Large-scale atmospheric singularities and summer long-cycle droughts-floods abrupt alternation in the middle and lower reaches of the Yangtze River

Chinese Science Bulletin 2006 Vol. 51 No. 16 2027 2034 DOI: 10.1007/s11434-006-2060-x Large-scale atmospheric singularities and summer long-cycle droughts-floods abrupt alternation in the middle and lower reaches of the Yangtze River WU Zhiwei 1,2, LI Jianping 2, HE Jinhai 1,3 & JIANG Zhihong 1 1. Key Laboratory of Meteorological Disaster, Nanjing University of Information Science & Technology, Nanjing 210044, China; 2. State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China; 3. Institute of Tropical and Marine Meteorology, China Meteorological Administration, Guangzhou 510080, China Correspondence should be addressed to Li Jianping (email: ljp@lasg.iap.ac.cn) Received January 18, 2006; accepted March 23, 2006 Abstract The daily precipitation data at 720 stations over China for the 1957 2003 period during summer (May August) are used to investigate the summer subseasonal long-cycle droughts-floods abrupt alternation (LDFA) phenomenon and a long-cycle droughts-floods abrupt alternation index (LDFAI) in the middle and lower reaches of the Yangtze River (MLYRV) is defined to quantify this phenomenon. The large-scale atmospheric circulation features in the anomalous LDFA years are examined statistically. Results demonstrate that the summer droughts-to-floods (DTF) in the MLYRV usually accompany with the more southward western Pacific subtropical high (WPSH), negative vorticity, strong divergence, descending movements developing and the weak moisture transport in the low level, the more southward position of the South Asia high (SAH) and the westerly jets in the high level during May June, but during July August it is in the other way, northward shift of the WPSH, positive vorticity, strong convergence, ascending movements and strong moisture transport in the low level, and the northward shift of the SAH and the westerly jets in the high level. While for the summer floods-to-droughts (FTD) in the MLYRV it often goes with the active cold air mass from the high latitude, positive vorticity, strong convergence, ascending movement developing and the strong moisture transport in the low level, and the SAH over the Tibetan Plateau in the high level, but during July August it is often connected with the negative vorticity, strong divergence, descending movements developing and the weak moisture transport in the low level, the remarkable northward shift of the WPSH, the SAH extending northeastward to North China and the easterly jets prevailing in the high level over the MLYRV. In addition, the summer LDFA in the MLYRV is of significant relationship with the Southern Hemisphere annual mode and the Northern Hemisphere annual mode in the preceding February, which offers some predictive signals for the summer LDFA forecasting in the MLYRV. Keywords: middle and lower reaches of Yangtze River, droughts-floods abrupt alternation, large-scale circulation singularities, long-cycle timescale. The summer droughts and floods in the middle and lower reaches of the Yangtze River valley (MLYRV) have often been one of the basic contents of the precipitation forecasting during the flooding period in China. Lots of studies suggest that summer rainfall in the MLYRV results from the advance of the East-Asian summer monsoon (EASM) [1,2] and the droughts and floods in the region are closely related with the global large-scale atmospheric singularities [3 7]. But most researches focus more on the total precipitation anomaly of summer, fewer on the subseasonal variation of rainfall that is also of great importance. In fact, the latter plays an important role in the water resource distributing, industry, agriculture, and people s daily lives. The droughts-floods coexistence (DFC) [8] and the droughtsfloods abrupt alternation (DFA) are the typical representatives of precipitation subseasonal anomalies and often occur in summer of the MLYRV [8]. Furthermore the DFC is composed by the DFA of different timescale [8] and this paper mainly discusses the summer long-cycle droughts-floods abrupt alternation (LDFA). The investigation of LDFA is not only to understand the linkage between the summer long-cycle subseasonal rainfall variation and the previous large-scale atmospheric singularities but also to give some comments on preparing work against summer droughts-floods. 1 Data and methods The daily precipitation data at 720 stations over www.scichina.com www.springerlink.com 2027

China for the 1957 2003 period are supplied by National Meteorological Information Center of China and the reanalysis data are taken from ERA-40 datasets. The indices of the Southern Hemisphere annual mode (SAM) [5] and the Northern Hemisphere annual mode (NAM) [9] for the same period are offered by state key laboratory of numerical modeling for atmospheric sciences and geophysical fluid dynamics of China. The magnitude of long-cycle in the paper is defined as 4 months, 2 months for the droughts and 2 months for the floods. So, the LDFA summers are of droughts (floods) during May June and floods (droughts) during July August. The summer in the MLYRV is defined as May August (MJJA) in this paper. According to the classification of precipitation area given by Ting and Wang [10], from the standard deviations of the summer precipitation of Chinese 720 stations for the 1957 2000 period, we select the station, Gaoyou, which has the largest rainfall variation in the MLYRV, as the base point. Based on the one-point correlation map of the summer precipitation between the base point, Gaoyou, and Chinese 720 stations (Fig. 1(a)), we use the mean precipitation at the 42 stations, which are of the same rainfall variation as that at the base point, their correlation coefficients exceeding the 95% confidence level based on the Monte Carlo test, to represent the summer precipitation in the MLYRV. To describe the LDFA phenomenon quantitatively, an LDFA index (LDFAI) is defined as follows: 56 78 LDFAI ( R78 R56 ) ( R56 R78 ) 1.8 R + = + R, where R 78 and R 56 refer to the normalized July August and May June precipitation of the MLYRV, respectively. (R78 R 56 ) represents the intensity term of DFA an (R + R ) denotes the magnitude of d 56 78 R56 + R78 the droughts and floods. 1.8 is the weight coefficient, to balance the LDFA summers and to reduce the weight of the severe floods or droughts summers. The absolute magnitude of those precipitation anomalies within 0.5 standard deviations is regarded as normal, while the anomalies over 0.5 standard deviations and under -0.5 standard deviations are defined as floods and droughts, respectively. 2 Choosing of the anomalous LDFA summers in the MLYRV and their temporal variation To verify the validation of LDFAI to describe the characteristics of the summer LDFA in the MLYRV, we calculate the normalized May June and July August precipitation of the 6 highest and 6 lowest LDFAI summers from 1957 to 2003 in the MLYRV (Table 1). The high LDFAI summers have more precipitation in July August than in May June whereas the low LDFAI summers have more precipitation in May June than in July August, which means that high LDFAI Fig. 1. (a) One-point correlation of summer (May August) seasonal mean precipitation over the eastern China using the base point, Gaoyou Station. Contour interval is 0.1. Those values exceeding the 95% confidence level based on the Monte Carlo test are shaded. (b) The normalized summer LDFAI time series in the MLYRV from 1957 to 2003. 2028 Chinese Science Bulletin Vol. 51 No. 16 August 2006

High LDFAI summer Table 1 Normalized precipitation of the 6 highest (lowest) LDFAI summers from 1957 to 2003 May June precipitation June July precipitation Low LDFAI summer May June precipitation ARTICLES June July precipitation 1982 1.18 1.33 1971 1.48 1.25 1969 1.09 2.12 1964 1.39 0.77 1965 1.93 1.03 1959 0.61 1.69 1987 0.63 1.78 1960 1.02 0.53 1958 1.65 0.27 1999 1.2 0.02 1962 0.41 1.04 2000 0.67 0.27 summers are corresponding to the increasing precipitation from May through August, and vice versa. In addition, the July August precipitation of all the high LDFAI summers is over 0.5σ, 5 summers of them even over 1.0σ, which indicates that they all appear as the anomalous floods, whereas the May June precipitation of 4 high LDFAI summers is under 1.0σ (anomalous droughts). Thus, DTF usually occurs in the high LDFAI summers. The situations for the low LDFAI summers are just opposite. The July August of 4 low LDFAI summers has anomalous floods and the July August of all the 6 low LDFAI summers is abnormally drought. So, low LDFAI summers are often related to FTD. According to the above analysis, LDFAI might be regarded as a quantitative index describing the summer LDFA phenomenon. We choose 6 high LDFAI summers (1958, 1962, 1965, 1969, 1982 and 1987) as the DTF summers and 4 low LDFAI ones (1959, 1960, 1964 and 1971) as FTD summers. Among DTF summers, although the July August precipitation anomaly of 1958 and the May June precipitation anomaly of 1962 are between 0.5σ and 0.5σ (not met the anomalous criterion), the May June of 1958 and the July August of 1962 have significantly anomalous precipitation (absolute value of precipitation anomaly over 1.0σ ). To get enough samples for composite analysis, these 2 years are chosen as DTF summers approximately. Fig. 1(b) shows the temporal variation of the summer LDFAI in the MLYRV from 1957 to 2003, to a certain extent, which exhibits the fluctuations on interannual and decadal timescales. The interannual variation is more remarkable before 1972 than after 1972, which means that the summer LDFA events had occurred more frequently in the MLYRV before 1972. According to the May June and July August precipitation in the MLYRV from 1957 to 2003 (not shown), there are 11 droughts-to-droughts (DTD) or floods-to-floods (FTF) summers for the period of 1972 2003 and only 3 DTD or FTF summers for the period of 1957 1971. It indicates that the occurrences of the summer LDFA in the MLYRV are decreasing and those of the summer DTD or FTF are increasing, which might be due to the smaller fluctuations of summer LDFAI in the MLYRV after 1972. In addition, the fluctuation trend of LDFAI is to strengthen after 1997. 3 Northern Hemisphere circulation features of the anomalous LDFA summers Fig. 2 shows the features of 500 hpa height of the LDFA summers in the MLYRV. In May June of the DTF summers (Fig. 2(a)), there is a large significant abnormal area over the western Pacific and the 585-line is located over the oceanic areas from the South China Sea to western Pacific, which indicates that the western Pacific subtropical high (WPSH) becomes more southward. The eastern China is controlled by the dry and cold continent high, which might attribute to the droughts in the MLYRV. In July August of the DTF summers (Fig. 2(b)), the significant abnormal area moves to the northern area of Taiwan and the 585-line shifts northward to the MLYRV simultaneously, which implies that the WPSH has an anomalously northward jump. The MLYRV is just at the northwest side of the WPSH where the moisture transport is abundant. It provides the necessary circulation condition for more rainfall. In May June of the FTD summers (Fig. 2(c)), there is a large area of significant anomaly over the middle and high latitude regions from Siberia to the MLYRV, which suggests that the northern cold air mass is very active and expands southward to the MLYRV. That the warm air meets the cold air in the MLYRV, benefits for more rainfall [11]. In July August of the FTD summers (Fig. 2(d)), the WPSH moves quite northward and the rainfall band is pushed to North China. The MLYRV is controlled by the WPSH and has less precipitation. www.scichina.com www.springerlink.com 2029

Fig. 2. The composite 500 hpa height (in 10 gpm) for (a) May June of the DTF summers, (b) July August of the DTF summers, (c) May June of the FTD summers, and (d) July August of the FTD summers in the MLYRV. The shaded areas exceed the 95% confidence level based on the t-test. The composite difference of the relative vorticity in 850 hpa between the DTF and FTD summers shows that in May June (Fig. 3(a)), negative difference center prevails from the Korean Peninsular to the MLYRV, which indicates that in May June, the DTF (FTD) summers are often associated with the divergence (convergence) and downdraft (updraft) in the low level. It supplies the necessary circulation condition for precipitation decreasing (increasing) in the MLYRV, while in July August, the situations are just opposite (Fig. 3(b)). Fig. 4 shows the features of 100 hpa height and wind fields of the anomalous LDFA summers in the MLYRV. In May June of the DTF summers (Fig. 4(a)), the South Asia high (SAH) is located over the region north of the Bay of Bengal, more southward, and westerly jets are also more southward, which is matching with the position of the 500 hpa WPSH (Fig. 2(a)). In July August (Fig. 4(b)), the SAH has a northwestward shift to the Tibetan Plateau and the westerly jets move to the MLYRV. At the same time the 500 hpa WPSH moves northeastward. This is consistent with the correlation between the WPSH and the SAH suggested by Tao [12]. In May June of the FTD summers (Fig. 4(c)), the SAH is located over the south side of the Tibetan Plateau, associated with more precipitation in the MLYRV, while in July August of the FTD summers (Fig. 4(d)), with the SAH significantly strengthening and expanding northeastward to North China, the easterly jets prevail in the MLYRV and the precipitation becomes less. The 850 hpa composite difference of the moisture transport between the DTF and FTD summers shows that in May June (Fig. 5(a)), southward moisture transport difference prevails in the MLYRV, which im- 2030 Chinese Science Bulletin Vol. 51 No. 16 August 2006

ARTICLES Fig. 3. The composite 850 hpa relative vorticity difference (in 10 5s 1; DTF minus FTD) for (a) May June and (b) July August. The shaded dark (light) areas exceed the 95% confidence level based on the t-test. Fig. 4. The composite 100 hpa height (in 10 gpm) and wind (in m s 1) for (a) May June of the DTF summers, (b) July August of the DTF summers, (c) May June of the FTD summers, and (d) July August of the FTD summers in the MLYRV. The shaded areas are the winds fields exceeding the 95% confidence level based on the t-test. plies that moisture transport becomes weak (strong) in May June of the DTF (FTD) summers. It supplies the necessary moisture for the precipitation decreasing (increasing) in the MLYRV[13,14], while in July August, the situations are on the contrary (Fig. 5(b)). The 700 hpa composite difference of vertical velocity displays www.scichina.com www.springerlink.com that in May June (Fig. 6(a)), there is a large area of significant difference in the MLYRV, which shows that downdraft (updraft) develops in May June of the DTF (FTD) summers. It is the necessary dynamic condition for precipitation decreasing (increasing) in the MLYRV while the situations are opposite (Fig. 6(b)) in July 2031

Fig. 5. The composite 850 hpa moisture transport difference (in g (cm s) 1 ; DTF minus FTD) for (a) May June and (b) July August. The shaded areas exceed the 95% confidence level based on the t-test. Fig. 6. The composite 700 hpa vertical velocity difference (in 10 2 hpa s 1 ; DTF minus FTD) for (a) May June and (b) July August. The shaded dark (light) areas exceed the 95% confidence level based on the t-test. August. 4 Large-scale atmospheric singularities in the preceding months of the LDFA summers To investigate the abnormity of the large-scale circulation in the preceding months of the LDFA summers, correlation coefficients between the summer LDFAI and the SAMI (NAMI) of the preceding 6 months are calculated, respectively (Table 2). It shows that February has the most significant correlation coefficient between the summer LDFAI and the SAMI reaching 0.32, which indicates that if the February SAM is strong (weak), the summer LDFAI in the MLYRV is likely to be low (high) and the summer FTD (DTF) might occur in the MLYRV. In addition, among the 6 preceding months, there is the most significant correlation in February between the summer LDFAI and the NAMI and the correlation coefficient is 0.25 but the significance is less than that of the SAM. The linkage between the preceding SAM and the summer LDFA in the MLYRV is revealed by Fig. 7, which illustrates the composite Southern Hemisphere 2032 Chinese Science Bulletin Vol. 51 No. 16 August 2006

Table 2 Correlation coefficients between the summer LDFAI and the SAMI (NAMI) in the 6 preceding months Month November December January February March April LDFAI and SAMI 0.27 0.05 0.12 0.32 a) 0.03 0.00 LDFAI and NAMI 0.13 0.15 0.23 0.25 0.10 0.23 a) The most significant correlation coefficient which exceeds the 95% confidence level based on the Monte Carlo test. Fig. 7. The composite Southern Hemisphere (10 90 S) sea level pressure (in Pa) of the preceding February for (a) the DTF summers and (b) the FTD summers in the MLYRV. The shaded dark (light) areas exceed the 95% confidence level based on the t-test. (SH) sea level pressure (SLP) pattern in the preceding February of the DTF and FTD summers. In the preceding February of the DTF summers (Fig. 7(a)), the polar cap is covered by the significantly positive anomaly but in the mid-latitude it appears the significantly negative anomaly. It means that the pressure gradient is small between the middle and high latitude regions in the SH in the preceding February, so the SAM is weak [5]. The situation of the FTD summers is just opposite (Fig. 7(b)). These results are consistent with those in Table 2. To further verify the correlation between the summer LDFA in the MLYRV and the preceding February SAM, the correlation pattern between the summer LDFAI and the composite SH SLP of the preceding February is showed by Fig. 8. The polar cap has the significantly positive correlation values whereas the mid-latitude is the significantly negative correlation region. It suggests that in the preceding February, the higher the summer LDFAI is, the smaller the pressure gradient between the middle and high latitude regions in the SH and the weaker the SAM is, and vice versa. This is in conformity with the above analysis. The pattern of the preceding February NAM is similar but with a weaker intensity. Therefore, the SAM and NAM in the preceding February might be regarded as a predictive signal for the summer LDFA in the MLYRV. Fig. 8. The correlation coefficients between the LDFAI and the composite SH (10 90 S) SLP of the preceding February. The shaded dark (light) areas exceed the 95% confidence level based on the Monte Carlo test. 5 Conclusions Summer LDFA frequently occurs in China, but there is little research on it before. This paper focuses on some basic features of atmospheric circulation abnormity of the anomalous LDFA years in the MLYRV. www.scichina.com www.springerlink.com 2033

Results show that the summer droughts-to-floods (DTF) in the MLYRV often go with the more southward western Pacific subtropical high (WPSH), negative vorticity, strong divergence, descending movements developing and the weak moisture transport in the low level, and the more southward position of the South Asia high (SAH),the westerly jets in the high level during May June; While during July August it usually accompanies with the northward shift of the WPSH, positive vorticity, strong convergence, ascending movements and strong moisture transport in the low level, and the northward shift of the SAH and the westerly jets in the high level. The summer floods-to-droughts (FTD) in the MLYRV is often related to the active cold air mass from the high latitude, positive vorticity, strong convergence, ascending movement developing and strong moisture transport in the low level, and the SAH over the Tibetan Plateau in the high level, while during July August it is often connected with negative vorticity, strong divergence, descending movements developing and weak moisture transport in the low level, the remarkable northward shift of the WPSH, the SAH extending northeastward to North China and the easterly jets prevailing in the high level over the MLYRV. In addition, the summer LDFA in the MLYRV is significantly related with the Southern Hemisphere annual mode and the Northern Hemisphere annual mode in the preceding February, which provides some predictive signals for the summer LDFA in the MLYRV. Acknowledgements The authors would like to thank the ECMWF and National Meteorological Information Center of China for releasing the related data, thank the two anonymous reviewers for their valuable suggestions and comments, and also thank Dr. Dion Manly and Dr. Baohua Chen of Illinois Institute of Technology from the USA for proofreading and revising the paper. This work was jointly supported by the National Natural Science Foundation of China (Grant Nos. 40523001 and 40221503) and the National Basic Research Program of China (Grant No. 2004CB418303). References 1 Zhu K Z. Southeastern monsoon and precipitation in China. Acta Geogra Sin (in Chinese), 1934, 1(1): 1 27 2 Tu C W, Huang S S. The advance and withdraw of Chinese Summer Monsoon. Acta Meteor Sin (in Chinese), 1944, 18(1): 1 20 3 Tao S Y, Xu S Y. Some Aspects of the Circulation during the Periods of the Persistent Drought and Flood in the Yangtze River and Huaihe River valleys in summer. Acta Meteor Sin (in Chinese), 1962, 32: 1 10 4 Gong D Y, Zhu J H, Wang S W. The significant correlation between summer rainfall in the Yangtze River valley and the preceding Arctic Oscillation. Chin Sci Bull, 2002, 47(7): 546 549 5 Nan S L, Li J P. The relationship between the summer precipitation in the Yangtze River valley and the boreal spring Southern Hemisphere annular mode. Geophys Res Lett, 2003, 30(24): 2266, doi: 10.1029/2003GL018381[DOI] 6 He J H, Zhou B, Wen M, et al. Vertical circulation structure, interannual variation features and variation mechanism of western Pa- cific subtropical high. Adv Atmos Sci, 2001, 18(4): 497 510 7 Ju J H, Lu J M, Cao J et al. Possible Impacts of the Arctic Oscillation on the Interdecadal Variation of Summer Monsoon Rainfall in East Asia. Adv Atmos Sci, 2005, 22(1): 39 48 8 Wu Z W, Li J P, He J H, et al. The occurrence of droughts and floods during the normal summer monsoons in the mid- and lower reaches of the Yangtze River. Geophys Res Lett, 2006, 33, L05813, doi: 10.1029/2005GL024487[DOI] 9 Li J P, Wang J. A modified zonal index and its physical sense. Geophys Res Lett, 2003, 30(12): 1632, doi: 10.1029/2003GL017441[DOI] 10 Ting M F, Wang H. Summer time U.S. precipitation variability and its relation to Pacific sea surface temperature. J Climate, 1997, 10: 1853 1873[DOI] 11 Wu Z W, He J H, Jiang Z H. The Compare Analysis of Flood and Drought Features among the First Flood Period in South China, Meiyu and Rainy Season in North China in the Past 50 Years. J Atmos Sci (in Chinese), 2006, 30(3): 391 401 12 Tao S Y, Zhu F K. The l00mb Flow Patterns in Southern Asia in Summer and its Relation to the Advance and retreat of the west-pacific Subtropical Anticyclone 0ver the Far East. Acta Meteor Sin (in Chinese), 1964, 34: 385 395 13 Xu X D, Tao S Y, Wang J Z, et al. The relationship between water vapor transport features of Tibetan Plateau-monsoon large triangle affecting region and drought flood abnormality of China. Acta Meteor Sin, 2002, 60(3): 257 267 14 Zhang R H. Relations of water vapor transport from Indian Monsoon with that over East Asia and the summer rainfall in China. Adv Atmos Sci, 2001, 18(5): 1005 1017 2034 Chinese Science Bulletin Vol. 51 No. 16 August 2006