J. Geogr. Sci. 2011, 21(4): 579-593 DOI: 10.1007/s11442-011-0865-2 © 2011

Science Press

Springer-Verlag

Interdecadal variation of East Asian summer monsoon and drought/flood distribution over eastern China in the last 159 years LI Qian1,2, *WEI Fengying2, LI Dongliang1 1. Nanjing University of Information Science and Technology, Nanjing 210044, China; 2. State Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, China

Abstract: Based on the drought/flood grades of 90 meterological stations over eastern China and summer average sea-level pressure (SLP) during 1850–2008 and BPCCA statistical methods, the coupling relationship between the drought/flood grades and the East Asian summer SLP is analyzed. The East Asian summer monsoon index which is closely related with interdecadal variation of drought/flood distribution over eastern China is defined by using the key areas of SLP. The impact of the interdecadal variation of the East Asian summer monsoon on the distribution of drought/flood over eastern China in the last 159 years is researched. The results show that there are four typical drought and flood spatial distribution patterns in eastern China, i.e. the distribution of drought/flood in southern China is contrary to the other regions, the distribution of drought/flood along the Huanghe River–Huaihe River Valley is contrary to the Yangtze River Valley and regions south of it, the distribution of drought/flood along the Yangtze River Valley and Huaihe River Valley is contrary to the other regions, the distribution of drought/flood in eastern China is contrary to the western. The main distribution pattern of SLP in summer is that the strength of SLP is opposite in Asian continent and West Pacific. It has close relationship between the interdecadal variation of drought/flood distribution patterns over eastern China and the interdecadal variation of the East Asian summer monsoon which was defined in this paper, but the correlation is not stable and it has a significant difference in changes of interdecadal phase. When the East Asian summer monsoon was stronger (weaker), regions north of the Yangtze River Valley was more susceptible to drought (flood), the Yangtze River Valley and regions south of it were more susceptible to flood (drought) before the 1920s; when the East Asian summer monsoon was stronger (weaker), the regions north of the Yangtze River Valley was prone to flood (drought), the Yangtze River Valley and regions south of it were prone to drought (flood) after the 1920s. It is indicated that by using the data of the longer period could get much richer results than by using the data of the last 50–60 years. The differences in the interdecadal phase between the East Asian summer monsoon and the drought/flood distributions in eastern China may be associated with the nonlinear feedback, which is the East Asian summer monsoon for the

Received: 2010-10-10 Accepted: 2010-12-01 Foundation: National Natural Science Foundation of China, No.40890053; Special Scientific Fund for Non-profit Public Industry (Meteorology), No.GYHY200906016; No.GYHY201006038 Author: Li Qian (1984–), Ph.D Candidate, specialized in the climate change and forecast. E-mail: [email protected] * Corresponding author: Wei Fengying, E-mail: [email protected]

www.geogsci.com

springerlink.com/content/1009-637X

580

Journal of Geographical Sciences

extrinsic forcing of solar activity. Keywords: eastern China; drought/flood distribution; East Asian summer monsoon; interdecadal variation

1

Introduction

Drought/flood as one of the most serious climate disasters has brought huge loss to human life, wealth, and economy in China. The characteristic of spatial distribution and temporal evolution of drought/flood were analyzed from different aspects. Especially, based on local annals, historical documents and so on, the atlas entitled Yearly Charts of Dryness/Wetness in China for the Last 500-year Period was completed during the 1970s (CMI, 1981), the dryness/wetness index was classified into five grades. The atlas provides great advantages for investigating long-term evolution of the drought/flood in China. The elementary characteristics of drought/flood in China were analyzed using the atlas and the related data (Zhang et al., 1983). The eastern China was divided into several climate regions and the characteristic of spatial distribution of drought/flood was explored by using the drought/flood grade data (Ye et al., 1997; Zhu, 2003). With the penetrating of the research of interdecadal variation on climate change, the atlas has aroused a lot of interest from more and more scholars at home and abroad. Based on the drought/flood grade data during 1470–1997, Hu et al. (2001) found that the overall variation of drought/flood in eastern China would have moved southward much more at centennial time scale. The distribution of drought/flood in China was discussed for different centuries by using the drought/flood grade data during 1470–1998 (Song, 2000). The relationship between the East Asian summer monsoon and the drought/flood distribution in eastern China was very significant (Zhu, 1934). Different definition standards, including utilizing teleconnection of East Asia–Pacific (e.g., Nitta, 1987; Huang, 2004) and the difference in atmospheric pressure gradient between the sea and land (Guo, 1983) result in diverse East Asian summer monsoon strength indexes (Wang et al., 2004; Gao et al., 2003). The indexes defined by Guo indicated that when the East Asian summer monsoon was stronger than normal, regions north of the Huaihe River and southern China were rainy, the Yangtze River Valley was short of rain; when the monsoon was weaker than normal, the Yangtze River Valley and regions north of the Yangtze River were short of rain; when the index was close to the annual average, the middle and lower reaches of the Yangtze River Valley were rainy. Huang et al. (2001, 2004, 2008) indicated that the East Asian summer monsoon system had a quasi-biennial oscillation at the interannual scale and the interdecadal variations of the system had presented a significant weakening trend since the 1970s, which influenced the distribution of drought/flood in China. The East Asian summer monsoon system was also characterized by teleconnection of the tropospheric temperature between Asia and the mid-latitude Pacific region and it had a significant indication of precipitation in East Asia (Zhao et al., 2007). Based on the grade datasets over northern China for the past 530 years, Zhu et al. (2003) studied the interdecadal variability of the East Asian summer monsoon. The result indicated that the 80a-oscillation of summer rainfall over northern China correlated closely with the long-term change in the East Asian summer monsoon intensity. However, many uncertainties still exist in how the spatial and temporal variability of the East Asian summer monsoon system influenced the anomalous distribution of drought/flood in China. Moreover, few studies focused on the relationship between the interdecadal vari-

LI Qian et al.: Interdecadal variation of East Asian summer monsoon and drought/flood distribution

581

ability of the East Asian summer monsoon and the interdecadal variability of drought/flood in China because of lack of sufficient data in longer period. The aim of this paper is to explore the interdecadal variation of distribution of drought/flood over eastern China responding to the interdecadal variation of the East Asian summer monsoon which is investigated in a longer time scale using the drought/flood grade and SLP data.

2

Data

Data used in this study are as follows: (1) yearly drought/flood grades of 90 stations in eastern China from 1850 to 2008 (Figure 1). Grade 1—very wet, grade 2—wet, grade 3—normal, grade 4—dry and grade 5—very dry. The reliability of data was testified in the literature (Zhang, 1983); (2) monthly mean SLP of June, July and August from 1850 to 2008 with a resolution of 5°×5° and covering areas from 0°N to 80°N and from 40°E to 180°E, being downloaded from the internet database (http://hadobs.metoffice.com); (3) annual relative sunspot number from 1861–2008, being downloaded from the internet database (http://sidc. oma.be/sunspot-data).

3 The relationship between drought /flood grades and SLP in the last 159 years The heating of underlying surfaces, the differences in longitudinal distribution of solar radiation, the difference in heating between the land and sea are the basic reasons for a Figure 1 Distribution of the 90 stations in eastern China monsoon. The difference in heating between the land and sea which is an important factor of monsoon is caused by different terrains between the sea and land surface. The East Asian summer monsoon indexes which were defined by using SLP could well control summer rainfall over eastern China (Guo, 1983; Wei, 2007a). Meanwhile, we obtained SLP and drought/flood grade data of the longer period, which offered advantages for analyzing the relationship between the East Asian summer monsoon and the drought/flood distribution in eastern China over a hundred years. 3.1

The spatial distribution of drought/flood over eastern China in the last 159 years

An EOF analysis is performed on the normalized drought/flood grades over eastern China for the period 1850–2008. The four leading EOF modes together account for 34% of the total drought/flood grade variance. Individually, they explain 12.3%, 8.8%, 7.8% and 5.1% of the variance and all passed North’ rule of thumb (North et al., 1983).

582

Journal of Geographical Sciences

The spatial patterns associated with the first four drought/flood modes are depicted in Figure 2. Figure 2a exhibits that positive anomalies occupies a large area over eastern China, while negative anomalies are located in southern China (Zhao, 2005), i.e. southern China is wetness (dryness) and the other areas of eastern China are dryness (wetness). Figure 2b displays the dryness (wetness) along the Huanghe River–Huaihe River Valley, and the wetness (dryness) in regions north of the Yellow River and regions south of the Yangtze River Valley. Figure 2c exhibits the wetness (dryness) mainly in the region between the Yellow River and Yangtze River, and the dryness (wetness) in the other areas of eastern China. Figure 2d shows that the eastern part of eastern China is prone to wetness (dryness), while the western part of eastern China is prone to dryness (wetness).

Figure 2

Four leading eigenvectors of drought/flood grades (a, b, c and d) during 1850–2008

LI Qian et al.: Interdecadal variation of East Asian summer monsoon and drought/flood distribution

583

Due to the tested consistency between the characteristics of the drought/flood distribution from the drought/flood grade data and that from observed precipitation, the EOF analysis is also performed on the normalized precipitation over eastern China from May to September for the period 1959–2008. The four leading EOF modes together account for 42% of the total drought/flood grade variance. It is clear that the eigenvectors of precipitation have faster convergence than that of the drought/flood grades (not shown). The characteristics of the four leading eigenvectors of precipitation and the drought/flood grades are distinct, and the other eigenvectors are disordered. The first eigenvector of precipitation also shows the distribution of drought/flood in the southern part of eastern China is contrary to the northern part, but the zero line moves northward of 2–3 latitudes compared to the first eigenvector of the drought/flood grades. The distribution of the second eigenvector of precipitation is similar to the third eigenvector of the drought/flood grades. The distribution of the third eigenvector of precipitation is consistent with the second eigenvector of the drought/flood grades. The characteristics of the fourth eigenvector of precipitation and the fourth eigenvector of the drought/flood grades are in longitudinal distribution. Therefore, the four leading eigenvectors of precipitation basically accord with that of the drought/flood grades and they represent the major rainfall patterns over eastern China. 3.2

The spatial distribution of SLP in the last 159 years

A similar EOF analysis is performed on the normalized SLP anomalies during 1850–2008. The fist four EOF modes account for 54.9% of the total SLP variance, while they individually explain fractions of 19.7%, 16.9%, 10.7% and 7.6%. Figure 3 shows the four spatial patterns of SLP. The first eigenvector (Figure 3a) shows the opposite distribution between the higher latitudes and middle and lower latitudes of the Northern Hemisphere, i.e. SLP in higher latitudes are positive anomaly and SLP in the middle and lower latitudes is negative anomaly. The pattern indicates that while SLP in the middle and lower latitudes is weaker (stronger), SLP in the higher latitudes is stronger (weaker). This is the most common mode in the Northern Hemisphere. The second eigenvector (Figure 3b) exhibits the weaker (stronger) SLP in China and the stronger (weaker) SLP in the North Pacific. The opposite distribution of SLP between sea and land is shown on the third eigenvector (Figure 3c), i.e. SLP in the Asian continent is positive anomaly and that in the Northeast Pacific is negative anomaly. The fourth eigenvector (Figure 3d) shows the “+ – +” pattern of SLP from north to south. In summary, the first and fourth eigenvectors of SLP are zonal pattern, and the second and third eigenvectors of SLP are meridional pattern. 3.3

The variability of the coupled fields: BPCCA

Based on BPCCA (Wei, 2007b), the relation between SLP and the drought/flood grades is analyzed. BPCCA is usually applied to two fields together in order to identify pairs of coupled spatial patterns, and it is simple in computation and clear in physical meaning. The sensitive region which reflects the coupled spatial patterns of two fields could be tested. The process consists of three steps. First, an EOF analysis is performed on the normalized fields, and the matrix which combined the time coefficients of the four leading eigenvectors of the drought/flood grades with SLP could be constructed. Then, according to the canonical correlation analysis, the series of canonical variable are obtained. Finally, the correlation coef-

584

Figure 3

Journal of Geographical Sciences

Four leading eigenvectors of SLP (a, b, c and d) during 1850–2008

ficients between the significant canonical variables and these original fields are calculated, the coupled spatial patterns of the two fields are also obtained. The related coefficients of the three leading coupled canonical variables are beyond the 0.05 significant level. Figures 4a and 4b depict the first typical spatial patterns of the coupled fields. The spatial pattern of SLP (Figure 4a) characterizes that the distinctive positive correlation is located in China and the north of 65°N, while the other areas are the negative correlations and the significant negative correlations lie in the Northwest Pacific and the correlation coefficient is up to 0.6. It is similar to macro-scale cell pattern over the Northern Hemisphere (Trenberth et al., 1981; Wallace, 2000), i.e., the variability of SLP has band seesaw structure between the change of atmospheric quality in the middle and high latitudes. The changes in the phase and the strength of this structure play an important role in climate change of China (Gong et al., 2002; Wei, 2006). The spatial pattern of drought/flood over eastern China (Figure 4b) shows a triple structure pattern. The centre of negative correlation is located in southern China and northern China, and the centre of positive correlation is located in the Yangtze River Valley, and they all passed the 0.01 significant level. This pair of coupled spatial patterns manifested that SLP is stronger than normal in the high latitudes and that is weaker than normal in the middle latitudes including the Northwest Pacific, while southern China and northern China are more susceptible to flood and the middle and lower reaches of the Yangtze River are more susceptible to drought, and vice versa. Figures 4c and 4d show the components of the second BPCCA mode of the coupled fields. The typical spatial pattern of SLP (Figure 4c) is characterized by the obvious difference

LI Qian et al.: Interdecadal variation of East Asian summer monsoon and drought/flood distribution

585

Figure 4 Three pairs of canonical correlation fields between SLP and drought/flood grades in 1850–2008 (a and b, c and d, e and f) (shaded areas indicate that the coefficient is above the 0.05 significant level; correlation coefficient×100)

between sea and land. The positive correlation lies over the east of 130°E, and the correlation coefficient is higher than 0.3 around the Northwest Pacific and beyond the 0.01 signifi-

586

Journal of Geographical Sciences

cant level; the negative correlation is located to the west of 130°E, and the most parts of China passed the 0.01 significant level. The spatial pattern of drought/flood over eastern China (Figure 4d) also exhibits a triple structure pattern. The significant negative correlation is located in the middle of Inner Mongolia and Hetao region and most parts of the southern Yangtze River Valley is beyond the 0.01 significant level. This pair of coupled spatial patterns indicates that SLP is stronger than normal in the Northwest Pacific and that is weaker than normal in China, while the middle of Inner Mongolia and Hetao region and the most parts of southern Yangtze River Valley are more susceptible to flood, and vice versa. Figures 4e and 4f depict the third typical spatial patterns of the coupled fields. The typical spatial pattern of SLP (Figure 4e) shows the negative correlation over the north of 45°N and the positive correlation in the other areas. The centre of negative correlation is located around Lake Baikal and the centre of positive correlation is located over the West Pacific region. It is indicated that the variation of pressure in high latitudes and subtropical West Pacific also play an important role in the distribution of drought/flood over eastern China. The spatial pattern of drought/flood over eastern China (Figure 4f) shows a dipole structure pattern. Regions north of the Huaihe River Valley present positive correlation and the centre of it is located in Hetao region, while regions south of the Huaihe River Valley present negative correlation and the centre of it is located in the middle and lower reaches of the Yangtze River. It is demonstrated that SLP is stronger than normal in the subtropical West Pacific and SLP is weaker than normal around Lake Baikal, while the middle and lower reaches of the Yangtze River is more susceptible to flood and Hetao region is more susceptible to drought, and vice versa. The above analysis illustrates that the drought/flood distribution over eastern China is obviously influenced by the variation of SLP difference between sea and land. Hence, the two areas of SLP which have notable influence on the drought/flood distribution over eastern China are chosen. They are region A (Land) from 25°N to 40°N and from 80°E to 100°E and region B (Sea) from 40°N to 60°N and from 140°E to 160°E (Figure 5). Then, the normalized SLP series averaged for region B is subtracted from the normalized SLP series averaged for region A to form the strength of the East Asian summer monsoon.

Figure 5

The selected regions A and B for defining East Asian summer monsoon

LI Qian et al.: Interdecadal variation of East Asian summer monsoon and drought/flood distribution

587

4 Relationship between the East Asian summer monsoon and the drought/flood distribution over eastern China on interdecadal time scale 4.1 The periodic analyses of the typical pattern of SLP and drought/flood over eastern China The time coefficients of the three leading eigenvectors of SLP and drought/flood grades are studied by using power spectrum analyses. The ratio of the calculated spectrum and standard spectrum (the level of significance is 0.05) is power spectrum ratio. If the corresponding period of the spectrum passed the 0.05 significance level, the power spectrum ratio is greater than 1. The power spectrum ratio of the first pattern of SLP (Figure 6a) shows the most clearly 40a oscillation, and 80a, 26a and 13a oscillations. Also the most significant period is 40a oscillation, and 10–20a, 5–6a, 80a and 3.5a oscillations in the power spectrum ratio of the second pattern of SLP (Figure 6c). The power spectrum ratio of the third pattern of SLP (Figure 6e) shows the most clearly 40a oscillation, and 80a, 10–20a oscillations. The 40a oscillation is also found in the first or second period of the power spectrum ratio of the three leading patterns of the drought/flood grades. The power spectrum ratio of SLP and the drought/flood grades have 40a oscillation, which is similarly to the conclusion by Guo et al.

Figure 6 The ratios of power spectrum density are the three leading typical modes of SLP (a, c and e) and the ratios of power spectral density are the three leading typical modes of drought/flood grades (b, d and f)

588

Journal of Geographical Sciences

(2004). So, the 40a oscillation component is chosen to represent the interdecadal variation. 4.2 Relationship between the East Asian summer monsoon and the drought/flood distribution on interdecadal time scale The normalized SLP series averaged for region B is subtracted from the normalized SLP series averaged for region A, the 40a oscillation component of that is to form the index of the East Asian summer monsoon interdecadal variation (called NMSI), which reflects the strength of the East Asian summer monsoon interdecadal variation. The interdecadal variation of NMSI and the three leading eigenvectors of drought/flood grades (Figure 7) show that the interdecadal variation of NMSI and the three leading eigenvectors of drought/flood grades are mainly the same during 1850–1890, but the differences is found after 1891. The variation of NMSI indicates that the East Asian summer monsoon was stronger than normal during 1861–1890, 1915–1929, 1954–1999, and it was weaker than normal during 1850–1960, 1891–1914, 1930–1953 and 2000–2008. A little abrupt of the strength of the East Asian summer monsoon occurred in 1976. The interdecadal variation of the three leading eigenvectors of drought/flood grades (FD1, FD and FD3) explored that they are also the same during 1850–1890, but there is a significant difference in phase, even the opposite phase, i.e. the relationship between NMSI and the typical patterns of drought/flood is extremely complex. In order to get a better understanding of the relation between the East Asian summer monsoon and the drought/flood distribution on interdecadal time scale, the coefficient correlation between NMSI and the typical patterns of drought/flood grades are calculated on 40a moving. The moving correlation coefficients (Figure 8) exhibit that the relationship between NMSI and the interdecadal component of the three leading eigenvectors has remarkable difference in phase. NMSI-FD1 presented positive correlation before 1875–1914, but the relationships are not significant at the other periods. NMSI-FD2 shows positive correlation before 1908–1947 and passes the significant test at 0.01 level, and then it turns negative correlation during 1908–1949 and 1950–1989 and also passes the significant test at 0.01 level, which becomes positive correlation again after 1951–1990. NMSI-FD3 experiences

Figure 7 NMSI and the interdecadal component of the three leading eigenvectors of drought/flood grades (FD1, FD2 and FD3)

LI Qian et al.: Interdecadal variation of East Asian summer monsoon and drought/flood distribution

589

Figure 8 The 40 moving correlation coefficient of NMSI and the interdecadal component of the three leading eigenvectors of drought/flood grades (dotted line: α = 0.01 level of significance)

modulation for four times. Clearly, based on the relatively stable correlation that experienced two adjustments in NMSI-FD2, the relationship between the interdecadal variation of NMSI and the drought/flood distribution is further analyzed. Based on the relationship between NMSI and FD2 and their turning when NMSI-FD2 presents obviously positive correlation and NMSI is stronger than normal during 1861–1890, the interdecadal component of drought/flood grades is constructed (Figure 9a); when NMSI-FD2 exhibited obviously positive correlation and NMSI is weaker than normal during 1891–1914, the interdecadal component of drought/flood grades is constructed (Figure 9c). To the contrary, while NMSI-FD2 shows significant negative correlation and NMSI is stronger than normal during 1954–1975, the interdecadal component of drought/flood grades is constructed (Figure 9d); while NMSI-FD2 shows significant negative correlation and NMSI is weaker than normal during 1930–1953, the interdecadal component of drought/flood grades is constructed (Figure 9d). As there is positive correlation between NMSI and FD2, Figure 9a shows that when NMSI is stronger than normal, Northeast China and regions south of the Yangtze River as well as areas along the Yangtze River are prone to flood, and the centre of them lies in southern China and Northeast China; Hetao region and northern China are prone to drought, and the centre of them lies in Shanxi Province. But, when NMSI is weaker than normal (Figure 9c), regions north of the Yangtze River Valley are prone to flood, and the centre of it is located in Shanxi and Shaanxi provinces; the central and western parts of southern China, Zhejiang Province and the eastern part of Inner Mongolia are prone to drought. For the purpose of a complete and in-depth understanding of the difference in the distribution of the drought/flood influenced by the strength of the East Asian summer monsoon, the difference between Figures 9a and 9c is drawn (Figure 9e). Figure 9e indicates that when the East Asian summer monsoon is stronger (weaker) than normal, regions north of the Yangtze River Valley is more susceptible to drought (flood) and the Yangtze River Valley and regions south of it are more susceptible to flood (drought). When there is negative correlation between NMSI and FD2, while NMSI is stronger

590

Figure 9 periods

Journal of Geographical Sciences

The distribution of interdecadal component of drought/flood grades in eastern China during different

(a) 1861–1890; (b) 1954–1975; (c) 1891–1914; (d) 1930–1953; (e) the difference between 1861–1890 and 1891–1914; (f) the difference between 1954–1975 and 1930–1953 (Shaded areas denote the difference beyond the 0.05 significance level in Figures e and f)

LI Qian et al.: Interdecadal variation of East Asian summer monsoon and drought/flood distribution

591

than normal (Figure 9b), the Huanghe–Huaihe River Valley and southern China are prone to flood, and the Yangtze River Valley is prone to drought. Conversely, while NMSI is weaker than normal (Figure 9d), the Huanghe–Huaihe River Valley is drier, and the Yangtze River Valley is wetter. The difference between Figures 9b and 9d is shown in Figure 9f. It is shown that when the East Asian summer monsoon is stronger (weaker) than normal, regions north of the Yangtze River Valley are more susceptible to flood (drought), and the Yangtze River Valley and regions south of it are more susceptible to drought (flood). As we know, based on the data since 1950, some researches suggest that when the East Asian summer monsoon index was stronger (weaker) than normal, the northern part of eastern China was rainy (short of rain), regions south of the Yangtze River Valley and regions along the Yangtze River were short of rain (rainy). But using the data of the longer period, it could be found that this correlation is not stable and it may change in interdecadal phase. As previously discussed, when the East Asian summer monsoon was stronger (weaker) than normal, regions north of the Yangtze River Valley was more susceptible to drought (flood), and the Yangtze River Valley and regions south of it were more susceptible to flood (drought) before the 1920s; when the East Asian summer monsoon was stronger (weaker) than normal, regions north of the Yangtze River Valley was prone to flood (drought), and the Yangtze River Valley and regions south of it were prone to drought (flood) after the 1920s.

5

Conclusions and discussion

(1) There are four typical drought and flood spatial distribution patterns in eastern China in the last 159 years, i.e. the drought/flood distribution in southern China is contrary to the other areas, the drought/flood distribution along the Huanghe River–Huaihe River Valley is contrary to the Yangtze River Valley and regions south of it, the drought/flood distribution along the Huaihe River Valley is contrary to the other areas, the drought/flood distribution in the eastern part is contrary to the western. The main distribution pattern of SLP in summer is that the strength of SLP are opposite in Asian continent and West Pacific. (2) The coupling relationship between the drought/flood grades and the East Asian summer SLP is analyzed by using the BPCCA statistical methods. The East Asian summer monsoon index which is related closely with the interdecadal variation of drought/flood distribution over eastern China is defined by using the key areas of SLP. It has close relationship between the interdecadal variation of drought/flood distribution patterns over eastern China and the interdecadal variation of the East Asian summer monsoon. When the East Asian summer monsoon was stronger (weaker), regions north of the Yangtze River was more susceptible to drought (flood), and the Yangtze River Valley and regions south of it was more susceptible to flood (drought) before the 1920s; When the East Asian summer monsoon was stronger (weaker), regions north of the Yangtze River was prone to flood (drought), and the Yangtze River Valley and regions south of it were prone to drought (flood) after the 1920s. (3) So, what influenced the changes of interdecadal characteristic on the East Asian summer monsoon? It is speculated that the interdecadal variation of the East Asian summer monsoon was mainly impacted by sea, solar activity and volcanic activity and so on (Gerard, 2005). The polynomial curve of normalized relative sunspot number and the coefficients of the three leading eigenvectors of drought/flood grades are shown during 1861–1914 (Figure

592

Journal of Geographical Sciences

10). According to the polynomial curve of relative sunspot number (Figure 10) and NMSI (Figure 7), when the East Asian summer monsoon is stronger than normal during 1861–1890, the sunspot number is more, and when the East Asian summer monsoon is weaker than normal during 1891–1914, the sunspot number is less, the relationship between the East Asian summer monsoon and the solar activity presents positive feedback. But, since the 1920s, when the East Asian summer monsoon is weaker than normal during 1930–1953, the sunspot number is larger, and when the East Asian summer monsoon is stronger than normal during 1954–1975, the sunspot number is less, they presented negative feedback. This means that the East Asian summer monsoon responding to forcing of solar activity presented differences in interdecadal phase, and both of them are nonlinear feedback. So, the differences in interdecadal phase between the East Asian summer monsoon and the drought/flood distributions in eastern China may be partly explained around the 1920s. With the phases lead or lag, the trends of FD1 and FD3 are opposite with the trend of solar activity, and the trend between FD2 and solar activity is basically the same. It is suggested that there is a relationship between solar activity and the East Asian summer monsoon and the interdecadal variation of drought/flood distributions. Certainly, the physical mechanisms of the relationship are still uncertain. It is preliminarily concluded that the interdecadal variation of the East Asian summer monsoon and solar activity might be associate with the Solar Cycle.

Figure 10 The polynomial curve of sunspot number and the eigenvectors of drought/flood grades on time coefficients during 1861–1975

References Central Meteorological Institute of Chinese Meteorological Administration, 1981. The Drought and Flood Charts in China for the Last 500 Years. Beijing: China Cartographic Publishing House, 1981: 332pp. Gao Hui, Zhang Fanghua, 2003. Comparison on East-Asian summer monsoon index. Journal of Tropical Meteorology, 19(1): 79–86. (in Chinese) Gerard J M Versteegh, 2005. Solar forcing of climate: 2. Evidence from the past. Space Science Reviews, 120: 243–286. Gong Daoyi, Zhu Jinghong, Wang Shaowu, 2002. The significant correlation between precipitation over Changjiang River Valley in summer and earlier stage AO. Chinese Science Bulletin, 47(7): 546–549. (in Chinese) Guo Qiyun, 1983. The summer monsoon intensity index in East Asia and its variation. Acta Geographica Sinica, 38(3): 207–217. (in Chinese)

LI Qian et al.: Interdecadal variation of East Asian summer monsoon and drought/flood distribution

593

Guo Qiyun, Cai Jingning, Shao Xuemei et al., 2004. Studies on the variations of East-Asian summer monsoon during AD 1873–2000. Chinese Journal of Atmospheric Sciences, 28(2): 206–215. (in Chinese) Hu Q, Feng S, 2001. A southward migration of centennial-scale variations of drought/flood in eastern China and the western United States. Journal of Climate, 15: 1323–1328. Huang Gang, 2004. An index measuring the interannual variation of the East Asian summer monsoon: The EAP index. Advances in Atmospheric Sciences, 21: 41–52. (in Chinese) Huang Ronghui, 2001. Decadal variability of the summer monsoon rainfall in East Asia and its association with the SST anomalies in the tropical Pacific. CLIVAR Exchange, 2: 7–8. Huang Ronghui, Gu Lei, Chen Jilong et al., 2008. Recent progresses in studies of the temporal-spatial variations of the East Asian monsoon system and their impacts on climate anomalies in China. Chinese Journal of Atmospheric Sciences, 32(4): 691–719. (in Chinese) Huang Ronghui, Huang Gang, Wei Zhigang, 2004. Climate variations of summer monsoon over China. In: Chang C P (ed.). World Scientific Series on Asia–Pacific Weather and Climate: East Asian Monsoon. World Scientific Publisher, 213–270. Nitta T, 1987. Convective activities in the tropical western Pacific and their impact on the northern hemisphere summer circulation. Journal of the Meteorological Society of Japan, 64: 373–390. North G, Bell T, Cahalan R et al., 1983. Sampling errors in the estimation of empirical orthogonal function. Mon. Wea. Rev., 110(7): 699–706. Song J, 2000. Changes in dryness/wetness in China during the last 529 years. International Journal of Climatology, 20: 1003–1015. Trenberth K E, Paolino D A, 1981. Characteristic pattern of variability of sea level pressure in the Northern Hemisphere. Mon. Wea. Rev., 109, 1169–1189. Wallace J. M, 2000. North Atlantic Oscillation/Annular Mode: Two paradigms – one phenomenon. Quart. J Royal Met. Soc., 126: 791–805. Wang B, Linho, Zhang Yongsheng et al., 2004. Definition of South China Sea monsoon onset and commencement of the East Asia summer monsoon. Journal of Climate, 17: 699–710. Wei Fengying, 2006. Relationships between precipitation anomaly over the middle and lower reaches of the Changjiang River in summer and several forcing factors. Chinese Journal of Atmosphere Sciences, 30(2): 202–211. (in Chinese) Wei Fengying, 2007a. The integration estimation models of summer rain belt distribution. Progress in Natural Science, 17(5): 639–645. (in Chinese) Wei Fengying, 2007b. Modern Climatic Statistical Diagnosis and Forecasting Technology. 2nd ed. Beijing: China Meteorological Press, 157–160. (in Chinese) Ye Jinlin, Wang Shaowu, Li Xiaodong, 1997. A study of drought/flood type in eastern China. Quarterly Journal of Applied Meteorology, 8(1): 70–77. (in Chinese) Zhang De’er, 1983. Reconstruction of the past 500a and the reliability of climate sequences. In: Chinese Academy of Meteorological Sciences. Meteorological Science and Technology Bulletin (Climate and Drought). Beijing: China Meteorological Press, 17–26. (in Chinese) Zhang Jiacheng, Zhang Xiangong, Xu Xiejiang, 1983. Wet and dry in China of past 500 years. In: Chinese Academy of Meteorological Sciences. Meteorological Science and Technology Bulletin (Climate and Drought), Beijing: China Meteorological Press, 1–16. (in Chinese) Zhao Nan, 2005. Zonally symmetric normal modes associated with the AO/NAM. Acta Meteorologica Sinica, 63(6): 857–863. (in Chinese) Zhao P, Zhu Y N, Zhang R H, 2007. An Asian–Pacific teleconnection in summer tropospheric temperature and associated Asian climate variability. Climate Dynamics, 29: 293–303. Zhu Jinhong, Wang Shaowu, Mu Qiaozhen, 2003. 80a-oscillation of summer summer rainfall over the east part of China and East-Asia summer monsoon. Advances in Atmospheric Sciences, 18(5): 1043–1051. (in Chinese) Zhu Kezhen, 1934. Southeast monsoon and precipitation over China. Acta Geographica Sinica, (1): 1–28. (in Chinese) Zhu Yafen, 2003. The regional division of dryness/wetness over eastern China and variations of dryness/wetness in northern China during the last 530 years. Acta Geographica Sinica, 58(suppl.): 100–107. (in Chinese)