Previous studies have shown that the variation of SWEA is characterized by seasonal northward (southward) migration and weakened (strengthened) intensity from winter to summer (summer to winter), influencing the position and intensity of the rain belt in eastern China (Lin and Lu, 2005). In this section, we try to discuss the variation of position, intensity, and period of the WP on different scales within the SWEA and their association with the summer rainfall in eastern China.
Based on the SWEA position index definition given in Section 2.2, the 56-yr mean daily SWEA position indexes for the WP on different scales are calculated from 1 June to 31 August, to analyze the daily evolution of the SWEA position index and the meridional displacement of the SWEA in association with the typical rainy season in eastern China. The solid blue and red lines in Fig. 3a illustrate daily evolution of the SWEA position index for the total WP and the planetary-scale WP, respectively. It is clear that the SWEA for the total WP and the planetary-scale WP synchronously and steadily moves from south to north and reaches the northernmost point of 43°N by the end of July, and then slowly retreats to the south again. Similar to the results in Lin and Lu (2009) about the relationship between the SWEA position and the rain belt in boreal summer in eastern China, the daily evolution of the SWEA position corresponds to seasonal variation of the rain belt, which also steadily moves from south in early June to north in late July and then slowly retreats to the south (Fig. 3c). Specifically, the positions of the SWEA for the total WP and the planetary-scale WP swing between 35° and 36°N during 1–16 June, corresponding to the pre-summer rainy season in South China when the WP in the SWEA is located to the north of the main rain belt by about 12 latitudes. The position of the SWEA for the total and planetary-scale WPs slowly shifts northward from 37° to 39°N during 16 June–16 July, when Meiyu rainfall occurs in the Yangtze–Huaihe region. During the Meiyu period, the position of the WPs in the SWEA is located to the north of the main rain belt by about eight latitudes. The position of the SWEA for the total and planetary-scale WPs further moves northward again from 40°N to the northernmost position around 43°N during 17–29 July and then gradually moves southward before early August, which coincides with the rainy period over North China. During this period, the position of the WPs in the SWEA is located to the north of the main rain belt by about five latitudes. Hence, along with the meridional shift of the SWEA and rain belt in eastern China, the latitude difference between the two gradually reduces during JJA. Furthermore, the influence mechanism of the SWEA on the three typical rainy seasons over eastern China is worth a further investigation in the future.
Figure 3. (a) Daily 56-yr mean position index (°N) for different scale perturbations in the SWEA from 1 June to 31 August; (b) annual correlation coefficients of the SWEA position index for the total WP with the position indexes for the planetary- and synoptic-scale WPs during 1960–2015 [the coefficients greater (less) than 0.205 (–0.205) denoted by the straight black lines are statistically significant at the 0.05 level from the two tailed t-test]; and (c) daily 56-yr mean precipitation (mm) in eastern China (110°–120°E) from 1 June to 31 August. The three red frames from the bottom to top respectively represent the pre-flood season in South China, the Meiyu season in the Yangtze–Huaihe region, and the rainy season in North China.
The SWEA position index for the planetary-scale WP (solid red line in Fig. 3a) displays a variation pattern similar to that of the position index for the total WP. However, the position index for the synoptic-scale WP (solid green line in Fig. 3a) oscillates within 38°–40°N, and does not show significant meridional displacement in JJA. Therefore, it is deduced that the meridional movement of summer rain belt in eastern China is mainly dominated by the SWEA position of the planetary-scale WP. In addition, it is worth noting that the position indexes for the total, planetary-scale, and synoptic-scale WPs coincide during 29 June–16 July. This period corresponds to the Meiyu period in the Yangtze–Huaihe region, which seems to imply that the combined effect of the planetary- and synoptic-scale waves along the SWEA is more significant in the Meiyu period over the Yangtze–Huaihe region. It may also possibly explain why the intensity of the SWEA during the Meiyu period in the Yangtze–Huaihe region is stronger than the summer-mean SWEA intensity (Jin et al., 2012).
Annual correlation coefficients of the SWEA position index for the total WP with those for the planetary- and synoptic-scale WPs at 200 hPa in JJA during 1960–2015 are shown in Fig. 3b. The coefficients greater than 0.205 are statistically significant at the 0.05 level by the two tailed t-test. The correlation coefficient of the SWEA position index between the total and planetary-scale WPs is positive with large values. Overall, there exists a significant interannual variation in the correlation coefficient between the total and planetary-scale WPs. The largest correlation coefficient of 0.97 appears in 1967, while the smallest correlation coefficient of 0.73 appears in 2001. Moreover, the correlation coefficient is above 0.8 in 49 out of the total 56 years. In contrast, the correlation of the SWEA position index between the total and synoptic-scale WPs is comparably weak as the correlation coefficient varies between –0.2 and 0.4 and does not pass the significance test in 30 out of the total 56 years. However, it can still be concluded that the contribution of synoptic-scale to total WPs of the SWEA is in general positive because the negative correlation coefficients are generally small and do not pass the significance test. Similar conclusion is also reported in other studies. For example, Xiang and Yang (2012) verified the positive feedback of transient vorticity forcing on the position variation of the EASJ, i.e., transient perturbations always make the EASJ located more northward or southward. This result is now confirmed by the positive feedback of the synoptic-scale WP to the SWEA position obtained in this study.
In summary, considering both the climatological mean and the daily evolution, the SWEA position is dominated by quasi-stationary wave associated with the planetary-scale WP, which also influences the position of the summer rain belt in eastern China. The contribution of the synoptic-scale WP to the position of the SWEA is comparably small. Meridional displacement of the SWEA on different scales vibrates within the latitude range of meridional displacement of westerly activity, presenting different amplitudes and corresponding oscillation periods. Based on this conclusion, Section 4.3 will further discuss the period characteristics of the SWEA perturbation on different scales.
Based on the SWEA intensity index definition, the 56-yr mean daily SWEA intensity indexes for different scale perturbations are calculated from 1 June to 31 August, to analyze the daily evolution and the intensity change of the SWEA in association with the typical rainy season in eastern China. The daily evolution of the SWEA intensity index for the total WP (blue solid line in Fig. 4a) shows that the SWEA intensity for the total WP steadily weakens and reaches the weakest by the end of July, and then slowly intensifies again. The SWEA intensity index for the planetary-scale WP (red solid line in Fig. 4a) changes simultaneously with that for the total WP. The planetary-scale WP is positively correlated with the total WP (Fig. 4b). The correlation coefficients are greater than 0.8 in 55 out of the total 56 years. The smallest correlation coefficient of 0.78 appears in 2002, while the largest correlation coefficient of 0.97 appears in 2015. The SWEA intensity index for the synoptic-scale WP oscillates with small amplitude within the range of 4–6 m s–1, and is weakly and negatively correlated with that for the total WP. There are only 7 out of the total 56 years when the correlation coefficients pass the significance test. The biggest negative correlation coefficient of –0.42 appears in 1981 (Fig. 4b).
Figure 4. (a) Daily 56-yr mean intensity index (m s–1) for the different scale perturbations in the SWEA from 1 June to 31 August, and (b) annual correlation coefficients of the SWEA intensity index for the total WP with those for the planetary- and synoptic-scale WPs during 1960–2015.
As shown in Figs. 3a, 4a, following the northward (southward) movement of SWEA for the total and planetary-scale WPs, their perturbation intensities weaken (enhance) correspondingly, whereas the intensity of the synoptic-scale WP enhances (weakens) and the intensity difference between the total and planetary-scale WPs increases (decreases). In addition, the configurations of the position and intensity of the planetary- and synoptic-scale WPs are different during different rainy periods in eastern China. Specifically, during the rainy period (late July to early August) over North China when the SWEA position is in the northernmost point (above 40°N), the intensity of the planetary-scale WP in the SWEA is weak within 10–12 m s–1, while the intensity of the synoptic-scale WP in the SWEA is the strongest with values greater than 6 m s–1. This indicates that the synoptic-scale wave exerts more important influence during this rainy period than that during other rainy periods. During the Meiyu period (16 June–16 July) over the Yangtze–Huaihe region, the situation is reversed. As the SWEA slowly moves to around 37°–39°N, the intensity of the planetary-scale WP in the SWEA varies within the range of 15–20 m s–1, while the intensity of the synoptic-scale WP is in its weak phase with values of around 5 m s–1. Apparently, the planetary-scale wave is dominant during this rainy period. Similar results are also found in the study of Yang and Zhang (2007) from the perspective of energy disturbance. They demonstrated that the EASJ is southward (northward) of the mean and stronger (weaker) than usual when the perturbation kinetic energy enhances (weakens).
The study of Lu et al. (2013) indicated that the upper-level jet stream in summer over East Asia tends to move southward with a tendency of increased intensity in the recent 50 years, but the intensity of the westerly has shown a weakening trend since the middle 1990s. As shown in Figs. 4a, b, the correlation of the intensity of the planetary- and synoptic-scale WPs with the intensity of the total WP in the SWEA also reveals different characteristics before and after the 1990s with a decreased correlation. Further investigation should be conducted to study whether the interdecadal variation of the SWEA position and intensity is related to the wave activities caused by the planetary- and synoptic-scale WPs in the SWEA.
Perturbations of the subtropical westerly influencing Meiyu in the Yangtze–Huaihe region are associated with both the planetary- and synoptic-scale WPs in the SWEA. Existing studies have shown that perturbations on different scales exert different impacts on Meiyu. Yang and Zhang (2007) claimed that during the year of strong disturbance of Rossby waves along the SWEA, the upper, middle, and lower tropospheric circulations and vertical velocity of the whole layer are all favorable for the strengthening of tropical monsoon circulation in East Asia and reinforcement of the Meiyu front as well as the southward displacement of the northwestern Pacific subtropical high (NPSH). In their case study of the anomalous Meiyu season of 1998, Mei and Guan (2009) found that there existed quasi-stationary wave trains (planetary-scale perturbations) that served as the background field for high-frequency waves (synoptic-scale perturbations) during the Meiyu period. Originating from the regions to the east of the Caspian Sea, the baroclinic waves in the upper troposphere were organized into the Rossby wave packets during the 1998 Meiyu period and then propagated downstream along the SWEA. The energy dispersion associated with the baroclinic wave packets led to the accumulation of energy for rainstorms in the Yangtze River valley.
In this section, we try to reveal the periodic oscillation characteristics of the SWEA perturbations on different scales, and answer the following questions: What are the exact oscillation periods for the positions of different scale WPs in the SWEA? What is the configuration of these periods for the positions of different scale WPs in the SWEA during the typical rainy season over eastern China?
By means of wavelet transform, the temporal variation of daily SWEA position index for 56-yr mean 200-hPa U-wind field (original field), total WP, planetary-scale WP, and synoptic-scale WP can be filtered on the time frequency domain from 1 June to 31 August. As shown in Fig. 5, three wavelet variance peaks are detected for the above four position indexes on the frequency domain during the 46 days. Wavelet variance curves of the original field, the total WP, and the planetary-scale WP agree well with the first two peaks occurring on 38 and 21 days, respectively. The third peak of the wavelet variance for the mean field occurs on 10 days, while the third peak for the total WP and the planetary-scale WP occurs on 13 days. The synoptic-scale WP presents a unique characteristic of the oscillation frequency, with the main period peaking on 16 days and the other two periods on 8 and 28 days, respectively. Actually, the periods of 16 and 28 days do not have any physical significance for the synoptic-scale WP of the SWEA, and the periods of 21 and 38 days make no sense for other perturbations as well.
Figure 5. Wavelet variances of the SWEA position index for 56-yr mean 200-hPa U-wind field (original field), total WP, planetary-scale WP, and synoptic-scale WP from 1 June to 31 August. The results have passed the red noise test at the 0.05 significant level.
Table 1 shows the first three main periods of annual position index for the different scale WPs and their corresponding statistical frequencies during 1960–2015. The first three main periods correspond to the three most frequently occurred cycles during the 56 years. It is clear that the first three periods of the original field are 8, 15, and 13 days, respectively. The first three periods of the total WP and the planetary-scale WP are 12, 13, and 14 days, respectively. The first three periods of the synoptic-scale WP are 5, 6, and 7 days, respectively. This reveals that the periods of the original field are actually the comprehensive results of the synoptic-scale WP superimposed on the planetary-scale WP. The temporal variation of the total WP exhibits quasi-biweekly (13 days) oscillation, consistent with that of position of the planetary-scale WP. Meanwhile, weekly (6 days) oscillation is the main period of the synoptic-scale WP. It can be probably inferred that quasi-biweekly (13 days) oscillation is significant in the original field due to the impact of the planetary-scale WP during its strongest stage, whereas weekly (6 days) oscillation becomes a bit more important when the synoptic-scale WP enhances from late July to early August.
Original field Total WP Planetary-scale WP Synoptic-scale WP Top three periods (day) 8/15/13 14/13/12 12/13/14 6/5/7 Frequency (yr) 13/10/9 12/12/11 14/10/8 15/14/12
Table 1. The first three main periods of annual position index for the different scale WPs and their corresponding statistical frequencies during 1960–2015 (periods illustrated are at the 0.05 significance level)
Previous studies revealed that the typical rainy season in boreal summer in China also evolves with significant periodic variation characteristics. For instance, quasi-biweekly oscillation is significant during the Meiyu (e.g., Lau et al., 1988; Chen et al., 2000). The high-frequency precipitation episodes are characterized with obvious 3–8-day oscillation, which continues from the very beginning to the end of the rainy season in North China (Liu, 2009). Such obvious periodic characteristics of precipitation in the typical rainy season are apparently related to impacts of the SWEA perturbations, especially the WPs on planetary and synoptic scales.
Hence, these results can provide references for medium-range forecasting of large-scale precipitation in boreal summer in East Asia. For example, during the Meiyu season in the Yangtze–Huaihe region when the SWEA moves northward to the geographic location around 37°–39°N, quasi-biweekly (13 days) oscillation is dominant because the planetary-scale WP contributes to the major variation of the SWEA. During the rainy period of North China when the SWEA moves to 40°N, besides the quasi-biweekly oscillation, weekly (6 days) oscillation should also be considered because the synoptic-scale WP intensifies during this period.
|Original field||Total WP||Planetary-scale WP||Synoptic-scale WP|
|Top three periods (day)||8/15/13||14/13/12||12/13/14||6/5/7|