The Role of Indian Ocean Warming on Extreme Rainfall in Central China During Early Summer 2020: Without El Niño Influence


 This study investigates the role of water vapor transport and sea surface temperature (SST) warming in the tropical Indian Ocean (TIO) on the heavy rainfall in central China during boreal early summer. In the past four decades, four significant rainfall events, in 1983, 1998, 2016, and 2020, occured in central China and caused severe floods, in which the year 2020 has the most extreme event. All four events are associated with significant TIO SST warming, associated with a strong and westward extending anomalous anticyclone on the western North Pacific (WNPAC). The anomalous winds in the northwestern flank of the WNPAC bring excess water vapor into central China. The water vapor, mainly carried from the central tropical Pacific, converges in central China and result in heavy rainfall. A theory of regional ocean-atmosphere interaction can well explain the processes, called the Indo-Western Pacific Ocean Capacitor (IPOC) effect. The WNPAC is usually associated with strong El Niño-Southern Oscillation (ENSO), except for the 2020 case. The 2020 event is extraordinary, without ensuring El Niño occurred in the previous winter. In 2020, the significant TIO warming sustained the anomalous WNPAC, inducing the most significant extreme rainfall event in central China. This study reveals that the IPOC effect can dramatically influence the East Asian climate even without involving the ENSO in the Pacific.

shows the time series of rainfall anomalies in central China during early boreal summer. Four signi cant rainfall events, which exceed one standard deviation of rainfall anomaly, occurred in 1983, 1998, 2016, and 2020.
The 2020 rainfall event was extremely strong, with its rainfall anomaly exceeded three standard deviations, making a historical record since 1979 (Fig. 1b) Ding et al. (2021) reported that the East Asian monsoon circulation system, including the high-level westerly jet, the low-level southwesterly jet, and WPSH, featured a quasi-biweekly oscillation, all contributing to this extreme event. Besides, the atmospheric circulation pattern in the middle to high latitudes facilitated the cold air intrusions to the central China  Benton et al. (1950) showed that even local evaporation is signi cant, the external moisture advection contributes the majority of rainfall. As the abnormal water vapor continuously enters the sink region, heavy rainfall can be formed (Trenberth et al., 2003). Water vapor transport is vital for determining the rainfall pattern of China (Simmonds et al. 1999;Ninomiya and Kobayashi 1998;Ding and Sun 2001).
For four signi cant rainfall events in central China in 1983China in , 1998China in , 2016, and 2020, the origin of the water vapor and the cause of such water vapor transport patterns are not always the same. This motivates us to distinguish the difference, especially for the 2020 event, from the aspect of water vapor transport. This study found that not like the other three events, no strong El Niño conditions occurred in the central to eastern equatorial Paci c in the previous winter of 2020. Only a weak warm state appeared in the centralwestern tropical Paci c. Our research investigates the 2020 case and demonstrates the greatly important role of TIO SST warming even without ensuring El Niño in the Paci c.
The rest of the study is organized as follows. Section 2 describes the dataset and methods utilized in this work. Section 3 analyzes the water vapor transport associated with four signi cant rainfall events. Section 4 explores the direct impact of WNAPC. Section 5 discusses the IPOC effect on extreme rainfall in central China during early summer 2020 in detail. Section 6 summarizes major points and provides further discussions.

Data And Methodology
In this study, monthly mean atmospheric data, including geopotential wind, sea level pressure, and speci c humidity, are from the fth generation European Centre for Medium-Range Forecasts (ECMWF) reanalysis (ERA5) (https://cds.climate.copernicus.eu) (Hersbach et al. 2019a(Hersbach et al. , 2019b. The horizontal resolution of the ERA5 datasets is 0.75°×0.75°. Global Precipitation Climatology Project (GPCP) precipitation data are provided by the NOAA/OAR/ESRL PSL, Boulder, Colorado, USA, from their website at https://psl.noaa.gov/data/gridded/data.gpcp.html. The sea surface temperature (SST) data on a 1°×1° grid are derived from the NOAA/OAR/ESRL PSL, Boulder, Colorado, USA, from their website at https://www.psl.noaa.gov/data/gridded/data.noaa.oisst.v2.html (Reynolds et al. 2002). All above data are available during the period 1979-2020. Anomalies are calculated relative to the climatological seasonal cycle based on the period 1979-2020. The Niño3.4 SST index is also used in this study, which is de ned as the SST anomaly in the Niño3.4 region (170°W-120°W, 5°S-5°N). This index is available from NOAA/ESRL, updated weekly from 1982 to 2020 (https://stateoftheocean.osmc.noaa.gov/sur/pac/nino34.php). In the present study, the statistical signi cance of the composite anomalies is assessed by a student's t test.
The ERA5 monthly mean vertically integrated zonal and meridional water vapor ux are de ned as follows.
( 1) is the surface atmospheric pressure, is the zonal and meridional wind, is the speci c humidity, is the gravitational acceleration. and are the pressures at the ground surface and top of the troposphere, respectively. Since moisture content above 200 hPa can be negligible (Trenberth 1991), the vertically integrated water vapor ux is integrated from 1000 hPa to 200 hPa in our study.
The divergence of vertically integrated water vapor ux can be expressed as: (2) and are the divergence, is vertically integral of zonal water vapor ux and is vertically integral of meridional water vapor ux.
We integrate the water vapor ux across each line segment along the boundaries of central China. The equation (Richter and Xie 2010) is Where is the moisture transport across the boundary, and the unit is Sverdrup (symbol: Sv), with 1 Sv equal to , is position along the boundary.

Water Vapor Transport Associated With Signi cant Rainfall Events
To illustrate the relationship between water vapor supply, ascending motion, and rainfall over central China, average mean rainfall, net regional water vapor budget, and p-velocity at 500 hPa (multiplied by − 1, -ω500) anomalies over central China are displayed in Fig. 2. The early summer rainfall anomalies over central China show remarkably interannual variation and coincide well with the water vapor budget and -ω500 (Fig. 2). In general, the year-to-year pulse of the rainfall anomalies is consistnet with the change of the water vapor budget and -ω500. The rainfall and net water vapor budget are highly correlated, and their correlation coe cient (r) is 0.75. The correlation coe cient between rainfall and -ω500 is slightly higher (r = 0.79). All these coe cients are statistically signi cant at the 95% con dence level. In four signi cant rainfall events (de ned in Fig. 1b; 1983, 1998, 2016, and 2020), positive rainfall is corresponding to positive net water vapor budget and -ω500, which indicates that the water vapor supply and local vertical motion are important for the rainfall in central China. It is reasonable to investigate the possible mechanisms governing heavy rainfall over central China from the aspect of external water vapor transport.
The vertically integrated water vapor transport ux (WVF) and the divergence for climatological conditions are presented in Fig. 3a. During MJJ, three branches of water vapor are converyed into central China. The rst one is water vapor from the northern Indian Ocean, which originates in the southern Indian Ocean. The second one is the warm and moist water vapor transported from the Southern Hemisphere and the South China Sea, and the third one is carried from the tropical western Paci c Ocean.
Warm and moist water vapor meets and converges over central China (Fig. 3a). The western and southern boundaries of central China are the main input boundaries. About 0.04 Sv and 0.23 Sv water vapor are brought into the region, respectively. The eastern boundary is the main out ow boundary (0.13 Sv). The water vapor carried out through the eastern boundary is much stronger than that via the northern boundary (0.05 Sv). In Fig. 3a, strong convergence regions also include western India and western South Asia, where the external water vapor supply is strong. As a result, more precipitation appears over these areas than in other places (Fig. 1a).  showed that rainfall anomalies coincide very well with entering water vapor anomalies in the south (r = 0.59). When central China receives more rainfall, more water vapor is transported through its southern boundary. In four signi cant early summer rainfall cases, water vapor entered through the southern boundary are particularly strong, all exceed one standard deviation, increasing about 30% (0.07Sv, 1983), 39% (0.09Sv, 1998), 22% (0.05Sv, 2016), and 43% (0.10Sv, 2020) over the mean state, respectively ( Fig. 4). It suggests that during four signi cant rainfall events, extra water vapor is carried through the southern boundary by the northeastward water vapor path, then converges in central China and leads to heavy rainfall nally.

Wnpac Impacts On Extreme Rainfall
To unravel the origin of water vapor, the composite of 850 hPa wind anomalies in four signi cant rainfall events is displayed in Fig. 5a. During four signi cant rainfall events, an anomalous strong anticyclone appears over the western North Paci c (WNPAC). Strong anomalous easterly winds blow from the tropical western Paci c to the Indian Ocean, which weakens the climatological westerly winds. Meanwhile, part of anomalous easterly winds turns to northeastwards and blow to East Asia (Fig. 5a).
Geopotential height anomalies at 500 hPa during four cases are shown in Fig. 5b. In four signi cant rainfall events, signi cantly increased geopotential height is observed in most regions (e.g., the tropical western Paci c, the Indian Ocean, and eastern China). The maximum enhancement appears over the western North Paci c. On the contrary, geopotential height decreases over Mongolia (Fig. 5b). The center of the WPSH is extended westward and southward compared to its climatological mean (Fig. 5b), which is consistent with the position of anomalous WNPAC (Fig. 5a). A large amount of water vapor is transported into central China by the southwesterly jet at the northwestern edge of the WPSH. It also means that the northwest ank of the WNPAC is the northeastward water vapor path mentioned in the last section.
To quantify the intensity of WNPAC, this study de nes a WNPAC index, as the 850 hPa zonal wind difference between a southern region (100°E-130°E, 5°N-15°N) and a northern region (110°E-140°E, 25°N-35°N). The de nition follows Huang et al. (2010), except that the northern region shifts northward. A positive (negative) WNPAC index indicates an anomalous anticyclone (cyclone) appears in the western North Paci c. The relationship between WNPAC and El Niño is shown in Fig. 6. In most El Niño events, a signi cant anticyclone appears in the western North Paci c during the decay phase (Fig. 6), consistent with previous studies. For four signi cant rainfall events, the values of the WNPAC index are positive and large, which indicates the enhanced WNPAC during these years. It is worth noting that 1983, 1998, and 2016 rainfall cases were associated with preceding strong El Niño events. However, no signi cant El Niño occurred in 2019/2020 winter. The WNPAC index is positively correlated with the Niño3.4 index with a correlation coe cient of 0.66. The correlation coe cient becomes slightly higher (0.69) when the 2020 event is removed. All these coe cients are statistically signi cant at the 95% con dence level. These results suggest that such enhanced and westward extending WNPAC supplies continuous water vapor to central China through its northwestern ank. The resultant strong converge over central China would, in turn, lead to heavy rainfall. During three signi cant rainfall cases, following strong El Niño events (1983,1998,2016), the WNPAC was highly related to El Niño. However, a strong WNPAC in 2020 occurred without El Niño. Therefore, one provoking question is what process sustained the WNPAC in 2020.

Ipoc Effect On 2020 Event Without El Niño
SST, 850 hPa winds, and rainfall anomalies during MJJ of three events (1983,1998,2016) and 2020 are compared and presented in Fig. 7. The western Paci c and the northern Paci c were warming when the southeastern Paci c was cooling in 2020, opposite the SST pattern of the other three events (Fig. 7a, c). In all signi cant rainfall events, the tropical Indian Ocean and the South China Sea are warming. Anomalous easterly winds blow from the tropical western Paci c to the Bay of Bengal, and an anticyclone appears on the western North Paci c (Figs. 7a, c). During four signi cant events, rainfall in the western North Paci c and South Asia decreased. On the contrary, more rainfall occurred over the tropical eastern Indian Ocean, East Asia, and Japan (Figs. 7b, d). In 2020, the SST warming dominated the TIO and even extended to the tropical western Paci c, while the eastern equatorial Paci c maintained a slight cooling state (Fig. 7c), the eastern equatorial Paci c SST cooling may play an important role in maintaining the WNPAC due to atmospheric Rossby wave response to its northwest (Chen et al., 2016). Induced by the TIO warming, anomalous easterly winds extended to the Arabian Sea; more rainfall occurred in the entire TIO basin in 2020 (Figs. 7c, d). According to the IPOC theory, the above processes were sustained by WNPAC (Xie et al., & 2016.

IPOC theory well explains the relationship between Indian Ocean SST warming and the western North
Paci c anomalous circulation. In 2020 early summer, the TIO rainfall is the key. The rainfall and SST anomalies averaged over the TIO region are displayed in Fig. 8. Despite that, the correlation of the year-toyear uctuation of the SST and rainfall anomalies over the TIO is not very high (r = 0.38). The positive correlation coe cient indicates the ocean SST forcing relates to rainfall in a particular year. SST of TIO has substantial interannual variation, and there are signi cant increasing trends for rainfall and SST since 1982 (Fig. 8, the linear ttings). For four signi cant rainfall events (1983,1998,2016,2020), the TIO SST and rainfall anomalies are distributed above the linear trend lines, indicating that the TIO SST warming matches abnormal positive rainfall in these years (Fig. 8). The most striking feature for four signi cant rainfall events is the excess rainfall in the TIO forced by the intense TIO warming. Among the four cases, the 2020 event is the most signi cant since the TIO rainfall reaches the maximum. In recent decades, the TIO warming trend may also provide a favorable background to maintain a stronger WNPAC (Fig. 8), and the 2020 extreme rainfall in central China was a manifestation. We also nd that signi cant TIO warming and enhanced rainfall appeared in 2010, associated with the delayed response to El Niño (Chen et al., 2016). In the summer of 2010, the TIO warming also played an essential role in maintaining the WNPAC (Figs. 6 and 8). As a result, abnormal positive rainfall appeared in central China (Fig. 1), due to the IPOC effect. Note, the WNPAC in 2010 was slightly southward, compared to the four signi cant rainfall events; thus, rainfall in central China was weaker.
Climatologically, in the low latitude, the zonal water vapor is transported from the northern Indian Ocean to the western Paci c, and in the middle latitude, water vapor is transported eastwards north of 30°N (Fig. 9a). In 2020 early summer, the zonal water vapor is enhanced north of 20°N, and the maximum enhancement appears around 30°N. The slight enhancement also appeared in the northern Arabian Sea. On the contrary, zonal water vapor is weakened over the Bay of Bengal, the South China Sea, and the tropical Paci c (Fig. 9a). As a result, less water vapor from the Bay of Bengal is transported to the tropical western Paci c. More eastwards water vapor is transported through the eastern boundary of central China. Climatologically, the meridional water vapor is transported from ocean to continent in most regions (Fig. 9b). In general, the zonal water vapor transport was larger than the meridional water vapor. In 2020 early summer, excessive northward meridional water vapor was transported from the South China Sea to central China (Fig. 9b), related to enhanced and westward extending WNPAC and TIO SST warming.

Conclusions
This study mainly reveals that the extreme rainfall of 2020 MJJ is attributed to the IPOC effect even without involving the ENSO in the Paci c. In the past four decades, four signi cant rainfall events (1983,1998,2016, and 2020) occurred in central China, resulting in catastrophic oods. The year 2020 has the most extreme rainfall among four signi cant events. This study investigated the cause for the 2020 extreme rainfall from water vapor transport and TIO SST warming. In climatology, three pathways carry water vapor to central China from tropical Indian and Paci c Oceans, and the schematic representation is displayed in Fig. 10a. In the rst pathway, the warm and moist water vapor is originated from the south Indian Ocean, transporting northeastwards to central China. In the second pathway, the water vapor is brought from the southern hemisphere and the South China Sea. In the third pathway, the water vapor is carried from the western Paci c (Fig. 10a). During 2020, the intense warming of the tropical Indian Ocean excited a strong and westward extending WNPAC in MJJ, coupled with the TIO SST warming (Fig. 10b). Such coupled mechanism persisted in the early summer. Meanwhile, TIO SST warming forced enhanced rainfall in the Indian Ocean basin. Excess water vapor was carried into central China mainly through the southern boundary by the northwestern ank of the WNPAC and convergence in central China, leading to record-breaking rainfall. Rainfall intensity is strengthened on the northern branch of the WNPAC. On the contrary, less rainfall appeared over the western Paci c due to the WNPAC and suppressed advection (Fig. 10b)   a Climatology of vertical integral of water vapor ux (WVF) (vectors, unit: kg/(m·s)) and divergence (shading, unit: kg/(m2·s)). b Composite of vertical integral of water vapor ux (WVF) (vectors, unit: kg/(m·s)) and divergence (shading, unit: kg/(m2·s)) anomalies in four extreme rainfall events (i.e., 1983,1998,2016,2020  a Composite of 850 hPa wind anomalies (vectors, unit: m/s) in four extreme rainfall events, the shading denotes the wind anomaly intensity, the black vectors denote the wind anomalies is signi cant at 95 % con dence level according to student's t test, the two red boxes denote the regions to de ne WNPAC index. b Composite of 500 hPa geopotential height anomalies (shading, unit: gpm) in four extreme rainfall events, the solid and dashed green lines (the 5880 gpm contour lines) represent the main body of the WPSH in climate mean and extreme events, respectively. The stippled areas denote the geopotential height anomalies are signi cant at 95 % con dence level according to student's t test which is presented in Fig.5. 2020 extreme rainfall event is tagged by red pentagram marker, other three events following strong El Niño events are tagged by green star markers In a, the stippled areas denote the SST anomalies are signi cant at 95 % con dence level according to student's t test, the black vectors denote the wind anomalies are signi cant at 95 % con dence level according to student's t test. In b, the stippled areas denote the rainfall anomalies are signi cant at 95 % con dence level according to student's t test Vertical integral of (a) zonal and (b) meridional water vapor ux (unit: kg/(m·s)). The contours and shading represent the climatology and anomaly in 2020, respectively