Advanced Search
Article Contents

Why Was the Strengthening of Rainfall in Summer over the Yangtze River Valley in 2016 Less Pronounced than that in 1998 under Similar Preceding El Niño Events?——Role of Midlatitude Circulation in August


doi: 10.1007/s00376-017-7003-8

Get Citation+

Export:  

Share Article

Manuscript History

Manuscript received: 04 January 2017
Manuscript revised: 29 April 2017
Manuscript accepted: 26 May 2017
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Why Was the Strengthening of Rainfall in Summer over the Yangtze River Valley in 2016 Less Pronounced than that in 1998 under Similar Preceding El Niño Events?——Role of Midlatitude Circulation in August

  • 1. Center for Monsoon System Research, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China
  • 2. State Key Laboratory of Numerical Modelling for Atmospheric Sciences and Geophysical Fluid Dynamics, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China
  • 3. Climate Change Research Center, Chinese Academy of Sciences, Beijing 100029, China

Abstract: It is widely recognized that rainfall over the Yangtze River valley (YRV) strengthens considerably during the decaying summer of El Niño, as demonstrated by the catastrophic flooding suffered in the summer of 1998. Nevertheless, the rainfall over the YRV in the summer of 2016 was much weaker than that in 1998, despite the intensity of the 2016 El Niño having been as strong as that in 1998. A thorough comparison of the YRV summer rainfall anomaly between 2016 and 1998 suggests that the difference was caused by the sub-seasonal variation in the YRV rainfall anomaly between these two years, principally in August. The precipitation anomaly was negative in August 2016——different to the positive anomaly of 1998. Further analysis suggests that the weaker YRV rainfall in August 2016 could be attributable to the distinct circulation anomalies over the midlatitudes. The intensified "Silk Road Pattern" and upper-tropospheric geopotential height over the Urals region, both at their strongest since 1980, resulted in an anticyclonic circulation anomaly over midlatitude East Asia with anomalous easterly flow over the middle-to-lower reaches of the YRV in the lower troposphere. This easterly flow reduced the climatological wind, weakened the water vapor transport, and induced the weaker YRV rainfall in August 2016, as compared to that in 1998. Given the unique sub-seasonal variation of the YRV rainfall in summer 2016, more attention should be paid to midlatitude circulation——besides the signal in the tropics——to further our understanding of the predictability and variation of YRV summer rainfall.

摘要: 一般认为, 在厄尔尼诺衰退年, 长江流域夏季降水会明显增强, 比如1998年灾难性的大洪水. 然而, 同样是超级厄尔尼诺衰退年, 2016年夏季降水则较1998年明显偏弱. 本文研究发现它们的差异主要来自长江流域降水的次季节(主要是8月份)变化. 2016年8月份长江流域降水有明显负异常, 与1998年相反. 本文进一步的研究发现中纬度显著的环流异常导致了2016年8月偏弱的长江流域降水. 在2016年8月, “丝绸之路遥相关”和乌拉尔山地区的位势高度异常显著, 为近40年来最强, 它们会导致中纬度东亚地区出现显著的反气旋式环流异常, 引起长江中下游地区异常的东风异常, 进而会减弱副高外围气候态风场和水汽输送, 导致长江流域降水偏少. 基于2016年独特的次季节变化特征, 本工作表明研究长江流域夏季降水可预测性和变化特征时, 不能仅仅考虑热带的信号, 更需要关注中纬度环流的特征.

1. Introduction
  • The summer rainfall anomaly over the Yangtze River valley (YRV) exhibits strong interannual variation, associated with tropical air-sea interactions (e.g., Tao and Chen, 1987; Ding, 1992; Chang et al., 2000). As one of the most significant sources of interannual variability and seasonal predictability, El Niño has a pronounced impact on YRV summer rainfall (Kawamura et al., 2001; Lee et al., 2011; Li et al., 2012, 2014). Generally, El Niño matures in winter and exerts a strong impact on the YRV climate during its decaying summer, due to the lagged response of atmospheric circulation to the El Niño signal (Huang and Wu, 1989; Wang et al., 2000; Chou et al., 2003; Wu et al., 2009). At that time, there is an anticyclonic circulation anomaly over the western North Pacific (hereafter referred to as the WNPAC). The southwesterly anomalies at the western edge of the WNPAC intensify the water vapor transport from the warm ocean to the YRV, and therefore result in greatly strengthened YRV summer rainfall (e.g., Huang and Sun, 1992; Zhang et al., 1999; Zhou and Yu, 2005).

    As a result of the above mechanism, serious flooding occurred over the YRV in the summer of 1998, following a super El Niño that occurred in the preceding winter. More than 3000 deaths, $25 billion of economic losses, and 15 million cases of homelessness were caused by this flood (Huang et al., 1998; Jiang et al., 2008). Another super El Niño event took place in the winter of 2015/16, with comparable intensity to the 1997/98 case. However, the YRV summer rainfall was overestimated, with most predictors pointing towards a similar YRV flooding event as that in 1998. The remarkable difference in precipitation experienced in summer 2016 compared to that in 1998 quickly stimulated the interest of academics (Guo et al., 2016).

    Previous studies have indicated that the impact of an El Niño event depends on the intensity of its mature phase (Wang et al., 2000; Li et al., 2007; Chen et al., 2014) and length of its decaying phase (Chen et al., 2012, 2016). Did these features differ between the 2015/16 and 1997/98 cases? Figure 1 compares the intensity and evolution of the two cases, in which the El Niño events are measured by the SST anomalies over the central and eastern equatorial Pacific (Niño3.4 index; 5°S-5°N, 190°-240°E). On the one hand, both matured in winter with similar intensity. The strongest intensity occurred in November, with a value of 2.25°C for the 1997/98 case and 2.29°C for the 2015/16 case —— both greater than two standard deviations of the Niño3.4 index in winter (1.03°C for one standard deviation). On the other hand, these two El Niño events both declined rapidly and terminated in June —— identifiable as short decaying El Niños, according to Chen et al. (2012, 2016).

    Generally, markedly strengthened YRV summer rainfall is expected throughout the entire summer season following a strong and short decaying El Niño. However, the YRV summer rainfall in 2016 was weaker than expected, as compared to that in 1998. Other factors must therefore have been responsible for the weak YRV summer rainfall. Previous studies have indicated that extratropical factors, including the Silk Road Pattern (SRP) (or circumglobal teleconnection in some studies) (Ding and Wang, 2005; Lin, 2014; Hong and Lu, 2016) and the meridional displacement of the upper-tropospheric jet (e.g., Liang and Wang, 1998; Lu, 2004; Lin and Lu, 2005; Kuang and Zhang, 2006; Xuan et al., 2011; Tian and Fan, 2013; Hong and Lu, 2016), play a crucial role in the rainfall over the YRV. It is suggested that these midlatitude circulation anomalies operate mainly through modifying the low-level circulations around the YRV. So, did these midlatitude circulation anomalies contribute to the distinct YRV summer rainfall in 2016?

    In the above context, the aim of the present study was to elucidate the difference in the YRV summer precipitation anomalies between 2016 and 1998, and the possible reasons for the difference. Following this introduction, the datasets used are described in section 2. The difference in the summer anomalies of YRV precipitation between 2016 and 1998 and the associated anomalous circulations are compared in section 3. Section 4 discusses the role of midlatitude circulation anomalies in the YRV summer rainfall. A summary follows in section 5.

    Figure 1.  SST evolution of the Niño3.4 index for 2015/16 (red) and 1997/98 (blue) El Niño events from the previous (-1) October to the following September. The Niño3.4 index was defined as the averaged SST anomaly over (5°S-5°N, 170°-120°W).

2. Data
  • Two sets of monthly precipitation data were employed in this study: 160-station rainfall data from the China Meteorological Administration, and the Climate Prediction Center's Merged Analysis of Precipitation (Xie and Arkin, 1997). The circulation and SST data were obtained from the National Centers for Environmental Prediction/National Center for Atmospheric Research reanalysis products (Kalnay et al., 1996) and version 4 of NOAA's monthly mean Extended Reconstructed SST dataset (Huang et al., 2015), respectively. Statistical analyses in this study were performed from 1980 to 2016, in which the monthly anomalies were calculated by removing the average of these 37 years.

3. Difference in the summer anomalies between 2016 and 1998
  • Figure 2 compares the summer (June-July-August, JJA) precipitation anomaly over the YRV (25°-35°N, 100°-120°E) between 1998 and 2016. For the whole of the summer season , the precipitation anomaly in 2016 was 0.28 mm d-1 —— about half of one standard deviation of the YRV summer precipitation anomaly (0.58 mm d-1). Moreover, the 2016 summer YRV precipitation anomaly was only one-sixth of that in 1998 (1.60 mm d-1), despite both occurring during the decaying phase of a super El Niño with similar intensity and rate of decline (Fig. 1). However, this considerable difference did not occur in every month of the summer season —— it mainly took place in August.

    For June and July, positive YRV precipitation anomalies occurred in both 2016 and 1998. The difference in the monthly mean precipitation between these two months was not as large as that for the seasonal mean. The intensity was 1.09 mm d-1 in June 2016, which was larger than one standard deviation of the YRV precipitation in June (0.80 mm\;d-1), although it was weaker than that in June 1998 (2.52 mm d-1). The YRV precipitation anomaly in July 2016 (1.11 mm d-1) was comparable to that in July 1998 (1.48 mm d-1), both being larger than one standard deviation of the YRV precipitation in July (0.78 mm d-1).

    Nevertheless, the YRV precipitation transformed into a negative anomaly in August 2016 (-1.36 mm d-1), but still persisted with a strong positive anomaly in August 1998 (0.80 mm d-1). The negative precipitation anomaly was remarkable in August 2016, exceeding 1.5 times of one standard deviation of the YRV precipitation in August (0.83 mm d-1). The greatly reduced precipitation in August 2016 led to the weak summer mean precipitation anomaly in 2016, even though there was a considerable increase in precipitation in June and July of that year.

    Figure 2.  Summer rainfall anomaly (mm d-1) over the YRV (25°-35°N, 100°-120°E), for the whole season (JJA mean) and the individual months of 2016 (red) and 1998 (blue), based on China station data.

    Figure 3 illustrates the spatial pattern of the precipitation anomaly in 2016 and 1998, as derived from the China station dataset. For the whole of the summer season (Figs. 2a and b), the strong precipitation centers in these two years were both over the YRV, but the intensity was weaker and the location shifted towards the west in 2016, as compared to those in 1998. For individual months, the distribution of the precipitation anomaly over the YRV was basically similar in June and July of these two years, but differed significantly in August. In June (Figs. 2c and d), the positive precipitation anomaly stretched from the middle to the lower reaches of the YRV in 2016 and 1998, but extended further into South China in 1998. In July (Figs. 2e and f), both the spatial pattern and intensity of the precipitation anomaly over the YRV were similar in these two years. In August, however, the distributions of the precipitation anomaly were almost opposite, particularly over the YRV (Figs. 4g and h). The spatial patterns of the precipitation anomaly lend further support to the conclusion that the different summer YRV rainfall in these two years was principally due to the situation in August.

    The lower-tropospheric circulation patterns for JJA-mean and individual months were associated with the precipitation anomalies in these two years. For the whole of the summer season, the circulation over the western North Pacific (WNP) was weak in 2016, which was related to the weak precipitation anomalies over the YRV (Fig. 4a). By contrast, in 1998 (Fig. 4b), there were strong anticyclonic circulation anomalies over the WNP, suggesting a strong WNP subtropical high. The anomalous southwesterly along the YRV induced by the WNPAC favored water vapor transport to the YRV and therefore led to the serious flooding during summer 1998.

    In June and July, the WNPAC was well-organized in both years (Figs. 4c-f), albeit a little stronger in 1998 than in 2016. The anomalous southwesterly associated with the WNPAC was also apparent in both years, which contributed to the positive precipitation anomalies over the YRV. In August, the WNPAC maintained in 1998 (Fig. 4h), but changed in 2016 (Fig. 4g), being largely weakened. Anomalous anticyclonic circulation anomalies occupied midlatitude East Asia, associated with easterly wind anomalies along the middle-to-lower reaches of the YRV. The anomalous easterly flow reduced the climatological southwesterly wind along the edge of the WNP subtropical high and prevented more water vapor transport to East Asia. Therefore, the reduced precipitation in August 2016 may be attributable to the unique midlatitude circulation anomalies.

    Figure 3.  Spatial distribution of the rainfall anomaly (mm d-1) over East China in (a, b) the whole summer season (JJA mean), (c, d) June, (e, f) July, and (g, h) August of 2016 (left) and 1998 (right), based on China station data.

    Figure 4.  As in Fig. 3 but for 850-hPa wind (m s-1) and precipitation (mm d-1) anomalies over the WNP and East Asia.

  • Having discovered a remarkable difference in the precipitation anomaly between the years in August, we further checked the nature of the midlatitude circulation anomaly in this month. Figure 5 shows the 300-hPa geopotential height anomaly in August 2016 and 1998. Significant positive anomalies appeared around the Ural Mountains in August 2016, where the geopotential height anomalies in the upper troposphere were about 200 m (Fig. 5a). These anomalous circulations around the Ural Mountains were more than double the standard deviation for the 300-hPa geopotential height detected from the last 37 years. Besides the anomalous high over the Urals region, geopotential height anomalies were also apparent around Mongolia and North China. However, in comparison, there were no significant 300-hPa geopotential height anomalies in August 1998 (Fig. 5b), exhibiting large differences to those in 2016.

    Anomalous midlatitude circulation in 2016 was also found in the upper-tropospheric meridional wind (Fig. 6a), displaying as an anomalous zonally wave-like pattern with alternating southerly and northerly flow along 40°N across Eurasia. This is generally recognized as the SRP teleconnection pattern (Lu et al., 2002; Enomoto et al., 2003), which is geographically phase-locked from the source region around the Caspian Sea and propagates eastwards to East Asia. The distinct anomalous SRP associated closely with strong anticyclonic anomalies over midlatitude East Asia, corresponding well to the lower-tropospheric anticyclonic wind anomalies over midlatitude East Asia (Fig. 4g) and deficient rainfall over the YRV (Fig. 3g) in August 2016. The impacts from the variation in the SRP in August and the associated midlatitude circulation on the YRV precipitation anomaly were therefore further investigated, as reported in the following section. By comparison, no significant upper-tropospheric meridional wind anomaly existed in August 1998 (Fig. 6b), suggesting quite weak midlatitude wave activity and thus a weak influence on YRV rainfall. These anomalies in August 1998 were consistent with the enhanced WNP subtropical high and increased rainfall over the YRV (Figs. 3h and 4h), which were basically modulated by the decaying effect of El Niño.

    The anomalous SRP was also reflected in the upper-tropospheric zonal wind anomaly (Fig. 7). A notable 200-hPa westerly around Mongolia, and easterly around Central Asia and western China, appeared in August 2016 (Fig. 7a). These anomalies, associated with the meridional wind, together suggested a strong anomalous anticyclone, which corresponded well to the enhanced geopotential height anomaly over midlatitude East Asia (Fig. 5a), favoring the easterly around the middle-to-lower reaches of the YRV. On the other hand, unlike that in 2016, an anomalous meridional tripole pattern for the upper-tropospheric zonal wind anomaly was found over the WNP and East Asia in August 1998 (Fig. 7b). This pattern corresponded well to the lower-tropospheric anticyclonic and cyclonic anomalies over the WNP and East Asia (Fig. 4h), and is generally recognized as the East Asia-Pacific or Pacific-Japan teleconnection pattern, which is also closely related to tropical SSTs and associated air-sea interaction (Huang and Sun, 1992; Lu, 2001; Kosaka et al., 2012).

    The midlatitude circulation anomalies in June and July 2016 were weaker than those in August, including the upper-tropospheric geopotential height and meridional wind anomalies (not shown), implying that ENSO and the following intensified WNP subtropical high played a dominate role in these two months for YRV summer rainfall. This also suggested that the midlatitude teleconnection changed significantly from June and July to August. Relatively, the change from June and July to August in 1998 was weak, in which a similar distribution of midlatitude circulation anomalies appeared in June and July, but with weaker anomalies in August, as compared with those in 2016, suggesting ENSO and the following intensified WNP subtropical high persistently dominated the YRV summer rainfall in August 1998.

    Figure 5.  The 300-hPa geopotential height (units: m) anomaly in August (a) 2016 and (b) 1998. Shading indicates the standard deviation of the anomaly larger (yellow) than 2 m or smaller (cyan) than -2 m. The solid (dashed) contours represent the positive (negative) anomaly, respectively, and the interval of the contours is 25 m.

4. Role of midlatitude circulation in the YRV rainfall anomaly
  • In view of the distinct nature of the midlatitude circulation in August following the super El Niño events in these two years, we detected that two key factors —— the anomalous upper-tropospheric geopotential height over the Urals region, and the SRP —— modulated the YRV summer rainfall in August. Thus, in this section, we offer further interpretation of the underlying processes related to these two anomalous midlatitude circulations in August.

    Figure 8 shows the normalized time series of the Urals anomalous geopotential height in August, which was defined as the averaged 300-hPa geopotential height anomalies over (50°-70°N, 40°-70°E). The anomaly in August 2016 was the largest of the last 37 years, whose normalized value was 2.76. This was about three times the interannual variability, which was 44.18 m for the corresponding standard deviation. By comparison, the anomaly was quite weak in August 1998, with a normalized anomaly of only 0.38.

    Figure 6.  As in Fig. 5 but for the 200-hPa meridional wind (units: m s-1) anomaly. Contour interval: 2 m s-1.

    Figure 9a shows the horizontal distribution of the SRP, manifested by the first leading EOF mode of the 200-hPa meridional wind anomalies in August within the domain (20°-60°N, 0°-150°E), following (Yasui and Watanabe, 2010). This mode explained 26.5% of the total variance and was well isolated from the other modes. According to Rossby wave ray theory (Hoskins and Ambrizzi, 1993), it had a wavelength of about 60° in longitude and tended to be geographically phase-locked to preferred longitudes (Lu et al., 2002; Ding and Wang, 2005; Kosaka et al., 2009). It was characterized by a clear wave-like pattern from western Eurasia to East Asia in the midlatitudes, consistent with previous studies (e.g., Lu et al., 2002; Yasui and Watanabe, 2010), although only August was used here. These anomalous cells of the SRP, together with an evident northerly anomaly over East Asia, corresponded well to the anomalous centers of 200-hPa meridional wind in August 2016 (Fig. 6a), implying a profound impact from the SRP in that year. The year-to-year variation of the SRP is shown by the associated principal component (PC) in Fig. 9b and defined as the SRP index. This also illustrated a remarkable anomaly in 2016 (normalized anomaly of 2.48) and was the largest of the last 37 years, since 1980. In contrast, the anomaly was quite weak in 1998 (0.19). The different anomalies matched well with the different spatial distributions of 200-hPa meridional wind anomalies in these two years (Fig. 6).

    Figure 7.  As in Fig. 5 but for the 200-hPa zonal wind (units: m\;s-1) anomaly. Contour interval: 2 m s-1.

    Figure 8.  Normalized time series of anomalous 300-hPa geopotential height over the Urals region (50°-70°N, 40°-70°E). The two orange dashed lines indicate the two anomalous years of 1998 and 2016.

    Based on linear regression of 300-hPa geopotential height onto the Urals anomalous geopotential height and SRP index, we identified anomalous positive geopotential height anomalies around midlatitude East Asia (Fig. 10). These positive geopotential height anomalies indicated an anomalous anticyclone in the upper troposphere, and this anticyclone showed close teleconnections with both the Urals anomalous geopotential height and the SRP. The correlation coefficient between this anticyclone, which was defined as the averaged 300-hPa geopotential height over (35°-45°N, 90°-120°E), and the Urals anomalous geopotential height (SRP index), was 0.51 (0.51). Both exceeded the 99% confidence level according to the Student's t-test, although the anticyclone related to the Urals anomalous geopotential height was shifted slightly to the southeast. Furthermore, the teleconnection patterns, both for the Urals anomalous geopotential height and the SRP, illustrated similar spatial distributions, suggesting they had a corporate impact on the anomalous anticyclone over midlatitude East Asia.

    The anomalous anticyclone over midlatitude East Asia in the upper troposphere was key in connecting the upper-tropospheric circulation in midlatitude Eurasia and the YRV summer rainfall. Figure 11a shows the normalized time series of this anomalous anticyclone, as defined by the averaged 300-hPa geopotential height over (30°-45°N, 90°-120°E). The corresponding standard deviation was 24.8 m. The anomaly was also largest in August 2016, with the standard anomaly being 3.46. Relatively, it was larger than the standard anomaly of both the Urals anomalous geopotential height and the SRP index (Figs. 8 and 9b), suggesting a joint effect of them on this anticyclone. In contrast, the standard anomaly was only -0.02 in August 1998. These anomalies corresponded well to those shown in Fig. 5. Moreover, consistent with Rossby wave ray theory (Hoskins and Ambrizzi, 1993) and the horizontal structure of the SRP (Lu et al., 2002; Ding and Wang, 2005), this anomalous anticyclone exhibited a clear barotropic structure and penetrated the lower troposphere (Fig. 11b). Related to this upper-tropospheric anticyclone, we also detected an anomalous anticyclone in the lower troposphere over midlatitude East Asia, with remarkable easterly flow over the middle-to-lower reaches of the Yangtze River. This reduced the climatological southwesterly wind along the WNP subtropical high, weakened the water vapor transport, and thus gave rise to reduced rainfall over the YRV. It corresponded well to the significant negative rainfall anomalies regressed onto the anticyclone over midlatitude East Asia (Fig. 11c). The correlation coefficient between this midlatitude anticyclone and summer rainfall over the YRV was -0.30, which was moderately statistically significant (90% confidence level). However, as this anomalous midlatitude anticyclone was more than threefold the standard deviation than in 2016, it thus projected a profound negative rainfall anomaly over the YRV.

    Figure 9.  The (a) spatial distribution and (b) its principal component (PC) of the leading EOF mode for the 200-hPa meridional wind anomaly in August 1980-2016. The EOF1 of 200-hPa meridional wind over (20°-60°N, 0°-150°E) was defined as the SRP, following Yasui and Watanabe (2010). The percentage value in the upper-right corner of (a) shows the percentage variance explained by this mode. The two orange dashed lines in (b) indicate the two anomalous years of 1998 and 2016.

    Figure 10.  Regression of 300-hPa geopotential height anomalies (Units: m) onto the (a) 300-hPa geopotential height over the Urals region and (b) the SRP Index. Shading indicates regions where anomalies exceed the 95% confidence level. Contour interval: 5 m.

    Unsurprisingly, the above regressed anomalies were similar in spatial distribution to our year of focus in 2016, including the midlatitude anticyclone in the upper troposphere and the lower-tropospheric wind anomalies. Taking a step further, the regressed anomalies after excluding 2016 also demonstrated a similar pattern (data not shown). This verified the importance of the SRP and the Urals anomalous geopotential height to the YRV rainfall in August. Therefore, as the strongest anomalous year, the SRP and the Urals anomalous geopotential height in August 2016 would definitely have contributed to the anticyclone over midlatitude East Asia in the upper troposphere, exciting a profound lower-tropospheric easterly and resulting in reduced rainfall over the YRV (Fig. 3g). Additionally, the impact of the above midlatitude circulation anomalies were also reflected in the anomalous lower-tropospheric anticyclone over the midlatitude WNP east of Japan (Fig. 11b), which was closely consistent with the lower-tropospheric circulation anomalies in August 2016 (Fig. 4g).

    Figure 11.  (a) Normalized time series of the anomalous anticyclone over East Asia in the upper troposphere, defined as the averaged 300-hPa geopotential height anomalies over (30°-45°N, 90°-120°E), and the regression of (b) 850-hPa wind and (c) precipitation anomalies onto it in August 1980-2016. The two orange dashed lines in (a) indicate the two anomalous years of 1998 and 2016.Shading in (b) indicates regions where anomalies exceed the 95% confidence level. Contour intervals: 0.5 m s-1 and 0.25 mm d-1.

5. Summary and discussion
  • Serious flooding tends to occur over the YRV in the decaying summer of El Niño events, as with the devastating floods of 1998, resulting in catastrophic damage to the livelihoods of many millions of people, and severe economic losses. Therefore, any credible progress in understanding the nature of the YRV summer rainfall following El Niño would be of benefit to economic planning and disaster mitigation. As one of the two strongest super El Niño events on record, we found quite similar SST evolution for the 2015/16 super El Niño as that of 1997/98. However, the following-summer rainfall over the YRV was much weaker in 2016 relative to 1998. Therefore, we investigated the differences in the YRV rainfall between these two summers and identified that they arose principally from the distinct YRV rainfall anomaly in August. The YRV rainfall anomaly was clearly negative in August 2016——the opposite to what we would usually expect to see during the decaying phase of El Niño, and what we did see in August 1998. Nonetheless, relatively, positive rainfall anomalies in June and July were observed in both years.

    Furthermore, we found that the negative YRV rainfall anomaly in August 2016 was associated with remarkable midlatitude circulation anomalies. The 200-hPa wind anomalies in August 2016 displayed a wave-like SRP over midlatitude Eurasia and a significantly enhanced upper-tropospheric geopotential height over the Urals region. These anomalies were strongest in August 2016 when examined with respect to the last 37 years (since 1980), but quite weak in August 1998. Further analysis highlighted that the anomalous SRP and strong geopotential height anomalies over the Urals region resulted in significant anticyclonic circulation anomalies over midlatitude East Asia. This anomalous anticyclone would have excited lower-tropospheric easterly wind anomalies along the middle-to-lower reaches of the YRV, reducing the climatological wind and therefore resulting in the negative YRV rainfall anomaly.

    It is widely recognized that the variation in YRV rainfall in summer is closely related to the activities of sub-seasonal oscillations from both tropical and midlatitude regions, and this was true for the particular summer of 1998 (Chan et al., 2002; Chen et al., 2005; Sun et al., 2016). In this study, we found profound sub-seasonal changes in midlatitude circulation in 2016 from June and July to August, and revealed their significant contributions to the variation in YRV rainfall. But from what do these changes arise? It has been revealed in previous studies that heating anomalies over northern India can result in the downstream propagation of midlatitude wave trains (Lu et al., 2002; Ding and Wang, 2005), suggesting convection over northern India may play an important role. Further investigation is required to uncover the detailed reasons.

    This study emphasizes the important role played by midlatitude circulation in the variation of YRV summer rainfall, which could modulate the YRV rainfall anomaly, even in the summer following a strong El Niño. The results achieved here imply that more attention should be paid to midlatitude circulation in order to better understand the predictability and variation of YRV summer rainfall, and thus increase the chances for skillful seasonal prediction of YRV summer rainfall. In addition, as the primary source of prediction skill for YRV summer rainfall and the subtropical high (Li et al., 2012, 2016), tropical air-sea interaction dominated and gave rise to the enhanced rainfall in June and July of 2016, and in the summer of 1998. Therefore, identifying sources from the variations in midlatitude circulation and combining that information with tropical factors via dynamical or statistical downscaling methods, would potentially be an effective way to improve the prediction of YRV summer rainfall.

Reference

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return