ORIGINAL PAPER

Stefan Becker Æ Marco Gemmer Æ Tong Jiang

Spatiotemporal analysis of precipitation trends in the Yangtze Rivercatchment

Published online: 7 February 2006� Springer-Verlag 2006

Abstract Precipitation trends in the Yangtze Rivercatchment (PR China) have been analyzed for the past50 years by applying the Mann-Kendall trend test andgeospatial analyses. Monthly precipitation trends of 36stations have been calculated. Significant positive trendsat many stations can be observed for the summermonths, which naturally show precipitation maxima.They were preceded and/or followed by negative trends.This observation points towards a concentration ofsummer precipitation within a shorter period of time.The analysis of a second data set on a gridded basis with0.5� resolution reveals trends with distinct spatial pat-terns. The combination of classic trend tests and spa-tially interpolated precipitation data sets allows thespatiotemporal visualization of detected trends. Monthswith positive trends emphasize the aggravation of severesituation in a region, which is particularly prone to flooddisasters during summer. Reasons for the observedtrends were found in variations in the meridional windpattern at the 850 hPa level, which account for an in-creased transport of warm moist air to the YangtzeRiver catchment during the summer months.

Keywords East Asian monsoon Æ Precipitation Æ

Climate trends Æ Mann-Kendall trend test Æ YangtzeRiver catchment

1 Introduction

Summer precipitation over central China is associatedwith the East Asian summer monsoon. It begins with atrough forming over the South China Sea in mid-May(Domroes and Peng 1988). It moves northward in con-cert with the expansion of the North Pacific high andreaches central China mid June (Ho et al. 2003). TheYangtze River flood events in the 1990s and theincreasing losses due to floods started a new discussionon possible implications of climatic change (Gemmer2004). Although, heavy monthly precipitation in theYangtze River catchment occurred regularly duringdistinct wet sequences in the twentieth century, summerprecipitation in the 1990s is claimed to be higher than inthe earlier decades (Li et al. 1999; Xu 2001). Simulta-neously, northern China became much drier accordingto Gong et al. (2001) and Xu (2001). Meanwhile, anincrease of rainfall over eastern China without a de-crease of the number of dry years has been predicted forthe actual and next decade (Gong and Wang 2000; Yangand Yuan 1996).

Causes and regional impacts of climatic change orvariability have been widely discussed under variousaspects. The latest IPCC report (IPCC 2001) indicates a30–50% increase of precipitation in southern China inthe winter months (December, January, February) from1900 to 1999. It describes an inconsistent pattern with anincrease in the western and a decrease in the easternYangtze River catchment in the summer months (June,July, August) from 1900 to 1999. Zhai et al. (1999) (fromIPCC 2001) showed a significant increase in precipita-tion over the middle and lower reaches of the YangtzeRiver (river course downstream Yichang hydrologicalstation and Hukou hydrological station, respectively, cp.Fig. 1 and Table 1) and western China during the latterpart of the twentieth century, while also detecting adeclining trend in precipitation over northern China.

Much attention has been paid to the evaluation of thesummer precipitation (May–August) in the Yangtze

S. Becker (&)Department of Geography and Urban Planning,University of Wisconsin Oshkosh, 800 Algoma Blvd.,Oshkosh, WI, 54901 USAE-mail: beckers@uwosh.edu

M. GemmerEuropean Commission, DG Joint Research Centre,Institute for the Protection and Security of the Citizen,TP 723, 21020 Ispra, Italy

T. JiangNanjing Institute of Geography and Limnology,Chinese Academy of Sciences,Nanjing, People’s Republic of China

Stoch Environ Res Risk Assess (2006) 20: 435–444DOI 10.1007/s00477-006-0036-7

River catchment and the East Asian monsoon as thesedata are strongly related to flood risks (Gong et al. 2001;Zhu and Wang 2001). Huang et al. (2001) gave anupdated contemplation of the progresses of the recentstudies on the variabilities of the East Asian monsoonand their causes. Various trends of summer precipitationamounts, such as a negative trend for eastern Chinafrom 1954 to 1976 and a positive trend between 1977and 1988 have been detected by Gong and Wang (2000).Qian and Zhu (2001) observed an increase of summerprecipitation in northern China for the same period.Jiang and You (1996) described an abrupt reduction ofsummer precipitation from the 1960s to the 1980s inmiddle China and a somewhat weaker jump from a dryto a wet period in the middle and lower Yangtze Rivercatchment, which started in the 1970s. Nevertheless,

below average precipitation fell in the same area from1981 to 1985 (Yang and Yuan 1996). Schaefer (2001)found positive trends of summer precipitation in south-eastern China and negative trends in the north since1951.

Fu and Wen (1999) mentioned a change in the EastAsia monsoon characteristics that started in the 1920s,which might contribute to the explanation of thosefindings. Precipitation in the lower and middle reachesof the Yangtze River is below average and deficient in astrong summer monsoon year such as 1992 or 1997whereas it is sufficient in the same area during weakmonsoon years such as 1991 and 1998 (Li et al. 2001;Zhang and Tao 1998). In general, the rain belt movesfaster northward during a strong summer monsoonyear. According to Xu (2001), a southward move of the

Table 1 Reaches of the Yangtze River

Upper reaches Middle reaches Lower reaches

Length (km) 4,504 955 938Beginning station(Province)–End station (Province)

Source(Qinghai)–Yichang (Hubei)

Yichang(Hubei)–Hukou (Jiangxi)

Hukou(Jiangxi)–Mouth(Jiangsu)

Catchment size of the reach (km2) 1,000,000 680,000 120,000Gradient (m) 5,100 45–50 <20

Fig. 1 Location of the measuring stations

436

monsoon rain belt started in the late 1970s and led to atrend of decreasing summer precipitation in northernChina and increasing summer precipitation over theYangtze River catchment (Xu 2001).

Most existing studies on precipitation trends inChina, such as Gong and Wang (2000) and Qian andZhu (2001), are based on regional analyses, which mergestation data information of the precipitation stations.However, monthly precipitation trends and their re-gional pattern should not be neglected by merging sev-eral stations to one region. Two underlying questions ofthe following analyses are (1) whether there is a signifi-cant change of summer precipitation towards the 1990s,and (2) how regional characteristics of precipitation inthe 1990s can be evaluated in the scope of last century.In order to find an appropriate answer, the nonpara-metric Mann-Kendall trend test is applied for stationdata and gridded data sets.

2 Data and methods

Monthly precipitation data of 36 stations in the YangtzeRiver catchment have been analyzed in the presentstudy. Fig. 1 shows the location of the stations. Unfor-tunately, only a few reliable data were available for theupper reaches of the Yangtze river, especially the2,700 km of the river upstream Yibin (cp. Fig. 1 andTable 1). The data were taken from the Global Histor-ical Climatology Network (GHCN) project version 2(Vose et al. 1992; Peterson et al. 1998). They have beenprovided by the Chinese Academy of Sciences (CAS)and collected at the Carbon Dioxide InformationAnalysis Center for the use in several studies (Kaiseret al. 1995; Tao et al. 1997). Original data from CAShave been used in earlier studies such as Wang et al.(2000). They are mainly available until 1993. The stationdata were individually tested and visually inspected forthe presence of spurious trends and jumps by Kaiseret al. (1995). The US Department of Energy’s (DOE)Carbon Dioxide Information Analysis Center (CDIAC)has conducted a quality assurance (QA) review of thedata, checking them for completeness, reasonableness,and accuracy.

Precipitation data have been published for the prov-ince capitals in the China Statistical Yearbooks (ChinaState Statistical Bureau 1988–2001) since the foundingof the PR China. Data gaps in the GHCN time-seriesfrom 1988 to 2000 have been filled in with data fromthese yearbooks or data from Yangtze River WaterResources Commission (2000).

The homogeneity of the precipitation records wasanalyzed by calculating the von Neumann ratio (N), thecumulative deviations (Q/n�0.5 and R/n�0.5), and theBayesian procedures (U and A) (Buishand 1982). Thedata sets of most stations proved to be homogeneouswith significance beyond the 95% confidence level. Onlydata sets of Chengdu, Guiyang, and Zhijiang showhomogeneity with significance slightly less than the 95%

confidence level according to some of the appliedcriteria.

The trends of monthly precipitation totals have beenanalyzed by applying the Mann-Kendall trend test forall of the 36 stations from 1950 to 1999. Confidencelevels of 90, 95, and 99% were taken as thresholds toclassify the significance of positive and negative precip-itation trends. Trends at significance below the 90%confidence level were not considered.

An additional data set was used for the calculationand illustration of the spatial distribution of precipita-tion trends. These evaluations are based on the ‘‘Ter-restrial Air Temperature and Precipitation: Monthlyand Annual Time Series (1950–1999), (Version 1.02)’’which was produced and documented by Willmott andMatsuura (1995) (see also Legates and Willmott 1990).The applied interpolation method is named Climato-logically Aided Interpolation (CAI). It is documented inWillmott and Matsuura (1995) and Willmott andRobeson (1995).

First of all, a centroid (mid-point) for each raster cellwas calculated by using a Geographical InformationSystem (GIS). The database resulting from this analysishad the spatial entities X and Y for each point and IDswere created. These IDs were connected with the pre-cipitation information from the raster cells. Then, theMann-Kendall trend test was applied for the attachedinformation of each of these points. Finally, raster cellswith a 0.5� resolution were built for the precipitationtrend information of each of the centroids by creating agrid surface for the Yangtze River catchment with aGIS.

Data on time series of mei-yu characteristics wereobtained from the Nanjing Bureau of Water Resources(Nanjing Hydrological Data, data from 1951 to 2001).They have been replenished by data from the YangtzeRiver Yearbooks. The Mann-Kendall trend test was alsoapplied to these data.

3 Observed precipitation trends

3.1 Station data

Figure 2 shows some examples of the precipitation timeseries in the Yangtze River catchment with data availablefor most of the last century. Considerable differencesbetween the stations can be observed regarding absolutevalues, variations, and temporal courses. The analysis ofthe 100-year records ofGHCN from 1900 to 2000 for fourstations reveals no significant trends. It can also be seenthat the annual precipitation sums at all stations are notexceptional in the 1990s in relation to the last century.Both, the annual values and the low-pass filtered data ofthe 1990s are well within the ranges of the precedingdecades. The high-pass filtered data also do not showany characteristic trends. Variabilities in the 1990s arerelatively high but not exceptional in comparison with theprevious decades. The increase of flood disasters in the

437

Yangtze River catchment can therefore be explainedneither by annual precipitation trends nor by increasedinter-annual variabilities.

Monthly precipitation shows maximum amounts inthe summer months June, July, and August. The maximaat the stations in the north and west of the Yangtze Rivercatchment (e.g. Chengdu, Hefei, Kunming, Nanjing, Xi-chang, Yichang) can be detected comparatively late inJuly or August. Maximum precipitation can be noted atmost other stations in June. This phenomenon is causedby the South East Asian summer monsoon that is de-scribed inDomroes andPeng (1988), Ju andSlingo (1995),Tao et al. (1998), and Zhang and Tao (1998).

Precipitation trends at the 36 stations from 1950 to1999 were investigated by applying the Mann-Kendalltrend test. The significance of trends was evaluatedaccording to the 90, 95, and 99% confidence levels.Figure 3 shows the results of the Mann-Kendall trendtests for each station and each month.

It is clearly visible, that monthly precipitation trendsin the Yangtze River catchment do not show a contin-uous pattern for all stations and investigated time frame.However, some characteristics are predominant formany stations from 1950 to 1999. A typical feature formany stations is the significant positive trend for winterprecipitation in January. Other striking features are thepositive trends during the summer months with theprecipitation maxima. These months are typically pre-ceded and/or followed by months with negative trends.Twenty-five of the 36 stations show this characteristicfeature. The precipitation totals during the summermonth which show positive trends can vary from stationto station, however, the underlying pattern is clearlyrecognizable. This observation is further backed up bythe relative monthly frequency of stations in the YangtzeRiver catchment with significant trends, which can betaken from Fig. 4. January precipitation has increasedsignificantly at more than 50% of the stations if the 90%confidence level is taken as a threshold. Even morestations show positive trends with significance slightlybelow this threshold. At about 30% of the stations, noor hardly any significant increase is noted for May andSeptember but a significant decrease can be detected. Onthe contrary, June and July are marked by a strikingprevalence of positive trends. This is an indication forthe tendency towards a concentration of summer pre-cipitation within a shorter period of time.

Besides this temporal feature of monthly precipita-tion trends, the spatial occurrence is an important fac-tor. It is analyzed on a gridded basis that allows thespatial and temporal visualization of precipitationtrends.

3.2 Gridded data

Monthly precipitation trends for time-series between1950 and 1999 based on data from Willmott andMatsuura (1995) were analyzed by applying the Mann-Kendall trend test to the interpolated data set. The trendtest was applied to every of the 0.5� raster cells in theYangtze River catchment. Figure 5 depicts the spatialdistribution of the observed trends in the Yangtze Rivercatchment for May, June, July, and August. The shadingof the maps refers to the confidence levels of theobserved trends at each grid point. The light shadesindicate significant negative trends or decrease of pre-cipitation whereas the darker shades indicate positivetrends or increase of precipitation. Precipitation trendsin the large area with medium gray shade are not sig-nificant beyond the 90% confidence level.

It is clearly visible again that large parts of theYangtze River catchment are dominated by negativetrends in May and by positive trends in June and July.Significant trends predominate mainly in parts of thesouthern and eastern catchment area. The concentrationof precipitation within a shorter period of time withincreased amounts of precipitation in the months ofprecipitation maximum in the summer months is obvi-ously one aspect that contributed to the aggravation offlood disasters in the Yangtze River catchment area inthe last decade. This conclusion corresponds somewhatto the findings of Ho et al. (2003), who observed ashorter relatively dry period between the primary andsecondary precipitation peak in Korea, which are alsoassociated with the South Asian monsoon.

The magnitude of the July precipitation increase inthe Yangtze River catchment is displayed in Fig. 6. Theinterpolated regions are based on the deviation of anaverage 1991–2000 July precipitation, in comparison to1951–1990 in percent. The values vary between �38.1%in Mianyang in the northern region and +119.6% inYueyang in the central region. The central and easternregions, which were showing significant positive trends

Nanjing

Shanghai

500

1900

1920

1940

1960

1980

2000

1900

1920

1940

1960

1980

2000

1900

1920

1940

1960

1980

2000

1900

1920

1940

1960

1980

2000

1000

1500

P [

mm

]P

[m

m]

P [

mm

]P

[m

m]

2000

Year

Wuhan

Year

Chongqing

500

1000

1500

2000Nanjing

Shanghai

-500

-250

0

250

500

Year

Wuhan

Year

Chongqing

-500

-250

0

250

500Fig. 2 Left AnnualPrecipitation (mm, original andGauss-low-pass filtered data),Right Annual Precipitation(mm, Gauss-high-pass filtereddata and absolute decadalmeans)

438

in July (Fig. 5, lower left), is dominated by anomaliesabove +41%. The increase of summer precipitation inthe region which is particularly affected by flood disas-ters has obviously been considerable.

4 Interpretation

Summer precipitation in the Yangtze River catchmentis directly connected with the southwesterly flow whichis a part of the summer monsoon in the East Asian

region. Heavy and long-lasting precipitation in thestudy area originates from a quasi-stationary frontwhich is commonly described as mei-yu. The averageonset of the mei-yu in the middle and lower YangtzeRiver catchment (catchment east of Yichang station) isusually in mid-June. These conditions last forapproximately one month, although the length of themei-yu season varied between 0 days (null mei-yuseason) and over 50 days since 1950 (Tao et al. 1998).The length of the mei-yu season is generally used todescribe the activity of the East Asian monsoon and

Anqing Changde Changsha Chengdu Chongqing Enshi

Ganzhou Guiyang Hangzhou Hanzhong Hefei

Hengyang Hueili Jian Kunming Lijiang Lingling

M ianyang Nanchang Nanchong Nanjing Nanyang Neijiang

Shanghai Shashi Xichang Yaan Yibin Yichang

Youyang Yueyang Yunxian Zhichang

Guilin

Zhongxiang Zunyi

Fig. 3 Significance of monthly precipitation trends based on theconfidence levels of 90, 95, and 99%. The abscissa marks the12 months and the ordinate marks the confidence levels of the

observed trends. The six gray lines running parallel the abscissamark the confidence levels of 90, 95 and 99% for positive trends(above the abscissa) and negative trends (below the abscissa)

439

the meteorological flood risk in the respective floodseason.

After observing a change in the precipitation char-acteristics in the Yangtze River catchment one wouldtherefore also expect a change in the southwesterly flowand the mei-yu characteristics. In fact, Fig. 7 showsdistinct trends toward an earlier onset and a longerduration of the mei-yu season at the example of Nanjingfrom 1950 to 2001. The Mann-Kendall trend test revealsa significant negative trend for the date of the onset ofthe mei-yu season (confidence level: 99.5%) and a

significant positive trend for the duration of the mei-yuseason (confidence level: 99.2%).

One reason for the described changes of precipitationcharacteristics in the summer months is therefore clearlya change in the monsoon circulation. This is backed upby Qian et al. (2003), who detected a significant changein 850 hPa southwesterly summer winds. The summers1954–1976 were characterized by strong southerly andsouthwesterly winds which reached north and northeastChina and brought abundant precipitation for thoseregions. Those winds were rather weak during thesummers 1977–1999, and they only reached the lowerYangtze River catchment. This pattern may explainsome of the observed changes of precipitation charac-teristics, especially the increase of precipitation in thesummer months in the Yangtze River catchment as wellas the negative precipitation trends in north and north-east China.

Additional information can be obtained by having acloser look at the summer wind patterns without aver-aging them over a long time. Figure 8 exemplifies thedifference between the v-wind patterns at the 850 hPalevel during two summers. The data were taken from theNOAA-CIRES/Climate Diagnostics Center. Little pre-cipitation was observed during the 1985 summerwhereas the 1998 summer was marked by heavy pre-cipitation in the Yangtze River catchment area and innortheast China. The summer 1985 is marked by rela-tively weak southerly winds over most of the research

-40

-30

-20

-10

0

10

20

30

40

50

60

1 2 3 4 5 6 7 8 9 10 11 12

Month

Fre

quency (

%)

Fig. 4 Relative monthly frequency of stations with significanttrends (based on the 90% confidence level)

Fig. 5 Interpolated precipitation trends in the Yangtze River catchment 1950–1999

440

area which are extending far to the north. The summer1998 is marked by relatively strong southerly winds inthe southern area which are met by northerly winds athigher latitudes. This leads to the extensive transport ofhumid air masses to central China. They do not reachany further to northern China which again explains theheavy precipitation in the Yangtze River catchment. Anincreasing number of wind patterns which are similar tothe 1998 summer can be observed since the 1970s. Thesestrong southwesterly winds can explain the observedpositive precipitation trends in that area.

Another aspect, which may account for the observedtrends is the coupling between precipitation in theYangtze River catchment and the sea surface tempera-tures (SST). Mao and Wu (2000) found a positive cor-relation between the SST anomalies over the equatorialIndian Ocean in May and precipitation in the YangtzeRiver valley in July. Deng et al. (1989) indicated thatsummer precipitation in the Yangtze River valley has atwo-month lagged positive correlation with the SST inthe Arabian Sea. Guo et al. (2002) observed positive SSTanomalies over large parts of the Indian Ocean and large

140

150

160

170

180

190

200

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

Onset of th

e M

eiY

u P

eriod (

Julia

n D

ate

)

0

10

20

30

40

50

60

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000L

ength

of th

e M

eiY

u P

eriod (

days)Fig. 7 Onset and length of the

mei-yu season in Nanjing1950–2001

Fig. 6 Precipitation anomaly of the 1991–2000 period in comparison to the period 1951–1990 in percent

441

moisture transports from the Bay of Bengal to theYangtze River valley during years with heavy precipi-tation in the Yangtze River catchment (e.g. 1998). Thefindings of these studies underline the strong connectionof v-winds with heavy precipitation in the Yangtze Rivercatchment.

Huang et al. (2003) trace summer precipitation vari-ability and flood events back to the variability of theEast Asian monsoon, and later again back to the vari-ability of SST, Southern Oscillation, and the thermaleffect of the Tibetan Plateau. Ho et al. (2003) suggestthat deforestation and the expansion of arid areas innorthern China and Mongolia lead to a cooling trend ofsurface temperatures and a significant increase ofdownward motion in the lower troposphere. This in-duced lower troposphere convergence and increasedsummer precipitation over central China. The sugges-tion is based on observed variations of geopotentialheights at 700 hPa which shoed a clear discontinuity inthe early 1970s. These factors might also be relevant for

the increase of summer precipitation in the YangtzeRiver catchment from 1950 to 1999.

5 Conclusion and outlook

Significant precipitation trends in the Yangtze Rivercatchment have been observed for the past 50 years. Themethods applied in this work are highly effective tovisualize these trends on a spatiotemporal scale. Themain advantage of this method is that it preserves muchof the computational efficiency associated with geosta-tistical analyses, while at the same time offering amethod for accessing spatially interpolated climate datasets (on a gridded basis). The information provided bythese data sets exceeds the information which can beobtained by analyzing the data of classical observationnetworks (station data basis). This applies especially forthe spatial resolution of 0.5� covering the entire YangtzeRiver catchment.

Fig. 8 June/July average V-Wind at the 850 hPa level in m/s of the years 1985 (upper) and1998 (lower)

442

Herewith, the analyses show that annual precipita-tion trends in the 1990s do not play an extraordinaryrole in explaining the increase of flood disasters in thesecond half of the twentieth century. However, summerprecipitation shows a distinct tendency towards higheramounts within a shorter period of time at several sta-tions. This is partly relevant for the increase of flooddisasters in the 1990s of the last century.

Numerous studies have been conducted and will stillbe necessary to explain the summer precipitation vari-ability and heavy precipitation events. A matter ofconcern is clearly the observed distinct trends towardsan earlier onset and a longer duration of the mei-yuseason which is initiated by strong and weak southerlyand southwesterly winds. The coupling between precip-itation in the Yangtze River catchment and the seasurface temperatures (SST) proves that many climato-logically and meteorological factors influence the ob-served precipitation trends. As can be seen in the reviewof section 4, also the variability of SST, SouthernOscillation, the thermal effect of the Tibetan Plateau,and the expansion of arid areas in northern China andMongolia can influence long-term and seasonal precip-itation in the Yangtze River catchment. Obviously, alarge number of factors are causing the observed trendsand a clear statement such as ‘‘Global Warming in-creases flood risk in China’’ cannot be made. Includingall drivers of a possible climate change in the YangtzeRiver catchment and assessing the underlying telecon-nections is a great challenge and is actually investigatedby the authors. An aggravation of the present floodhazard and also other weather driven natural hazardshas to be expected if the detected trends of the last50 years will prevail in the future.

Nevertheless, the effect of climate change must not beoverestimated, as the reasons for monetary losses due toflood disasters at the Yangtze River as well as alongother rivers are chiefly man-made interferences with theriver flow and in the catchment area (King et al. 2001;Gemmer 2004). Even if trends of precipitation andrunoff in the Yangtze River catchment can be detected,the main reason for flood losses is human activity in thefloodplain.

Acknowledgement This study was initiated for the preparation of aSino-German project which is now funded by the German Re-search Foundation (Deutsche Forschungsgemeinschaft, DFG,Project Number BE 1962/6-1). The original work was funded by akey project of the Chinese Academy of Sciences (CAS, Grant No.KZCX3-SW-331) and the Chinese Meteorological Administration(CMA, Climate Change Project ccsf2005-2-qh4).

References

Buishand TA (1982) Some methods for testing the homogeneity ofrainfall records. J Hydrol 58:11–27

China State Statistical Bureau (1988–2001) China StatisticalYearbook 1988–2001. China Statistical Publishing House,Beijing

Deng A, Tao S, Cheng L (1989) The temporal and spatial charac-teristics of Indian Ocean SST and it’s relationship with summerrainfall in China (in Chinese). Sci Atmos Sin 13:393–399

Domroes M, Peng G (1988) Climate of China. Springer, BerlinHeidelberg New York

Fu C, Wen G (1999) Variation of ecosystems over East Asia inassociation with seasonal, interannual and decadal monsoonclimate variability. Clim Change 43:477–494

Gemmer M (2004) Decision support for flood risk management atthe Yangtze River by GIS/RS-based flood damage estimation.Shaker, Aachen

Gong DY, Wang SW (2000) Severe summer rainfall in Chinaassociated with enhanced global warming. Clim Res 16:51–59

Gong DY, Zhu JH, Wang SW (2001) Flooding 1990s along theYangtze River, has it concern of global warming? J Geogr Sci11(1):43–52

Guo Y, Zhao Y, Wang J (2002) Numerical simulation of therelationships between the 1998 Yangtze River valley flood andSST anomalies. Adv Atmos Sci 19(3):391–404

Ho C-H, Lee J-Y, Ahn M-H, Lee H-S (2003) A sudden change insummer rainfall characteristics in Korea during the late 1970s.Int J Climatol 23:117–128

Huang R, Zhou L, Chen W (2003) Progresses of recent studies onthe variabilities of the East Asian monsoon and their causes.Adv Atmos Sci 20(1):55–69

IPCC (Intergovernmental Panel on Climate Change) (2001) Cli-mate change 2001: the scientific basis. Cambridge UniversityPress, Cambridge

Jiang J-M, You X-T (1996) Where and when did an abrupt climaticchange occur in China during the last 43 years? Theor ApplClimatol 55:33–39

Ju J, Slingo J (1995) The asian monsoon and ENSO. Quart J RMeteor Soc. 121:1133–1168

Kaiser D, Tao S, Fu C, Zeng Z, Zhang Q, Wang W-C, Karl T(1995) Climate Data Bases of the People’s Republic of China1841–1988. ORNL/CDIAC-102, ORNL-TR 055. CarbonDioxide Information Analysis Center, Oak Ridge NationalLaboratory, U.S. Department of Energy, Oak Ridge, Tennessee

King L, Gemmer M, Wang R (2001) Hochwasserschutz undLandnutzung am Yangtze (In German). GeographischeRundschau 53(10):28–34

Legates DR, Willmott CJ (1990) Mean Seasonal and Spatial Var-iability in Gauge-Corrected, Global Precipitation. Int J Cli-matol 10:111–127

Li C, Long Z, Zhang Q (2001) Strong/weak monsoon activity overthe south China sea and atmospheric intraseasonal oscillation.Adv Atmos Sci 18(6):1146–1160

Li J, Chen J, Wang X (1999) Comparative analysis on 1998/1954Precipitation along the Yangtze River and discussion on thecauses to the 1998 flood. Natural disaster reduction in China(8)2:55–58

Mao J, Wu G (2000) Variations of Indian Ocean SST and theirrelationship with atmospheric circulation. In: Sun A et al (eds)Diagnostic study and prediction method of serious drought andflooding and low temperatures (in Chinese). China Meteoro-logical Press, Beijing, pp 41–49

Peterson TC, Vose R, Schmoyer R, Razuvaev V (1998) Globalhistorical climatology network (GHCN) quality. Int J Climatol18:1169–1179

Qian W, Zhu Y (2001) Climate Change in China from 1880 to 1998and its Impact on the Environmental Condition. Clim Change50:419–444

Qian W, Hu Q, Zhu Y, Lee D-K (2003) Centennial-scale dry-wetvariations in East Asia. Clim Dyn 21(1):77–89

Schaefer D (2001) Rezente Klimaaenderungen in China? (In Ger-man) Geographische Rundschau 53(10):42–46

Tao S, Fu C, Zeng Z, Zhang Q (1997) Two Long-Term Instru-mental Climatic Data Bases of the People’s Republic of China.ORNL/CDIAC-102, NDP-039. Carbon Dioxide InformationAnalysis Center, Oak Ridge National Laboratory, USDepartment of Energy, Oak Ridge

443

Tao S, Li J, Wang A (1998) The East Asian Monsoon andFlooding in China. Natural Disaster Reduction in China(7)1:14–18

Vose RS, Schmoyer RL, Steurer PM, Peterson TC, Heim R, KarlTR, Eischeid JK (1992) The global historical climatologynetwork: long-term monthly temperature, precipitation,ORNL/CDIAC-102, NDP-041. Carbon dioxide informationanalysis center, Oak ridge national laboratory, US departmentof energy, Oak Ridge

Wang SW, Gong DY, Ye JL, Chen ZH (2000) Seasonal precipi-tation series over China since 1880 and the variability (in Chi-nese). Acta Geogr Sin 55(3):281–293

Willmott CJ, Matsuura K (1995) Smart Interpolation of AnnuallyAveraged Air Temperature in the United States. J Appl Meteor34:2577–2586

Willmott CJ, Robeson SM (1995) Climatologically aided interpo-lation (CAI) of terrestrial air temperature. Int J Climatol15:221–229

Xu Q. (2001) Abrupt change of the mid-summer climate in centralEast China by the influence of atmospheric pollution. AtmosEnviron 35:5029–5040

Yang F-L, Yuan C-G (1996) Comparative tests and analyses onmonsoon and regional precipitation. Theor Appl Climatol55:163–176

Yangtze River Water Resources Commission (2000) Yearbook forYangtze River Improvement. The Publishing House of theCommission of Water Resources of the Yangtze River, Beijing.Original issue from 1992 added with unpublished data of thecommission (in Chinese)

Zhang Q, Tao S (1998) Relationship between tropical and sub-tropical summer monsoon over East Asia and summer rainfallover East China (in Chinese). Quart J Appl Meteor 9:16–23

Zhu JH, Wang SW (2001) 80a-Oscillation of summer rainfall overthe east part of China and east-Asian summer monsoon. AdvAtmos Sci 5(18):1044–1051

444