Unusual heavy rainfall, accompanied with severe flooding and landslides, affected millions of people over large parts of the East Asian countries in summer 2020. However, the detailed patterns of rainfall in July and August exhibit some differences. In July 2020, the observed rainfalls in East Asia around the latitude of 30°N were well above normal and the highest recorded one (Fig. 1a, b). The rainfall anomaly of July 2020 in East Asia (105°-150°E, 25°-40°N) is exceeding 3.27 standard deviation for the 42-year period. In the tropics, a zonally extended rainfall decrease occurred from the Indian subcontinent to subtropical WNP, and increased in the Indian Ocean.
A relatively weak but broad rainfall increase in East Asia is the predominant feature in August 2020, evolving further to the north (Fig. 1c) compared to that of July (Fig. 1a). Because of the northward rainband, the different East Asian regions are selected in July and August based on their maximum rainfall. The rainfall magnitude of East Asia (100°-135°E, 30°-50°N) in August 2020 is above 3.0 standard deviation during the 42 years. In comparison with July, the rainfall decrease still remained in the subtropical WNP but extended further north to around Japan in August. In contrast, the rainfall anomalies from the Indian subcontinent to the South China Sea became positives in August.
The question remains, what is responsible for the rainfall extremes in East Asia during July and August? To address this, the SST and rainfall anomalies correlated with the East Asian rainfall indices in July and August are shown in Fig. 2. Recently, Kim and Kug (2021) suggested the delayed impact of Indian Ocean warming on the East Asian surface temperature variation in summer. The Indian Ocean warming in June is responsible for significant cooling over the Korea-Japan region that peaks in July with a 1-month delay. From the lead-lag correlation with July East Asian rainfall, positive correlations are evident over most of the Indian Ocean and South China Sea in June with a maximum center in the Arabian Sea and Bay of Bengal (Fig. 2a). This implies that there is a potential role of Indian Ocean SST during June in driving the rainfall increase in East Asia during July. The correlation field of rainfall has a similar spatial pattern to that of July 2020 (Fig. 1a), appearing as west-to-east elongated rainfall decrease from the eastern Indian Ocean to subtropical central Pacific and increase in the western Indian Ocean, yet only significant around the subtropical WNP and central Pacific (Fig. 2c).
The correlations between the August East Asian rainfall index and Indian Ocean SST are weaker than that of June (Fig. 2a), but still significant in the Arabian Sea (Fig. 2b). The negative SST anomalies in the adjacent region to East Asia may be a result of the increased rainfall by reducing shortwave radiation, rather than a cause. Interestingly, the East Asian rainfall anomalies in August are highly correlated with the negative rainfall anomalies in the subtropical WNP (Fig. 2d). In particular, the WNP rainfall decrease was evident in August 2020 (Fig. 1c), suggesting an important role in modulating the East Asian rainfall increase.
As shown in Fig. 2, the highly-correlated factors in potentially explaining the East Asian rainfall are to some extent different for July and August. Given the above results, we choose two possible factors that are responsible for the East Asian summer rainfall: the tropical Indian Ocean SST in June and subtropical WNP rainfall anomalies in August. Therefore, the question arises of how the Indian Ocean SST and WNP rainfall anomalies affect the rainfall variability in East Asia during July and August, respectively. Here, we elucidate the role of two indicators on the development of rainfall in East Asia for July and August.
The delayed impact of tropical Indian Ocean SSTs in June on the East Asian rainfall anomalies in July is investigated using regression analysis (Fig. 3). Anomalous SST warming in the Indian Ocean, centered over the Arabian Sea and eastern Bay of Bengal, are quite similar that in June, suggesting strong persistency (Fig. 3a). Note that standardized coefficient refers to the regression coefficients simply multiplied by the value of the Indian Ocean SST anomaly in June 2020. In addition, the significant negative SST anomalies are zonally confined near East Asia latitude between 30°-50°N. This regressed SST pattern in July onto 1-month leading Indian Ocean warming (Fig. 3a) is quite similar to the SST anomalies of July 2020 (Fig. 3b), and has a spatial correlation (45°E-165°W, 5°S-55°N) of 0.52.
Traditionally, the tropical Indian Ocean basin warms the season after El Niño (Weare 1979), indicating a passive response to El Niño. The relative importance of remote forcing from the Pacific to the Indian sector (i.e., tropical atmospheric bridge; Klein et al. (1999)) and internal variability of the Indian Ocean is still debated within the research community (Allan et al. 2001; Baquero-Bernal et al. 2002; Hastenrath 2002; Lau and Nath 2003; Schott et al. 2009). The Indian Ocean warming was observed in the summer of 2020 possibly in accordance with the super IOD in fall 2019 (Takaya et al. 2020; Zhou et al. 2021) and/or El Niño event in the winter of 2019/20 (Qiao et al. 2021).
The rainfall responses in July onto the Indian Ocean warming in June clearly show the zonally elongated rainband in East Asia (Fig. S1). The Indian Ocean warming enhances convective activity, particularly in the Arabian Sea, because the surface moisture convergence in that region is presumably stronger in comparison with the other Indian Ocean basins (Roxy et al. 2013). Enhanced convective heating stimulates the tropospheric Kelvin waves that propagate eastward to the western Pacific. Subsequently, the Kelvin wave-induced boundary layer divergence suppresses the local convection in the WNP, and thus low-level anticyclonic anomalies develop through the atmospheric Rossby wave response, namely the IPOC mode (Xie et al. 2009).
The resultant low-level anticyclonic anomalies in the WNP are clearly seen (Fig. 3c) as a result of the Indian Ocean warming (Wu et al. 2009; Xie et al. 2009), and are a part of the meridional wave train propagating northward, the PJ teleconnection pattern (Nitta 1987). Associated with the southwesterlies on the western side of the WNP anticyclonic circulation, large amounts of water vapor can be transported from the tropics to East Asia. Based on the moisture budget analysis, the dynamic effect due to changes in the atmospheric circulation is dominant in July (not shown). Therefore, the Indian Ocean warming in June may be responsible for the positive rainfall anomalies in East Asia in July, exhibiting a 1-month leading role. Note that the Indian Ocean warming was also observed in July 2020 (Fig. 3b), indicating a persistent SST which continuously plays a role in decreasing the WNP rainfall. The low-level atmospheric circulation anomalies in July 2020 (Fig. 3d) resemble the regressed pattern onto the June Indian Ocean warming (Fig. 3c), and the spatial correlation coefficient with geopotential height at 850 hPa is 0.66. The temporal correlation of the Indian Ocean SST during June and the East Asian rainfall anomalies during July is also strong (0.56; Fig. 3e). Importantly, the magnitude of the Indian Ocean SST in June 2020 ranks among the top four since 1979. These results imply that the tropical Indian Ocean warming in June is one of the dominant factors leading to the rainfall extremes in East Asia during July 2020.
In August, associated with the subtropical WNP rainfall index the negative local rainfall anomalies are dominant (Fig. 4a). Concurrently, pronounced rainfall increases in East Asia due to the southwesterlies corresponding to anticyclonic anomalies (Fig. 4c). Although the amplitude of regressed rainfall anomalies in East Asia is weaker than that during August 2020 (Fig. 4b), the signal is statistically significant. Importantly, a rainfall decrease was also evident in the subtropical WNP during the August 2020 (Fig. 4b). This indicates that the subtropical WNP rainfall anomalies can play an important role in increasing rainfall in East Asia during August. The simultaneous impact of Indian Ocean SST on the East Asian rainfall in August is relatively weaker than that of Indian Ocean SST in July (0.41) with a correlation coefficient of 0.34, but possibly contributes to the East Asian region.
In response to the subtropical WNP rainfall forcing in August, the anticyclonic anomalies centered at 137.5°E, 22.5°N (Fig. 4c) can be interpreted as a Rossby wave response. The subtropical diabatic forcing in August is far from the equator, thus the wave response can be established at a relatively higher latitude. The resultant anticyclonic flow is accompanied by southwesterly anomalies from the off-equatorial towards East Asia and contributes to the rainfall anomalies there in August. Overall, the atmospheric pattern in August 2020 is quite similar to the regressed result (Fig. 4c), with anticyclonic anomalies evidently centered south of Japan (Fig. 4d). The spatial correlation coefficient between these low-level patterns represented as the geopotential heights at 850 hPa is 0.66. The distribution of subtropical WNP rainfall and East Asian rainfall anomalies for the period 1979-2020 also indicates a significant relationship with a correlation coefficient of -0.56 (Fig. 4e). In August 2020, the magnitude of the subtropical WNP rainfall anomaly ranks 4th over the last 42 years. This implies that in August the subtropical WNP rainfall decrease may contribute to the development of anticyclonic anomalies, affecting the extreme East Asian rainfall.
To examine atmospheric response to the subtropical diabatic forcing, the LBM (Watanabe and Kimoto (2000)) experiment is carried out (Fig. 5). In this experiment, the prescribed forcing is obtained from the linearly regressed local rainfall with respect to the subtropical WNP rainfall index in August (Fig. 4a), using August basic state during 1979-2020. It is evident that the negative rainfall forcing over the subtropical WNP region (Fig. 5a) is critical in developing the anticyclonic circulation (Fig. 5b), matching well the observational pattern (Fig. 4c). Therefore, the subtropical WNP anticyclonic anomalies can be explained by the negative local rainfall anomalies through the atmospheric Rossby wave response, and then play an important role in modulating the East Asian rainfall in August.
Importantly, the rainfall extremes in July and August 2020 showed different features in East Asia, as a zonally-elongated rainband located near 30°N in July (Fig. 1a), and a relatively broad one confined further north of 30°N in August (Fig. 1c). At the same time, the dominant negative rainfall anomalies over the WNP region are also distinct in July and August, showing that it moves further northeastward in accordance with the northward shifted rainfall in East Asia during August (Fig. 2d), compared to the pattern of July (Fig. 2c). Both the WNP rainfall anomalies in July and August may play an important role in modulating the rainfall variability in East Asia, respectively, although their magnitude and location are different.
However, it is not certain whether the July WNP rainfall evolves to the August one, or another physical process induces the August rainfall independently. To understand the rainfall variability over the subtropical WNP region in August, its preceding SST and rainfall patterns are investigated (Fig. S2). The positive local SST anomalies are dominant in June then persist until July, linking to the subtropical WNP rainfall increase in August. However, in August the local SST correlations turn negatives located near 30°N, which can be regarded as a result of rainfall increase rather than a cause. Interestingly, the north tropical Atlantic (NTA) warming in June is significantly correlated with the subtropical WNP rainfall anomalies in August with 2-month leading role, but the correlations over the NTA region become weaker after June. As suggested in Ham et al. (2013), the NTA warming during the spring can induce the low-level cyclonic flow over the eastern Pacific that in turn triggers the low-level anticyclonic flow over the western Pacific in the following months through a subtropical teleconnection. Therefore, the NTA SST anomalies possibly have an impact on the East Asian summer variability through the WNP anticyclonic anomalies (Jin and Huo 2018).
The lagged regressions of SST, rainfall and wind responses at 850 hPa to the NTA (50°-15°W, 0°-20°N) SST anomalies in June are shown in Fig. 6. In June, the positive SST anomalies in the NTA region (Fig. 6a) enhances the local convective activity (Fig. 6d). The resultant diabatic heating gives rise to the low-level cyclonic anomalies over the subtropical eastern Pacific in July as a Gill-type response, with the northerly anomalies on its western side (Fig. 6e). In July, the zonally-elongated rainband in East Asia are quite similar with the observed July 2020 (Fig. 1a), suggesting the potential role of NTA SST anomalies. The northerlies in the subtropical Pacific induce the surface cooling through the enhanced wind speed and cold/dry advection from higher latitudes. As a result, the negative rainfall anomalies in the subtropical Pacific induce the subtropical WNP anticyclonic flow until August through the strong air-sea coupling (Fig. 6f). The associated rainfall decrease and anticyclonic circulation in the subtropical WNP region possibly affect the summer rainfall in East Asia during August. The NTA-regressed rainfall increase in East Asia during August is significant (Fig. 6f) and quite similar with that of August 2020 (Fig. 1c). This implies that the NTA warming in June can play a role in modulating the East Asian rainfall in August through the development of subtropical WNP anticyclonic anomalies with 2-month time lag. In June 2020, the NTA region anomalously warmed that is above 0.92 standard deviation during 1979-2020.
Additionally, relatively stronger mean and variability of rainfall exist in the subtropical WNP region during August compared to that during July (not shown) due to the northward migration of the Pacific Intertropical Convergence Zone (ITCZ). Therefore, the interannual variability of subtropical WNP rainfall may well be sensitive to the relatively enhanced climatological local rainfall from July to August. It provides favorable conditions such that enhancement of convective instability in the subtropical WNP extends further north during August. As a result, this strengthened rainfall variability in the subtropical WNP during August may play a more dominant role in modulating local anticyclonic anomalies, and have a profound impact on the East Asian rainfall anomalies.