Many studies have shown that the northward (southward) displacement of the East Asian westerly jet (EAWJ) drastically reduces (increases) summer rainfall in the Yangtze River valley (YRV). However, the effect of the jet’s intensity on interannual variation in summer rainfall has not been systematically studied. The present study investigates the effect of the EAWJ’s intensity on this interannual variation and analyzes the mechanism by which this process occurs. In early summer, the EAWJ consists of two branches: one located over northern continental East Asia [western branch (EAWJWB)] and one extending from southern China to the northern Pacific [eastern branch (EAWJEB)]. The former merges into the latter over the Yellow Sea. A stronger EAWJEB leads to increased rainfall in the YRV, while the EAWJWB does not significantly affect rainfall in the YRV. The faster EAWJEB directly strengthens midtropospheric warm advection over the YRV because the corresponding changes in the meridional wind and horizontal temperature gradient are insignificant. The strengthened warm advection increases rainfall in the YRV by accelerating both adiabatic ascent and the ascent associated with diabatic heating primarily generated by convection. In midsummer, the EAWJ has no branches and is located over the midlatitudes of Asia. The strengthening of the EAWJ reduces rainfall in the YRV in midsummer through the Pacific–Japan (PJ) pattern. As the EAWJ strengthens, the PJ pattern turns to its positive phase. This results in the deceleration of the midtropospheric westerly wind and a reduction in the meridional temperature contrast, which weakens midtropospheric warm advection. The weakened warm advection in turn reduces rainfall in the YRV, following the process outlined for early summer.
Eastern China, especially the Yangtze River valley (YRV; 27°–33°N, 110°–120°E in the present study), is one of the most populated and developed regions in China. Annual precipitation in this region, particularly summer precipitation, is closely related to the regional economic development. Frequent floods and droughts often cause massive losses in the region. For example, both the severe drought in the summer of 2013 and the continuous heavy rainfall in 1998 caused severe damage to properties throughout the YRV. These disasters are usually the result of the high variability of the East Asian summer monsoon. Therefore, the effect of the East Asian summer monsoon on rainfall in eastern China, especially in the YRV, has attracted substantial attention for many years, and many relevant reports have been generated (e.g., Neyama 1963; Zhao and Zhang 1996; Wang et al. 2013; Huang et al. 2013).
The variation of the East Asian westerly jet (EAWJ), a component of the East Asian summer monsoon, drastically affects summer rainfall in the YRV, and the jet’s meridional displacement has the largest impact. When the EAWJ shifts southward, the rainband over the YRV also shifts southward (e.g., Fang et al. 2009; Dong et al. 2011; Liang and Wang 1998), favoring increased rainfall in the YRV (e.g., Kuang and Zhang 2006; Xuan et al. 2011). Lu (2004) found that the southward movement of the EAWJ strengthens the convective activities along the rainband over eastern East Asia, which may lead to increased rainfall in the YRV. The zonal position of the jet core also affects the rainfall in this area. Existing research indicates that in years when the EAWJ’s core shifts westward earlier than the jet’s axis jumps northward, the process of the northward movement of the rainband clearly differs from that in other years (Dong et al. 2011). Recently, the correlation between the EAWJ’s intensity and rainfall in the YRV has attracted the attention of a number of researchers. Li and Zhang (2014) found that on a daily scale, the changing configuration of the polar and subtropical jets’ intensities affects both the intensity and the position of the rainband in the YRV from late June to early July. Lu (2004) observed that the strengthening of the EAWJ in July enhances convective activities along the rainband over East Asia, while in August, it has the opposite effect. In addition, the EAWJ connects other climate variables and rainfall in the YRV. When the Arctic Oscillation has a positive polarity, the EAWJ is pushed north, which in turn causes a drier summer in the YRV, and vice versa (Gong and Ho 2003). An El Niño event in winter and spring affects rainfall in eastern China during the following summer by shifting the EAWJ’s meridional position (Liang and Wang 1998).
Despite the attention given to the effect of the EAWJ’s variation on summer rainfall in the YRV, most previous studies have focused on the role of the jet’s position. A systematic analysis of the effect of the EAWJ’s intensity on the interannual variation of summer rainfall in the YRV is still needed. Such an analysis is performed in this work, with emphasis on the jet in the middle troposphere. The choice to study the midtropospheric jet was made based on the following consideration. In the upper troposphere, the level on which most previous studies regarding the EAWJ have focused, the horizontal wind is drastically influenced by condensation heating generated by rainfall in eastern China (Tao and Chen 1987; Zhang et al. 2006). As a result, it is difficult to determine the causal relationship between the jet’s intensity and YRV rainfall. Sampe and Xie (2010) noted that the diabatic heating in the middle troposphere has little effect on the horizontal wind at 500 hPa, and the circulation at this level generates large-scale environmental forcing on the rainband in eastern China independent of the local diabatic heating feedback. Therefore, a study of the EAWJ’s intensity in the middle troposphere makes it easier to identify the causality and understand the role of large-scale circulation and associated climate variables [e.g., sea surface temperature (SST)] in shaping YRV rainfall.
We also analyzed the dynamic process by which the EAWJ’s intensity affects YRV rainfall. Climatologically, the EAWJ advects warm air from the Tibetan Plateau eastward at 500 hPa in early summer (from mid-June to mid-July), generating a midtropospheric warm advection center over the YRV. This in turn induces upward motion and anchors the rainband in the YRV (Sampe and Xie 2010). Kosaka et al. (2011) generalized this process and noted that the intensity and meridional position of the midtropospheric warm advection center over the YRV determine the intensity and meridional position of the rainband in both early and midsummer (from mid-July to mid-August). The present study investigates how the midtropospheric warm advection center over the YRV connects the midtropospheric jet’s intensity to rainfall in the YRV.
The rest of this article is organized as follows. Section 2 describes the data and methods used in the study. Section 3 analyzes the effect of the EAWJ’s intensity on the interannual variation of YRV rainfall in early and midsummer. A further investigation of the dynamic process connecting EAWJ intensity and YRV rainfall is presented in section 4, followed by brief conclusions and additional discussion in section 5.
2. Data and methods
The 6-hourly horizontal wind velocity, vertical velocity, and air temperature data were obtained from the ERA-Interim reanalysis dataset (Dee et al. 2011), and daily interpolated outgoing longwave radiation (OLR) data were obtained from the Earth System Research Laboratory (Liebmann and Smith 1996) for use in the present study. The horizontal resolution of the data is 2.5° latitude × 2.5° longitude, and daily data for the ERA-Interim variables were calculated by averaging the 6-hourly data at 0000 and 1200 UTC each day. Rainfall observations from 756 stations were provided for the period from 1951 to 2009 by the China Meteorological Data Sharing Service System.
Climatologically, the rainband suddenly weakens and shifts northward over the YRV and North China in mid- to late July (Ding and Chan 2005; Huang et al. 2008), and the circulation simultaneously and abruptly changes over East Asia (Zhang et al. 2006; Sampe and Xie 2010). This suggests that EAWJ intensity may affect YRV rainfall differently at different periods during the summer. Therefore, we split the season into early summer (16 June–15 July) and midsummer (16 July–14 August), and we analyze the effects during the two periods separately. Existing studies indicate that many climate variables and their interrelationships exhibited a notable regime shift in approximately 1979 (e.g., Gong and Ho 2002; Wu et al. 2003; Wang and He 2012). The ERA-Interim reanalysis dataset covers the period since 1979. Thus, this study focuses on patterns that occurred in the 31 years from 1979 to 2009. Average daily rainfall in the YRV (hereafter referred to as ARY) is calculated by averaging the interpolated station rainfall with 1° × 1° horizontal resolution over the YRV (27°–33°N, 110°–120°E) prior to calculating the 30-day accumulation for early summer or midsummer.
The jet axis occurrence frequency was statistically analyzed in the area surrounding East Asia. If a specific point satisfies the following conditions, the location of the point is identified as a jet axis: 1) the algebraic value of the zonal wind velocity exceeds 6 m s−1 and 2) the zonal wind velocity at the point is larger than those at the two adjacent points along the same longitude of the specific point. The choice of the westerly velocity threshold (6 m s−1) used to identify the presence of a jet axis is based on the fact that westerly winds are weak along the jet axis through summer, and the maximum zonal wind velocity barely exceeds 6 m s−1 along some longitudes over East Asia (Fig. 1). The jet axis occurrence frequency is calculated as the number of years for which a jet axis occurs at a selected point divided by the total number of years (31) in the study period (expressed as a percentage).
3. The effect of EAWJ intensity on YRV rainfall
a. Climatology of the EAWJ
This section describes the climatology of the EAWJ (Figs. 1 and 2) to identify the jet’s climatological-mean position and structure. The EAWJ consists of two branches in early summer (Fig. 1a, contours and vectors). One branch, hereafter referred to as the western branch (EAWJWB), is located on the northern flank of the Tibetan Plateau and the Loess Plateau. The jet axis is centered on 42.5°N. The other branch, hereafter referred to as the eastern branch (EAWJEB), forms along the northern rim of the North Pacific subtropical high. It originates at approximately 30°N, 110°E, and its maximum zonal wind velocity exceeds 15 m s−1 and occurs east of Japan. The EAWJWB merges into the EAWJEB over the Yellow Sea. In midsummer, the EAWJ has no branches (Fig. 1b, contours and vectors). It extends from northern continental East Asia to the North Pacific and is centered near 45°N. Figure 1 also shows the jet axis occurrence frequency at 500 hPa (shaded). In early summer, two regions with high occurrence frequency exist over East Asia, and these regions correspond well with the locations of the EAWJEB and EAWJWB (Fig. 1a, contours). In midsummer, only one region with high occurrence frequency exists over East Asia, which corresponds well to the EAWJ (Fig. 1b, contours). Figure 2 shows the vertical structure of EAWJ. Its characteristics are similar to that of Sampe and Xie (2010), which pointed out that the jet extends from the lower though middle to upper troposphere both in early summer and midsummer.
b. Effects of the EAWJ’s intensity on the rainfall in YRV
The zonal wind velocity anomalies at 500 hPa associated with early and midsummer rainfall in the YRV are illustrated in Fig. 3 to demonstrate the effect of EAWJ intensity on rainfall in the YRV. In early summer, a significant positive anomaly center and a significant negative anomaly center (Fig. 3a) are observed south and north of the climatological-mean EAWJEB axis (Fig. 1a), respectively. The two centers form a meridional dipole of the zonal wind anomalies centered around the climatological-mean EAWJEB. Thus, the southward displacement of EAWJEB is closely related to increased rainfall in the YRV, as suggested in previous studies (e.g., Kuang and Zhang 2006; Xuan et al. 2011). Comparing Figs. 3a and 1a also reveals that positive zonal wind anomalies occur along the climatological-mean EAWJEB axis and that they are significant at the 0.95 confidence level. Furthermore, it is notable that the dipole of the zonal wind anomalies is asymmetric: the negative value center is substantially farther from the EAWJEB axis and is weaker than the positive value center (Fig. 3a). These findings indicate that the intensified EAWJEB is related to increased rainfall in the YRV. For the EAWJWB, the zonal wind velocity anomalies (Fig. 3a) are not significant along the climatological-mean axis of the EAWJWB (Fig. 1a), implying that the relationship between the intensity of the EAWJWB and rainfall in the YRV is not significant. In midsummer, anomalous increased rainfall in the YRV is associated with negative zonal wind velocity anomalies (significant at the 0.95 confidence level) observed in northern continental East Asia (Fig. 3b), where the climatological-mean jet axis is located (Fig. 1b). It is also notable that the negative anomaly’s center is located slightly north of the jet axis (Figs. 1b and 3b). These observations imply that the weakening and slight southward displacement of the EAWJ result in the increased rainfall observed in the YRV in midsummer.
Based on the significant zonal wind anomalies associated with ARY and the location of the jet, as determined by the jet axis occurrence frequency and the zonal wind velocity, two regions over 30°–37.5°N, 110°–150°E and 37.5°–47.5°N, 65°–120°E are identified as key locations of the EAWJEB and EAWJWB during early summer. The average zonal wind velocities at 500 hPa over the two sites are defined as the intensity indices for the EAWJEB (EJI) and EAWJWB (WJI). In midsummer, one key location (40°–50°N, 90°–150°E) is identified, and the average zonal wind velocity at 500 hPa over this area is defined as the intensity index for the EAWJ (JI).
The phenomenon revealed by regression analysis can also be obtained from correlation analysis. The correlation coefficient between the EJI and ARY is 0.43 in early summer, significant at the 0.95 confidence level. When both the EJI and ARY are detrended, the correlation coefficient increases slightly to 0.46, which is significant at the 0.99 confidence level. The correlation coefficient between the WJI and ARY is −0.15 (−0.15) before (after) detrending and is not significant, even at the 0.9 confidence level. In midsummer, the correlation coefficient between ARY and JI is −0.65 (−0.66) before (after) detrending, which is significant at the 0.99 confidence level.
The rainfall anomalies in eastern China associated with jet intensity indices for early and midsummer are shown in Figs. 4 and 5, respectively. These diagrams confirm the impact of the jet intensity on YRV rainfall. In early summer, significant positive EJI-associated rainfall anomalies are concentrated in a zonally elongated region along the climatological rainband (Figs. 4a and 4c), indicating that the strengthening of the EAWJEB intensifies the rainband and thereby increases rainfall in the YRV. However, few significant WJI-associated rainfall anomalies are observed along the climatological rainband (Figs. 4b and 4c), suggesting that the effect of EAWJWB intensity on rainfall over the YRV is not significant. In midsummer, the JI-associated negative rainfall anomalies are located along the axis and on the south flank of the climatological rainband. This reveals that the weakening of the EAWJ pushes the rainband southward and increases its intensity, resulting in increased rainfall in the YRV. This analysis confirms the effect of jet intensity on rainfall in the YRV.
Overall, these results demonstrate that a significant positive relationship exists between EAWJEB intensity and early summer rainfall anomalies in the YRV, while no significant relationship exists between EAWJWB intensity and early summer rainfall anomalies in the YRV. In midsummer, however, the relationship between jet intensity and YRV rainfall anomalies is significant and negative.
4. Dynamic processes governing the effect of jet intensity on YRV rainfall
a. Early summer
The dynamic processes governing the effect of EAWJ intensity on rainfall in the YRV are studied in this section. The horizontal temperature advection anomalies and vertical wind anomalies at 500 hPa associated with the EJI are shown in Figs. 6a and 6b, respectively. Temperature advection is calculated from the daily data before being averaged over 30 days in early summer. Anomalous warm advection and ascent associated with an anomalously strong EAWJEB are observed in the YRV, which is significant at the 0.95 confidence level for a large portion of the YRV (Fig. 6a). We can also see that the anomalous ascent is collocated with the anomalous warm advection over the YRV (Figs. 5a, 6a, and 6b). To validate the relationship between anomalous warm advection and anomalous ascent, correlations between the two variables at the same point are analyzed. Significant negative correlations are observed over the YRV (figure not shown), indicating that stronger warm advection is related to faster ascent over the YRV.
The above results demonstrate the close relationship between jet intensity, the anomalous midtropospheric temperature advection, and the anomalous vertical velocity. Two issues remain unclear: how jet intensity affects the midtropospheric temperature advection and how variations in the midtropospheric temperature advection affect the vertical velocity over the YRV. We aim to resolve these issues in the rest of section 4a.
The horizontal temperature advection is calculated as
where T is the air temperature, u is the zonal wind velocity, and υ is the meridional wind velocity. The variation of horizontal temperature advection is clearly controlled by the zonal wind–temperature gradient and the meridional wind–temperature gradient.
Climatologically, southwesterly winds prevail in the middle troposphere over the YRV, and the air temperature is low in the northeast and high in the southwest (figure not shown). This configuration of air temperature and horizontal wind generates both zonal and meridional warm advection and, thus, the midtropospheric warm advection band over the YRV in early summer. The westerly anomalies at 500 hPa associated with the anomalously strong EAWJEB are significant over the YRV (Fig. 7a), while the zonal temperature gradient anomalies at 500 hPa are not significant over the YRV (Fig. 7c). According to (1), these anomalies favor stronger midtropospheric zonal warm advection (Fig. 7e). Both the meridional wind velocity anomalies at 500 hPa and the meridional temperature gradient anomalies are not significant over most of the YRV (Figs. 7b and 7d), which reveals that the effect of the EAWJEB intensity on midtropospheric meridional temperature advection is not significant (Fig. 7f). Thus, the strengthening of the EAWJEB directly enhances the midtropospheric warm advection over the YRV.
The thermodynamic equation in the isobaric coordinate system can be written as
where T is the air temperature; J is the rate of diabatic heating per unit mass from radiation, conduction, and latent heat release; ω is the vertical velocity for the isobaric system; cp is the specific heat of dry air at constant pressure; and Sp is the static stability parameter for the isobaric system (Holton 2004). The local change rate of the air temperature is negligible on the seasonal time scale. Equation (2) can be rewritten as
The variables Sp, H, and ω are now divided into constant basic-state portions (denoted by overbars) and perturbation portions (denoted by primes):
It can also be rewritten as
The standard deviation of the daily data of the static stability parameter for the 31-yr study period is an order of magnitude smaller than the climatology of the static stability parameter over the YRV, while the standard deviation of the daily data of vertical velocity and the climatology of vertical velocity share the same order of magnitude (figures omitted). Based on these observations, we can obtain the following:
We can then approximate (6) by
Recalling that , , and are all constants and there are only two variables (H′ and ω′) in (8), we can obtain that the variance of vertical velocity is primarily controlled by that of H and the variance of static stability is negligible. The standard deviation of the daily data of horizontal temperature advection and diabatic heating share the same order of magnitude (figures omitted). Here diabatic heating has been derived as a residual of (2) with daily data for the ERA-Interim variables and the same below. This result suggests that the variance of horizontal temperature advection and diabatic heating together determine the variance of H. Thus, the variance of vertical velocity is primarily determined by the variance of horizontal temperature advection and diabatic heating over YRV.
Enhanced warm advection can adiabatically accelerate upward motion. Whether diabatic heating is related to the identified horizontal temperature advection is investigated. The correlation coefficient between diabatic heating and YRV-mean adiabatic vertical velocity (ADV/Sp, controlled primarily by horizontal temperature advection) is significantly negative over the YRV (Fig. 8). Thus, when the adiabatic ascent is faster (or the warm advection is stronger), the diabatic heating over the YRV is increased. To identify the causal relationship between anomalous adiabatic ascent and diabatic heating, lead–lag correlations between the daily fluctuations of diabatic heating and YRV-mean adiabatic vertical velocity are calculated (Fig. 9). It is clear that the negative lead–lag correlation in the high correlation coefficient region surrounding the YRV when adiabatic vertical velocity leads by one day is equivalent to the contemporaneous correlation and is much stronger than when it lags. This difference suggests that faster adiabatic ascent causes increased diabatic heating over the YRV. The difference in the location of the high correlation region surrounding the YRV when adiabatic vertical velocity leads, lags, or is concurrent with diabatic heating reveals the eastward movement of the weather systems surrounding the YRV (Figs. 9a, 9b, and 9c). Many existing studies used OLR to characterize the intensity of convective activities (e.g., Knutson et al. 1986; Lu 2004; Gadgil et al. 2004). If we substitute OLR for diabatic heating and repeat the above analyses, similar results are obtained (figures omitted). Overall, these results reveal that stronger warm advection intensifies convection over the YRV by inducing faster adiabatic ascent, which helps trigger convection by lifting air parcels. Additionally, the intensified convection enhances the faster ascent through the resultant increase in diabatic heating.
Figures 8 and 9 illustrate that the area with a strong negative correlation between diabatic heating and YRV-mean adiabatic vertical velocity is centered slightly north of the YRV. This finding is consistent with the deviation of the vertical wind anomaly’s center from the temperature advection anomaly’s center (Fig. 6). A similar deviation is also observed if OLR replaces diabatic heating. The effect of the interaction between cold and warm air over the YRV may cause this difference (e.g., Li and Zhang 2014), although this issue requires further study.
The above analyses reveal the mechanisms between the EAWJEB’s intensity and the rainfall in the YRV. Similar analyses for the EAWJWB are presented as follows (figures omitted). The zonal wind anomalies and zonal temperature gradient anomalies at 500 hPa associated with the anomalously stronger EAWJWB are not significant over most of the YRV. Thus, zonal temperature advection anomalies are not obvious over the YRV. Although significant southerly and meridional temperature gradient weakening anomalies at 500 hPa are observed over the YRV, the anomalies of the two variables offset each other’s effects on meridional temperature advection over the YRV. Thus, midtropospheric horizontal temperature advection anomalies associated with anomalously stronger EAWJWB and the resultant rainfall anomalies are not significant over the YRV.
The regression analysis of the EAWJEB in early summer was repeated for the EAWJ in midsummer. Significant cold advection and descent anomalies at 500 hPa associated with the anomalously strong EAWJ are observed over the YRV (Fig. 10), and the descent anomalies lead to reduced rainfall in the YRV (Fig. 5a). The anomalous descent is almost collocated with the anomalous cold advection and rainfall decrease over the YRV. The correlation between vertical velocity and horizontal temperature advection at 500 hPa is significantly negative over the YRV, indicating that the cold advection anomalies are related to the descent anomalies over the YRV (figure omitted). At this point, the issue of how midtropospheric temperature advection over the YRV is correlated with EAWJ intensity in the midlatitudes remains unresolved.
The relationship between relative vorticity and the intensity of the EAWJ is investigated through regression analysis to identify the possible systems connecting the EAWJ’s intensity to circulation over the YRV. A west–east-oriented band of significant cyclonic circulation anomalies associated with the anomalously strong EAWJ is centered at approximately 52.5°N (Fig. 11a). To its south, a band of significant anticyclonic circulation anomalies is centered at approximately 42.5°N. Another west–east-oriented band of significantly cyclonic circulation anomalies is located at approximately 26°N. These three bands form a meridional tripole of vorticity anomalies. This anomaly tripole extends into the lower troposphere, where the pattern is not as clear (Fig. 11b, the 500-hPa pattern is similar to that at 200 hPa and is thus omitted). At this location, a notable displacement toward the equator relative to its upper-tropospheric counterpart exists. These phenomena are characteristic of the Pacific–Japan (PJ) pattern in its positive phase (Nitta 1987; Xue et al. 2004; Kosaka et al. 2011). These results suggest that the intensified EAWJ is closely correlated with the positive-phase PJ in midsummer. Significant easterly anomalies at 500 hPa associated with the anomalously strong EAWJ are observed over the YRV; these anomalies are located between the anomalous anticyclonic circulation north of the YRV and the anomalous cyclonic circulation south of the YRV (Figs. 12a and 11a). Thus, an intensified EAWJ weakens westerly winds over the YRV via the positive-phase PJ pattern. Significant cold (warm) anomalies are almost collocated with the anomalous cyclonic (anticyclonic) circulation south (north) of the YRV (Figs. 12c and 11a), which is a feature of the barotropic structure of the PJ pattern. The temperature anomalies surrounding the YRV reduce the meridional temperature gradient over the YRV (Figs. 12b and 12c). The zonal temperature gradient anomalies and the meridional wind velocity anomalies are not significant over the YRV (figures omitted). Thus, according to (1), the anomalous weakening of the westerly and meridional temperature gradients, which is related to the anomalously strong EAWJ via the positive-phase PJ pattern, weakens the midtropospheric warm advection over the YRV, in recognition of the fact that the climatogical-mean air temperature is low in the east and high in the west and climatogical-mean southerly wind prevails over the YRV (figure omitted).
One may ask why the intensity of the EAWJWB, which is also located in the midlatitudes, is not closely related to the temperature advection over the YRV in early summer. This question can be answered by the WJI-related relative vorticity anomalies at 200 and 850 hPa (figures omitted). The anomalies’ features do not resemble the PJ pattern in early summer (Kosaka et al. 2011), suggesting that the intensity of the EAWJWB is not closely related to the PJ pattern. Thus, the PJ pattern does not connect the EAWJWB’s intensity and the temperature advection over the YRV in early summer. In midsummer, the wavelike circulation anomalies along coastal East Asia associated with the PJ pattern extend from 100° to 160°E (Kosaka et al. 2011), where most of the EAWJ is located. In early summer, however, the PJ-related wavelike circulation anomalies are located east of 110°E (Kosaka et al. 2011), and most of the EAWJWB is not in that region. This may be the reason why the intensity of the EAWJWB is not closely related to the PJ pattern in early summer.
5. Conclusions and discussion
The present study investigates the effect of the EAWJ’s intensity on rainfall in the YRV in early and midsummer and suggests the possible mechanism underlying this effect. In early summer, the EAWJ consists of two branches: the EAWJWB over northern continental East Asia and the EAWJEB extending from around the YRV to the east of Japan. The EAWJWB merges into the EAWJEB over the Yellow Sea. In midsummer, the EAWJ has no branches and extends from northern continental East Asia to the northeast of Japan. The regression coefficient of the zonal wind velocity at 500 hPa onto ARY suggests that an intensified EAWJEB increases rainfall in the YRV in early summer, while the intensity of the EAWJWB is not closely related to rainfall in the YRV. The regression analysis also reveals that an intensified EAWJ reduces rainfall in the YRV in midsummer. Based on the area displaying a significant regression coefficient and an analysis of the jet axis occurrence frequency and zonal wind velocity, the area-average zonal wind velocities at 500 hPa over 30°–37.5°N, 110°–150°E and 37.5°–47.5°N, 65°–120°E are defined as the intensity indices for the EAWJWB and EAWJEB, respectively. In midsummer, the average zonal wind velocity at 500 hPa over 40°–50°N, 90°–150°E is defined as the intensity index for the EAWJ. The correlation between ARY and the intensity indices and the regression coefficient of rainfall in eastern China onto the intensity indices further confirm the effect of the jet’s intensity on rainfall in the YRV.
In early summer, the intensified EAWJEB, which surrounds the YRV, directly strengthens the midtropospheric warm advection, considering the fact that the meridional wind velocity and the meridional and zonal temperature gradients do not change significantly over the YRV. The intensified warm advection not only induces a faster adiabatic ascent over YRV, but also reinforces a faster diabatic ascent by enhancing convection and intensifying the resultant diabatic heating. The faster ascent causes increased rainfall in the YRV. The intensity of the EAWJWB is not closely related to warm advection over the YRV, and thus, it is not related to rainfall in the YRV. In midsummer, the intensified EAWJ is related to the positive-phase PJ pattern. An anomalous warm anticyclone associated with the positive-phase PJ pattern is located north of the YRV, and an anomalous cold cyclone is located south of the YRV, both of which decelerate the midtropospheric westerly winds and reduce the meridional temperature gradient over the YRV. Thus, midtropospheric warm advection is weakened, which reduces the rainfall in the YRV in the same way that strengthened warm advection increases rainfall in early summer.
The present study reveals that the intensity of the EAWJ significantly affects the interannual variation of rainfall over the YRV in early and midsummer. Compared with previous studies, which focused on the meridional displacement of the EAWJ, the present study provides a new understanding of the effect of the interannual variation in the EAWJ on YRV rainfall. Furthermore, the effect of the jet’s intensity on rainfall in the YRV in early summer is distinctly different from that in midsummer, which may facilitate the theoretical analysis and practical prediction of abrupt drought–flood alterations in YRV rainfall patterns.
The present study also notes that there is a close relationship between the PJ pattern and the EAWJ, which is similar to previous researches (e.g., Kosaka et al. 2011). However, further study is needed to identify the causal relationship between the two variables. The correlation coefficient between 500-hPa vertical velocity and the YRV-mean horizontal temperature advection has an amplitude of about 0.5 both in early summer and midsummer (figure omitted), suggesting that the midtropospheric temperature advection anomalies can explain the primary part of the variance of vertical wind and the variance of rainfall over the YRV. The changes of other factors, such as lower-tropospheric warm southerly in the south and cold northerly in the north, may also play a role in causing the rainfall anomalies over the YRV, which needs further study to confirm. Previous studies suggest that many climate variables, such as diabatic heating associated with the Tibetan Plateau, the meridional tropospheric temperature gradient shift, and barotropic energy conversion, can impact the variation of the EAWJ (e.g., Kuang and Zhang 2005; Zhang and Huang 2011; Lu and Ye 2011). Further study is needed to identify the large-scale climate variables and variations affecting the intensity of the EAWJ to increase the relevance of the results found in the present study for the prediction of summer rainfall in the YRV.
The authors thank the editor and three anonymous reviewers who provided valuable comments and suggestions for improving the manuscript. This work was jointly supported by the National Basic Research Program of China (973 Program, Grant 2012CB956200) and the National Natural Science Foundation of China (Grants 41275019 and 41475009).