of the mid–upper-tropospheric air temperature over East Asia. Embedding in the subtropical westerly basic flow, this ISO originates over western TP, then propagates eastward to South China. On the 10–30-day time scale, the coupling between atmospheric heat source over TP and spring rainfall over South China acts as an amplifier to enhance the extratropical ISO when it crosses the large topography over East Asia.
The 10–30-day extratropical ISO in the mid–upper troposphere shows effective regulation on onset process of SCSSM. The SCSSM onset date is greatly different from the transition time of the MTG seasonal cycle when the ISO is more active. When the ISO moving toward South China, the meridional wind anomaly on the east of the anomalous circulation first induces PV advection in the mid–upper troposphere over northern SCS to modulate the upper-level pumping effect in situ. As the ISO arriving at South China, the anomalous air temperature in the mid–upper troposphere to the north of the SCS directly affects the negative-to-positive transition of the MTG, along with the anomaly of low-level zonal wind and monsoon onset convection over the SCS. The influences of the 10–30-day extratropical ISO on the SCSSM onset, which is relatively separated from the ENSO effects, provide additional subseasonal predictability of SCSSM onset dates.
b. Discussion
The MTG-based onset dates of the SCSSM (MTGOD) are significantly correlated with the onset dates defined by U850 (U850OD; e.g., Kajikawa and Wang 2012), which have been commonly used in many previous studies. The climate-mean U850OD during 1980–2019 is 21 May, which is one week later than the climatological MTGOD of 14 May. We thus can treat the MTGOD as a precursor for the U850OD of the SCSSM. Furthermore, although the 10–30-day period is evident in the subseasonal component of the U850 index, the reconstructed onset dates including both seasonal cycle and 10–30-day ISO do not improve evidently comparing with the ones solely defined by the seasonal cycle component (figure not shown). Therefore, the MTGOD exhibits higher predictability involving both tropical and mid–high-latitude influences.
The 10–30-day MTG ISO can explain the difference between MTGOD and MTGSCT in some specific years, regardless of the ENSO events. For instance, the MTGOD (15 May) was earlier than the MTGSCT (23 May) in 2004 (neutral ENSO year). Before MTGSCT, the 10–30-day ISO of 500–200-hPa averaged air temperature was warmer to north of the SCS, along with the anomalous anticyclone inducing more positive PV onto the SCS, especially on 15 May. In 2019 with an El Niño event, the MTGOD (1 May) was advanced from the MTGSCT (10 May), accompanied by the mid–upper-tropospheric warmer air propagating from the TP to the northern SCS during 25 April–5 May. Such configuration advanced the MTGOD from the MTGSCT over the SCS, consistent with the upscaling effect of Typhoon Fani reported by Liu and Zhu (2020). In 2018 with a La Niña event, however, the MTGOD (2 June) was much later than the MTGSCT (15 May). A cold anomaly persisted to the north of the SCS in the 10–30-day ISO of mid–upper-tropospheric air temperature after the MTGSCT, along with an anomalous cyclone over the northern SCS. A similar situation occurred in the neutral ENSO year of 1991, but presented stronger cooling to the north of the SCS, corresponding to the larger delaying of the MTGOD from the MTGSCT after 25 May.
Although the 10–30-day MTG ISO can regulate the onset process of the SCSSM, what determines the interannual variability of this ISO is a remaining issue. We do not observe a statistically significant relationship between the phase intensity of the ISO and Niño-3.4 index in May, indicating that the ISO intensity may depend on other factors except for the ENSO events. For instance, the underlying conditions over land, such as snow cover and soil moisture over the TP and Eurasian continent, may act as predictors of the intensity of this ISO. These land processes are likely to provide the opportunity for more skillful prediction of the SCSSM onset in a longer lead time (Mariotti et al. 2020). Future studies will be necessary to validate the above hypothesis.
Another major period of the subtropical variability over the SCS is 31–70 days (corresponding to the 5th–12th harmonics), which becomes significant in the MTG ISO index in some years. Some previous study proposed that the influences of 10–25-day and 30–60-day OLR ISO over the SCS were distinct on the SCSSM onset. The former tended to delay the monsoon onset, while the latter would advance the onset of SCSSM (Kajikawa and Yasunari 2005). As to the MTG ISO index, both the more pronounced 10–30- and 31–70-day variability could enlarge the difference between MTGOD and MTGSCT (figure not shown). This is probably because the subseasonal fluctuation, acting as a “noise” of the seasonal predictability ascribed to the ENSO events, is exactly separated from the interannual variation of the seasonal evolution in the present study. But it is still an open question about from where this lower-frequency MTG ISO stems and how it affects the SCSSM onset in collaboration with the 10–30-day ISO, especially against the ENSO background.
Acknowledgments
This work was jointly funded by the National Key R&D Program (2018YFC1505904), the National Natural Science Foundation of China (41775052, 41830969), the Second Tibetan Plateau Scientific Expedition and Research (STEP) program (2019QZKK0105), and the Science and Technology Development Foundation of Chinese Academy of Meteorological Sciences (2020KJ009, 2020KJ012). The authors declare that they have no conflicts of interest.
REFERENCES
Bingham, C., M. Godfrey, and J. Tukey, 1967: Modern techniques of power spectrum estimation. IEEE Trans. Audio Electroacoust., 15, 56–66, https://doi.org/10.1109/TAU.1967.1161895.
Bretherton, C. S., M. Widmann, V. P. Dymnikov, J. M. Wallace, and I. Bladé, 1999: The effective number of spatial degrees of freedom of a time-varying field. J. Climate, 12, 1990–2009, https://doi.org/10.1175/1520-0442(1999)012<1990:TENOSD>2.0.CO;2.
Chang, C. P., and G. T. J. Chen, 1995: Tropical circulations associated with southwest monsoon onset and westerly surges over the South China Sea. Mon. Wea. Rev., 123, 3254–3267, https://doi.org/10.1175/1520-0493(1995)123<3254:TCAWSM>2.0.CO;2.
Chen, M., W. Shi, P. Xie, V. B. S. Silva, V. E. Kousky, R. Wayne Higgins, and J. E. Janowiak, 2008: Assessing objective techniques for gauge-based analyses of global daily precipitation. J. Geophys. Res., 113, D04110, https://doi.org/10.1029/2007JD009132.
Ding, Y., and Y. Liu, 2001: Onset and the evolution of the summer monsoon over the South China sea during SCSMEX field experiment in 1998. J. Meteor. Soc. Japan, 79, 255–276, https://doi.org/10.2151/jmsj.79.255.
Ding, Y., and J. C. L. Chan, 2005: The East Asian summer monsoon: An overview. Meteor. Atmos. Phys., 89, 117–142, https://doi.org/10.1007/s00703-005-0125-z.
Fujinami, H., and T. Yasunari, 2009: The effects of midlatitude waves over and around the Tibetan Plateau on submonthly variability of the East Asian summer monsoon. Mon. Wea. Rev., 137, 2286–2304, https://doi.org/10.1175/2009MWR2826.1.
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.
Harada, Y., and Coauthors, 2016: The JRA-55 reanalysis: Representation of atmospheric circulation and climate variability. J. Meteor. Soc. Japan, 94, 269–302, https://doi.org/10.2151/jmsj.2016-015.
He, B., Y. Zhang, T. Li, and W.-T. Hu, 2017: Interannual variability in the onset of the South China Sea summer monsoon from 1997 to 2014. Atmos. Ocean. Sci. Lett., 10, 73–81, https://doi.org/10.1080/16742834.2017.1237853.
He, J., and Z. Zhu, 2015: The relation of South China Sea monsoon onset with the subsequent rainfall over the subtropical East Asia. Int. J. Climatol., 35, 4547–4556, https://doi.org/10.1002/joc.4305.
Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc., 111, 877–946, https://doi.org/10.1002/qj.49711147002.
Hu, P., S. Chen, W. Chen, Y. Liu, and R. Huang, 2020: Extremely early summer monsoon onset in the South China Sea in 2019 following an El Niño event. Mon. Wea. Rev., 148, 1877–1890, https://doi.org/10.1175/MWR-D-19-0317.1.
Huang, B., and Coauthors, 2017: Extended Reconstructed Sea Surface Temperature, version 5 (ERSSTv5): Upgrades, validations, and intercomparisons. J. Climate, 30, 8179–8205, https://doi.org/10.1175/JCLI-D-16-0836.1.
Jiang, N., W. Qian, and J. C.-H. Leung, 2016: The global monsoon division combining the k-means clustering method and low-level cross-equatorial flow. Climate Dyn., 47, 2345–2359, https://doi.org/10.1007/s00382-015-2967-5.
Jiang, X., Z. Wang, and Z. Li, 2018: Signature of the South China Sea summer monsoon onset on spring-to-summer transition of rainfall in the middle and lower reaches of the Yangtze River basin. Climate Dyn., 51, 3785–3796, https://doi.org/10.1007/s00382-018-4110-x.
Ju, J., and J. M. Slingo, 1995: The Asian summer monsoon and ENSO. Quart. J. Roy. Meteor. Soc., 121, 1133–1168, https://doi.org/10.1002/qj.49712152509.
Kajikawa, Y., and T. Yasunari, 2005: Interannual variability of the 10–25- and 30–60-day variation over the South China Sea during boreal summer. Geophys. Res. Lett., 32, L04710, https://doi.org/10.1029/2004GL021836.
Kajikawa, Y., and B. Wang, 2012: Interdecadal change of the South China Sea summer monsoon onset. J. Climate, 25, 3207–3218, https://doi.org/10.1175/JCLI-D-11-00207.1.
Kajikawa, Y., T. Yasunari, S. Yoshida, and H. Fujinami, 2012: Advanced Asian summer monsoon onset in recent decades. Geophys. Res. Lett., 39, L03803, https://doi.org/10.1029/2011GL050540.
Kobayashi, S., Y. Ota, and Y. Harada, 2015: The JRA-55 Reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 5–48, https://doi.org/10.2151/jmsj.2015-001.
Li, C., and M. Yanai, 1996: The onset and interannual variability of the Asian summer monsoon in relation to land–sea thermal contrast. J. Climate, 9, 358–375, https://doi.org/10.1175/1520-0442(1996)009<0358:TOAIVO>2.0.CO;2.
Li, C., Z. Zheng, D. Gu, and A. Lin, 2019: Asymmetric effects of atmospheric circulation on the South China Sea summer monsoon onset. Dyn. Atmos. Oceans, 87, 101099, https://doi.org/10.1016/j.dynatmoce.2019.101099.
Li, Y., S. Yang, Y. Deng, and B. Zheng, 2020: Signals of spring thermal contrast related to the interannual variations in the onset of the South China Sea summer monsoon. J. Climate, 33, 27–38, https://doi.org/10.1175/JCLI-D-19-0174.1.
Liang, Z.-R., Z.-H. Jiang, H. Yang, Z.-Y. Liu, and T.-T. Gu, 2013: Application of HYSPLIT model in definition of the onset of South China Sea summer monsoon. J. Trop. Meteor., 19, 130–137.
Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77, 1275–1277, https://doi.org/10.1175/1520-0477-77.6.1274.
Lin, A., and R. Zhang, 2020: Climate shift of the South China Sea summer monsoon onset in 1993/1994 and its physical causes. Climate Dyn., 54, 1819–1827, https://doi.org/10.1007/s00382-019-05086-4.
LinHo, and B. Wang, 2002: The time–space structure of the Asian–Pacific summer monsoon: A fast annual cycle view. J. Climate, 15, 2001–2019, https://doi.org/10.1175/1520-0442(2002)015<2001:TTSSOT>2.0.CO;2.
Liu, B., and C. Zhu, 2016: A possible precursor of the South China Sea summer monsoon onset: Effect of the South Asian high. Geophys. Res. Lett., 43, 11 072–11 079, https://doi.org/10.1002/2016GL071083.
Liu, B., and C. Zhu, 2019: Extremely late onset of the 2018 South China Sea summer monsoon following a La Niña event: Effects of triple SST anomaly mode in the North Atlantic and a weaker Mongolian cyclone. Geophys. Res. Lett., 46, 2956–2963, https://doi.org/10.1029/2018GL081718.
Liu, B., and C. Zhu, 2020: Boosting effect of tropical cyclone “Fani” on the onset of the South China Sea summer monsoon in 2019. J. Geophys. Res. Atmos., 125, e2019JD031891, https://doi.org/10.1029/2019JD031891.
Liu, B., Y. Liu, G. Wu, J. Yan, J. He, and S. Ren, 2015: Asian summer monsoon onset barrier and its formation mechanism. Climate Dyn., 45, 711–726, https://doi.org/10.1007/s00382-014-2296-0.
Liu, B., C. Zhu, Y. Yuan, and K. Xu, 2016: Two types of interannual variability of South China Sea summer monsoon onset related to the SST anomalies before and after 1993/94. J. Climate, 29, 6957–6971, https://doi.org/10.1175/JCLI-D-16-0065.1.
Liu, P., Y. Qian, and A. Huang, 2009: Impacts of land surface and sea surface temperatures on the onset date of the South China Sea summer monsoon. Adv. Atmos. Sci., 26, 493–502, https://doi.org/10.1007/s00376-009-0493-2.
Lu, M.-M., C.-H. Sui, J.-R. Sun, and P.-H. Lin, 2020: Influences of subseasonal to interannual oscillations on the SCS summer monsoon onset in 2018. Terr. Atmos. Ocean. Sci., 31, 197–209, https://doi.org/10.3319/TAO.2020.02.25.01.
Mao, J., and J. C. L. Chan, 2005: Intraseasonal variability of the South China Sea summer monsoon. J. Climate, 18, 2388–2402, https://doi.org/10.1175/JCLI3395.1.
Mao, J., and G. Wu, 2007: Interannual variability in the onset of the summer monsoon over the eastern Bay of Bengal. Theor. Appl. Climatol., 89, 155–170, https://doi.org/10.1007/s00704-006-0265-1.
Mao, J., J. C. L. Chan, and G. Wu, 2004: Relationship between the onset of the South China Sea summer monsoon and the structure of the Asian subtropical anticyclone. J. Meteor. Soc. Japan, 82, 845–859, https://doi.org/10.2151/jmsj.2004.845.
Mariotti, A., and Coauthors, 2020: Windows of opportunity for skillful forecasts subseasonal to seasonal and beyond. Bull. Amer. Meteor. Soc., 101, E608–E625, https://doi.org/10.1175/BAMS-D-18-0326.1.
Martin, G. M., A. Chevuturi, R. E. Comer, N. J. Dunstone, A. A. Scaife, and D. Zhang, 2019: Predictability of South China Sea summer monsoon onset. Adv. Atmos. Sci., 36, 253–260, https://doi.org/10.1007/s00376-018-8100-z.
Matsuno, T., 1966: Quasi-geostrophic motions in the equatorial area. J. Meteor. Soc. Japan, 44, 25–43, https://doi.org/10.2151/jmsj1965.44.1_25.
Pan, W., J. Mao, and G. Wu, 2013: Characteristics and mechanism of the 10–20-day oscillation of spring rainfall over southern China. J. Climate, 26, 5072–5087, https://doi.org/10.1175/JCLI-D-12-00618.1.
Qian, W., and D.-K. Lee, 2000: Seasonal march of Asian summer monsoon. Int. J. Climatol., 20, 1371–1386, https://doi.org/10.1002/1097-0088(200009)20:11<1371::AID-JOC538>3.0.CO;2-V.
Qian, Y., S. Y. Wang, and H. Shao, 2001: A possible mechanism effecting the earlier onset of southwesterly monsoon in the South China Sea compared to the Indian monsoon. Meteor. Atmos. Phys., 76, 237–249, https://doi.org/10.1007/S007030170032.
Shao, X., P. Huang, and R.-H. Huang, 2015: Role of the phase transition of intraseasonal oscillation on the South China Sea summer monsoon onset. Climate Dyn., 45, 125–137, https://doi.org/10.1007/s00382-014-2264-8.
Straub, K. H., G. N. Kiladis, and P. E. Ciesielski, 2006: The role of equatorial waves in the onset of the South China Sea summer monsoon and the demise of El Niño during 1998. Dyn. Atmos. Oceans, 42, 216–238, https://doi.org/10.1016/j.dynatmoce.2006.02.005.
Tong, H. W., J. C. L. Chan, and W. Zhou, 2008: The role of MJO and mid-latitude fronts in the South China Sea summer monsoon onset. Climate Dyn., 33, 827–841, https://doi.org/10.1007/s00382-008-0490-7.
Wang, B., and X. Xu, 1997: Northern Hemisphere summer monsoon singularities and climatological intraseasonal oscillation. J. Climate, 10, 1071–1085, https://doi.org/10.1175/1520-0442(1997)010<1071:NHSMSA>2.0.CO;2.
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., LinHo, Y. Zhang, and M.-M. Lu, 2004: Definition of South China Sea monsoon onset and commencement of the East Asia summer monsoon. J. Climate, 17, 699–710, https://doi.org/10.1175/2932.1.
Wang, H., F. Liu, B. Wang, and T. Li, 2018: Effects of intraseasonal oscillation on South China Sea summer monsoon onset. Climate Dyn., 51, 2543–2558, https://doi.org/10.1007/s00382-017-4027-9.
Wang, M. R., and A. M. Duan, 2015: Quasi-biweekly oscillation over the Tibetan Plateau and its link with the Asian summer monsoon. J. Climate, 28, 4921–4940, https://doi.org/10.1175/JCLI-D-14-00658.1.
Webster, P. J., and S. Yang, 1992: Monsoon and ENSO: Selectively interactive systems. Quart. J. Roy. Meteor. Soc., 118, 877–926, https://doi.org/10.1002/qj.49711850705.
Webster, P. J., V. O. Magana, T. N. Palmer, J. Shukla, R. A. Tomas, M. Yanai, and T. Yasunari, 1998: Monsoons: Processes, predictability, and the prospects for prediction. J. Geophys. Res., 103, 14 451–14 510, https://doi.org/10.1029/97JC02719.
Wen, Z., J.-L. Chan, and L. Chongyin, 2005: South China Sea summer monsoon onset in relation to the off-equatorial ITCZ. Adv. Atmos. Sci., 22, 665–676, https://doi.org/10.1007/BF02918710.
Wu, G., and Y. Zhang, 1998: Tibetan Plateau forcing and the timing of the monsoon onset over South Asia and the South China Sea. Mon. Wea. Rev., 126, 913–927, https://doi.org/10.1175/1520-0493(1998)126<0913:TPFATT>2.0.CO;2.
Wu, G., B. He, Y. Liu, Q. Bao, and R. Ren, 2015: Location and variation of the summertime upper-troposphere temperature maximum over South Asia. Climate Dyn., 45, 2757–2774, https://doi.org/10.1007/s00382-015-2506-4.
Wu, R., and B. Wang, 2000: Interannual variability of summer monsoon onset over the western North Pacific and the underlying processes. J. Climate, 13, 2483–2501, https://doi.org/10.1175/1520-0442(2000)013<2483:IVOSMO>2.0.CO;2.
Xiang, B., and B. Wang, 2013: Mechanisms for the advanced Asian summer monsoon onset since the mid-to-late 1990s. J. Climate, 26, 1993–2009, https://doi.org/10.1175/JCLI-D-12-00445.1.
Xie, P., A. Yatagai, M. Chen, T. Hayasaka, Y. Fukushima, C. Liu, and S. Yang, 2007: A gauge-based analysis of daily precipitation over East Asia. J. Hydrometeor., 8, 607–626, https://doi.org/10.1175/JHM583.1.
Yuan, Y., W. Zhou, J. C. L. Chan, and C. Li, 2008: Impacts of the basin-wide Indian Ocean SSTA on the South China Sea summer monsoon onset. Int. J. Climatol., 28, 1579–1587, https://doi.org/10.1002/joc.1671.
Zhang, Y., and Y. Qian, 2002: Mechanism of thermal features over the Indochina Peninsula and possible effects on the onset of the South China Sea monsoon. Adv. Atmos. Sci., 19, 885–900, https://doi.org/10.1007/s00376-002-0053-5.
Zhou, W., and J. C. L. Chan, 2005: Intraseasonal oscillations and the South China Sea summer monsoon onset. Int. J. Climatol., 25, 1585–1609, https://doi.org/10.1002/joc.1209.
Zhou, W., and J. C. L. Chan, 2007: ENSO and the South China Sea summer monsoon onset. Int. J. Climatol., 27, 157–167, https://doi.org/10.1002/joc.1380.
Zhu, Z., and T. Li, 2017: Empirical prediction of the onset dates of South China Sea summer monsoon. Climate Dyn., 48, 1633–1645, https://doi.org/10.1007/s00382-016-3164-x.
