Warming of the Kuroshio Current Over the Last Four Decades has Intensified the Meiyu‐Baiu Rainband

In recent decades, rise of sea surface temperature (SST) in the Kuroshio region of the East China Sea (ECS), which is associated with global warming, has attracted considerable attention. Despite its relevance to air‐sea interaction phenomena, the atmospheric consequences of this SST increase remain largely unexplored. Using the ERA5 reanalysis data set in conjunction with a moisture budget analysis, we found that during 1979–2022, warming in the ECS‐Kuroshio has contributed to the intensification of the East Asian Meiyu‐Baiu rainband in June. This intensification is attributed to augmented wind convergence in the low‐level free troposphere (950–700 hPa). Importantly, the atmospheric responses to ECS‐Kuroshio warming penetrated the deep troposphere (to approximately 300 hPa), suggesting an enhancement of deep convection. Furthermore, ECS‐Kuroshio warming likely strengthened the overlying atmospheric low pressure, resulting in wind convergence enhancement. The findings clarified the important role of the ECS‐Kuroshio in driving East Asian climate change amid global warming.

Supporting Information may be found in the online version of this article.

ERA5 Reanalysis Data Set
ERA5 is the fifth iteration of the global climate and weather reanalysis undertaken by the European Centre for Medium-Range Weather Forecasts (ECMWF), covering a temporal span of eight decades.The ERA5 reanalysis data set provides hourly and monthly variables at single levels and across pressure levels, that is, 37 pressure levels between 1 hPa and 1,000 hPa, all at a horizontal resolution of 0.25°.Data for the period 1979-2022 were used in this study.For the single-level variables, we used total precipitation (TP), convective precipitation (CP), large-scale precipitation (LSP), SST, surface latent heat flux (SLHF), surface sensible heat flux (SSHF), evaporation, mean sea level pressure (SLP), 10-m wind vector, and 10-m wind speed.For the pressure-level variables, we used specific humidity, wind vector, and temperature.
The high spatial and temporal resolutions of the ERA5 reanalysis data set make it possible to investigate the overlying atmospheric responses to the ECS-Kuroshio with a narrow current width (∼200 km).Further, the ERA5 data set provides comprehensive variables, which makes it possible to clarify the mechanisms by which the ECS-Kuroshio affects the Meiyu-Baiu rainband on a multidecadal timescale.However, to validate the increasing rainfall trend during the Meiyu-Baiu season using ERA5, we employed supplementary precipitation products, as outlined in Section 3.1.

Quantitative Diagnosis of Moisture Budget
Moisture budget analysis is an effective quantitative method for clarifying the physical processes that contribute to precipitation.Following Seager and Henderson (2013), the column-integrated moisture budget equation can be expressed as: where P is precipitation, E is evaporation, q is specific humidity, u is a horizontal wind vector, ∇ h is a horizontal divergence operator, g is gravity, ρ w is water density, t is time, p is pressure, p s is surface pressure, q s is surface specific humidity and u s is a surface horizontal wind vector.For the trend in the time period from 1979 to 2022, the first term on the right-hand side of Equation 1, which represents the local temporal tendency, is negligible on a multidecadal timescale.The third term on the right-hand side, which represents the surface quantities, can be ignored because its magnitude is considerably smaller than that of the second term (i.e., lateral moisture flux convergence).Using these simplifications, Equation 1 yields: where, hourly data of q and u are used,  (⋅) expresses the monthly mean (i.e., the mean value in June in this study), and 〈 .Note that, using the ERA5 data set on pressure levels, the vertical integration was QIAO ET AL.

10.1029/2023GL107021
3 of 11 calculated in the range between 1 hPa and 1,000 hPa.Applying  =  +  ′ and   =  +  ′ , the right-hand term in Equation 2 can be decomposed as follows: where the subscripts c and a denote the long-term monthly climatology and monthly anomaly, respectively.− 〈∇ h • q a u a 〉 in Equation 4 can be neglected because of its small magnitude.The anomaly of , can be derived by moving the term  −⟨∇ ⋅ ⟩ to the left-hand side, as follows: (5) in the precipitation trend, then Equation 5 is used to further clarify which terms on the right-hand side contributed effectively to the precipitation trend.

Intensified Meiyu-Baiu Rainband in June
The monsoon governs the formation of the Meiyu-Baiu rainband, as the dominant southwesterly winds transport substantial amounts of moisture from the tropics, which initiates onset of the rainy season in East Asia.The Meiyu-Baiu rainband emerges in East Asia in May, develops in June along with northward migration, and gradually weakens in July. Figure 1a shows the climatological TP, wind speed at 850 hPa, and SST for June.The precipitation associated with the Meiyu-Baiu rainband is primarily concentrated in two regions: the southeastern regions of China, and the marine areas around the ECS-Kuroshio and western Japan.The precipitation over the warm tongue of the ECS-Kuroshio, which is comprised of warm water advected by the Kuroshio, is considerably stronger than that in the surrounding areas.Based on the pressure adjustment mechanism proposed by Lindzen and Nigam (1987), Sasaki et al. (2012) suggested that the ECS-Kuroshio warm tongue generated a low-pressure atmospheric region overlying the current.This, in turn, leads to wind convergence along the ECS-Kuroshio, resulting in significant accumulation of rainfall over the ECS-Kuroshio region.Figure 1b, which shows the rate of change in average TP for the month of June during the period 1979-2022, reveals an intensification of precipitation along the Meiyu-Baiu rainband, particularly in the marine areas around the ECS-Kuroshio and to the south of Japan.Increasing trends in precipitation were also observed in May and July (data not shown), but their magnitudes and areas were considerably smaller than those in June.In June, particularly, the air-sea conditions around the ECS-Kuroshio satisfy high SSTs, a strong SST front, and abundant overhead vapor due to the Meiyu-Baiu rainband; this leads to a more pronounced atmospheric response to the ECS-Kuroshio in June compared to other months from early spring to summer (Sasaki et al., 2012).Hence, we only focused on the increasing trend in precipitation observed in June.
To validate the June precipitation trend identified using ERA5, precipitation data from three independent sources were utilized: the CPC Merged Analysis of Precipitation (CMAP), Global Precipitation Climatology Project (GPCP) and NCEP/DOE Reanalysis 2. All of the data sets corroborated the intensification of precipitation over the marine areas around the ECS-Kuroshio and to the south of Japan from 1979 to 2022, as shown in Figure 1b (data not shown).In addition, rainfall gauge data from Naha in Okinawa Island also showed an increasing trend in the June precipitation during the period 1979-2022 (data not shown).
Next, we performed the moisture budget analysis.Using hourly specific humidity and wind data, we found that the trend for lateral moisture flux convergence, that is, S1a in Supporting Information S1), corresponded geographically with the trend observed for the TP in June (Figure 1b).In addition, the difference between the trends in TP and − along the Meiyu-Baiu rainband was mainly contributed by  −⟨∇ ⋅ ⟩ , suggesting that the enhanced precipitation along the Meiyu-Baiu rainband in June was mainly caused by the convergent trends of wind.We also confirmed that local evaporation supplied little moisture to the atmosphere, implying that it did not contribute markedly to observed increase in the precipitation amount (Figure S2a in Supporting Information S1).In the region overlying the warm tongue of the ECS-Kuroshio, the amplitude of the upward trend in evaporation was approximately one-tenth of that of precipitation (Figure 1b).Furthermore, the trends in evaporation over the marine area to the south of Japan were statistically insignificant.

Potential Effect of ECS-Kuroshio Warming on Convective Precipitation
Figures 2a and 2b show the trends in the CP and LSP in June from 1979 to 2022, respectively.The CP was observed to intensify around the ECS-Kuroshio while the LSP mainly intensified over the ocean areas to the south of Japan.Since the aim of this study was to clarify the roles of ECS-Kuroshio warming in precipitation trends along the Meiyu-Baiu rainband, we focused on the driving mechanisms underlying the trends in CP (Figure 2a).  Figure 2c shows the rate of change in SSTs in June.The SST exhibited an increasing trend along the warm tongue of the Kuroshio in the ECS and to the south of Japan.Although a warming trend in SST was observed throughout the entire ECS, the area of warming within the ECS-Kuroshio was separated from the increase in SST along the Chinese coast.As a result, the SST front to the west of the ECS-Kuroshio was enhanced in June.Indeed, as shown in Figure S2b in Supporting Information S1, which shows the rate of change in the absolute SST gradient (|∇SST|) in June, the SST fronts surrounding the ECS-Kuroshio were enhanced.According to previous studies, acceleration of the ECS-Kuroshio (Sasaki & Umeda, 2021) and the increase in SSTs to the east of the Philippines, which is the upstream area of the Kuroshio Current (Wang et al., 2016;Wang & Wu, 2022), likely contributed to the observed warming of the ECS-Kuroshio in June. Figure 2a shows that the increasing trend in CP was most pronounced along the warm tongue of the ECS-Kuroshio (see the black isoline for 0.06 mm • day −1 yr −1 ).Interestingly, the core tends to meander along the ECS-Kuroshio in the Tokara Strait and extends slightly to the south of Japan.These phenomena imply that warming of the ECS-Kuroshio in June likely intensified the CP within the Meiyu-Baiu rainband by effectively enhancing the overlying wind convergence (Figure 1c).
The TP in June that was area-averaged around the ECS-Kuroshio (see the black isoline in Figure 2a) showed a trend of 1.455 × 10 −1 mm • day −1 yr −1 (significant at 95% confidence level) (Figure 2d).The CP (Figure 2e) and LSP (Figure 2f) accounted for 54.57% and 45.43% of this trend, respectively.These results suggest that, around the ECS-Kuroshio, the increasing trend observed for CP is central to the intensification of the Meiyu-Baiu rainband in June, although positive feedback might exist between the increasing trends of the CP and LSP.Furthermore, we observed that the SST around the ECS-Kuroshio had a trend of 2.01 × 10 −2 ℃ • yr −1 (Figure 2g).A numerical experiment of Sasaki et al. (2012) indicated that, in June, an SST change in the ECS-Kuroshio with the range of 0.5-1.5℃can induce a precipitation change ranging from 1 to 3 mm • day −1 around the ECS-Kuroshio (see their Figure 11c).Based on this result, we can estimate that the precipitation trend solely caused by the SST increase in the Kuroshio (Figure 2g) potentially accounted for 27.63% of TP (Figure 2d) and 50.63% of CP (Figure 2e).The remained precipitation trends around the ECS-Kuroshio, excluding those induced by SST increasing, may be explained by changes in the atmosphere, such as the intensification of the Pacific subtropical high which was known to have enhanced the moisture transport to East Asia in summer during 1998-2019 (Takahashi & Fujinami, 2021).In addition, the LSP trends over the south of Japan (Figure 2b) may be mainly caused by the enhanced moisture transport (Takahashi & Fujinami, 2021), however, contributions of the local subtropical SST front over the south of Japan to these trends should be investigated in future (Figure 2b).

Vertical Atmospheric Structure: Intensification of Deep Convection Overlying the ECS-Kuroshio
A distinct core is evident for the rising trend in TP over the ECS-Kuroshio to the northwest of Okinawa Island (126.5°E,27.5°N in Figure 1b).Such a feature is also visible in the trends obtained for  −⟨∇ ⋅ ⟩ (Figure 1c) and CP (Figure 2a).Thus, along the zonal section at 27.5°N, we visualized the vertical distributions in the trends of  −∇ ⋅  , reversed vertical velocity (−w), and diabatic heating (Q d ) in June (Figures 3a-3c).Following Yamamoto (2013), we estimated Q d as: where θ is potential temperature.To elucidate the effect of the ECS-Kuroshio on the atmosphere, we plotted the climatological SSTs (black line in Figure 3d) and the rates of SST increase (red line in Figure 3d) along the 27.5°N section for the month of June. Figure 3d shows that the SSTs over the warm tongue of the ECS-Kuroshio had increased significantly and that the rate of increase was markedly larger at its center (i.e., maximum position of the climatological SSTs) compared to the surrounding areas.These features are consistent with the horizontal distribution of the trends in SST shown in Figure 2c. Figure 3a shows that the positive trends of  −∇ ⋅  , that is, moisture flux convergence induced by convergence of wind anomalies, were generally confined below 600-500 hPa.Consequently, the maximum upward trend in vertical winds appeared in the range of 600-500 hPa (Figure 3b).Importantly, unlike the trends observed in the surrounding areas, the positive trend observed in  −∇ ⋅  overlying the warm tongue of the ECS-Kuroshio was pronounced in the 950-700 hPa layer; however, this trend was insignificant at the 95% confidence level in layers below 950 hPa, which included almost the entire marine atmospheric boundary layer (MABL).Figure 3a  the upward trend of the vertical wind overlying the warm tongue of the ECS-Kuroshio was larger than those over the same horizontal levels in the free troposphere above the MABL.These results suggest that the atmospheric circulation in the free troposphere exhibits a sensitive response to the increase in SST over the ECS-Kuroshio in June, leading to the enhanced moisture flux convergence above the warm tongue of the ECS-Kuroshio.
Furthermore, the positive trend observed in diabatic heating over the warm tongue of the ECS-Kuroshio, which was characterized by a core at 600-350 hPa (Figure 3c), was markedly higher than the trends observed in the surrounding areas.We also compared diabatic heating with the wind direction in the vertical section.The streamlines in Figure 3c were based on the changing rates of −w and zonally 7-degree high-pass filtered eastward wind.
The application of a high-pass filter to the eastward wind is necessary as the localized responses of the horizontal  wind to the narrow ECS-Kuroshio and SST fronts were obscured by large-scale atmospheric circulations.While the wind trends below 900 hPa layer were predominantly horizontal, in the free troposphere above this layer, vertical trends were more pronounced, especially over the warm tongue of the ECS-Kuroshio.Figure 3c shows that the upward trends of wind were collocated with positive trends in diabatic heating.In conjunction with the intensification of the Meiyu-Baiu rainband, these increasing trends in upward motions and diabatic heating pervaded almost the entire free troposphere, reaching a maximum in the mid-troposphere (Figures 3b and 3c).The elevation in SST over the ECS-Kuroshio likely contributed to amplifying these trends.
The vertical distributions in  −∇ ⋅  , −w, and diabatic heating shown in Figures 3a-3c imply that the CP intensification (Figure 2a) was associated with the SST increase over the ECS-Kuroshio region.This intensification appears to occur through conditional instability of a second kind (CISK), a mechanism involving positive feedback between low-level wind convergence and latent heat release due to condensation, which results in the development of deep convection (Charney & Eliassen, 1964).Figure 3e, showing the trends in −w in June for the layer at 300 hPa, indicates that the relatively large positive trends for −w at 300 hPa layer (≥1.5 × 10 −3 Pa • s −1 yr −1 ) meandered along the Kuroshio in the Tokara Strait and extended slightly to the southern coast of Japan.This horizontal pattern in −w resembles that of the increasing rate of CP (Figure 2a) and suggests that increases in the SSTs of the ECS-Kuroshio contribute to the deep convection process.Trends of −w at 600 hPa level did not exhibit meandering patterns along the Kuroshio, but they did resemble the horizontal pattern of the Meiyu-Baiu rainband (Figure S3 in Supporting Information S1).The spatial distributions of trends in diabatic heating in June paralleled the horizontal distributions of −w at both the 300 hPa and 600 hPa levels (data not shown).Consequently, the response of the atmosphere to increases in the SSTs of the ECS-Kuroshio penetrated the deep troposphere in June, suggesting that the warming of the ECS-Kuroshio intensifies CP with enhancing deep convection.

Possible Mechanism of Wind Convergence Enhancement Over the ECS-Kuroshio
Both the Gulf Stream and ECS-Kuroshio were considered to affect the atmospheric conditions above the respective currents through the pressure adjustment mechanism (Minobe et al., 2008(Minobe et al., , 2010;;Sasaki et al., 2012).Briefly, this mechanism suggests that the warm SSTs over the current, coupled with SST gradients on either side of the current, result in low-level (≥700 hPa) wind convergence by inducing low atmospheric pressure (Lindzen & Nigam, 1987).Other studies further suggested that atmospheric responses to the SST front depend on the wind directions relative to the SST front (Bai et al., 2019;Kilpatrick et al., 2016;Schneider & Qiu, 2015;Xu & Xu, 2015).Xu and Xu (2015) suggested that the ECS-Kuroshio can enhance the rainfall and cumulus convection by adjusting the pressure when the wind blows along its western SST front.As shown in Figure 1a, the prevailing wind at 850 hPa in June blows nearly parallel to the SST front in the ECS, implying that the ECS-Kuroshio warming may intensify the CP through the pressure adjustment mechanism.Minobe et al. (2008) employed a relationship between near-surface wind convergence and the Laplacian of SLP, ∆(SLP), or −SST Laplacian, ∆(−SST), to examine the importance of pressure adjustment mechanism in the vicinity of the Gulf Stream.Because we observed that the positive trends of  −∇ ⋅  were pronounced in the 950-700 hPa layer overlying the ECS-Kuroshio (Figure 3a), ∆(SLP) and −∇ h • u a at 850 hPa averaged around the warm tongue of ECS-Kuroshio were scattered in Figure S4 in Supporting Information S1.It can be seen that −∇ h • u a at 850 hPa was significantly raised with ∆(SLP) increasing, consistent with Figure 1f in Minobe et al. (2008).Furthermore, Figures 3f and 3g show that the western flank of the ECS-Kuroshio warm tongue exhibited positive trends in both ∆(SLP) and ∆(−SST).Additionally, the SST front in this region was also intensified (Figure S2b in Supporting Information S1).These findings imply that the increase in the SSTs of the ECS-Kuroshio induced the development of a low-pressure region above the current, strengthening low-level wind convergence overhead (Figure 3a).
Figure S5 in Supporting Information S1 shows that the horizontal distributions of the changing rates for SLHF and SSHF were congruent across the ECS.However, the positive trend in the SLHF was about five times greater in magnitude than that in the SSHF in the downstream region of the ECS-Kuroshio.This result indicates that the positive trend in evaporation due to the positive trend in SST over the ECS-Kuroshio was an important factor in the pressure adjustment mechanism (Figure S2a in Supporting Information S1).Nevertheless, its influence on the supply of moisture to the intensified precipitation was relatively minor, as mentioned in Section 3.1.The SLHF (Figure S5a in Supporting Information S1) exhibited a negative trend over the upstream part of ECS-Kuroshio, even though the SST increased in this region (Figure 2c).This negative trend could be ascribed to the negative QIAO ET AL. Figure 3a shows that the positive changing rates in  −∇ ⋅  below the 950 hPa layer, which nearly corresponds to the MABL, were statistically insignificant over the warm tongue of the ECS-Kuroshio.This suggests that, overlying the ECS-Kuroshio warm tongue, the free troposphere was more responsive to increases in the SSTs of the ECS-Kuroshio compared to the MABL.Additionally, Figure 3a shows that the positive trends in  −∇ ⋅  in the MABL were more concentrated in ocean areas defined by SST fronts (see black line in Figure 3d), implying that the vertical mixing mechanism (Chelton et al., 2004;Wallace et al., 1989;Xu & Xu, 2015) may operate in the MABL.We discussed in more details in Supporting Information S1 regarding the vertical mixing mechanism (Text 1, Figures S6 and S7 in Supporting Information S1), because this mechanism had little effect on the precipitation (Xu & Xu, 2015).

Conclusion
Using ERA5 reanalysis data set from 1979 to 2022, we found that precipitation along the Meiyu-Baiu rainband intensified markedly in June, particularly over the ocean areas around the ECS-Kuroshio and areas to the south of Japan.Moisture budget analysis revealed that  −⟨∇ ⋅ ⟩ , that is, moisture flux convergence arising from wind anomaly convergence, played a dominant role in this intensification of precipitation.Approximately 54.57% of the enhanced precipitation around the ECS-Kuroshio was accounted by CP.This CP had an interesting horizontal distribution, with relatively large positive trends that appeared to meander along the ECS-Kuroshio and extend slightly to the south of Japan through the Tokara Strait.This phenomenon implied that the increase in SSTs over the ECS-Kuroshio likely contributed to the intensification of the Meiyu-Baiu rainband in June through effective enhancement of CP.Furthermore, according to the numerical experiment of Sasaki et al. (2012), we estimated that the precipitation trend solely caused by the SST increase potentially accounted for 27.63% of TP and 50.63% of CP; the remained precipitation trends around the ECS-Kuroshio was probably caused by the enhanced moisture transport to East Asia in summer due to changes in the atmosphere, such as the intensification of the Pacific subtropical high (Takahashi & Fujinami, 2021).
Moreover, the vertical distribution of  −∇ ⋅  revealed that the convergent trends in the 950-700 hPa layer were amplified significantly, with a peak at approximately 900 hPa overlying the ECS-Kuroshio warm tongue where the SST increased significantly.Correspondingly, in the vertical section, the positive trends in −w and diabatic heating over the warm tongue of the ECS-Kuroshio were considerably larger than the trends observed over the surrounding regions.In addition, the increasing rates for −w and diabatic heating at 300 hPa also meandered along the ECS-Kuroshio warm tongue around the Tokara Strait, exhibiting a spatial distribution similar to that observed for the CP.Based on these findings, we conclude that the atmospheric responses to the increasing ECS-Kuroshio SSTs penetrated the deep troposphere in June, implying that ECS-Kuroshio SST augmentation intensified the Meiyu-Baiu rainband with enhancing deep convection.Finally, given the observed intensification of ∆(SLP), ∆(−SST) and SST fronts on the western flank of the ECS-Kuroshio warm tongue, it was probably that the ECS-Kuroshio SST elevation contributed to atmospheric pressure adjustments over the current, enhancing low-level free tropospheric moisture flux convergence in the 950-700 hPa layer; however, the convergent trends in the MABL below 950 hPa could possibly be ascribed to the vertical mixing mechanism.

•
The Meiyu-Baiu rainband in June intensified over the marine areas around the Kuroshio and south of Japan during 1979-2022 • The intensified Meiyu-Baiu rainband was caused mainly by the wind convergence accompanied by enhanced deep convection • Kuroshio warming in the East China Sea likely intensified the Meiyu-Baiu rainband by a pressure adjustment mechanism Supporting Information: ) where (•)′ denotes the hourly deviation from the monthly mean.− ⟨ ∇ℎ ⋅ ⟩ and  − ⟨ ∇ ⋅  ′  ′ ⟩ express the moisture flux convergences induced by the monthly mean and its fluctuation states, respectively.Applying   =  +  and   =  +  ,

⟩
was small (data not shown), implying that the intensified precipitation in June depended considerably on the moisture supply due to − S1c in Supporting Information S1), revealed that an increasing trend of the precipitation along the Meiyu-Baiu rainband was driven by − precipitation along the Meiyu-Baiu rainband.We also calculated the trends for the four components of −
also shows a prominent core in the positive trend of  −∇ ⋅  overlying the warm tongue of the ECS-Kuroshio at about 900 hPa.Corresponding to these features,

Figure 3 .
Figure 3. (a-c) Vertical distributions in (a)  −∇ ⋅  , (b) reverse vertical wind (−w), and (c) diabatic heating (Q d ) along a zonal section at 27.5°N showing their rates of change in June from 1979 to 2022.Stippled areas indicate areas where statistical significance exceeds the 95% confidence level.Stream lines in (c) denote trends in wind directions, based on the changing rates of −w and zonally 7-degree high-pass filtered eastward wind.(d) Climatological SST (black line, left Y-axis) and SST changing rate (red line, right Y-axis) along the zonal section at 27.5°N in June from 1979 to 2022; thick red lines indicate areas where the statistical significance exceeds the 95% confidence level.(e-g) Rates of change in (e) −w at 300 hPa, (f) ∆(SLP) and (g) ∆(−SST) in June from 1979 to 2022; hatched areas indicate areas where the statistical significance exceeds the 95% confidence level; contours denote the climatological SST (unit: °C) in June.
the overlying near-surface wind speed (FigureS2cin Supporting Information S1), which may have reduced the evaporation.