A Review of Seasonal Climate Prediction Research in China

WANG Huijun,FAN Ke,SUN Jianqi,LI Shuanglin,LIN Zhaohui,ZHOU Guangqing, CHEN Lijuan,LANG Xianmei,LI Fang,ZHU Yali,CHEN Hong,and ZHENG Fei

1Nansen-Zhu International Research Center,Institute of Atmospheric Physics,Chinese Academy of Sciences,Beijing100029

2Institute of Atmospheric Physics,Chinese Academy of Sciences,Beijing100029

3Climate Change Research Center,Chinese Academy of Sciences,Beijing100029

4National Climate Center,China Meteorological Administration,Beijing100081

A Review of Seasonal Climate Prediction Research in China

WANG Huijun∗1,3,FAN Ke1,3,SUN Jianqi1,3,LI Shuanglin1,3,LIN Zhaohui1,2,ZHOU Guangqing2,3, CHEN Lijuan4,LANG Xianmei1,3,LI Fang2,3,ZHU Yali1,3,CHEN Hong2,3,and ZHENG Fei2,3

1Nansen-Zhu International Research Center,Institute of Atmospheric Physics,Chinese Academy of Sciences,Beijing100029

2Institute of Atmospheric Physics,Chinese Academy of Sciences,Beijing100029

3Climate Change Research Center,Chinese Academy of Sciences,Beijing100029

4National Climate Center,China Meteorological Administration,Beijing100081

The ultimate goal of climate research is to produce climate predictions on various time scales.In China,efforts to predict the climate started in the 1930s.Experimental operational climate forecasts have been performed since the late 1950s, based on historical analog circulation patterns.However,due to the inherent complexity of climate variability,the forecasts produced at that time were fairly inaccurate.Only from the late 1980s has seasonal climate prediction experienced substantial progress,when the Tropical Ocean and Global Atmosphere project of the World Climate Research program(WCRP)was launched.Thispaper,followingabrief descriptionof thehistoryof seasonal climatepredictionresearch,provides anoverview of these studies in China.Processes and factors associated with the climate variability and predictability are discussed based on the literature published by Chinese scientists.These studies in China mirror aspects of the climate research effort made in other parts of the world over the past several decades,and are particularly associated with monsoon research in East Asia.As the climate warms,climate extremes,their frequency,and intensity are projected to change,with a large possibility that they will increase.Thus,seasonal climate prediction is even more important for China in order to effectively mitigate disasters produced by climate extremes,such as frequent f oods,droughts,and the heavy frozen rain events of South China.

seasonal prediction,climate variability,predictability

1. Introduction

In 1934,a well-known Chinese meteorologist,ZHU Kezhen,published a paper to document the southeastern Asian monsoon’s circulation characteristics and its association with rainfall in China(Zhu,1934).Following that,TU Changwang,who was later the administrator of the China Meteorological Administration in the 1950s,studied the relationship between precipitation in China and atmospheric oscillations and attempted to provide an outlook for summer precipitation in China by considering the early signals of atmosphericoscillations,particularlythe SouthernHemisphere oscillation(Tu,1937).This was the f rst attempt to address seasonal climate prediction.Tu and Huang(1944)further revealed that the monsoon in China is characterized by seasonal migration from South China to North China during the spring-summer and a retreat during late summer-autumn. Therefore,an eff cient outlook for precipitation may be obtained only on the basis of suff cient understanding of theabove seasonal movement of the rain belt and its associated atmospheric circulation features,both locally and remotely.

Operational long-range weather forecasting(referring to seasonal climate prediction or short-term climate prediction) started in 1958,based on various empirical methods.YANG Jianchu,a senior climatological expert at the Institute of Atmospheric Physics(IAP),Chinese Academy of Sciences (CAS),was the pioneer in this f eld.He proposed making short-term climate predictions by considering the persistence ofmeteorologicalquantities,historicalanalogpatterns,atmospheric periodicities,and atmospheric teleconnection.Following that,various statistical methods were developed and employedin short-termclimate predictionsforboth scientif c research and operational applications.

In1988,the climate modeldevelopmentgroupat the IAP, led by ZENG Qingcun,conducted a dynamical seasonal climate prediction experiment by running the IAP atmospheric general circulation model(AGCM)coupled with a tropical Pacif c oceanic general circulation model(OGCM).The results were encouraging and were published in 1990(Zeng et al.,1990).The philosophy of the experimental research was based on the impacts of the tropicaloceanographicanomalieson the global atmosphere,which is the theory of the Tropical Ocean and Global AtmosphereProgram(TOGA).Validation of the AGCM’s capability to“forecast”the summer rainfall anomaly,given the observed SST,was performed(Wang, 1997).The results revealeda generallow predictabilityof the summer rainfall anomaly over a large part of China,particularly in the northern areas.Therefore,a number of correction techniques were developed to improve the GCM output (Zeng et al.,1994;Wang et al.,2000b).Variable effects were achieved by applying correction schemes.

There have also been scientif c endeavors to improve OGCM and ENSO predictions(Zhou et al.,1998;Zhou and Zeng,2001)and to improve land surface process modeling (Dai and Zeng,1996).The comprehensive land surface process model developedat the IAP by Dai andZeng(1996)was adoptedasthemajorbasisforcodingthecommonlandmodel at the National Center for Atmospheric Research(NCAR)in theUS(Zengetal.,2002).Meanwhile,thecoupledmodeldevelopedfor seasonal climate predictionwas later widely used for various purposes,including a doubling CO2-induced climate change simulation,paleoclimate simulations for the last glacial maximum and the mid-Holocene,and so on(Wang and Zeng,1992a,1992b;Wang et al.,1993).

Since global GCMs had course resolutions,dynamical downscaling predictions were tested by several groups in China(e.g.,Liu et al.,2005a).Liu et al.(2005a)employed a modif ed regional climate model(RegCM)to perform a 10-yr(1991-2000)hindcast experiment forced by hindcast results from a coupled GCM(CGCM)that was jointly developedby the IAP andthe National Climate Center of China on the basis of the EuropeanCenter for Medium-RangeWeather Forecasting(ECMWF)model.The anomaly pattern correlation coeff cient between the observed and simulated summer precipitation demonstrated variable scores from year to year, but with overall limited skill.A study by Gao et al.(2006) demonstrated that the spatial resolution of regional climate models plays an important role in reproducing the summer precipitation anomaly.They indicated that a grid point spacing of at least 60 km is needed to reproduce the observed precipitation patterns reasonably well(Gao et al.,2006).

The role of land surface processes in short-term climate variability in China was f rst tested through AGCM sensitivity experiments.The results of early experiments by Yeh et al.(1983,1984)suggested that initial soil moisture or snow cover anomalies may lead to anomalous summer climates. This f nding was subsequently demonstrated by several other numerical experiments(e.g.,Lin et al.,2001).Therefore,efforts have been made to design soil moisture and temperature assimilation schemes in order to improve seasonal climate prediction skill(Zhan,2008;Zhan and Lin,2011).Snow and sea ice cover changes as well as dynamical vegetation processes have also been found to play roles in seasonal climate variability(Zhao et al.,2004b;Chen et al.,2009;Wu et al., 2009a,2009b;Liu et al.,2012;Ma et al.,2012;Li and Wang, 2012).

Statistical downscalingforrainfallpredictioninChinausing GCM outputs of large-scale circulation(e.g.geopotential height of 500 hPa or sea-level pressure)has been another topic of interest in recent years(Kang et al.,2007;Wang et al.,2007b;Zhu et al.,2008).However,the effectiveness of statistical downscaling is weak when applied to rainfall in China as compared to Southeast Asia(e.g.the Philippines). This is likely attributable to the fact that precipitation in China is substantially affectedby mid-and high-latitudevariability,which is quite diff cult to predict,in addition to lowlatitude variability,whereas precipitationin Southeast Asia is mainly modulated by tropical variability.Thus,new statistical downscaling schemes suitable for use in China should be developedin the future to achieve better prediction scores. A recent scheme proposed by Wang and Fan(2009)has the potential to result in higher prediction skill.In their scheme, the summer precipitation anomaly patterns in the historical analog years are considered,together with the model output precipitation anomaly,to compose the f nal prediction. Outputs from six European CGCMs for the years of 1979-2001 were employed to verify the effectiveness of the new scheme.Cross-validation for summer precipitation prediction shows that the anomaly pattern correlation coeff cient increases and the root-mean-square error(RMSE)reduces. Based on the above CGCMs,Liu and Fan(2012a)and Chen et al.(2012)also developed statistical downscaling models to predict spring and summer precipitation in China,separately.Both the statistical downscaling models outperformed the original CGCMs.Also,from the results of Liu and Fan (2012a),it is indicated that a combinationof the synchronous and preceding information involved in the statistical downscaling model is an effective method to improve the predictive capability.All of these previous studies developed statistical downscaling schemes to improve regional climate prediction.Recently,Sun and Chen(2012)showed that the statistical downscaling method can also be used to improve global climate prediction if the better predictedand more stable predictors can be successfully selected from the CGCM.

Since 2003,both dynamical and statistical models have been developed to make seasonal forecasts of the climate background conditions for dust weather in North China and tropical cyclones over the western North Pacif c(WNP) (Wanget al.,2003;Wang et al.,2006;LangandWang,2008). The results are promising,particularlywhen the GCM output is combined with the preceding observed SST anomalies,as well as the major atmospheric modes including the Antarctic Oscillation,North Pacif c Oscillation etc.

Another interesting f nding is associated with the selection of the object that is to be predicted.Traditionally,the precipitation anomaly(or another quantity)in comparison to its multi-yearaverage is selected as the predictand.The yearto-year increment for a quantity is chosen as the new predictand(Wang et al.,2000b;Fan et al.,2008).When new statistical models established for the new predictand,such as those for summer precipitation,surface air temperature,and WNP tropical cyclone frequency,are applied,higher prediction scores can be achieved.

In the context of the above brief history of seasonal climate prediction research,the present paper reviews Chinesestudies on the factors and processes associated with precipitationand monsoons,and seasonal predictionsforthe climate background of dust weather in North China and typhoon activity.A summary and an indication of prospects for future research is provided in the concluding section.

2. Factors and processes associated with summer precipitation and monsoon

2.1.Major characteristics of the East Asian summer monsoon and precipitation

A substantialpartofChina’spopulationlivesinthebasins ofthegreatriverssuchastheYangtzeRiver.Inaddition,agricultural productionplays a signif cant role in China’s domestic economy.As a result,climate,particularly in the summer,exerts a signif cant inf uence on the country.Anomalous climate events,particularly f oods and droughts resulting from the East Asian summer monsoon(EASM)anomaly, often cause severe disasters.For example,the devastating f oods in the middle and lower basin of the Yangtze River in the summer of 1998 cost over two thousand human lives and resulted in a direct loss of 145.09 billion RMB(～20 billion US dollars).The severe drought in West China in the summer of 1996 resulted in a def ciency of drinking water for a population of 18 million and accounted for a loss of 15 billion RMB(～2 billion US dollars).Thus,the prediction of summer precipitation and the EASM is a major priority for seasonal climate research in China.

It is well knownthat the formationof monsoonsis related to the seasonal transition of the structure of atmospheric circulation.The circulation system forming the EASM is different from that of the Indian summer monsoon,albeit they are linked to each other to some extent,as summarized by Tao and Chen(1987).In comparison to the Indian summer monsoon,the moisture source for the EASM area is not only the Bay of Bengal,but also the South China Sea and the west subtropical Pacif c(Wang and Chen,2012).In contrast,the southwesterly airf ow bringing moisture to the Indian summer monsoon rainfall solely originates from the ITCZ,blowing along the Somali coast across the Arabian Sea to reach South Asia.Both the EASM and the South Asian monsoon are strongly regulated by ENSO and the mechanical/thermal effects of the Tibetan Plateau,and this is demonstrated in many recent publications(e.g.,Wang et al.,2000a;Zhao et al.,2007;Wang and Chen,2012).The EASM can be affected by extratropical processes from the Northern Hemisphere(e.g.the Arctic Oscillation,Arctic sea ice extent, North Atlantic Oscillation,North Pacif c Oscillation,and the interdecadal modes in the Atlantic and Pacif c),as supported by numerous recent evidence(e.g.,Sun et al.,2008a,2008b). In addition,the EASM is affected by synoptic systems of the Southern Hemisphere,including the Mascarene High,Australian High,and Antarctic Oscillation.These features illustrate that complicated systems are involved in the formation of the EASM,and this explains the strong interannual variability of China’s rainfall as well as the great challenges encountered in predicting summer rainfall.

Concurrently,the climatological summer seasonal rainfall in China also exhibits some particularfeatures.The onset of the EASM usually begins in South China in May,with a strong rainfall band in coastal southeastern China.Following that,rainfall moves seasonally to the north until mid-August, with a rainfall band shifting to the middle and lower basins of the Yangtze River in mid-June and a further shift northward to North China in mid-July.After mid-August,the EASM retreats swiftly to the south.However,the year-toyear variations of this pattern are considerable,with the sequence of events often being disturbed and causing substantial rainfall anomalies.The variability of rainfall in China usually exhibits two leading patterns,as indicated by the analytical Empirical Orthogonal Function(EOF).The f rst one is a triple pattern with more/less rainfall in both southeastern China and North China sandwiched between less/more rainfall in the middle and lower basins of the Yangtze River. The second one is a dipolar pattern with more/less rainfall in southern China along with less/more rainfall in northern China.The key to summer rainfall prediction is estimating the occurrence probability of the rainfall pattern.

2.2.Role of SST

Due to their longer persistence relative to atmospheric internal anomalies,SSTs play an important role in summer precipitation prediction.As such,the SST is also a practical factor in seasonal climate prediction.El Ni˜no/La Ni˜na events are the strongest signal of global climate on interannual timescales.Naturally,their impacts on the EASM have been investigated in a great number of studies.As early as the 1930s,before a coherent link between El Ni˜no and the Southern Oscillation(SO)had been established,and before the word ENSO had been coined,Chinese scientists had already noted a relationship between summer rainfall in China and the SO.Tu(1937)found that the SO is positively correlated to the rainfall in the middle-lower basin of the Yangtze River,but negatively correlated to the rainfall in southeastern China.Decades later,when a record strong ENSO event occurred in 1982/83,the link between the EASM and ENSO was extensively investigated.Some studies suggested that the impact of ENSO depends on the timing of warm/cold event occurrence or the location where warming/cooling f rst emerges(Tao et al.,1988;Li,1990).The impact of ENSO was further clarif ed and found to depend on the phase stage of the event.In the summer following the mature phase of an El Ni˜no event,such as those in 1983 and 1998,the middle and lower basin of the Yangtze River is expected to receive morerainfall(HuangandWu,1989).Thetime-laggedimpact of ENSO on the EASM occurs through a lower-tropospheric anticyclone around the Philippine Sea in the western Pacif c. The southwesterly in the rear of the anticyclone overlaps and intensif es the climatological southwesterly in the summer and transports more moisture to East China,resulting in more rainfall in the basin of the Yangtze River.However, the formation of the lower-level anticyclone is controversial. One viewpoint is that it is induced by a tropical Rossby wavepropagatingwestwardandthatitself-intensif esthroughlocal air-sea interactions near the Philippines(Wang et al.,2000a). The Rossby wave is excited by the ENSO-induced diabatic heating anomalies over the central-eastern Pacif c.When the wave reaches the western Pacif c,it induces an anticyclone near the Philippine Sea.The associated lower-level winds modify the seasonal trade wind,and the local feedback between evaporation and wind favors the enhancement of the anticyclone.The other viewpoint is that the anticyclone is induced by Indian Ocean basin-scale warming following an ENSO event(Yang et al.,2007;Li et al.,2008).The warming is often excited by the ENSO event,with a lag of several months.This process is analogous to charging an electric capacitor.Nonetheless,the underlying physics needs to be further revealed.More recently,the impact of ENSO was found to be nonlinearly dependent on the amplitude of the event itself.However,a stronger ENSO event tends to have a larger impact on summer rainfall in China(Xue and Liu, 2007).On the other hand,there are some studies that have found the relationship between ENSO and East Asian monsoon is unstable(Wang,2002;Wang and He,2012;Wang et al.,2013;He et al.,2013;He,2013;He and Wang,2013a, 2013b).Over some periods,their relationship is close,but over other periods it is broken,indicating that caution should be applied when making predictions of East Asian monsoon using the ENSO signal.

In comparison with ENSO,the tropical western Pacif c exerts a more direct inf uence on the EASM.The western tropical Pacif c is part of the largest warm pool of the world’s oceans.Previous studies have revealed that summer rainfall in North China is negatively correlated with the SST in the western Pacif c Ocean,and this relationship is opposite for rainfall in central China(Tao et al.,1988).This appears reasonable,since the warmer western tropical Pacif c tends to trigger stronger local convection.Enhanced convection intensif es the subtropical anticyclone over the western Pacif c,favoring a stronger northward-stretching southwesterly reaching a northerly location.Soon after,the underlying mechanismwas extensivelyexploredbyHuangandLi(1987) and Nitta(1987),who both proposed a wave train originating from the convection region and propagating to the northeast.They respectivelycoinedthe wave train the East-Asian-Pacif c(EAP)and Pacif c-Japan(PJ)pattern.Since there is a seesaw relationship between the SSTs of the central-eastern and western Pacif c,the western Pacif c can potentially act as a bridge linking ENSO to East Asia.

In addition to the tropical Pacif c,the SST in the WNP, particularly in the region of the Kuroshio Current,is another factor that inf uences summer rainfall.As early as the 1970s, studies suggested that the winter SST in the region is well correlated with the following summer’s rainfall in the middle and lower basin of the Yangtze River.When the primary axis of the region with maximum SST is located in a southerly position,it favors more summer rainfall in central China,and vice versa.Due to the coherenceof the Pacif c basin SST,the extentand mannerin which ENSO acts as a role in this link is unclear.Recently,Zhuet al.(2011)furtherfoundthat a phase change in the Pacif c Decadal Oscillation can contribute to a decadal shift in East Asian monsoon rainfall.

As for the Indian Ocean,Chinese scientists have found that its SST is closely correlated to the June-July mean rainfall in the Yangtze River basin(Luo et al.,1985;Jin and Shen,1987;Deng et al.,1989).An SST structure with warm anomalies in the eastern Indian Ocean and South China Sea along with cold anomalies in the western part of the Indian Ocean is linked to more rainfall in central China,and vice versa(Chen,1988).This point was also conf rmed by studies on the link between the rainfall in China and the SST contrast between the South China Sea and Arabian Sea(Deng et al.,1989),as well as studies on the impact of the Indian SST dipolar mode(Li and Mu,2001).In addition to this structure,the basin-scale warmth/coldness,which is the leading mode on interannual timescales,is another SST pattern thatinf uencestheEASM.NumericalexperimentsbyWuand Liu(1995)revealedthat theIndianOceanbasin-scaleanomalies excite not only neighborhood rainfall responses but also large-scale circulation anomalies in the western tropical Pacif c.In view of the excitation role of western Pacif c heating on the EAP or PJ wave train,the hypothesis that the IndianOcean SST inf uencesEast Asian summerrainfallseems reasonable.As mentioned above,recent studies have illustrated that this pattern indeed favors an enhancement of summer rainfall in central China by forcing an anticyclone over the Philippine Sea(Yang et al.,2007;Li et al.,2008).This provides an alternative explanation for the increased summer rainfall in El Ni˜no events that follow these summers.

As for the Atlantic Ocean,the winter North Atlantic SST can be a precursor to the following summer’s rainfall in the middle and lower basin of the Yangtze River.An early warm North Atlantic favors an early onset of the following summer’s Mei-yu in central China,and vice versa(Xu et al., 2001).Additionally,a dipolar SST pattern with warm water in the latitudes north of Newfoundland and cold water in the south favors the emergenceof meridionalcirculation patterns in both the winter and the following summer,causing more summer rainfall in northeasternChina(Bai,2001).These Atlantic impacts are physically reasonable,because two wave trains,anarchedextratropicaloneandanotherjet-guidedsubtropical one,have been found to link East Asia to the North Atlantic(Li,2004;Sun et al.,2009).These impacts have already been verif ed in recent studies on the Atlantic Multidecadal Oscillation(Lu et al.,2006;Wang et al.,2009).

2.3.Role of land surface processes

The impact of land surface processes on the East Asian monsoon climate has been illustrated in many studies(e.g., Lin et al.,1996;Yang and Lau,1998).By introducing an improved climate model with modif ed land surface parameterization,Lin et al.(1999)found that the prediction skill can be improved to some extent.Wang(2001)conducted ensemble integrations with the IAP AGCM to investigate the mechanism of severe f ooding over the Yangtze River valleys in 1998and found that the excessive snow coveroverthe Tibetan Plateau during winter and springtime can lead to aweakening of the South Asian High,followed by a weakening of the East Asian monsoon circulation and its associated monsoon precipitation over the lower reaches of the Yangtze River valleys.This result was furtherdemonstratedin similar research(Zhang and Tao,2001).

Meanwhile,the impact of soil moisture on the simulation of EASM has also been revealed by Wang(2001);he suggested that wet soil conditions during springtime over South China can lead to a weakening of the EASM followed by increased monsoon precipitation over the lower reaches of the Yangtze River valleys.Using observational data over the Huaihe River basin from HUBEX(Huaihe River Basin Experiment in China),Lin et al.(2001)further indicated that the sensitivity of the land surface model to the initial soil moisture is quite strong in late spring and summer over the Huaihe River basin;however,this sensitivity becomes relatively weak in autumn.Using the IAP AGCM,Guo and Wang(2003)suggested that the prediction skill of the 1998 summer rainfall anomalies could be increased by introducing more realistic initial soil moisture.

Furthermore,Zhan and Lin(2011)investigated the impact of initial soil moisture anomalies over the Yangtze and Huaihe river basins that were induced by the extreme freezing-rain disasters of January 2008,on the seasonal predictive skill of spring rainfall anomalies over China using the IAP AGCM,whose land surface process scheme had been replaced with Common Land Model(CoLM)(Liu,2007). They suggested that,when the“observed”initial soil moisture anomalies are taken into account,the model’s predictive skill in terms of the spring precipitation anomalies in 2008 can be substantially increased.In particular,over North and Northwest China,the model predicted rainfall anomalies that agreedquite well with the observations,and the anomalypattern correlation coeff cient(ACC)between the observed and predicted spring precipitation over the whole of China increased from-0.02 to 0.11.

Recently,the inf uence of land surface conditions on dynamical seasonal forecasts was systematically assessed through two sets of June-July-August(JJA)hindcasts for the 20-yr period from 1981 to 2000 by an IAP AGCM that has been coupled with CoLM(Zhan,2008).In these two sets of experiments,10 ensemble members were conducted each year using the model atmospheric initial conditions from 00 UTC 10-19 May,and then the model was integrated from 15 May to 31 August;the observed SST data from the UK MeteorologicalOff ce HadiSST dataset was applied in both sets of experiments.The only difference between the two experiments was that,in the second set,when driving the land surface model CoLM in the coupled experiment,the model-simulated rainfall at each time-step was discarded and replaced by the observed GPCP(Global Precipitation Climatology Project)precipitation data.Given that rainfall is the primary factor determining land surface conditions(soil moisture,in particular),forcing the surface model with the observed rainfall at each time-step ensured that realistic representation of the land surface conditions during the model hindcast integrations;this experiment was designated as EXP SSTs+SM,and the one without rainfall replacement was EXP SSTs.

Results from the spatial anomaly correlations between the hindcast rainfall anomalies and observed results for 20-yr JJA ensemble hindcasts during 1981-2000 over China show that,when only considering the SSTA as the predictor,the seasonal prediction skill of summer rainfall anomalies over China is relatively low,with a spatial anomalycorrelationcoeff cient(ACC)of-0.05 averaged during 1981-2000,and the ACCs are even less than-0.5 for the years 1992 and 1993.However,by taking into account more realistic land surface conditions during the hindcast period,the seasonal predictability of summer rainfall anomalies over China is largely increased;the 20-yr averaged ACCs of summer rainfall anomalies over China reach～0.19,with a positive correlation for most years during 1981-2000(Zhan,2008).

2.4.Role of atmospheric teleconnection

Atmospheric teleconnection refers to a recurring,persistent,and large-scale pattern of atmospheric circulation, which links the climate variability in different regions.The identif cation of atmospheric teleconnection is of major importance for understanding and predicting a region’s climate.

In the last two decades,teleconnection patterns that are related to the EASM have received increasing attention.Results have indicated that the variability of the summer monsoon and the corresponding precipitation are signif cantly inf uenced by atmospheric teleconnection from the tropics,the Northern Hemispheric mid-to-high latitudes,and even the Southern Hemisphere.In this section,we review the roles of atmospheric teleconnection over these three regions.

2.4.1.Teleconnection from the tropics

At tropical latitudes,there are four teleconnection patterns that have been shown to have an impact on the EASM. One is the PJ or EAP teleconnection pattern(Nitta,1987; Huang and Li,1987;Lau,1992).Observational data,reconstructed data,theoretical work,and modeling analyses have indicated that this teleconnectionpattern results from anomalousheatingoverthe tropicalwesternwarmpool,andits variability has a close connection to the circulation and rainfall of the EASM(e.g.,Nitta,1987;Huang and Lu,1989;Huang and Sun,1992;Lau,1992;Shi and Zhu,1993;Huang and Yan,1999;Han and Zhang,2009).With a positive-phase (negative-phase)EAP pattern,the EASM-related rainfall is less(more)than normal.

The ENSO,acting as a major source of interannual variability in the tropics,has a profound inf uence on the tropical climate and even on the climates of some extratropical regions.The EASM is one such climate system subjected to ENSO impact,as has been revealed by many previous studies,from current observational data to historical reconstructed proxy data(Huang and Wu,1989;Chen et al.,1992; Li,1992;Shen and Lau,1995;Zhang et al.,1996;Lau and Weng,2001).However,ENSO is most active in the winter time;therefore,how does it exert a delayed impact on the following summer’s rainfall in East Asia?By investigat-ing the observed evolution of ENSO-related circulation from winter to summer and by conducting a numerical sensitivity experiment,Wang et al.(2000a)pointed out that the EAP teleconnection is responsible for the link between ENSO and the East Asian climate.

Condensationheatingin the South Asian monsoonregion is an important source of atmospheric variability.Dai et al. (2002)foundthat the latent heat of condensationof the South Asian summer monsoon can excite a teleconnection pattern, which can result in an in-phase variability of precipitation over India and North China.Recently,Liu and Ding(2008) demonstratedthat the onset of SouthAsian summermonsoon over Kerala in the southwestern coast of the Indian Peninsula can lead to a so-called southern teleconnection of the Asian summer monsoon,which emanates from the western coast of India,across the Bay of Bengal,to the Yangtze River valleys and southern Japan.Via this teleconnection pattern,the date of South Asian summer monsoon onset over Kerala on the southwestern coast of the Indian Peninsula provides a potentially valuable signal for predicting the onset of the Mei-yu over the Yangtze River valleys two weeks later.

2.4.2.Teleconnection fromthe extratropical Northern Hemisphere

Over the extratropical Northern Hemisphere,the North Atlantic Oscillation(NAO)/Arctic Oscillation(AO)is the largest hemisphere-scale interannual mode.Signals of the NAO/AO in the East Asian summer climate have been indicated by several studies.Xu et al.(2001)revealed that in a strong(weak)winter NAO year,the onset of the Mei-yu is advanced(delayed).Sung et al.(2006)demonstrated a signif cantly delayed impact of the winter North Atlantic Oscillation(NAO)on the precipitation of the following EASM. Using 530-yr(1470-1999)reconstructed NAO and eastern China drought and f ood indices,Fu and Zeng(2005)also indicated that the winter NAO can inf uence summer droughts and f ooding for the following one to three years in East China.Furthermore,Ju et al.(2005)indicated that the NAO can also modulate the decadal variability of the East Asian summerclimate.Thus,thevariabilityoftheNAO canprovide some useful information for seasonal and long-term prediction of the EASM and corresponding droughts or f ooding. In addition,other studies have revealed that the NAO/AO in other seasons is also related to the East Asian summer climate.Forexample,GongandHo(2003)showedthattheMay AO is highly negatively correlated with the summer rainfall in the Yangtze River valleys.Sun et al.(2008a)and Yuan and Sun(2009)found that the summer NAO can modulate the midsummer air temperature via the excitation of a zonal Rossby wave train along the Asian upper-level jet.However, this impact experienceda decadal shift aroundthe late 1970s, with a strong impact after the late 1970s and a weak impact before.Recently,Sun andWang(2012)furtherfoundthat the summer NAO also has a dominantrole in the summer precipitation over the mid and high latitudes of East Asia.

The upper-level jet is the major Rossby wave conduit. Along this upper level jet,several zonal teleconnection patterns have been revealed. Using upper-level meridional winds,Lu et al.(2002)identif ed the existence of a teleconnection pattern in July,which emerges from North Africa to East Asia along the Asian westerly jet in the midlatitudes. Further analysis revealed the possible role of this upper-level teleconnection in linking the EASM to the South Asian summer monsoon,andevento fartherwestward heatsources over the Atlantic.Enomoto et al.(2003)proposed a“silk road”teleconnection along the East Asian jet.The identif cation of this teleconnection presents one possible mechanism for the formation of the Bonin high,a predominant subtropical anticyclone near Japan in the summer that is associated with the Mei-yu frontal zone.By synthetically analyzing the summer midlatitude circulation of the Northern Hemisphere, Ding and Wang(2005)found that there is a recurrent circumglobal teleconnection(CGT)pattern,and the two upperlevel teleconnection patterns over the Eurasian continent are regional manifestations of the CGT pattern.The CGT pattern is signif cantly associated with the rainfall and surface air temperature in Western Europe,European Russia,India, East Asia,and North America.

Althoughthe North Pacif c is located downstream of East Asia,the teleconnection variability over this region is also important to EASM variability.Zhang et al.(2007)indicated that the boreal winter North Pacif c Oscillation(NPO)has a delayed impact on the following summer’s rainfall over the Huaihe River valleys,with a strong(weak)NPO corresponding to less(more)rainfall.Lian(2007)suggested that the preceding winter NPO shows a negative correlation with the EASM,resulting in a dipole pattern of anomalous summer rainfall over eastern China.

Recently,two kinds of teleconnection patterns over the Asian continent and the North Pacif c region have been revealed.Zhao et al.(2007)pointed out that the upper-level summer air temperature over the Asian continent and the North Pacif c co-varies out of phase.They named this pattern the Asia-Pacif c Oscillation and found that this teleconnection has an impact on the Asian summer climate.In investigating the lower-level circulation variability over the Asia-NorthPacif c region,Sunet al.(2008b)identif edtheArabian Peninsula-North Pacif c Oscillation(APNPO).This teleconnection essentially ref ects the co-variability of the North Pacif c high and the South Asian summer monsoon.The APNPO is closely related to the Asian summer monsoon circulation in the upper and lower levels,the moisture transportation,and rainfall on both interannualand decadal time scales. In addition,the APNPO is found to have a persistent feature from spring to summer,which provides potentially valuable information for Asian summer monsoon predictions.

2.4.3.Teleconnection from the Southern Hemisphere

Although the EASM occurs over the Northern Hemispheric subtropics,a number of recent studies have indicated that Southern Hemispheric teleconnection patterns are also important in the variability of the EASM.In investigatingthe roleoftheMascarenehighandtheAustralianhighinthevariability of the EASM,Xue et al.(2003)found that these twohighsare closelyrelatedto the variabilityof the AntarcticOscillation(AAO).Thus,they proposed the hypothesis that the AAO may be an important far-reaching external forcing for the year-to-year variability of the EASM.Some later studies conf rmed this hypothesis,revealing the existence of a positive relationship between the boreal spring AAO and summer rainfall in the Yangtze River valleys(Gao et al.,2003; Nan and Li,2003;Xue et al.,2004;Fan,2006)and a negative relationship between the AAO and summer precipitation in central North China(Wang and Fan,2005).In addition to the AAO,Wang and Fan(2006)suggested that the upperlevel meridional teleconnection of zonal wind between the middle and high latitudes in the Southern Hemisphere is negatively correlated to the EASM circulation on an interannual scale.Zhu et al.(2009)compared the AAO-EASM link in the NCAR/NCEP reanalysis with that in the ERA40 reanalysis,and a similar relationship was found.

Different from the tropical and Northern Hemispheric teleconnection patterns,the possible mechanism underlying the impact of the Southern Hemispheric teleconnection pattern on the EASM is quite complicated.Up to this point,four kinds of mechanisms have been proposed.Xue et al.(2003) suggested that the Southern Hemispheric teleconnection patterns inf uence the EASM via changing the cross-equatorial f ows.Wang(2005)revealed that the meridional wind shows a circum-Pacif c teleconnection pattern,which could be one way to couple the Southern and Northern Hemisphere’s climate.Wang and Fan(2006)suggested that the meridional teleconnection in the zonal wind f eld,with its main part in the Eastern Hemisphere,from the mid and high latitudes in theSouthernHemispheretothetropicalregion,isresponsible for the link between the Southern Hemispheric upper-level zonal wind teleconnection and the EASM.These three possible mechanisms are all related by a simultaneous connection between the Southern Hemispheric teleconnection and the East Asian climate.For the seasonal delayed impact of the boreal spring AAO on the EASM,Sun et al.(2009)proposed that an anomalous AAO can affect the convection activity over the Maritime Continent,via the excitation of anomalous meridional circulations along the central South Pacif c and two meridional teleconnection wave train patterns.Thereafter,the anomalous convection propagates northward along the seasonal cycle and then changes the western Pacif c subtropical high in the following seasons,consequently impacting the EASM’s rainfall.

The identif cation of the above atmospheric teleconnection patterns has greatly deepened our understanding of EASM variability and vastly improved EASM prediction.

3. Seasonal prediction of precipitation and monsoons

3.1.Two-tier prediction approaches

3.1.1.Methods

The two-tiered method can save computer resources and was thus developed quite early in China(Zeng et al., 1990).To make real-time dynamical seasonal predictions in a two-tiered fashion,the SST is predicted by a coupled atmospheric-oceangeneralcirculationmodel(CGCM)in the f rst step.Then,an AGCM is forced with the predicted SST given by the f rst step.Because systematic errors in the predicted SST can be corrected in a statistical way,adverse impacts on the atmosphere predictions that arise from errors in the SST can be minimized.

Focusing on climate anomalies in the summer and winter as well as the spring,the integrations start from different dates in February and October for seasonal prediction and interannual prediction,respectively,in China.Usually,the predicted SST is available over the tropical Pacif c,and therefore,a linear combination of the observed SSTA in the initial month and the predicted monthly SSTA are used as the lower boundaryconditions of the AGCM.The role of the observed(predicted)SSTA is gradually decreased(increased) in the tropical Pacif c in the integration.Over other oceanic regions,the observedSSTA in the initial monthis maintained throughout the integration process.

3.1.2.SST forecast using air-sea coupled general circulation models

Forecasting of the tropical Pacif c SST has been performed since the early 1990s in China by using dynamical models or mathematical statistics models.Mathematical statistics has evolved signif cantly in recent years(e.g.,Ding et al.,1998).However,even more achievements have been made through dynamical prediction methods(Zhou et al., 1998;Wu and Ni,1999;Zhou and Li,2000;Shi et al.,2001; Zhou and Zeng,2001).Several ENSO prediction systems have been established in China since 1999,such as the IAP ENSO forecast system(Zhou et al.,1999;Zhou and Zeng, 2001),established by using the Tropical Pacif c-Global Atmosphere Coupled Model,incorporatedwith an initialization scheme in which only the observed or analyzed SSTA is inserted into the coupled model.The National Climate Center (NCC)set up a simplif ed ENSO prediction model(Zhao et al.,2000),and the IAP established a probabilistic ENSO ensemble prediction system based on an intermediate coupled model(ICM)developed by Keenlyside and Kleeman(2002) and Zhang et al.(2005a).In general,the prediction skill for the SST gradually decreases with time.As for the dynamical model,the prediction is more reliable than that achieved by means of persistence predictions,if the integrated time is longer than three months,which highlights the importance of ocean data assimilation.Zhang et al.(2006)concluded that theincorporationofSSTassimilationwithARGO(thebroadscale global array of temperature/salinity prof ling f oats)observation data is greatly benef cial for the prediction skill of the summer rainfall anomaly in China.Zheng et al.(2006) found that prediction skills can be signif cantly improved by assimilating SST observations into an intermediate coupled model.In addition,the accuracy of ENSO simulation is disturbed to a large degree by the“climate drift”.Recently,a regressive correction method was presented with the primary goal of improving ENSO simulations in a regional coupledGCM;it was shown that it is superior to the anomaly coupling both in reducing the coupled model climate drift and in improvingthe ENSO simulation in the tropical Pacif c Ocean (Fu and Zhou,2007).

More recently,dynamical predictions of SST have experienced new achievements in China.For example,using the parameterization proposed by Zhang et al.(2005),Zhu et al. (2013)established a hybrid coupled model by embedding an SSTA model in an OGCM.Such an embedded approach has proved to be effective in improving ENSO simulations and forecasts(Zhu et al.,2009),and its skill is notable in predicting the whole process of the 1997/98 El Ni˜no,which has notbeenrepresentedwell bymanyforecastsystems(Landsea and Knaff,2000).Moreover,to overcome the shortcomings of the ICM in some ENSO cases(such as the 2007/08 La Ni˜na event),a coupled assimilation scheme was preliminarily developed to be capable of assimilating the atmospheric data into the ENSO ensemble prediction system to improve theinitialatmosphericstate(ZhengandZhu,2010).An11-yr retrospective forecast comparison showed that the prediction skill of assimilating wind observations was better than that of assimilating SST observations.The assimilation of wind observations for the 2007/08 La Ni˜na event triggered better predictions,while that of SST observations failed to provide an early warning for that event.

3.1.3.Ensemble atmospheric forecast

The earliest experimental extraseasonal prediction of summer monsoon rainfall anomalies by GCMs in China was carried out by Zeng et al.(1990)at the IAP.Verif cations showed that,in general,the model possesses relatively high prediction skill in eastern China,especially for the Yangtze River and northern China,but it is insuff cient for northeastern and northern China(Wang,1997;Zeng et al.,1997). Modif cations of the surface albedo parameterization(Lin and Zeng,1997),the horizontal resolution(Zhang et al., 2004),and the correction system(Zhao et al.,1999)have resulted in improvementsin seasonal climate prediction.As for the summer rainfall anomaly in 1998,patterns in most parts of China have been successfully predicted by the system in advance(Lin et al.,1998;Chen and Lin,2006).

Based on the original two-levelAGCM mentionedabove, a nine-level AGCM was developed in the late 1990s(Zeng et al.,1987;Zhang,1990;Liang,1996).The resolution was 5°in longitude by 4°in latitude,with nine vertical levels with a top at 10 hPa.The model has been used in real-time extraseasonalpredictionforsummerclimateanomaliesinChinasince 2002(Langet al.,2004a).Thedroughtconditionsovernorthern northeast China and the Yellow River valleys and positive rainfall anomalies over eastern northeast China,South China,and most parts of western China are well captured by the model.

There have been some successful cases of real-time seasonal prediction of summer precipitation in China and monsoons via the dynamical approach.However,the overall skill of GCM-based seasonal precipitation prediction in China thus far is generally low.Considering that interactive oceanatmosphere coupling is neglected in the two-tier method,a one-tier approach was proposed and applied to seasonal climate prediction.

3.2.One-tier approach

Supported by the National 9th Five-Year Development Plan(1996-2000),the f rst generation of a dynamical climate model prediction operation system was established at the NCC in 2001(Ding et al.,2004).The seasonal prediction model system consists of the Coupled Global Atmosphere-Ocean Model(CGCM/NCC)and the Global Ocean Data Assimilation.The CGCM consists of a T63L16 AGCM(Dong, 2001)and L30T63 OGCM(Zhao et al.,2000).The T63L30 AGCM and L30T63 OGCM are coupled through a daily f ux anomaly coupling scheme(Yu and Zhang,2000).The observational data used in this system comprise both the Globl Temperature-Salinity Prof le Program(GTSPP)data(from 1981)and ARGO data(from 1998;Liu et al.,2005b).

Since the spring of 2003,four-season climate prediction produced by dynamical climate models has been in operation(Li et al.,2005).Two years later,monthly-moving seasonal climate productswere released at the beginningof each month with a lead time of 0-6 months.Results from 20-yr hindcasting experiments and routine operation of seasonal climate prediction since 2003 show that this system has variable capability of seasonal prediction from year to year in East Asia.

Withthe developmentofclimate systemmodelsandearth system models,two one-tier prediction systems were established at the IAP based on version 3.0 and version 4.0 of the Community Climate System Model(CCSM)developed by the NCAR and a self-designed initialization system.These prediction systems are hereafter referred to as PCCSM3 and PCCSM4,respectively(Liu andWang,2014aLiu,S.,and H.J.Wang,2014:Extra-seasonal short-term prediction systems based on CCSM3 and their evaluation.Int.J.Climatol.,under review.;Ma andWang, 2014).PCCSM3 is composed of an ocean-atmosphere coupled model that consists of complete air-sea interaction and twoinitializationsystemsfortheoceancomponent,onebeing based on observationalSST,similar to Zhou et al.(1999)and ZhouandZeng(2001),and theotherbasedonNCEP Climate Forecast System Reanalysis(CFSR)data,which can import deep ocean information.A detailed discussion on the performance of PCCSM3 is available in Liu and Wang(2014)aLiu,S.,and H.J.Wang,2014:Extra-seasonal short-term prediction systems based on CCSM3 and their evaluation.Int.J.Climatol.,under review.. PCCSM4 is based on CCSM4,the later version of CCSM3, with a mixed-layer ocean model that involves complete airland interaction,partial air-sea interaction,and an ensemble initialization scheme(Ma and Wang,2014).Retrospective summer prediction experiments have been carried out and demonstrate that both PCCSM3 and PCCSM4 possess good prediction skill in terms of summer climate,especially so for SST and over the tropical zone.In addition,PCCSM3 and PCCSM4 showgoodcapabilityinsimulatingtheyear-to-year variability of the Asian summer monsoon(data not shown). Both PCCSM3 and PCCSM4 have begun real-time summer climate prediction since 2013.

Precipitation forecasting for the rainy season(JJA)is a crucial task for short-term climate prediction in China.Results show that among 160 stations,there are 69 stations with ACCs between predictions and observation that are above zero,accounting for 43.12%(Ding et al.,2002).Therefore, the accuracies of the prediction are generally low,but might begoodin someparticularyearsovereasternChina.Forrealtime prediction of the summer climate in 2006,the predictions of summer rainfall anomalies are quite similar to observations(data not shown).

The temperature in West China(such as Xinjiang,Tibet, Qinghai,Gansu,Ningxia,and Shaanxi),North China,the regions between the Yangtze and Yellow rivers,and the southeast part of Northeast China,can in part be successfully predicted,with the central values of the ACCs reaching or exceeding the 90%signif cance level.Among the 160 stations in China,77 stations’ACCs of temperature are greater than zero,which accounts for 48.12%of the total.

In order to effectivelyimprovenumerical predictionskill, the so-called“dynamical analogue prediction”approach has been recently investigated(Ren,2006).Based on the atmospheric analogy principle,information from historical analogue data is utilized to estimate the model errors(Ren and Chou,2006;Gaoet al.,2006).Whenappliedto extraseasonal prediction experiments on an operational atmosphere-ocean CGCM,resultsshowedthatit canreducethepredictionerrors and successfully improve predictive skill(Ren,2006).

In addition,some statistical downscaling methods have been applied,which also result in improved skill of seasonal climate prediction in China(e.g.,Chen,2008).This is discussed in more detail in the following section.

3.3.Downscaling and correction of model results

As a simplif cationof arealistic climate system,theGCM inevitably has model errors resulting from the parameterization of physical processes,the effects of unresolved scales, and imperfect boundary conditions.In particular,model errors are considerable for simulations of precipitation and monsoons,due to their intricate mechanisms.Therefore,statistical downscaling or model error correction is essential for dynamical seasonal prediction in operation.

Model error correction is an approach to combine both statistical and dynamical methods together.It is a model output statistics technique,which regards the GCM as a black box and reduces the model error based on a certain statistical relation between model output f elds and historical observed data.The model error correction is not only much easier to implement than the advancement of the model parameterization of physical processes and dynamical cores,but it is also much cheaper than increasing the model resolution.Commonlyusedcorrectionmethodsincludemeanbias correction, regression analysis,coupled f eld techniques,analogy analysis,magnitude correction of predicted interannual variation, and statistical downscaling.

Zeng et al.(1994,1997)outlined four methods for correcting model error:mean bias correction,maximum similarity,minimum difference,and a coupled pattern technique based onEOF.The maximumsimilarity and minimumdifference are actually two methods for deriving the coeff cients of a regressionequationandthus belongto a regressionscheme. Some studies have used the mean bias correction method to reduce the model error of the IAP GCM(Zhao et al.,1999; Lang,2003;Chen,2003)and improve the forecast skill of summer seasonal rainfall over China.

One shortcoming of the mean bias correction method is thatthemagnitudeofthecorrectedanomaliesismuchsmaller than the observation due to the weak interannual variation of the original simulation(Wang et al.,2000b).In orderto overcome this shortcoming,some effective methods have been developed to correct the magnitude of the predicted interannual variation,which multiply the year-to-year or seasonal variation by the ratio of their observed to predicted standard deviation or standard difference(Wang et al.,2000b;Lang, 2003).

Li et al.(2005)introduced a coupled pattern technique basedonsingularvaluedecompositionanalysis(SVD)(Ward and Navarra,1997)and EOF(Feddersen et al.,1999)to correct the model error in the IAP GCM.The EOF correction method is a combination of the minimum difference and the coupledpatterntechniquebased on EOF in Zenget al.(1994, 1997).Since the summer rainfall over China is mostly subtropical monsoon rainfall,the chaotic component is greater than that of tropical rainfall,due to the larger contribution of the internal atmospheric process(Wang et al.,1997;Lang et al.,2004b).Therefore,Li(2008)introduced a double cross-validationmethod to obtain the prior informationabout the feasibility of the correction method and to determine the number of truncation modes of SVD and EOF,which improved the seasonal forecast skill signif cantly.In addition, Li(2008)explained the phenomenon pointed out by Feddersen et al.(1999)that the post-processed results were not sensitive to the choice of methods based on the SVD and EOF. The signif cant improvementof forecast skill achieved by the coupled pattern technique suggested that there was a shift in thespatialpatternofvariabilityofsummerrainfalloverChina in the IAP GCM,which resulted in a substantial drop in the skill score.

Analogycorrectionis also a widely used model error correction method.The analogy scheme always assumes that a similar spatial pattern of the predictand for the target year has already occurred in past years.Based on the fact that different model error patterns of the IAP GCM for summer rainfall over China appear in ENSO warm years,cold years and normal years,separately,Chen and Lin(2006)developed a method to estimate the model error in the target year.Ren and Chou(2007)proposed an analogue correction method based on the assumption that the errors in the predicted results with similar initial conditions are the same.In addition, due to the high spatial pattern correlation between summer rainfall anomalies overthe East Asian and western Pacif c region(EAWM)and tropical region and successfully predicted tropical precipitation,Wang and Fan(2009)developed a correction method with the predicted tropical precipitation as an analogicalindex.This new scheme can substantially improvethe forecast skill of summer rainfall with six models in the Development of a European Multimodel Ensemble System for Seasonal to Interannual Prediction(DEMETER)system.

Recently,the statistical downscaling technique,which is based on the assumption that atmospheric variability on local scales is conditioned,though not determined,by large scales(von Storch et al.,1993;von Storch,1999),was introduced to correct model error.Statistical downscaling correction methods aim at specifyingthe local f eld(the predictand, e.g.,the precipitation or its model error)statistically from a large-scale f eld(the predictor)that is skillfully predicted by the dynamical model.The choice of predictor depends on the predictand and should satisfy two conditions.First,it mustbepossibletoaccuratelyspecifythepredictandfromthe predictor.Second,the predictor should be well predicted by the dynamical model.Large-scale f elds such as geopotential height,sea level pressure,air and surface temperature,velocity,and tropical SST have been used as predictors to improve the forecast skill of precipitation(Zhang et al.,2005;Ren, 2008;Zhu et al.,2008;Chen,2008;Ke et al.,2009;Kang et al.,2009;Li et al.,2009).

In statistical downscaling correction schemes,the statistical relation between large-scale f elds and local prediction can be identif ed by coupled pattern techniques(Zhang et al., 2005),analogy analysis(Ren and Chou,2007),regression analysis(Chen,2008;Ke et al.,2009;Kang et al.,2009;Sun and Chen,2012;Liu and Fan,2013;2014;Liu and Li,2014), and Bayesian schemes(Li et al.,2009).The predictands include the summer rainfall and its model error.For the former, only the large-scale f elds are used.In this case,the corrected rainfall f elds obtained by statistical models with large-scale f elds as predictors are regarded as more valuable than those achievedwiththerainfallf eldsimulatedbya nonlinearGCM as predictors.This is usually true in view of the poor forecast skill of the GCM for summer rainfall over China.A new scheme proposedby Wang and Fan(2009)achieved a largely improved hindcast of summer precipitation over China.In their scheme,they attempted to make adequate use of tropical summer precipitation predictability to reinforcethe extratropical precipitation predictabilityin China,and their results were encouraging.The new scheme of Wang and Fan(2009) is based on the fact that the forecast skill of local f elds in the GCM is usually related to the different states of large-scale f elds.

Besides statistical downscaling,the dynamic downscaling method using a high-resolution regional climate model can be applied to regional climate prediction.Over East Asia where the topography is complex,a high-resolution regional climate model could be more capable of depicting the smallscaleweathersystemsandcapturinglocalclimatedetails,due to the more reasonable terrain and higher resolution compared to a GCM(Yu et al.,2010).The extreme f ooding overthemiddle-lowerreachesoftheYangtzeRivervalleysin the summer of 1998 was selected as a case to investigate the performanceof the Weather Research and Forecasting model (WRF)in retrospective summer precipitation prediction(Ma et al.,2014).The results were encouraging.The WRF model not only gave a better hindcast for the exclusive rainfall over the middle-lowerreaches of the Yangtze River valleys on the seasonal scale,but also for the two occurrences of Mei-yu rainfall,which was the main characteristic of the 1998 summer climate on the sub-seasonal scale.

Moreover,a rare catastrophic f ood occurred in central eastern China in 2003,and it was not predicted in advance by any of the prediction departments in China except for the NCC using the regional climate model(Ding et al.,2006). The f ood and drought distribution of the summer rainfall in eastern China derived from the regional climate model is quite consistent with observations.

3.4.Development of new statistical prediction models for summer precipitation

Current operational skill for summer precipitation in China,with both statistical and dynamical model methods, is quite limited.Recently,a year-to-year incremental approachwas proposedfor forecastingsummer rainfalloverthe middle-lower reaches of the Yangtze River valleys(RYRV) (Fan et al.,2008).In this prediction procedure,the year-toyear increase or decrease(represented by DY)of RYRV is f rstly forecasted to yield a RYRV or the percentage anomalies of the RYRV.Thus,the statistical forecast model for the DY of RYRV is established f rst and then applied to forecast the RYRV.Five predictors of the DY of RYRV include the DY of the March-April-May(MAM)mean AAO index,the index of the DY of the Ural circulation in MAM,the index of the DY of the East Asian circulation in MAM at 500 hPa,the DY of the meridional wind shear between 850 hPa and 200 hPa in MAM over Indo-Australia,and the index of the winter DY of the southern Pacif c sea level pressure.The model captures the interannual variability of the observed percentage anomalies of the RYRV well,with a correlation coeffcient between the observed and modeled percentage anomalies of the RYRV of 0.79 during the training period of 1965-96(32 years),with a RMSE of 20%.The model hindcast for 1997-2006 shows a high level of accuracy for the percentage anomalies of the RYRV,with an RMSE of 18%.It even reproduces the upward and downward trends of the percentage anomalies of the RYRV,respectively,during the periods from 1985 to 1998and 1999to 2006.In particular,the model successfully reproduces the abnormal precipitation observed from 1997 to 1999,with relative errors below 10%.

The year-to-year increment was also applied to forecasting summer rainfall over North China(RNC)(Fan et al., 2009).Fan et al.(2009)identif ed only fve predictors for the DY of the RNC,including the DY of the DJF North East Asiasea levelpressurecirculation,theDYoftheMAMNorth Pacif c circulation at 500 hPa,the DY of the intensity of the June South Asian high,the DY of the intensity of the June polar vorticity over the Northern Hemisphere,and the DY of the June Ni˜no3 index.Results demonstrated that the correlation coeff cient between the observed and simulated percentage anomalies of the RNC is 0.8 during the training period of 1965-99,with an RMSE of 18%.It was noted that the model reproduces the downward trend of percentage anomalies ofthe RNC.The model shows good prediction capacity for the hindcast period of 2000-06,with an RMSE of 21%.The modelcapturestheabnormaldroughtyearin2002,with arelative error of 10%for the precipitationprediction.Therefore, itappearsthattheyear-to-yearincrementapproachhaspotential to signif cantly improve the operational forecast skill of summer rainfall over China due to its advantages of:(1)using merely fve or six predictors that can explain 60%-80% ofthe interannualvariancesof a predictand;and(2)capturing the decadal trend of a predictand.

Liu and Fan(2012b)also applied the year-to-year increment prediction approach to develop an effective statistical downscaling scheme for summer rainfall prediction at the station-to-station scale in southeastern China.The independent sample test and the cross-validation test showed that the downscaling scheme yields better predicted skill for summer precipitation at most stations over southeastern China than the original GCM outputs,with greater correlation coeffcients,as well as lower RMSEs.

4. Seasonal prediction of the climate background for dust weather and typhoon activity

4.1.Climate background of dust weather

Dust weather is quite frequent in North China in the winter and spring,resulting in air pollution and climatic consequences.Recently,dust weatheranomalies have dramatically changed on seasonal,interannual,and decadal timescales (Wang and Dong,1996;Zhou and Zhang,2003;Wang et al.,2012).This is related to many land surface processes as well as atmospheric circulation.Thus,the prediction of dust weather frequency(DWF)is a challenging but nevertheless worthwhile task.

Dust weather in North China has been revealed to be closelyrelatedtosurfacewindvelocity,humidity,andsurface air temperature.On seasonal and interannual timescales,surface wind and precipitation are regarded as having the most important roles in DWF(Ye et al.,2000;Zhang and Ren, 2003).In addition,the preceding SST anomalies in the Pacif c,the ENSO cycle,and climatic factors in the northern middle and high latitudes(e.g.,geopotential height at 500 hPa,polarvortexintensity,AO,and westerlies in the northern high latitudes)are also documented as signif cantly inf uencing the DWF in China(Zhao et al.,2004a;Kang and Wang, 2005).Recently,a signif cant negative correlation between the AAO and spring DWF in North China was reported(Fan and Wang,2004),which provides a new important potential predictor for DWF prediction.

In 2003,a GCM-based numerical prediction method for spring DWF was proposed and tested by Wang et al.(2003). The prediction was realized according to the dynamically predicted anomalous cold air intensity,precipitation,surface air temperature,etc.,in winter and spring.Encouraging results were obtained by performing a real-time seasonal prediction experiment utilizing the IAP GCM.Subsequently,real-time predictions for the spring DWF have been performed every year with a lead time of several months.A comparisonofthe predictionwiththe observationshowsgeneral agreement for four years but disagreement for two years.

With the goal to improve the prediction,Lang(2008)developed a new method by considering major predictors that come from either the precedingobservedatmosphere or from the GCM output by applying multiple linear regression techniques.The predictors identif ed include the seasonal mean anomalous surface air temperature,precipitation,the AO, AAO,SO,near surface meridional wind,and the Eurasian westerly index in the preceding winter and current spring.

Based respectively on the observed data and corresponding forecast experimental data,two prediction models (model-I and model-II,hereafter)were set up for the spring DWF using regression analysis.Validation of the models indicates that the correlation coeff cient between the observed and modeled time series is larger than 0.93 for the calibration period of 1955-2001for both models.An examinationof the retrospective forecasts shows that both models possess high prediction skill for the spring DWF in North China.However,there are two def ciencies embedded in model-I.One is a time limitation that may arise in the real-time prediction because the DJF mean observational data are used in the prediction,andtheotheris thatthe synchronousclimatesignalin the spring is not taken into account.However,in general,the prediction skill of the spring DWF in North China could be largely reinforced if these two models,especially model-II, were applied in practice.

Besides the sophisticated inf uencing processes and factors,thespringdustweatherinChinapossessesregionalcharacteristics.Although much progress has been made in frequencypredictionin China,thepredictionforsynopticspring dust weather in China is still diff cult.New prediction methods are urgently needed in future studies,such as a regional climate model to make more reliable predictions of both the DWF and individual dust events.

4.2.Climate background of typhoon activity

Typhoon activity signif cantly inf uences China.It is important and diff cult to forecast interannual typhoon activity in the WNP.Major scientif c achievements associated with seasonal forecasts of WNP typhoon activity are reported in this subsection.One aspect is the relationship between typhoon activity and atmospheric circulation,and the other is the developmentof statistical and dynamical predictionmodels of typhoon activity.

4.3.Factors associated with typhoon activity

The selection of predictors is based on the relationship between typhoon activity and atmospheric circulation.Xie et al.(1963)indicated that 80%of typhoons develop in the eastern partof theequatorialconvergencezoneoverthe western Pacif c.They further identif ed that when the equatorial westerly is enhanced and extended northward and eastward,the typhoon genesis frequency is increased.Ding et al.(1977)further provided a synoptic conceptual model forthe formation of multiple typhoons in the ITCZ,and they emphasized the role of interaction of the tropical circulations of both hemispheres in the development of the ITCZ. Dong and Zhang(1979)suggested that there is a stronger jet at the middle-low level before the formation of typhoons in the ITCZ.In this sense,the strengthened jet at the middlelow level on both sides of the ITCZ could be an indicator for the formation of typhoons in the ITCZ.Chen(1965)suggested that the movementof typhoonsis led by three to seven days by the high-middle latitude circulation over East Asia. Tao et al.(1962)pointed out that there are meridional and zonal f ow patterns over the tropical and subtropical area in Asia,which cause distinct characteristics of typhoon activity through the position of the zonal wind belt over the Northern Hemisphere,the distribution of long waves,and the locations of subtropical highs and the ITCZ.Xu and Gu(1978) further indicated that only when both the upper-and lowerlevel zonal-dominated circulations synchronously turn into meridional-dominated circulation over the tropical western Pacif c do multiple typhoons and stronger typhoons occur.

Previous studies have also suggested that tropical and subtropical circulations in the Southern Hemisphere play an importantrole in the developmentof typhoons(Li,1956;Tao et al.,1962,Ding et al.,1977;Xu and Gu,1978;Wang and Leftwich,1984).These studies speculate that outbreaks of cold air along with enhanced low-level cross-equatorial f ow near Australia would result in an enhanced ITCZ,which favors the developmentof typhoons over the WNP.These studies were based on case studies and provide potential indicators of medium-range weather forecasts for typhoon activity.By employing long-term datasets,it has been indicated that the ENSO and stratospheric Quasi-Biennial Oscillation (QBO)can inf uence the interannual variability of typhoon activity over the WNP via changes in thermal and dynamical conditions over the typhoon genesis region(Pan,1982; Ding and Wright,1983,Li,1985;Dong and Zhong,1989; Chan,1995).Therefore,ENSO is one of the main predictors for seasonal prediction models for typhoon activity over the WNP(Chan et al.,1998,2001).

However,the interannual variability of typhoon activities is much more complicated than previously recognized.Thus, the search for other predictors of typhoon activities is an essential aspect in research on the variability and predictability of typhoon activities.In this context,the role of high-latitude atmosphere modes has been stressed recently.

The AAO is an extratropical annular mode in the Southern Hemisphere.Recent analysis has indicated that the AAO has close relationships with the DWF in North China and summer rainfall over North China and the Yangtze River valleys(Gao et al.,2003;Fan and Wang,2004;Wang and Fan,2005;Fan,2006;Sun et al.,2009).Wang and Fan (2007)showed that the WNP typhoon frequency(WNPTF) is also signif cantly negatively correlated with the AAO in June-July-August-September(JJAS)in the period 1949-98. Based ontheabovestudies,a positivephaseofthe JJAS AAO corresponds to a larger magnitude of vertical zonal wind shear,anomalous low-level anticyclonic circulation,anomalous high-level cyclonic circulation,and lower SSTs in the major typhoon genesis region,thus providing an unfavorable environment for typhoon genesis,and vice versa.Through the meridional teleconnection in the high troposphere existing primarily in the Pacif c sector,the convective activities over the equatorial region of the western Pacif c are connected to the AAO,and in turn are connected to the convective activities in the WNP,which are associated with typhoon activities.

The NPO is a major mode in the interannual atmospheric variability over the North Pacif c,as indicated by a seesaw pattern between the high-and low-latitude sea-level pressure variability.Wang et al.(2007a)discussed the relationship between the NPO and typhoonactivity,as well as hurricanefrequencies.The correlation coeff cient between the NPO index in JJAS and the WNPTF is 0.37 for the period 1949-98.The NPO is also correlated with the annual hurricane number in the tropical Atlantic at-0.28 for the same period.The variability of the NPO is found to be concurrent with changes in the thermal and dynamical conditions of the WNPTF and the annual hurricane number via atmospheric teleconnections.

Fan(2007)further suggested that the sea ice cover in the North Pacif c both in DJF and in MAM is negatively correlated with the annual WNPTF during the period 1965-2004. Anomalous sea ice cover may modulate the WNPTF through variations in the NPO or through meridional teleconnections in the high troposphere existing primarily in the Pacif c sector.

Zhou et al.(2008)reported that the spring Hadley Cell is negativelycorrelated to the summer WNPTF.In addition,the teleconnectionpatternovertheextratropicalAsian-Pacif cregion,which can give rise to anomalous large-scale atmospheric circulations such as the western North Pacif c subtropicalhigh,the SouthAsian high,andwesterly jets,is identif ed as co-varying with the WNPTF(Zhou and Cui,2008).

4.3.1.Development of statistical models for typhoon activity prediction

Chan et al.(1998)developed an operational statistical forecast model for seasonal tropical cyclone activity over the WNP and the South China Sea using the projection pursuit regression technique.Predictors include monthly values of indices representing the ENSO and environmentalconditions over East Asia and the WNP for the months from April of the previous year to March of the current year.However, the scheme partially failed in 1997 and 1998,during which a warm and a cold ENSO event occurred,respectively.Chan et al.(2001)improvedthe schemebycoveringnewpredictors relatedto ENSO andupdatedthe schemeto includeApril and May predictors.

Fan(2007)considered the inf uence of high-latitude circulation,including newly identif ed predictors such as sea ice cover over the North Pacif c and the NPO in the winterspring.Multi-linear regression was applied to establishing a forecastmodelfortheWNPTF byusingthedatasetsof1965-99.The forecast model shows a high correlation coeff cient (0.79)between the model-simulated and the actual typhoonfrequencies for the period of 1965-99.The forecast model also exhibits reasonable hindcasts for the typhoon frequencies for the years 2000-06.Therefore,this work demonstrates that the new predictors are signif cant for the prediction of the interannual variability of the WNPTF.Crossvalidation of the model performance demonstrated the high forecast skill of the model established.

Later,Fan and Wang(2009)proposed a new approach to forecasting the WNPTF.In their work,the year-to-year increase or decrease in typhoon frequency is f rst forecasted to yield a net typhoon frequency prediction.This new approach has been successfully applied in establishing forecast models of the summer/spring rainfall over the Yangtze River valleys and North China,the temperature in North China,and the EASM(Fan et al.,2008,2009;Fan and Wang,2009;Fan, 2012;Fan et al.,2012).Results demonstrate that the new approach can capture not only the interannual variability but alsothe lineartrendsofa predictand,showinghighprediction capability.

Only fve predictors were used in the model set up by Fan and Wang(2009),including winter sea ice cover over the North Pacif c and other indices of thermal and dynamical conditions over the WNP.The model performs with reasonable accuracy for the periods of calibration and validation.The model successfully captures the larger typhoon frequency anomalies in the WNP during 1997 and 1998 and the smaller anomalies of the WNPTF during 2002-07.The cross-validation test of the prediction model showed that the new approach and the prediction model result in better prediction skill,as compared to models established based on the typhoon frequency itself.Thus,it appears that this new approach has the potential to improve the operational forecast skill for typhoon frequency in the WNP.

Sun and Chen(2011)developed a statistical downscaling method to predict the WNPTF using six predictors from DEMETER CGCMs during 1974-2001.The prediction skill of the statistical downscalingmethod is much better than that of the raw CGCMs.Similar improvement can also be found in the prediction of the number of landfalling Chinese typhoons(Sun and Ahn,2011).In addition,the multi-model ensemble has the best prediction performance.Therefore, combining a multi-model ensemble and statistical downscaling greatly improves the CGCM prediction skill.

4.3.2.Dynamical approach for typhoon activity prediction

Wang et al.(2006)carried out a real-time numerical experimentontheclimatebackgroundconditionsovertheWNP associated with typhoon activity for 2006 by using the IAP GCM.They reasonably predicted less than normal typhoon genesis in 2006 according to the prediction of the climate background for the WNPTF,including reduced convective activity,an increased magnitude of vertical zonal wind shear, and low-level divergence over the WNP.

Lang and Wang(2008)further evaluated the capability of the IAP GCM to forecast the climate background for typhoon activity over the WNP.They focused on the vertical shear of zonal wind,outgoing longwave radiation,and divergence f elds in the lower and upper troposphere during the summer,which are related to the thermal and dynamical conditions for typhoon genesis.After analyzing their 34-yr(1970-2003)ensemble hindcast experiment results,it was found that the temporal correlation coeff cients between the hindcast and observation were 0.70 and 0.62 for the vertical shear of zonal wind and the divergence f eld,respectively,in the key region of the WNP.These results suggested that the model possesses good potential skill for predicting the largescale climate background for WNP typhoon activity.

5. Summary and future prospects

Frequent f ooding in South China and serious droughts in North China often cause signif cant damage to human life, the nation’s economy,and public security.Such disasters have recently resulted in economic losses of about 20 million Euros per year in China.In addition,the issue of water supply has become increasingly serious in North China, where the economy and society are undergoing rapid development and the population and level of urbanization are increasing at present,and are predicted to continue to increase over the next 30 years.Therefore,short-term climate prediction will be a continuous focus for atmospheric research in China.This is even more important when we consider that the frequency of climate extremes in China will likely be reinforced under the global warming background,based on the observed climate variability in the 20th century and GCM-projected climate change in the 21st century.

Despite tremendous social demands,short-term climate prediction for the country is scientif cally a diff cult issue, basically because the link between the climate variability and the tropical oceanic variability is complicated,unstable,and generally weak,as well as the fact that the internal variance in the mid-and high-latitude atmosphere is large.In addition,large uncertainty remains in understanding the roles of various land surface processes,such as soil moisture,undergroundwatertables,snowandseaicecoveretc.,inshort-term climate variability in China.

Therefore,GCM-based seasonal climate predictions,either via the“two-tiered”or“one-tiered”approach,have limited skill,even when the state-of-the-art GCM is employed. The spatial correlation coeff cient between the simulated and observed JJA mean precipitation anomalies in China south of 40°N is less than 0.2 on average for the years of 1979-2001 for all models in the DEMETER program of Europe(Wang and Fan,2009).There have also been validationsof domestic models for seasonal prediction in China,and similar results have been obtained.

Therehavebeennumerousanalysesandmodelingstudies aimed at understandingvarious processes and factors responsible for the seasonal climate variability in China.Besides the local elements,some remote factors have been identif ed. Among these,snow cover over the Tibetan Plateau and Eurasia,sea ice cover in the Northern Hemisphere,the SST in the tropical ocean,the NAO,AAO,and cross-equatorial f owin the Indian-Pacif c sector are the most notable.Studies of these aspects have helped to establish more rational and effective forecast models for the seasonal climate variability in China.

However,much effort has been devoted to developing new techniques to improve seasonal climate prediction,particularly for the summer precipitation.These endeavors involve dynamical and statistical downscaling,year-to-year incremental forecasting,and new statistical models that adopt new sets of predictors including preceding observational information and/or GCM output.

A successful case in dynamical downscaling was achieved by the NCC in its seasonal precipitation prediction fortheyear2003(Dinget al.,2006).However,therehas been noreports thus farto showan improvedpredictionskill based onmulti-yeardynamicaldownscalingseasonalforecasts.Regardless,such an effort would be worthwhile,and the results will likely be promising.

Statistical downscaling and correction techniques developed for climate prediction in China have achieved signifcant improvements for precipitation forecasting.In a recent new scheme developed by Wang and Fan(2009),the precipitation anomalies of the most similar year and the most dissimilar year,judged by the spatial anomaly pattern similarity between the observed and modeled summer precipitation in the tropical Eastern Hemisphere,are used to compose new predicted precipitation anomalies in East Asia and the western Pacif c region.This new scheme was tested by using the DEMETER multi-model hindcasts from Europe,and signifcant improvement was obtained by applying the new scheme for all six models and the multi-modelensemble for the years of 1979-2001.Thus,the developmentandapplicationofnew downscaling techniques may produce substantially increased forecast skill in operational seasonal climate prediction.

Another successful new technique is related to the selection of the prediction objective.Traditionally,the anomaly of a quantityas comparedto its multi-yearaverage is adoptedas the predictand.Fan et al.(2008)and Fan and Wang(2009) proposedto substitute the“anomaly”by the“year-to-yearincrement”for the prediction objective and re-established forecast models for the year-to-year increment.Applications in the seasonal forecast of summer rainfall in the Yangtze River valleys and central North China,as well as in the seasonal prediction of the WNP typhoon frequency,have demonstrated the apparent advantage of this new technique.All of the applications show improved forecast skill and decreased forecast errors.

Correct simulation of local soil moisture and snow cover over the Tibetan Plateau and Eurasia has been demonstrated to be signif cant for improving the seasonal prediction for precipitation as well as other quantities.A study by Zhan andLin(2011)documentedanexampleofimprovedseasonal predictionskill in the Yangtze and Huaihe river basins for the spring of 2008 by better describing the initial soil moisture.

With all of the above progress that has been achieved and problems that still remain,future work in the context of seasonal climate prediction in China should focus on several aspects such as process-based studies,GCM improvement, downscaling techniques,new statistical models based on the identif cation of new predictors and new processes,new integration schemes for seasonal prediction,and the development of various data assimilation schemes for land surface and oceanic processes.Multi-model ensembles always give high scores of seasonal prediction.Thus,the design of new schemes associated with the method of multi-model ensemble prediction is another important research focus in the future.

Acknowledgements.This work was supported by theNational Natural Science Foundation of China(Grant Nos.41130103 and 41210007).

REFERENCES

Bai,R.H.,2001:Relations between the anomaly of sea surface temperature in the Atlantic and the precipitation in summer over northeast China.Marine Science Bulletin,20,23-29.(in Chinese)

Chan,J.C.L.,1995:Tropical cyclone activityinthewesternNorth Pacif c in relation to the stratospheric quasi-biennial oscillation.Mon.Wea.Rev.,123,2567-2571.

Chan,J.C.L.,J.E.Shi,and C.M.Lam,1998:Seasonal forecasting of tropical cyclone activity over western North Pacif cand the South China Sea.Wea.Forecasting,13,997-1003.

Chan,J.C.L.,J.E.Shi,and K.S.Liu,2001:Improvements in the seasonal forecasting of tropical cyclone activity over the western North Pacif c.Wea.Forecasting,16,491-498.

Chen,H.,2003:IAP dynamical extraseasonal-interannual climate prediction system and its real-time prediction.Ph.D.thesis, Instituteof Atmospheric Physics,Graduate Universityof Chinese Academy of Sciences.(in Chinese)

Chen,H.,and Z.H.Lin,2006:A correction method suitable for dynamical seasonal prediction.Adv.Atmos.Sci.,23,425-430, 10.1007/s00376-006-0425-3.

Chen,H.P.,J.Q.Sun,and H.J.Wang,2012:A statistical downscaling model for forecasting summer rainfall in China from DEMETER hindcast datasets.Wea.Forecasting,27,608-628.

Chen,J.M.,P.Zhao,X.Y.Guo,and H.L.Liu,2009:Modeling impacts of vegetation in western China on the summer climate of northwestern China.Adv.Atmos.Sci.,26,803-812, doi:10.1007/s00376-009-9018-2.

Chen,L.J.,2008:Seasonal predictability and downscaling methods based on CGCMs.Ph.D.thesis,Institute of Atmospheric Physics,Graduate University of Chinese Academy of Sciences.(in Chinese)

Chen,L.S.,1965:Flow patterns in westerlies in relation to the East Asia typhoon tracks.Acta Meteor.Sinica,35,476-485. (in Chinese)

Chen,L.T.,1988:The zonal anomaly between tropical Indian and Pacif c Oceans and its impact on Asian summer monsoon.J. Atmos.Sci.,12,142-148.

Chen,L.X.,M.Dong,and Y.N.Shao,1992:Thecharacteristics of interannual variations on the East Asian monsoon.J.Meteor. Soc.Japan,70,397-421.

Dai,X.G.,J.F.Chou,and G.X.Wu,2002:The teleconnection relationship between Indian monsoon and East Asian summercirculation.Acta Meteor.Sinica,60,544-552.(in Chinese)

Dai,Y.J.,and Q.C.Zeng,1996:A land surface model(IAP94) for climate studies.Part I:Formulation and validation in off-line experiments.Adv.Atmos.Sci.,14,433-460,doi: 10.1007/s00376-997-0063-4.

Deng,A.J.,S.Y.Tao,and L.T.Chen,1989:The temporal and spatial distributions of Indian Ocean SST and its relationships with China rainfall.Chinese J.Atmos.Sci.,13,393-399.(in Chinese)

Ding,Q.H.,and B.Wang,2005:Circumglobal teleconnection in the Northern Hemispheric summer.J.Climate,18,3483-3505.

Ding,Y.G.,Z.H.Jiang,and Y.F.Zhu,1998:Experiment on short-term climate prediction models on SSTA over the Nino oceanic region.J.Trop.Meteor.,5,1-8.(in Chinese)

Ding,Y.H.,and E.R.Wright,1983:The large scale circulation condition for the western Pacif c typhoon genesis.Acta Oceanol.Sinica,5,561-574.(in Chinese)

Ding,Y.H.,H.J.Fan,Q.F.Xue,and G.X.Chen,1977:Preliminary study on the multityphoon development in the tropical convergence zone.Chinese J.Atmos.Sci.,2,89-98.(in Chinese)

Ding,Y.H.,Y.M.Liu,Y.J.Song,and Q.Q.Li,2002:Research and experiments of the dynamical model system for Short-Term climate prediction.Climatic and Environmental Research,7,236-246.(in Chinese)

Ding,Y.H.,and Coauthors,2004:Advance in seasonal dynamical prediction operation in China.Acta Meteor.Sinica,62, 598-612.(in Chinese)

Ding,Y.H.,Y.M.Liu,X.L.Shi,Q.Q.Li,Q.P.Li,and Y. Liu,2006:Multi-Year simulations and experimental seasonal predictions for rainy seasons in China by using a nested Regional Climate Model(RegCM NCC)Part II:The experimental seasonal prediction.Adv.Atmos.Sci.,23,487-503, doi:10.1007/s00376-006-0323-8

Dong,K.Q.,and W.P.Zhang,1979:On the relations between the formation of typhoon in ITCZ and the JET in the middlelower troposphere.Acta Meteor.Sinica,37,44-52.(in Chinese)

Dong,K.Q.,and S.Zhong,1989:The correlation between SST for equatorial eastern Pacif c and frequency of occurrence for tropical storms in the South China Sea.J.Trop.Meteor.,5, 345-350.(in Chinese)

Dong,M.,2001:The Basic Principle and Use Method of AGCM/NCC.China Meteorological Press,Beijing,152 pp. (in Chinese)

Enomoto,T.,B.J.Hoskins,and Y.Matsuda,2003:The formation mechanism of the Bonin high in August.Quart.J.Roy. Meteor.Soc.,129,157-178.

Fan,K.,2006:Atmospheric circulation in Southern Hemisphere and summer rainfall over Yangtze River valleys.Chinese J. Geophys.,49,599-606.(in Chinese)

Fan,K.,2007:North Pacif c sea ice cover,a predictor for the western North Pacif c typhoon frequency?Sci.China:Earth Sci., 50,1251-1257.

Fan,K.,2012:A statistical prediction model for spring rainfall over northern China based on the interannual increment approach.J.Trop.Meteor.,18,543-550.

Fan,K.,and H.J.Wang,2004:Antarctic oscillation and the dust weather frequency in North China.Geophys.Res.Lett.,31, L10201,doi:10.1029/2002GL019465.

Fan,K.,and H.J.Wang,2009:A new approach to forecasting typhoon frequency over the western North Pacif c.Wea.Forecasting,24,974-986,doi:10.1175/2009WAF2222194.1.

Fan,K.,H.J.Wang,and Y.-J.Choi,2008:A physically-based statistical forecast model for the middle-lower reaches of the Yangtze River valleys summer rainfall.Chinese Sci.Bull.,53, 602-609

Fan,K.,M.J.Lin,and Y.Z.Gao,2009:Forecasting the summer rainfall in North China using the year-to-year increment approach.Sci.China:Earth Sci.,52,532-539.

Fan,K.,Y.Liu,and H.P.Chen,2012:Improving the prediction of the East Asian summer monsoon:New approaches.Wea. Forecasting,27,1017-1030.

Feddersen,H.,A.Navarra,and M.N.Ward,1999:Reduction of model systematic error by statistical correction for dynamical seasonal predictions.J.Climate,12,1974-1989.

Fu,C.B.,and Z.M.Zeng,2005:Correlations between North Atlantic Oscillation index in winter and eastern China f ood/drought index in summer in the last 530 years.Chinese Sci.Bull.,50,2505-2516.

Fu,W.W.,and G.Q.Zhou,2007:Improved ENSO simulation in regional coupled GCM using regressive correction method.Sci.China:Earth Sci.,50,1258-1265.

Gao,H.,F.Xue,and H.J.Wang,2003:Inf uence of interannual variability of Antarctic Oscillation on Meiyu along the Yangtze and Huaihe River Valley and its importance to prediction.Chinese Sci.Bull.,48,61-67.

Gao,L.,H.L.Ren,J.P.Li,and J.F.Chou,2006:Analogue correction method of errors and its application to numerical weather prediction.Chinese Phys.,15,882-889.

Gong,D.Y.,and C.H.Ho,2003:Arctic oscillation signals in the East Asian summer monsoon.J.Geophys.Res.,108,4066, doi:10.1029/2002JD002193.

Guo,W.D.,and H.J.Wang,2003:A case study of the improvement of soil moisture initialization in IAP-PSSCA.Adv.Atmos.Sci.,20,845-848,doi:10.1007/BF02915411.

Han,J.P.,and R.H.Zhang,2009:The dipole mode of the summer rainfall over East China during 1958-2001.Adv.Atmos.Sci., 26,727-735,doi:10.1007/s00376-009-9014-6.

He,S.P.,2013:Reduction of the East Asian winter monsoon interannual variability after the mid-1980s and possible cause.Chinese Sci.Bull.,58(12),1331-1338.

He,S.P.,and H.J.Wang,2013a:Oscillating relationship between the East Asian Winter Monsoon and ENSO.J.Climate, 26(24),9819-9838.

He,S.P.,and H.J.Wang,2013b:Impact of the November/ December Arctic Oscillation on the following January temperature in East Asia.J.Geophys.Res.,118,12981-12998.

He,S.P.,H.J.Wang,J.P.Liu,2013:Changes in the relationship between ENSO and Asia-Pacif c Midlatitude Winter Atmospheric Circulation.J.Climate,26,3377-3393,doi: http://dx.doi.org/10.1175/JCLI-D-12-00355.1

Huang,G.,and Z.W.Yan,1999:The East Asian summer monsoon circulation anomaly index and its interannual variations.Chinese Sci.Bull.,44,1325-1329.

Huang,R.H.,and W.J.Li,1987:Inf uence of the heat source anomaly over the tropical western Pacif c on the subtropical high over East Asia,April 10-15,1987.Proc.International Conference on the General Circulation of East Asia, Chengdu,China,40-51.

Huang,R.H.,and L.Lu,1989:Numerical simulation of the relationship between the anomaly of subtropical high over East Asia and the convective activities in the western trop-ical Pacif c.Adv.Atmos.Sci.,6,202-214,doi:10.1007/BF 02658016.

Huang,R.H.,and Y.F.Wu,1989:The inf uence of ENSO on the summer climate change in China and its mechanism.Adv. Atmos.Sci.,6,21-32,doi:10.1007/BF02656915.

Huang,R.H.,and F.Y.Sun,1992:Impacts of the tropical western Pacif c on the East Asian summer monsoon.J.Meteor.Soc. Japan,70,243-256.

Jin,Z.H.,and R.G.Shen,1987:A comparison of the SST distribution between the f oods and drought years in the middle and lower basin of the Yangtze River and the associated circulation systems.Assembly of Meteor.Sci.and Tec.,China Meteorology Press,83-88.

Ju,J.H.,J.M.Lu,J.Cao,and J.Z.Ren,2005:Possible impact of the ArcticOscillation on the interdecadal variation of summer monsoon rainfall in East Asia.Adv.Atmos.Sci.,22,39-48, doi:10.1007/BF02930868.

Kang,D.J.,and H.J.Wang,2005:Analysis on the decadal scale variation of the dust storm in North China.Sci.China(D),48, 2260-2266.

Kang,H.,C.K.Park,S.N.Hameed,and K.Ashok,2009:Statistical downscaling of precipitation in Korea using multimodel output variables as predictors.Mon.Wea.Rev.,137,1928-1938.

Kang,H.W.,K.H.An,C.K.Park,A.L.S.Solis,and K.Stitthichivapak,2007:Multimodel output statistical downscaling prediction of precipitation in the Philippines and Thailand.Geophys.Res.Lett.,34,L15710,doi:10.1029/2007GL 030730.

Ke,Z.,P.Zhang,W.Dong,and L.Li,2009:A new way to improve seasonal prediction by diagnosing and correcting the Inter model systematic errors.Mon.Wea.Rev.,137,1898-1907.

Keenlyside,N.,and R.Kleeman,2002:Annual cycle of equatorial zonal currents in the Pacif c.J.Geophys.Res.,107,8-1-8-13.

Landsea,C.W.,and J.A.Knaff,2000:How much skill was there in forecasting the very strong 1997-98 El Ni˜no?Bull.Amer. Meteor.Soc.,81,2107-2119.

Lang,X.M.,2003:IAP dynamical theory and methodology of short-term climatic prediction and study on ensemble prediction experiment.Ph.D.dissertation,Institute of Atmospheric Physics,Graduate University of Chinese Academy of Sciences.(in Chinese)

Lang,X.M.,2008:Prediction model for spring dust weather frequency in North China.Sci.China(D),51,709-720.

Lang,X.M.,and H.J.Wang,2008:Can the climate background of western North Pacif c typhoon activity be predicted by climate model?Chinese Sci.Bull.,53,2392-2399.(in Chinese)

Lang,X.M.,H.J.Wang,and D.B.Jiang,2004a:Extraseasonal short-term predictions of summer climate with IAP9LAGCM.Chinese J.Geophys.,47,22-28.(in Chinese)

Lang,X.M.,H.J.Wang,G.Q.Zhou,and D.B.Jiang,2004b:Prediction of summer climate over China in 2002 with IAP9LAGCM and its performance verif cation.Journal of Nanjing Institute of Meteorology,27,101-107.(in Chinese)

Lau,K.M.,1992:East Asian summer monsoon rainfall variability and climate teleconnection.J.Meteor.Soc.Japan,70,211-242.

Lau,K.M.,and H.Y.Weng,2001:Coherent modes of global SST and summer rainfall over China:An assessment of the regional impacts of the 1997-98 El Ni˜no.J.Climate,14,1294-1308.

Li,C.Y.,1985:El Nino and the western Pacif c typhoon activity.Chinese Sci.Bull.,30,1087-1089.(in Chinese)

Li,C.Y.,1990:On Interaction Between Anomalous Circulation Climate in East Asia and El Ni˜no Climate Change Dynamics and Modeling.China Meteorological Press,101-126.(in Chinese)

Li,C.Y.,and M.Q.Mu,2001:The dipole in the equatorial Indian Ocean and its impacts on climate.Chinese J.Atmos.Sci.,25, 433-443.(in Chinese)

Li,F.,2008:Methodology of seasonal prediction and its application in the prediction of summer rainfall over East Asian monsoon region.Ph.D.dissertation,Institute of Atmospheric Physics,Chinese Academy of Science,96 pp.(in Chinese)

Li,F.,and H.J.Wang,2012:Autumn sea ice cover,winter Northern Hemisphere annular mode,and winter precipitation in Eurasia.J.Climate,26,3968-3981.

Li,F.,Z.D.Lin,R.T.Zuo,and Q.C.Zeng,2005:The methods for correcting the summer precipitation anomaly predicted extraseasonal over East Asian monsoon region based on EOF and SVD.ClimaticandEnvironmental Research,10,658-668.(in Chinese)

Li,F.,Q.C.Zeng,and C.F.Li,2009:A bayesian scheme for probabilistic multi-model ensemble prediction of summer rainfall over the Yangtze River Valley.Atmos.Ocean.Sci.Lett.,2, 314-319.

Li,J.,R.C.Yu,and T.J.Zhou,2008:Teleconnection between NAO and climate downstream of the Tibetan Plateau.J.Climate,21,4680-4690.

Li,S.L.,2004:Impact of northwest Atlantic SST anomalies on the circulation over the Ural Mountains during early winter.J.Meteor.Soc.Japan,82,971-988.

Li,W.J.,and Coauthors,2005:Research and operational application of dynamical climate model prediction system.J.Appl. Meteor.Sci.,16,1-11.(in Chinese)

Li,X.Z.,1956:A comprehensive theory for the typhoon genesis.Acta Meteor.Sinica,27,87-89.(in Chinese)

Li,Y.H.,1992:A preliminary analysis on the relationship between Southern Oscillation and dryness/wetness in the eastern China during the last 400 years.Mar.Sci.,4,37-40.(in Chinese)

Lian,Y.,2007:Correlation between North Pacif c Oscillation and East Asian Summer Monsoon.Sci.Geogr.Sinica,27(Sl),19-26.(in Chinese)

Liang,X.Z.,1996:Description of a nine-level grid point atmospheric general circulation model.Adv.Atmos.Sci.,13,269-298,doi:10.1007/BF02656847.

Lin,Z.H.,and Q.C.Zeng,1997:Simulation of East Asian summer monsoon by using an improved AGCM.Adv.Atmos.Sci., 14,513-526,doi:10.1007/s00376-997-0069-y.

Lin,Z.H.,Q.C.Zeng,and B.Ouyang,1996:Sensitivity of the IAP two-level AGCM to surface albedo Variations.Theor. Appl.Climatol.,55,157-162.

Lin,Z.H.,and Coauthors,1998:An improved short-term climate prediction system and its application to the extra-seasonal predictionof rainfall anomalyinChinafor1998.Climaticand Environmental Research,3,339-348.(in Chinese)

Lin,Z.H.,X.Li,and Y.Zhao,1999:The impact of the land surface processes on thepredictive skill of IAPprediction system for short-term climate anomalies.Study of energy and water cycle over Huaihe River Basin(I),B.L.Zhao and Y.H.Ding, Eds.,China Meteorological Press,187-200.(in Chinese)

Lin,Z.H.,X.S.Yang,and Y.F.Guo,2001:Sensitivity of land sur-face model to the initial condition of soil moisture.Climatic and Environmental Research,6,240-248.(in Chinese)

Liu,J.P.,J.A.Curry,H.J.Wang,M.R.Song,and R.M.Horton, 2012:Impact of declining Arctic sea ice on winter snowfall.Proc.Natl.Acad.Sci.USA,109,4074-4079.

Liu,N.,and S.L.Li,2014:Predicting summer rainfall over the Yangtze-Huai region based on timescale decomposition statistical downscaling.Wea.Forecasting,29,162-176.

Liu,S.F.,2007:The development of IAP coupled atmosphereland-vegetation modeland the quantif cation of landatmosphere coupling strength.Institute of Atmospheric Physics,Graduate University of Chinese Academy of Sciences.(in Chinese)

Liu,Y.,and K.Fan,2012a:Prediction of spring precipitation in China using a downscaling approach.Meteor.Atmos.Phys., 118,79-93.

Liu,Y.,and K.Fan,2012b:Improve the prediction of summer precipitation in the southeastern China by a hybrid statistical downscaling model.Meteor.Atmos.Phys.,117,121-134.

Liu,Y.,and K.Fan,2013:A new statistical downscaling model for autumn precipitation in China.Int.J.Climatol.,33,1321-1336.

Liu,Y.,and K.Fan,2014:An application of hybrid downscaling model to forecast summer precipitation at stations in China.Atmos.Res.,143,17-30,doi:10.1016/j.atmosres.2014.01. 024.

Liu,Y.M.,W.J.Li,and P.Q.Zhang,2005a:A global 4-dimensional ocean data assimilation system and the studies on its results in the tropic Pacif c.Acta Oceanol.Sinica,27, 27-35.(in Chinese)

Liu,Y.M.,R.H.Zhang,Y.H.Yin,and T.Niu,2005b:The application of ARGO data to the Global Ocean Data Assimilation Operational System of NCC.Acta Meteor.Sinica,19,354-365.(in Chinese)

Liu,Y.Y.,and Y.H.Ding,2008:Teleconnection between the Indian summer monsoon onset and the Meiyu over the Yangtze River valleys.Sci.China(D),51,1021-1035.

Lu,R.Y.,J.H.Oh,and B.J.Kim,2002:A teleconnection pattern in upper-level meridional wind over the North African and Eurasian continent in summer.Tellus A,54,44-55.

Lu,R.Y.,B.W.Dong,and H.Ding,2006:Impact of the Atlantic Multidecadal Oscillation on the Asian summer monsoon.Geophys.Res.Lett.,33,L24701,doi:10.1029/2006GL 027655.

Luo,S.H.,Z.H.Jin,and L.T.Chen,1985:The analysis of correlation between sea surface temperature in the Indian South-China Sea and precipitation in the middle and lower reaches of the Changjiang River.Chinese J.Atmos.Sci.,9,314-320. (in Chinese)

Ma,J.H.,and H.J.Wang,2014:Design and testing of a global climate prediction system based on a coupled climate model.Sci.China:Earth Sci.,doi:10.1007/s11430-014-4875-7.

Ma,J.H.,H.J.Wang,and Y.Zhang,2012:Will boreal winter precipitation over China increase in the future?The AGCM simulation under summer“ice-free Arctic”conditions.Chinese Sci.Bull.,57,921-926.

Ma,J.H.,H.J.Wang,and K.Fan,2014:Dynamic downscaling test of summer precipitation prediction over China in 1998 using WRFand CCSM4.Adv.Atmos.Sci.,doi:10.1007/ s00376-014-4143-y.(in press)

Nan,S.L.,and J.P.Li,2003:The relationship between the summer precipitation in the Yangtze River valleys and the boreal spring Southern Hemisphere Annular Mode.Geophys.Res. Lett.,30,2266,doi:10.1029/2003GL018381.

Nitta,T.,1987:Convective activities in the tropical western Pacif c and their impacts on the Northern Hemisphere summer circulation.J.Meteor.Soc.Japan,64,373-390.

Pan,Y.H.,1982:The effect of the thermal state of equatorial eastern Pacif con the frequency of typhoons over western Pacif c.Acta Meteor.Sinica,40,24-34.(in Chinese)

Ren,H.L.,2006:Strategy and methodology of dynamical analogue prediction.Department of Atmospheric Sciences, Lanzhou University.(in Chinese)

Ren,H.L.,2008:Predictor-basederror correction method inshortterm climate prediction.Prog.Nat.Sci.,18,129-135.

Ren,H.L.,and J.F.Chou,2006:Introducing the updating of multi reference states into dynamical analogue prediction.ActaMeteor.Sinica,64,315-324.

Ren,H.L.,and J.F.Chou,2007:Strategy and methodology of dynamical analogue prediction.Sci.China(D),50,1589-1599.

Shen,S.,andK.M.Lau,1995:Biennial oscillationassociated with theEast Asian summer monsoon and tropical sea surface temperatures.J.Meteor.Soc.Japan,73,105-124.

Shi,L.,Y.H.Yin,and Y.Q.Ni,2001:Experimental ENSO predictions with a simply global tropical air-sea couple model.Chinese J.Atmos.Sci.,25,628-640.(in Chinese)

Shi,N.,and Q.G.Zhu,1993:Studies on the northern early summer teleconnection patterns,their interannual variations and relation to drought/f ood in China.Adv.Atmos.Sci.,10,155-168.

Sun,J.Q.,and J.B.Ahn,2011:A GCM-based forecasting model for the landfall of tropical cyclones inChina.Adv.Atmos.Sci., 28,1049-1055,doi:10.1007/s00376-011-0122-8.

Sun,J.Q.,and H.P.Chen,2011:Predictability of western North Pacif c typhoon activity and its factors using DEMETER coupled models.Chinese Sci.Bull.,56,3474-3479.

Sun,J.Q.,and H.P.Chen,2012:A statistical downscaling scheme to improve global precipitation forecasting.Meteor.Atmos. Phys.,117,87-102,doi:10.1007/s00703-012-0195-7.

Sun,J.Q.,and H.J.Wang,2012:Changes of the connection between the summer North Atlantic Oscillation and the East Asian summer rainfall.J.Geophys.Res.,117,D08110,doi: 10.1029/2012JD017482.

Sun,J.Q.,H.J.Wang,and W.Yuan,2008a:Decadal variations of the relationship between the summer North Atlantic Oscillation and middle East Asian air temperature.J.Geophys.Res., 113,D15107,doi:10.1029/2007JD009626.

Sun,J.Q.,W.Yuan,and Y.Z.Gao,2008b:Arabian Peninsula-North Pacif c Oscillation and its association with the Asian summer monsoon.Sci.China(D),51,1001-1012.

Sun,J.Q.,H.J.Wang,and W.Yuan,2009:A possible mechanism for theco-variabilityof theboreal springAntarcticOscillation and the Yangtze River valleys summer rainfall.Int.J.Climatol.,29,1276-1284,doi:10.1002/joc.1773.

Sung,M.K.,W.Kwon,H.Back,K.Boo,G.Lim,and J.Kug, 2006:A possible impact of the North Atlantic Oscillation on the East Asian summer monsoon precipitation.Geophys.Res. Lett.,33,L21713,doi:10.1029/2006GL027253.

Tao,S.Y.,and L.X.Chen,1987:A review of recent research on the East Asian summer monsoon in China.Monsoon Meteorology,C.P.Chang and T.N.Krishnamuti,Eds.,Oxford University Press,60-92.

Tao,S.Y.,S.Y.Xu,and Q.Y.Guo,1962:Thecharacteristics of the zonal and meridional circulation over tropical and subtropicalregions in eastern Asia in summer.Acta Meteor.Sinica,32, 91-103.(in Chinese)

Tao,S.Y.,F.Zhu,and W.Zhao,1988:On the year-to-year variation of Meiyu.Chinese J.Atmos.Sci.,12,13-21.(in Chinese)

Tu,C.W.,1937:Relationship between the atmospheric circulation and world precipitation.Acta Geogr.Sinica,4,833-844.

Tu,C.W.,and S.S.Huang,1944:The seasonal migration and retreat of monsoon in China.Acta Meteor.Sinica,18,81-92.

von Storch,H.,1999:The global and regional climate system.Anthropogenic Climate Change,H.von Storch and G.Flosser, Eds.,Springer Verlag,3-36.

von Storch,H.,E.Z.H.,and U.Cubasch,1993:Downscaling of climate change estimates to regional scales:An application to winter rainfall in the Iberian Peninsula.J.Climate,6,1161-1171.

Wang,B.,R.Wu,and X.Fu,2000a:Pacif c-East Asia teleconnection:How does ENSO affect East Asian climate?J.Climate, 13,1517-1536.

Wang,H.J.,1997:On the uncertainty of the short-term climate prediction.Climatic and Environmental Research,2,333-338.(in Chinese)

Wang,H.J.,2002:The instability of the East Asian summer monsoon ENSO relations.Adv.Atmos.Sci.,19,1-11,doi: 10.1007/s00376-002-0029-5.

Wang,H.J.,2005:The Circum-Pacif c teleconnection pattern in meridional wind in the high troposphere.Adv.Atmos.Sci.,22, 463-466,doi:10.1007/BF02918759.

Wang,H.J.,and Q.C.Zeng,1992a:The seasonal simulation of the climate of 9000 years BP by using the IAP AGCM.Adv. Atmos.Sci.,9,451-457,doi:10.1007/BF02677077.

Wang,H.J.,and Q.C.Zeng,1992b:Modeling of the ice age climate.Acta Geogr.Sinica,50,279-289.(in Chinese)

Wang,H.J.,and K.Fan,2005:Central-north China precipitation as reconstructed from the Qing dynasty:Signal of the Antarctic Atmospheric Oscillation.Geophys.Res.Lett.,32,L24705, doi:10.1029/2005GL024562.

Wang,H.J.,and K.Fan,2006:Southern Hemisphere mean zonal wind in upper troposphere and East Asian summer monsoon circulation.Chinese Sci.Bull.,51,1508-1514.

Wang,H.J.,and K.Fan,2007:Relationship between the Antarctic Oscillation in the western North Pacif c typhoon frequency.Chinese Sci.Bull.,52,561-565.

Wang,H.J.,and K.Fan,2009:A new scheme for improving the seasonal prediction of summer precipitation anomalies.Wea. Forecasting,24,548-554.

Wang,H.J.,and H.P.Chen,2012:Climate control for Southeastern China moisture and precipitation:Indian or East Asian Monsoon?J.Geophys.Res.,117,D12109,doi:10.1029/2012 JD017734.

Wang,H.J.,and S.P.He,2012:Weakening relationship between East Asian Winter Monsoon and ENSO after mid-1970s.Chinese Sci.Bull.,57,3535-3540.doi:10.1007/s11434-012-5285-x.

Wang,H.J.,Q.C.Zeng,and X.H.Zhang,1993:The numerical simulation of the climatic change caused by CO2doubling.Sci.China(B),36,451-462.(in Chinese)

Wang,H.J.,F.Xue,and X.Q.Bi,1997:The interannual variability and predictability of a global climate model.Adv.Atmos. Sci.,14,554-562,

Wang,H.J.,G.Q.Zhou,and Y.Zhao,2000b:An effective method for correcting the seasonal-interannual prediction of summer climate anomaly.Adv.Atmos.Sci.,17,234-240,doi: 10.1007/s00376-000-0006-9.

Wang,H.J.,X.M.Lang,G.Q.Zhou,and D.J.Kang,2003:A preliminary report of the model prediction on theforthcoming winter and spring dust climate over China.Chinese J.Atmos. Sci.,27,136-140.(in Chinese)

Wang,H.J.,X.M.Lang,K.Fan,J.Q.Sun,and G.Q.Zhou,2006: Real-time climate prediction experiment for the typhoon frequency in the western North Pacif c for 2006.Climate Environ.Res.,11,133-137.(in Chinese)

Wang,H.J.,J.Q.Sun,and K.Fan,2007a:Relationships between the North Pacif c oscillation and the typhoon/hurricane frequencies.Sci.China(D),50,1409-1416.

Wang,H.J.,L.J.Chen,W.J.Li,P.Q.Zhang,and L.L.Liu, 2007b:Predictability of DERF on monthly mean temperature and precipitation over China.Acta Meteor.Sinica,65,725-731.(in Chinese with English abstract)

Wang,H.J.,andCoauthors,2012:ExtremeclimateinChina:Facts, simulation and projection.Meteorologische Zeitschrift,21, 279-304.doi:10.1127/0941-2948/2012/0330.

Wang H.J.,S.P.He,and J.P.Liu,2013:Present and future relationship between the East Asian winter monsoon and ENSO: Results of CMIP5.J.Geophys.Res.Ocean,118,5222-5237, doi:10.1002/jgrc.20332.

Wang,J.Z.,and P.W.Leftwich,1984:A major low-level crossequatorial current at 110°E during the Northern summer and its relation to typhoon activities.Sci.Atmos.Sinica,8,443-449.(in Chinese)

Wang,S.G.,and G.R.Dong,1996:A study on sand-dust storms over the desert region in North China.J.Nat.Disast.,5,86-94.(in Chinese)

Wang,Y.,2001:Research on mechanism of 1998’s catastrophic f oods by numerical simulation.Institute of Atmospheric Physics,Graduate University of Chinese Academy of Sciences.(in Chinese)

Wang,Y.M.,S.L.Li,and D.H.Luo,2009:Seasonal response of Asian monsoonal climate to the Atlantic Multidecadal Oscillation.J.Geophys.Res.,114,D02112,doi:10.1029/2008 JD010929.

Ward,M.N.,and A.Navarra,1997:Pattern analysis of SST-forced variability inensemble GCM simulations:Examples over Europe and the tropical Pacif c.J.Climate,10,2210-2220.

Wu,A.M.,and Y.Q.Ni,1999:A hybrid coupled oceanatmosphere model and ENSO prediction study.Adv.Atmos. Sci.,16,405-418,doi:10.1007/s00376-999-0019-y.

Wu,B.Y.,K.Yang,and R.H.Zhang,2009a:Eurasian snow cover variability and its association with summer rainfall in China.Adv.Atmos.Sci.,26,31-44,doi:10.1007/s00376-009-0031-2.

Wu,B.Y.,R.H.Zhang,and B.Wang,2009b:On the association between spring arctic sea ice concentration and Chinese summer rainfall:A further study.Adv.Atmos.Sci.,26,666-678, doi:10.1007/s00376-009-9009-3.

Wu,G.X.,and H.Z.Liu,1995:Neighbourhood response of rainfall to tropical sea surface temperature anomalies.Part I:numerical experiment.Chinese J.Atmos.Sci.,19,422-434.(in Chinese)

Xie,Y.B.,S.J.Chen,Y.L.Zhang,and Y.L.Huang,1963:A preliminarily statistic and synoptic study about the basic currents over southeastern Asia and the initiation of typhoons.Acta Meteor.Sinica,33,207-217.(in Chinese)

Xu,H.M.,J.H.He,and M.Dong,2001:Interannual variability of the Meiyu onset and its association with North Atlantic Oscil-lation and SSTA over North Atlatic.Acta Meteor.Sinica,59, 694-706.(in Chinese)

Xu,J.M.,and M.R.Gu,1978:The relationship between the circulation features and the typhoon genesis over the western Pacif c in summer.Chinese J.Atmos.Sci.,2,174-178.(in Chinese)

Xue,F.,and C.Z.Liu,2007:The inf uence of moderate ENSO on summer rainfall in eastern China and its comparison with strong ENSO.Chinese Sci.Bull.,53,791-800,doi: 10.1007/s11434-008-0002-5.

Xue,F.,H.J.Wang,and J.H.He,2003:Interannual variability of Mascarene high and Australian high and their inf uences on summer rainfall over East Asia.Chinese Sci.Bull.,48,492-497.

Xue,F.,H.J.Wang,and J.H.He,2004:Interannual Variability of Mascarene High and Australian High and their Inf uences on East Asian Summer Monsoon.J.Meteor.Soc.Japan,82, 1173-1186.

Yang,J.L.,Q.Y.Liu,S.P.Xie,Z.Y.Liu,and L.X.Wu, 2007:Impact of the Indian Ocean SST basin mode on the Asian summer monsoon.Geophys.Res.Lett.,34,L02708, doi:10.1029/2006GL028571.

Yang,S.,and K.M.Lau,1998:Inf uence of sea surface temperature and Ground wetness on Asian Summer Monsoon.J.Climate,11,3230-3246.

Yeh,T.-C.,R.T.Wetherald,and S.Manabe,1983:A model study of the Short-Term climatic and hydrologic effects of sudden Snow-Cover removal.Mon.Wea.Rev.,111,1013-1024.

Yeh,T.-C.,R.T.Wetherald,and S.Manabe,1984:The effect of soil moisture on the Short-Term climate and hydrology change—A numerical experiment.Mon.Wea.Rev.,112,474-490.

Ye,T.-C.,J.F.Chou,J.Y.Liu,Z.X.Zhang,Y.M.Wang,Z.J. Zhou,H.B.Ju,and Q.Huang,2000:Causes of sand-stormy weather in northern China and control measures.Acta Geogr. Sinica,55,513-521.(in Chinese)

Yu,E.T.,H.J.Wang,and J.Q.Sun,2010:A quick report on a dynamical downscaling simulation over China using the nested model.Atmos.Oceanic Sci.Lett.,3,325-329.

Yu,Y.Q.,and X.H.Zhang,2000:FluxAnomalyCoupling scheme of atmosphere-ocean.Research on DCMPS,China Meteorological Press,201-207.(in Chinese)

Yuan,W.,and J.Q.Sun,2009:Enhancement of the Summer North Atlantic Oscillation inf uence on the Northern Hemispheric air temperature.Adv.Atmos.Sci.,26,1209-1214,doi: 10.1007/s00376-009-8148-x.

Zeng,Q.C.,C.G.Yuan,X.H.Zhang,X.Z.Liang,and N.Ban, 1987:A global grid-point general circulation model.Collection of papers presented at the WMO/IUGG NWP Symposium,Tokyo,4-8 August 1987,421-430.

Zeng,Q.C.,C.G.Yuan,W.Q.Wang,and R.H.Zhang,1990: Numerical extra-seasonal prediction experiment for climate anomalies.Scientia Atmospherica Sinica,14,10-25.(in Chinese)

Zeng,Q.C.,B.L.Zhang,C.G.Yuan,P.S.Lu,F.L.Yang,X.Li, and H.J.Wang,1994:A note on some methods suitable for verifying and correcting the prediction of climate anomaly.Adv.Atmos.Sci.,11,121-127,doi:10.1007/BF02666540.

Zeng,Q.C.,and Coauthors,1997:Seasonal and Extraseasonal predictions of summer monsoon precipitation by GCMs.Adv. Atmos.Sci.,14,163-176,doi:10.1007/s00376-997-0017-x

Zeng,X.B.,M.Shaikh,Y.Dai,R.E.Dickinson,and R.Myneni, 2002:Coupling of the common land model to the NCAR community climate model.J.Climate,15,1832-1854.

Zhan,Y.L.,2008:Impact of Soil Moisture on the Seasonal Predictability over East Asia.Institute of Atmospheric Physics, Graduate University of Chinese Academy of Sciences.(in Chinese)

Zhan Y.L.,and Z.H.Lin,2011:Impact of Initial Soil Moisture Anomalies on the seasonal prediction skill over China.Presentation at 28th Annual Meeting of Chinese Meteorological Society,Xiamen,Nov 2-4.

Zhang,F.,H.Chen,Z.H.Lin,and Q.C.Zeng,2004:Improvement of horizontal resolution of IAP AGCM-I and its inf uence on the simulations of global and East Asian climate.Climatic and Environmental Research,9,396-408.(in Chinese)

Zhang,J.,W.J.Zhu,and Z.X.Li,2007:Relationship between winter North Pacif c Oscillations and summer precipitation anomalies in the Huaihe river basin.Journal of Nanjing Institute of Meteorology,30,546-550.(in Chinese)

Zhang,L.,and G.Y.Ren,2003:Change in dust storm frequency and the climatic controls in northern China.Acta Meteor. Sinica,61,744-750.(in Chinese)

Zhang,R.H.,A.Sumi,and M.Kimoto,1996:Impact of El Ni˜no on the East Asian monsoon:A diagnostic study of the 86/87 and 91/92 events.J.Meteor.Soc.Japan,74,49-62.

Zhang,R.H.,S.E.Zebiak,R.Kleeman,and N.Keenlyside,2005: Retrospective El Ni˜no forecast using an improved intermediate coupled model.Mon.Wea.Rev.,133,2777-2802.

Zhang,R.H.,Y.H.Yin,Q.Q.Li,Y.M.Liu,and T.Niu,2006: Utilizing ARGO data to improve the prediction of ENSO and Short-term climate prediction of summer rainfall in China.J. Appl.Meteor.Sci.,17,538-547.(in Chinese)

Zhang,S.L.,and S.Y.Tao,2001:Inf uence of snow cover over the Tibetan Plateau on Asian summer monsoon.Chinese J. Atmos.Sci.,25,372-390.(in Chinese)

Zhang,X.H.,1990:Dynamical framework of IAP nine-level atmospheric general circulation model.Adv.Atmos.Sci.,7,66-77,doi:10.1007/BF02919169.

Zhao,C.Z.,X.Dabu,and Y.Li,2004a:Relationship between climatic factors and dust storm frequency in inner Mongolia of China.Geophys.Res.Lett.,31,L01103,doi:10.1029/ 2003G1018351.

Zhao,P.,X.D.Zhang,X.J.Zhou,M.Ikeda,and Y.H.Yin,2004b: The sea ice extent anomaly in the North Pacif c and its impact on the East Asian summer monsoon rainfall.J.Climate,17, 3434-3447.

Zhao,P.,Y.N.Zhu,and R.H.Zhang,2007:An Asian-Pacif c teleconnection in summer tropospheric temperature and associated Asian climate variability.Climate Dyn.,29,293-303.

Zhao,Y.,X.Li,C.G.Yuan,and Y.F.Guo,1999:Quantitative assessment and improvement to correction technology on prediction system of short term climate anomaly.Climatic and Environmental Research,4,353-364.(in Chinese)

Zhao,Z.C.,Q.Q.Li,and Z.Q.Zhang,2000:A model for forecasting the interannual variation in ENSO events.Res.Seas. Fore.,Project,Off ce,Expert,and Team,Eds.,China Meteorological Press,323-333.(in Chinese)

Zheng,F.,and J.Zhu,2010:Coupled assimilation for an intermediated coupled ENSO prediction model.Ocean Dyn.,60, 1061-1073.

Zheng,F.,J.Zhu,R.H.Zhang,and G.Q.Zhou,2006:Ensemble hindcasts of SSTanomalies in the tropical Pacif cusing an intermediate coupled model.Geophys.Res.Lett.,33,L19604,doi:10.1029/2006GL026994.

Zhou,B.T.,and X.Cui,2008:Hadley circulation signal in the tropical cyclone frequency over the western North Pacif c.J. Geophys.Res.,113,D16107,doi:10.1029/2007JD009156.

Zhou,B.T.,X.Cui,and P.Zhao,2008:Relationship between the Asian-Pacif c oscillation and the tropical cyclone frequency in the western North Pacif c.Sci.China(D),51,380-385.

Zhou,G.Q.,and X.Li,2000:A global oceanic data assimilation system based on an ocean general circulation model.Development of Operational Dynamical Models for Short-term Climate Predictions,China Meteorological Press,393-400.(in Chinese)

Zhou,G.Q.,and Q.C.Zeng,2001:Predictions of ENSO with a coupled atmosphere-ocean general circulation model.Adv. Atmos.Sci.,18,587-603,doi:10.1007/s00376-001-0047-8.

Zhou,G.Q.,X.Li,and Q.C.Zeng,1998:A coupled Ocean-Atmosphere general circulation model for ENSO prediction and 1997/1998 ENSO forecast.Climatic and Environmental Research,3,349-357.(in Chinese)

Zhou,G.Q.,Q.C.Zeng,and R.H.Zhang,1999:An improved coupled ocean-atmosphere general circulation model and its numerical simulation.Prog.Natural Sci.,9,374-381.

Zhou,Z.J.,and G.C.Zhang,2003:Typical severe dust storms in northern China during 1954-2002.Chinese Sci.Bull.,48, 2366-2370.

Zhu,C.W.,C.K.Park,W.S.Lee,and W.T.Yun,2008:Statistical downscaling for multi-model ensemble prediction of summer monsoon rainfall in the Asia-Pacif c region using geopotential height f eld.Adv.Atmos.Sci.,25,867-884,doi: 10.1007/s00376-008-0867-x.

Zhu,J.S.,G.Q.Zhou,R.H.Zhang,and Z.B.Sun,2009:An improved hybrid coupled model:ENSO simulations.Chinese J. Atmos.Sci.,33,657-669.(in Chinese)

Zhu,J.S.,G.Q.Zhou,and R.H.Zhang,2013:An improved ENSO prediction in a hybrid coupled model with an embedded empirical entrainment temperature parameterization.Int. J.Climatol.,33,343-355.

Zhu,K.Z.,1934:The southeastern monsoon and the precipitation in China.Acta Geogr.Sinica,1,1-27.

Zhu,Y.L.,2009:The Antarctic oscillation-East Asian summer monsoon connection in NCEP-1 and ERA-40,Adv.Atmos. Sci.,26,707-716,doi:10.1007/s00376-009-8196-2.

Zhu,Y.L.,H.J.Wang,W.Zhou,and J.H.Ma,2011:Recent changes in the summer precipitation pattern in East China and the background circulation.Climate Dyn.,36,1463-1473, doi:10.1007/s00382-010-0852-9.

:Wang,H.J.,and Coauthors,2015:A review of seasonal climate prediction research in China.Adv.Atmos.Sci., 32(2),149-168,

10.1007/s00376-014-0016-7.

(Received 25 May 2014;revised 27 June 2014;accepted 18 July 2014)

∗Corresponding author:WANG Huijun

Email:wanghj@mail.iap.ac.cn