Multi-model Projection of July August Climate Extreme Changes over China under CO 2 Doubling. Part I: Precipitation

Similar documents
Multi-Model Projection of July August Climate Extreme Changes over China under CO 2 Doubling. Part II: Temperature

East China Summer Rainfall during ENSO Decaying Years Simulated by a Regional Climate Model

Projected change in extreme rainfall events in China by the end of the 21st century using CMIP5 models

Projection of Future Precipitation Change over China with a High-Resolution Global Atmospheric Model

The Interdecadal Variation of the Western Pacific Subtropical High as Measured by 500 hpa Eddy Geopotential Height

Impact of overestimated ENSO variability in the relationship between ENSO and East Asian summer rainfall

Monsoon Activities in China Tianjun ZHOU

FUTURE PROJECTIONS OF PRECIPITATION CHARACTERISTICS IN ASIA

Future pattern of Asian drought under global warming scenario

Using observations to constrain climate project over the Amazon - Preliminary results and thoughts

SUPPLEMENTARY INFORMATION

SUPPLEMENTARY INFORMATION

Altiplano Climate. Making Sense of 21st century Scenarios. A. Seth J. Thibeault C. Valdivia

The Formation of Precipitation Anomaly Patterns during the Developing and Decaying Phases of ENSO

fm``=^n_!"#$%&'()*+!"#$%&'()*

Changes in Daily Climate Extremes of Observed Temperature and Precipitation in China

Decrease of light rain events in summer associated with a warming environment in China during

Oceanic origin of the interannual and interdecadal variability of the summertime western Pacific subtropical high

Low-level wind, moisture, and precipitation relationships near the South Pacific Convergence Zone in CMIP3/CMIP5 models

Weakening relationship between East Asian winter monsoon and ENSO after mid-1970s

SUPPLEMENTARY INFORMATION

A Quick Report on a Dynamical Downscaling Simulation over China Using the Nested Model

Changing links between South Asian summer monsoon circulation and tropospheric land-sea thermal contrasts under a warming scenario

NOTES AND CORRESPONDENCE. Seasonal Variation of the Diurnal Cycle of Rainfall in Southern Contiguous China

Interdecadal and Interannnual Variabilities of the Antarctic Oscillation Simulated by CAM3

Comparison of the seasonal cycle of tropical and subtropical precipitation over East Asian monsoon area

Future Changes of Drought and Flood Events in China under a Global Warming Scenario

A Preliminary Analysis of the Relationship between Precipitation Variation Trends and Altitude in China

Recent weakening of northern East Asian summer monsoon: A possible response to global warming

The increase of snowfall in Northeast China after the mid 1980s

Dynamic and thermodynamic changes in mean and extreme precipitation under changed climate

The Coupled Model Predictability of the Western North Pacific Summer Monsoon with Different Leading Times

Spatiotemporal patterns of changes in maximum and minimum temperatures in multi-model simulations

Supplementary Figure 1 Observed change in wind and vertical motion. Anomalies are regime differences between periods and obtained

Changes in the El Nino s spatial structure under global warming. Sang-Wook Yeh Hanyang University, Korea

The ENSO s Effect on Eastern China Rainfall in the Following Early Summer

Will a warmer world change Queensland s rainfall?

INVESTIGATING THE SIMULATIONS OF HYDROLOGICAL and ENERGY CYCLES OF IPCC GCMS OVER THE CONGO AND UPPER BLUE NILE BASINS

SUPPLEMENTARY INFORMATION

18. ATTRIBUTION OF EXTREME RAINFALL IN SOUTHEAST CHINA DURING MAY 2015

the 2 past three decades

How Well Do Atmospheric General Circulation Models Capture the Leading Modes of the Interannual Variability of the Asian Australian Monsoon?

Long-term climate variations in China and global warming signals

Trends of Tropospheric Ozone over China Based on Satellite Data ( )

Climate Change Scenarios in Southern California. Robert J. Allen University of California, Riverside Department of Earth Sciences

Evaluation of the Twentieth Century Reanalysis Dataset in Describing East Asian Winter Monsoon Variability

Somali Jet Changes under the Global Warming

Attribution of anthropogenic influence on seasonal sea level pressure

Impact of the Atlantic Multidecadal Oscillation on the Asian summer monsoon

How Will Low Clouds Respond to Global Warming?

Changes in Mean and Extreme Temperature and Precipitation over the Arid Region of Northwestern China: Observation and Projection

Yuqing Wang. International Pacific Research Center and Department of Meteorology University of Hawaii, Honolulu, HI 96822

4.4 EVALUATION OF AN IMPROVED CONVECTION TRIGGERING MECHANISM IN THE NCAR COMMUNITY ATMOSPHERE MODEL CAM2 UNDER CAPT FRAMEWORK

Sensitivity of Precipitation in Aqua-Planet Experiments with an AGCM

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

Abstract: The question of whether clouds are the cause of surface temperature

Climate change in China in the 21st century as simulated by a high resolution regional climate model

Projections of the 21st Century Changjiang-Huaihe River Basin Extreme Precipitation Events

New proofs of the recent climate warming over the Tibetan Plateau as a result of the increasing greenhouse gases emissions

EL NIÑO MODOKI IMPACTS ON AUSTRALIAN RAINFALL

Climate change uncertainty for daily minimum and maximum temperatures: A model inter-comparison

Long-term changes in total and extreme precipitation over China and the United States and their links to oceanic atmospheric features

Increased Tibetan Plateau Snow Depth An Indicator of the Connection between Enhanced Winter NAO and Late- Spring Tropospheric Cooling over East Asia

22. DO CLIMATE CHANGE AND EL NIÑO INCREASE LIKELIHOOD OF YANGTZE RIVER EXTREME RAINFALL?

Interdecadal variability in the thermal difference between western and eastern China and its association with rainfall anomalies

Impacts of Climate Change on Autumn North Atlantic Wave Climate

Does the model regional bias affect the projected regional climate change? An analysis of global model projections

Consistent changes in twenty-first century daily precipitation from regional climate simulations for Korea using two convection parameterizations

NARCliM Technical Note 1. Choosing GCMs. Issued: March 2012 Amended: 29th October Jason P. Evans 1 and Fei Ji 2

Instability of the East Asian Summer Monsoon-ENSO Relationship in a coupled global atmosphere-ocean GCM

Spatial patterns of probabilistic temperature change projections from a multivariate Bayesian analysis

SST forcing of Australian rainfall trends

Duration and Seasonality of Hourly Extreme Rainfall in the Central Eastern China

Assessment of Snow Cover Vulnerability over the Qinghai-Tibetan Plateau

Simulation and Projection of the Western Pacific Subtropical High in CMIP5 Models

Global Monsoons Modeling Inter-comparison Project (GMMIP) and Challenges from Observational Data Perspective. ZHOU Tianjun.

Light rain events change over North America, Europe, and Asia for

20. EXTREME RAINFALL (R20MM, RX5DAY) IN YANGTZE HUAI, CHINA, IN JUNE JULY 2016: THE ROLE OF ENSO AND ANTHROPOGENIC CLIMATE CHANGE

Development of Super High Resolution Global and Regional Climate Models

CHAPTER 8 NUMERICAL SIMULATIONS OF THE ITCZ OVER THE INDIAN OCEAN AND INDONESIA DURING A NORMAL YEAR AND DURING AN ENSO YEAR

IAP Dynamical Seasonal Prediction System and its applications

Chapter 7 Projections Based on Downscaling

More extreme precipitation in the world s dry and wet regions

ENSO amplitude changes in climate change commitment to atmospheric CO 2 doubling

On Improving Precipitation Diurnal Cycle and Frequency in Global Climate Models

Explaining Changes in Extremes and Decadal Climate Fluctuations

Subseasonal Characteristics of Diurnal Variation in Summer Monsoon Rainfall over Central Eastern China

Ozone hole and Southern Hemisphere climate change

19. RECORD-BREAKING HEAT IN NORTHWEST CHINA IN JULY 2015: ANALYSIS OF THE SEVERITY AND UNDERLYING CAUSES

Effects of Large Volcanic Eruptions on Global Summer Climate and East Asian Monsoon Changes

Robust Arctic sea-ice influence on the frequent Eurasian cold winters in past decades

The Tropospheric Land Sea Warming Contrast as the Driver of Tropical Sea Level Pressure Changes

Contrasting impacts of spring thermal conditions over Tibetan Plateau on late-spring to early-summer precipitation in southeast China

4. Climatic changes. Past variability Future evolution

Hadley Centre for Climate Prediction and Research, Met Office, FitzRoy Road, Exeter, EX1 3PB, UK.

Prediction Research of Climate Change Trends over North China in the Future 30 Years

1 Ministry of Earth Sciences, Lodi Road, New Delhi India Meteorological Department, Lodi Road, New Delhi

Climate Change Scenario, Climate Model and Future Climate Projection

!"#$%&'()*+,-./ I!"#$%&

Interannual Relationship between the Winter Aleutian Low and Rainfall in the Following Summer in South China

Transcription:

ADVANCES IN ATMOSPHERIC SCIENCES, VOL. 28, NO. 2, 2011, 433 447 Multi-model Projection of July August Climate Extreme Changes over China under CO 2 Doubling. Part I: Precipitation LI Hongmei 1,2 ( ), FENG Lei 1,2 ( ), and ZHOU Tianjun ( ) 1 State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029 2 Graduate University of Chinese Academy of Sciences, Beijing 100049 (Received 16 April 2010; revised 16 August 2010) ABSTRACT Potential changes in precipitation extremes in July August over China in response to CO 2 doubling are analyzed based on the output of 24 coupled climate models from the Twentieth-Century Climate in Coupled Models (20C3M) experiment and the 1% per year CO 2 increase experiment (to doubling) (1pctto2x) of phase 3 of the Coupled Model Inter-comparison Project (CMIP3). Evaluation of the models performance in simulating the mean state shows that the majority of models fairly reproduce the broad spatial pattern of observed precipitation. However, all the models underestimate extreme precipitation by 50%. The spread among the models over the Tibetan Plateau is 2 3 times larger than that over the other areas. Models with higher resolution generally perform better than those with lower resolutions in terms of spatial pattern and precipitation amount. Under the 1pctto2x scenario, the ratio between the absolute value of MME extreme precipitation change and model spread is larger than that of total precipitation, indicating a relatively robust change of extremes. The change of extreme precipitation is more homogeneous than the total precipitation. Analysis on the output of Geophysical Fluid Dynamics Laboratory coupled climate model version 2.1 (GFDL-CM2.1) indicates that the spatially consistent increase of surface temperature and water vapor content contribute to the large increase of extreme precipitation over contiguous China, which follows the Clausius Clapeyron relationship. Whereas, the meridionally tri-polar pattern of mean precipitation change over eastern China is dominated by the change of water vapor convergence, which is determined by the response of monsoon circulation to global warming. Key words: extreme precipitation, projection, coupled climate model, CO 2 doubling Citation: Li, H. M., L. Feng, T. J. Zhou, 2011: Multi-model projection of July August climate extreme changes over China under CO 2 doubling. Part I: Precipitation. Adv. Atmos. Sci., 28(2), 433 447, doi: 10.1007/s00376-010-0013-4. 1. Introduction Corresponding author: ZHOU Tianjun, zhoutj@lasg.iap.ac.cn The global and regional climate changes, especially in extremes, induced by increasing greenhouse gases are important topics of study for the climate research community. The World Climate Research Program s (WCRP) phase 3 of the Coupled Model Intercomparison Project (CMIP3) provides an opportunity for model comparison and multi-model ensemble strategy. Coordinated by the WCRP CMIP3 project, projections of climate change under different scenarios were done for Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC AR4). Numerous analyses on the output of these scenario projections (IPCC AR4) have been carried out. Coupled model simulations show that precipitation extremes may increase over most regions in a future warmer climate due to increasing greenhouse gases (Kharin and Zwiers, 2000, 2005; Semenov and Bengtsson, 2002; Wilby and Wigley, 2002; Meehl et al., 2005; Tebaldi et al., 2006), and extreme precipitation has been projected to increase more significantly than mean precipitation in future climate change on a global scale (Kharin and Zwiers, 2000; Emori and Brown, 2005; Kharin and Zwiers, 2005). However, consensus and significance of these sim- China National Committee for International Association of Meteorology and Atmospheric Sciences (IAMAS), Institute of Atmospheric Physics (IAP) and Science Press and Springer-Verlag Berlin Heidelberg 2011

434 PROJECTION OF EXTREME PRECIPITATION CHANGE OVER CHINA: PRECIPITATION VOL. 28 ulations are weaker when regional patterns are considered. Meehl et al. (2005) showed that the increases of precipitation intensity do not have a uniform spatial distribution under future warmer climate scenario and different processes work over different areas. Turner and Slingo (2009a) showed that the uncertainties among model simulations in spatial distribution and magnitude changes of precipitation extremes over India monsoon region under CO 2 doubling scenarios, i.e., the magnitudes of simulated extreme precipitation changes in 6 of 15 models are close to the theoretical results based on Clausius Clapeyron relationship, while the predictions of another six models exceed the theoretical results. Furthermore, the predictions in the remaining three models are unrelated to local surface warming. Using simulation of the Hamburg Atmosphere-Ocean Coupled Circulation Model (ECHO-G), Min et al. (2006) projected that East Asian precipitation would increase in the 21st century with larger amplitudes than those of global means. Kimoto (2005) also predicted a weakened winter monsoon and increased activity of East Asian monsoonal rain band under the global-warming scenario. How would climate change, especially in terms of extremes over China in the future with a higher concentration of greenhouse gases? Projections of future climate extremes over the Yangtze River valley based on the ensemble mean of CMIP3 models simulation were analyzed by Xu et al. (2009), and the results modeled a larger fraction of heavy precipitation in the 21st century than is occurring in the present. Due to China s complex topography and coastal line distribution, the simulation of climate over China has been a great challenge for state-of-the-art global climate models (Zhou and Li, 2002; Zhou and Yu, 2006; Zhou et al., 2009a, b; Li et al., 2010; Chen et al., 2010). Using a regional climate model nested within a global ocean atmosphere coupled model has been a useful downscaling method (Gao et al., 2008). Regional climate model simulations also showed that precipitation in China would increase remarkably and extreme events tend to increase under increased greenhouse gases (Gao et al., 2001, 2002; Xu et al., 2006; Zhang et al., 2006). Because the output of global coupled models has been widely used to drive regional climate models as well as hydrologic models for downscaling of dynamic processes, the uncertainties among the global models deserve further research. The main motivation for this study is to investigate the potential changes of summertime climate extremes over China under a CO 2 doubling scenario using the output of CMIP3 models. Rather than presenting only the results of the multi-model ensemble (MME), the differences among the individual models are also discussed. The first part of the study focuses on precipitation. Our analysis shows that the majority of CMIP3 models fairly reproduce the observed spatial pattern of July August mean and extreme precipitation over contiguous China. However, all of the models underestimate the amount of extreme precipitation by about 50%. Models with higher resolutions generally yield better performances. The projected change of extreme precipitation under a CO 2 doubling scenario is more homogeneous than the total precipitation. The change in extreme precipitation follows the Clausius Clapeyron relationship and is determined by changes in surface temperature and water vapor content, while the change in mean precipitation is greatly affected by advection and dynamical processes associated with monsoon circulation. The remainder of the paper is organized as follows: section 2 describes the models, dataset, and methods used in this study. In section 3, the evaluations of the performance of the models in simulating the precipitation over China are presented using the Twentieth-Century Climate in Coupled Model (20C3M) experiment simulations. Potential changes of extreme precipitation over China were projected using the 1% per year CO 2 increase experiment (to doubling) (1pctto2x), and the results of these analyses are presented in section 4. The possible mechanisms for changes in mean and extreme precipitation are examined in section 5. Discussion and summary are provided in section 6. 2. Data, models, and methodology The observational data used in this study consists of the daily precipitation readings from 740 stations covering contiguous China for the period 1961 2000. This dataset was developed by the National Climate Center of the China Meteorological Administration and has been widely used in climate variability studies (Li et al., 2008). In our study, the data underwent quality-control procedures according to Alexander et al. (2006). The simulated daily precipitation data were extracted from the WCRP CMIP3 multi-model dataset. The WCRP CMIP3 multi-model dataset consists of the outputs from the current state-of-the-art coupled climate system models and is archived at the Program for Climate Model Diagnosis and Inter-comparison (PCMDI). In this study, we compare the 20C3M simulations against the observations. The 20C3M climate simulations were made with various combinations of observational forcing agents including greenhouse gases (GHGs), sulfate aerosols, ozone, volcanic aerosols, and solar variability; thus they could be com-

NO. 2 LI ET AL. 435 Table 1. Description of coupled climate models used in this study. They are ordered by the number of atmospheric horizontal grids. Interpolation grid systems are 5.0 4.0, 2.8125 2.8125, and 1.875 1.875 for LOW, MED and HIGH, respectively. The last column shows each model s resolution groups. Atmosphere Resolution Model Institute/Country (number of grids) Group INM-CM3.0 1,2 INM/Russia 5.0 4.0 L21 (3240) LOW GISS-EH 1 NASA/USA 5.0 4.0 L20 (3312) LOW GISS-ER 2 NASA/USA 5.0 4.0 L20 (3312) LOW CGCM3.1(T47) 1,2 CCCma/Canada 3.75 3.75 L31 (4608) LOW ECHO-G 1,2 MIUB/Germany/Korea 3.75 3.75 L19 (4608) LOW GISS-AOM 1 NASA/USA 4.0 3.0 L12 (5400) LOW IPSL-CM4 1,2 IPSL/France 3.75 2.5 L19 (6912) MED UKMO-HadCM3 1 Metoffice/UK 3.75 2.5 L38 (7008) MED FGOALS-g1.0 1 IAP/China 2.8125 3.0 L26 (7680) MED BCCR-BCM2.0 1,2 BCCR/Norway 2.8125 2.8125 L31 (8192) MED CGCM3.1(T63) 1 CCCma/Canada 2.8125 2.8125 L31 (8192) MED CNRM-CM3 1,2 CNRM/France 2.8125 2.8125 L45 (8192) MED MIROC3.2(medres) 1,2 CCSR/Japan 2.8125 2.8125 L20 (8192) MED PCM1 1 NCAR/USA 2.8125 2.8125 L26 (8192) MED MRI-CGCM2.3.2 1,2 MRI/Japan 2.8125 2.8125 L30 (8192) MED GFDL-CM2.0 1,2 GFDL/USA 2.5 2.0 L24 (12960) MED GFDL-CM2.1 1,2 GFDL/USA 2.5 2.0 L24 (12960) MED CSIRO-MK3.0 1,2 CSIRO/Australia 1.875 1.875 L18 (18432) HIGH CSIRO-MK3.5 1,2 CSIRO/Australia 1.875 1.875 L18 (18432) HIGH ECHAM5 1 MPI/Germany 1.875 1.875 L31(18432) HIGH UKMO-HadGEM 1 Metoffice/UK 1.875 1.25 L38(27840) HIGH CCSM3 1 NCAR/USA 1.40625 1.40625 L26(32768) HIGH INGV-SXG 1,2 INGV/Italy 1.125 1.125 L19 (51200) HIGH MIROC3.2(hires) 1,2 CCSR/Japan 1.125 1.125 L56 (51200) HIGH 1 20C3M experiment data are available. 2 1pctto2x experiment data are available. pared with the observed climate changes (Zhou and Yu, 2006). The performance of 20C3M models in simulating surface air temperature was assessed by Zhou and Yu (2006). The 1pctto2x and pre-industrial control simulations were used to model climate changes in China under the scenario of CO 2 doubling. In the 1pctto2x experiment, the atmospheric CO 2 was transiently increased at a rate of 1% per year. The concentration of atmospheric CO 2 doubled around the 70th model year. Then the atmospheric CO 2 was fixed at the doubled concentration, and an extended 150 years of model integration was done. The preindustrial control run is a long enough experiment to construct a non-drifting climate state under preindustrial conditions, i.e., the concentrations of model CO 2 were fixed at the level of the year 1860. The time period considered in this study is 1961 2000 for observations and the 20C3M simulation, and the last 20 year simulations of both the 1pctto2x run and the preindustrial control run. The difference between the 1pctto2x run and the preindustrial control run is used in the projection of CO 2 doubling. Information about the models used in this study is provided in Table 1. The models are ranked by the number of atmospheric horizontal grids and are classified into three categories according to their horizontal resolution: low, medium, and high. The last column shows each model s resolution group. Although models are independent, some models are actually different versions of a same system: MIROC3.2 (hires) and MIROC3.2 (medres) are only different in resolution; GFDL-CM2.0 and GFDL-CM2.1 are different only in numerical schemes for atmospheric advection. Precipitation days are defined as days with daily total precipitation >1 mm d 1 in both models and observations, because climate models generally rain too frequently (Sun et al., 2006). In addition to the total precipitation amount (hereafter PRCPTOT), three indices are used to describe precipitation extremes: extreme precipitation amount, which is defined as accumulated precipitation with daily precipitation amount greater than the 95th percentile of all precipitation days (hereafter R95p); extreme precipitation frequency, which is defined by dividing the number of days with extreme precipitation by the number of all days (hereafter R95pF); the percentage contri-

436 PROJECTION OF EXTREME PRECIPITATION CHANGE OVER CHINA: PRECIPITATION VOL. 28 bution from extreme precipitation to the total precipitation (hereafter R95pT). These indices were widely used in studies of extreme climate events (Frich et al., 2002; Zhai et al., 2005; Alexander et al., 2006; Li et al., 2008). Our analysis focuses on July August mean condition instead of the whole June August period. A few studies have noted that the past change of July August climate is different from that of June (Xu, 2001; Yu et al., 2004; Yu and Zhou, 2007; Li et al., 2008). Yu et al. (2004) also found a distinctive strong tropospheric cooling trend in East Asia during July August. Both mean and extreme precipitation amounts in July August have significantly increased along the Yangtze River valley and have evidently decreased in North China from 1958 2000, but this variation pattern was not significant in June (Li et al., 2008). Wang et al. (2009) also argued that it is reasonable to discuss the June and July August climate respectively. Following Zhou and Yu (2006) and Li and Zhou (2010), standard deviation is used to quantitatively measure the spread among the model simulations. The ensemble mean is taken with respect to individual models. Details of the calculation were given by Zhou and Yu (2006). The spread among the models represents the uncertainty aspect, which can be seen as the noise, and the result of the multi-model ensemble can be seen as the signal. When the ratio between the absolute value of precipitation change and model spread was >1, the precipitation change could be considered as relatively robust. 3. Evaluation of the model performance Before presenting the projected climate changes under global warming scenario, we evaluate the performance of the models in simulating the present climate. The simulated mean and extreme precipitation are compared with the observations. The 1961 2000 mean PRCPTOT in July August from the observation and the 20C3M simulations are shown in Fig. 1. The individual model simulations are shown according to the model resolution, from low to high. As shown in the MME (Fig. 1b), the models well reproduce the observed broad spatial precipitation pattern, i.e., precipitation over contiguous China decreases from southeast to northwest, with the largest precipitation located over southwestern China (the eastern flank of the Tibetan Plateau) and southeastern China, and the smallest precipitation seen over northwestern China. However, the consistency between the observation and the simulation is weaker when regional patterns are considered. Most models overestimate precipitation over the Tibetan Plateau and underestimate precipitation over eastern China, especially in South China and coastal areas of northern China (Fig. 1c). In individual simulations (Figs. 1d x), most models miss the large precipitation over North China. The models with higher resolutions generally have better performances than those with low or medium resolutions in simulating the spatial pattern of precipitation over China. The rain belt north to the middle and lower reaches of the Yangtze River valley is reasonably reproduced by the models with higher resolutions. Furthermore, the artificial precipitation center located to the east of Tibetan Plateau is less evident in high-resolution models. Horizontal resolution is an important factor for improving the simulation of summer mean precipitation over China. The overestimated precipitation over the Tibetan Plateau shown in low resolution models such as CMIP3 models (Fig. 1u) and CAM3 model (Chen et al., 2010) was greatly reduced in the high-resolution version (T319, corresponding to a horizontal resolution of 40 km) of ECHAM5 (Feng et al., 2010). Using a regional climate model, Gao et al. (2006, 2008) also emphasized that a spatial resolution adequate to resolve the physical and dynamical processes is important for accurately simulating the distribution of precipitation over China. The R95p in July August (Fig. 2) has similar spatial distribution as PRCPTOT (Fig. 1) in both the observation and the specific model simulation. The coupled models generally underestimate the R95p over eastern China by 50% (Figs. 2a c). As noted in previous studies, precipitation occurred too frequently at reduced intensity in many climate models, and most models produced too much convective and too little stratiform precipitation (Dai and Trenberth, 2004; Dai, 2006; Sun et al., 2006). For extreme precipitation, it seems that resolution is also an essential factor for improving the model simulation. The models with higher resolutions generally produce more extreme precipitation over eastern China than those employing low to medium resolutions (Figs. 2d x). For example, the extreme precipitations simulated by CSIRO-MK3.5, INGV-SXG, and MIROC3.2 (hires) models are generally higher than the other models. Another study also found that the simulated extreme precipitation in a high-resolution version of ECHAM5 is even larger than the observation (Feng et al., 2010) in contrast to the underestimation of CMIP3 models shown in Fig. 2u. To quantitatively evaluate the resemblance between the observations and model simulations, we employ the Taylor diagram here (Taylor, 2001) (Fig. 3). Each number in the Taylor diagram represents the re-

NO. 2 LI ET AL. 437 Fig. 1. 1961 2000 mean July August precipitation amount over China, units: 100 mm. (a) observation, (b) multi-model ensemble mean (MME), (c) difference between MME and observation, (d) (x) individual model simulation; Light (dark) shading indicate areas where the value >300 (500) mm. semblance between the precipitation patterns from a particular model and observation. A better simulation result would be that the correlation is close to the horizontal axis and the ratio is near 1.0. The Taylor diagram shows that most of the pattern correlations between the specific model simulation and the observation are between 0.3 and 0.6. All are statistically significant at the 5% level. This indicates that the coupled climate models had reasonable performances in simulating the spatial distribution of precipitation. The ratios of standard deviation of simulated PRCP- TOT in July August calculated from the spatial structure relative to that of observation are generally >1.0, which indicates that the simulated spatial variation is larger than observation. In contrast, all the ratios for R95p in July August are <1.0, which indicates that

438 PROJECTION OF EXTREME PRECIPITATION CHANGE OVER CHINA: PRECIPITATION VOL. 28 Fig. 2. As Fig. 1, but for July August extreme precipitation amount (units: mm). Light (dark) shading indicate areas where the value is >30 (60) mm. the simulated spatial variation of extreme precipitation amount is smaller than the observation. This is partly related to the underestimation of extreme precipitation in the coupled models. 4. Projected future changes Although discrepancies exist between the observations and the simulations, the evaluation of model simulations shows that coupled models are able to reproduce the broad characteristics of precipitation distribution over contiguous China. In the following section, we examine how future precipitation would change over China in response to CO 2 doubling. The simulated changes of PRCPTOT in July August over China under a CO 2 doubling condition relative to preindustrial control run are shown in Fig. 4. The MME shows an increase of precipitation over most parts of China, especially for northeastern and southwestern China (Fig. 4a). A decrease of precipitation is seen only over small parts of northwestern China and eastern China. The spread among the mod-

NO. 2 LI ET AL. 439 Fig. 3. Taylor diagram of 1961 2000 mean July August total and extreme precipitation amount, the number marks represent the resemblance between the spatial pattern from single model simulation and observation. The radial distance from the origin indicates the standard deviation of each model simulation, normalized by the observed value. The angle from the horizontal axis represents the inverse cosine of the spatial correlation between the simulation of certain model and the observation. The dots represent July August total precipitation amount, and the stars represent extreme precipitation amount. els is 10 50 mm, with the largest value over the Tibetan Plateau, which is consistent with another analysis of A1B scenario that also revealed a spread among the coupled models in projecting future change of precipitation over contiguous China (Li and Zhou, 2010). Compared with the result projected by ECHAM5 with a high resolution of T319 (Feng et al., 2010), the precipitation changes projected by CMIP3 models with lower resolutions in this study showed coarser spatial patterns. The lack of regional features in CMIP3 models is attributed to the limitation of low-resolution models in describing the complex topography over East Asia. The changes of R95p in July August over China under CO 2 doubling relative to preindustrial control run are shown in Fig. 5. The pattern of MME is similar to that of PRCPTOT but more homogeneous. For each individual model, the July August total and extreme precipitation change patterns are also similar, with the latter being more homogeneous. Relative to total precipitation changes, the area of extreme precipitation increase over China is larger. The spread among models in the extreme precipitation change (5 29 mm) is smaller than that in the total precipitation (10 50 mm). The ratio between the absolute value of extreme precipitation change and model spread is larger than that of total precipitation, indicating the relatively robust change of extreme precipitation under CO 2 doubling. As the total precipitation change, the extreme precipitation change projected by CMIP3 models also lacks regional features, in comparison with the result projected by a high-resolution model ECHAM5 (Feng et al., 2010). The change of extreme precipitation is determined by both the extreme precipitation intensity and the extreme precipitation frequency. Further analysis shows that the changes of R95pF in July August (Fig. 6) are similar to that of R95p, which indicates that extreme precipitation frequency over China tends to increase under CO 2 doubling relative to the preindustrial control run. The MME result shows that the R95pF would increase by >1% over southern China, with the largest increase over the Tibetan Plateau. The frequency of precipitation is a major factor

440 PROJECTION OF EXTREME PRECIPITATION CHANGE OVER CHINA: PRECIPITATION VOL. 28 Fig. 4. Simulated changes of July August precipitation amount over China under CO 2 doubling relative to preindustrial control run (units: mm): (a) MME, (b) (p) individual model simulation. Light (dark) shading indicate areas where the value is smaller (larger) than 20 (20) mm. worth considering in hydrologic cycle changes. If the precipitation rates increase faster than precipitation amount, the frequency of rainfall occurrence could decrease (Trenberth, 1998). We also examined the change of mean precipitation frequency, and the result exhibits a decreasing tendency over the entire contiguous China (figure not shown here). Since both extreme precipitation intensity and extreme precipitation frequency exhibit an increasing tendency under the CO 2 doubling scenario, the total precipitation PRCPTOT also tends to increase. The decreasing tendency of mean precipitation frequency should be dominated by rainfall events with intensity less than extreme precipitation. This analysis shows that the change of extreme precipitation generally follows that of total precipitation in spatial pattern, but has more homogeneous spatial features. How about the percentage contribution of extreme precipitation amount to total precipitation? As shown in Fig. 7, although there is a large spread among models, the MME (Fig. 7a) shows that the R95pT tends to increase after CO 2 doubling. Given the stronger impact of extreme precipitation on societal activity than mean precipitation, more attention should be paid to the projected increase of R95pT. To further investigate the regional features of precipitation change, we divide contiguous China domain into five subregions, as shown in Fig. 8f. Region 1 represents Northwest China (35 50 N, 80 100 E); region 2 represents the Tibetan Plateau (28 35 N, 80 100 E); region 3 represents Northeast China (43 54 N, 117.5 130 E); region 4 represents North China (35 43 N, 100 122.5 E); and region 5 represents Southeast China (22.5 35 N, 100 122.5 E). For each subregion, we calculate the area-averaged precipitation change and then count the number of models with different ranges in total and extreme precipitation changes (Fig. 8). The CMIP3 models are highly consistent in quantitatively projecting the R95pF change over all five subregions. The spread in quantitatively projecting the PRCPTOT and R95p changes is generally larger, especially in Northwest China, North

NO. 2 LI ET AL. Fig. 5. As Fig. 4, but for extreme precipitation amount. Fig. 6. As Fig. 4, but for extreme precipitation frequency (units: %). Light (dark) shading indicate areas where the value is smaller (larger) than 1% (1%). 441

442 PROJECTION OF EXTREME PRECIPITATION CHANGE OVER CHINA: PRECIPITATION VOL. 28 Fig. 7. As Fig. 4, but for the percentage contribution from extreme precipitation to the total precipitation (units: %). Light (dark) shading indicate areas where the value is smaller (larger) than 5% (5%). China, and Southeast China. In Northwest China (Fig. 8a), 9 of 15 models show a decrease of total precipitation, but most of the models show an increase of extreme precipitation. For the Tibetan Plateau (Fig. 8b), 12 models yield a projection of more total precipitation, and almost all the models show that there would be more extreme precipitation. Only two models (GISS-ER and GFDL- CM2.0) exhibit a negative change of total precipitation amount, and all of the models exhibit a positive change of extreme precipitation over Northeast China (Fig. 8c). Most models project that there would be more total and extreme precipitation over North China (Fig. 8d). For Southeast China (Fig. 8e), 9 of 15 models show that there would be more total precipitation, and all the models project that there would be more extreme precipitation. 5. Possible mechanism for precipitation change The above analyses indicate that the total precipitation change shows an inhomogeneous spatial change pattern; however, the extreme precipitation shows a consistent increase over contiguous China under CO 2 doubling. As the GFDL-CM2.1 model exhibits a similar spatial pattern with the MME, with the spatial pattern correlation coefficient >0.60, and as well provides enough data for our further diagnosis, here we focus on this model to further examine the possible reason for the precipitation change. 5.1 Total precipitation As shown in Fig. 4a, in response to CO 2 doubling, the total precipitation amount in July August would increase over South China, most of Northwest China, and the southern Tibetan Plateau, and it would decrease over most of the middle and lower reaches of Yangtze River. The summer climate of China is dominated by monsoon activity (Zhou et al., 2009c). The monsoon precipitation change is closely related to the changes of water vapor convergence, which is caused by dynamic and thermodynamic processes associated with global warming (Zhou and Yu, 2005; Meehl et al., 2005). To examine the monsoon circulation change in

NO. 2 LI ET AL. 443 Fig. 8. Number of model which simulated total and extreme precipitation changes in different change value bin over China, (a) (e) shows the results from five subregions over China, the locations of sub-regions are shown in (f). The light blue, red, blue, and green bars indicate changes of PRCPTOT, R95p, R95pF and R95pT, respectively. response to CO 2 doubling, we show the change of surface wind at 850 hpa and sea level pressure (SLP) simulated by GFDL-CM2.1 model in Fig. 9a. The most significant change of SLP is the negative center over the Western Pacific Ocean. The SLP change is associated with an anomalous cyclone at 850 hpa. The surface wind change indicates a weakened summer monsoon circulation south to the Yangtze River valley. The anomalous north wind on the western edge of the anomalous cyclone prevents the northward transportation of water vapor from the South China Sea. Following the change of monsoon circulation, a divergence (convergence) of water vapor supply is seen in central East China along the Yangtze River valley (South China) (Fig. 9b). Therefore, the precipitation change over eastern China exhibits a tri-polar anomaly pattern in the GFDL-CAM2.1 model (Fig. 4l). 5.2 Extreme precipitation In contrast to the total precipitation change, the extreme precipitation shows a consistent increase over entire China (Fig. 5l). Previous studies have indicated

444 PROJECTION OF EXTREME PRECIPITATION CHANGE OVER CHINA: PRECIPITATION VOL. 28 air exhibit a consistent increasing pattern over entire China. The surface warming shows northwest southeast-orientated decreasing pattern, and the increase of water vapor content is most evident over Southeast China, due to the favorable water vapor condition there. The spatially consistent changes of warming temperature and water vapor content lead to a uniform increasing pattern of extreme precipitation. Along the middle and lower reaches of the Yangtze River valley, the total precipitation decreases due to the monsoon circulation change, but the extreme precipitation still increases. The decrease in total precipitation may be attributed to the prolonged consecutive dry days or the decrease of wet days there. The magnitude of change in extreme precipitation is entirely predictable based on the degree of surface warming (climate sensitivity) and the increase in moisture content of the atmospheric column in climate models (Allen and Ingram, 2002; Turner and Slingo, Fig. 9. (a) GFDL-CM2.1 model simulated changes of 850 hpa wind (vector, units: m s 1 ) and the sea-level pressure (shaded, units: hpa) in July August under doubled CO 2 concentration relative to preindustrial control run, units: mm. (b) Same as (a), but for the water vapor convergence (units: mm d 1 ). Light (dark) shading indicate areas where the value is smaller (larger) than 30 (30) hpa in (a) and 0.05 (0.05) mm d 1 in (b). that the extreme precipitation change is closely tied to mean precipitation change and that 80% of the spatial pattern of change in the extreme precipitation (90th percentile) can be explained by changes in the mean, for broader South Asian region (Meehl et al., 2005; Turner and Slingo, 2009a). Furthermore, the extreme precipitation change mainly follows the Clausius Clapeyron relationship and is constrained by water vapor content in the air, which is greatly affected by the change in surface temperature (Trenberth, 1999; Allen and Ingram, 2002; Meehl et al., 2005; Turner and Slingo, 2009a). To examine the mechanism responsible for the extreme precipitation change over contiguous China, we analyze changes of surface temperature and integrated water vapor content under CO 2 doubling (Fig. 10). Both the surface temperature and the water vapor content in the Fig. 10. (a) GFDL-CM2.1 model simulated changes of the surface temperature (units: K) in July August under doubled CO 2 concentration relative to preindustrial control run, shading area highlights the spatial distribution of the surface temperature. (b) Same as (a), but for integrated water vapor content (units: kg kg 1 ).

NO. 2 LI ET AL. 445 2009a, b). The extent to which the change in extreme precipitation can be predicted by the model being close to the predicted result based on Clausius Clapeyron relationship depends on the choice of convective parameterization in climate models (Turner and Slingo, 2009a). For the model with bulk mass-flux schemes, the surface temperature change provides a strong constraint, and hence the measured extreme precipitation change is closer to the theoretical result. The GFDL-CM2.1 model employs the relaxed Arakawa Schubert scheme, which parameterizes deep convection with plume ensembles able to initiate from several levels in the vertical direction. Hence the extreme precipitation change is not so constrained by changes in surface conditions, and the predicted extreme precipitation change might be higher than the theoretical result predicted by Clausius Clapeyron relationship. In summary, the extreme precipitation change under CO 2 doubling is mainly constrained by the moisture-holding capacity of the atmospheric column and local surface warming. The consistent increase of surface temperature and water vapor content contribute to the larger increase of extreme precipitation over contiguous China. Whereas, the mean precipitation change over East Asian monsoon areas mainly depends on the changes of water-vapor convergence, which is greatly affected by the advection and dynamical feedbacks related to the changes of monsoon circulation. The decrease of sea-level pressure over the western Pacific Ocean and the associated anomalous low level of cyclones lead to a deficient monsoon rainfall along the central East China under CO 2 doubling. That is, for East Asian monsoon area, the mean precipitation in July August is dominated by dynamic effects associated with monsoon circulation change, while the extreme precipitation change is dominated by thermodynamic effects and follows the Clausius Clapeyron relationship. This is different from that over South Asia monsoon area, where the mean precipitation changes in global warming scenarios primarily occur from thermodynamic effects, but dynamic effects have a small influence (Meehl and Arblaster, 2003). 6. Conclusion Using daily precipitation data from the 20C3M simulations of CMIP3, we evaluate the performances of several current state-of-the-art coupled climate models in simulating the precipitation extremes over China during July August. Potential changes in precipitation extremes over China under the CO 2 doubling scenario are analyzed with the output of 1pctto2x simulations. Both the 20C3M and 1ptto2x experiments were coordinated by the WCRP CMIP3 for IPCC AR4. The main results are summarized below: (1) Although there is a large spread of simulations among the models used, the majority fairly reproduce the observed July August precipitation characteristics over China. Almost all the spatial-pattern correlation coefficients of mean total and extreme precipitation in July August between the simulations and observations are between 0.3 and 0.6; all are statistically significant at the 5% level. For individual models, the indices for means and extremes have similar spatial distributions. However, they generally underestimate the extreme precipitation amount by 50%. Models with higher horizontal resolutions generally have better performance than those employing low or medium resolutions, especially for extreme precipitation simulation. (2) Analyses on potential changes of extreme precipitation over contiguous China under the CO 2 doubling scenario reveal a spread of 5 29 mm among the models. The multi-model ensemble mean shows that both the total precipitation and extreme precipitation over most part of China would increase by >20 mm in July August. The consistency of CMIP3 models in projecting extreme precipitation change is higher than that of total precipitation amount. The change of extreme precipitation is more homogeneous than the total precipitation. A uniform increase of extreme precipitation is seen in most models. (3) The projected changes of precipitation extremes across five sub-regions are documented. For each specific region, there would be more total and extreme precipitation under CO 2 doubling relative to preindustrial control run. The consistency of CMIP3 models in projecting the R95pF change is higher than that of PRCPTOT and R95p. (4) The changes in extreme and mean precipitation share many common features in spatial pattern, albeit with more homogeneous features in extreme precipitation change. However, the mechanisms are different. The mean precipitation change, with an increase over South China but a decrease along most of the middle and lower reaches of Yangtze River valley, is dominated by change of water vapor convergence, which is induced by the monsoon circulation change. This is different from the South Asia monsoon area, where the mean precipitation changes in global warming scenarios primarily occur from thermodynamic effects and dynamic effects have a little influence. For extreme precipitation, however, the consistent warming and increase of water vapor content lead to a homogeneous increase of extreme precipitation over contiguous China. Our study demonstrates that the current state-of-

446 PROJECTION OF EXTREME PRECIPITATION CHANGE OVER CHINA: PRECIPITATION VOL. 28 the-art global coupled climate models have acceptable levels of performance in simulating the largescale precipitation systems over East Asia; however, there are limitations in simulating the regional-scale climate variations. This is consistent with Zhou and Yu (2006), whose study revealed the uncertainties of global coupled models in reproducing the past regional features of surface climate over eastern China. The limitation of CMIP3 global models implies that regional climate models may be useful tools for accurately projecting regional climate changes (Gao et al., 2006, 2008). Since the boundary conditions of regional models are from global models, the large spread among the global models shown here indicate the need of multi-model ensemble in regional modeling, which uses an identical regional climate model but different boundary forcings from global coupled models. In addition, increasing the resolution of global climate model should be an effective method for improving climate simulation and projection in regions with complex topography such as East Asia. For example, by using a high-resolution atmospheric global model ECHAM5 (T319), Feng et al. (2010) found that the simulation of both present mean and extreme precipitation has been greatly improved in terms of spatial pattern and magnitude. In addition, the projection of future precipitation change with the high-resolution model also provides many regional signals that are useful for hydrologic cycle management and research. Acknowledgements. This work is founded by National Key Technologies R&D Program under Grant No. 2007BAC29B03, R&D Special Fund for Public Welfare Industry (meteorology) (GYHY200806010), and China- UK-Swiss Adapting to Climate Change in China Project (ACCC) Climate Science. REFERENCES Alexander, L., and Coauthors, 2006: Global observed changes in daily climate extremes of temperature and precipitation. J. Geophys. Res., 111, 1 22. Allen, M., and W. Ingram, 2002: Constraints on future changes in climate and the hydrologic cycle. Nature, 419, 224 232, doi: 10.1038/nature01092. Chen, H., T. Zhou, R. Neale, X. Wu, and G. Zhang, 2010: Performance of the new NCAR CAM3.5 in East Asian summer monsoon simulations: Sensitivity to modifications of the convection scheme. J. Climate, 23, 3657 3675. Dai, A., 2006: Precipitation characteristics in eighteen coupled climate models. J. Climate, 19, 4605 4630. Dai, A., and K. Trenberth, 2004: The diurnal cycle and its depiction in the community climate system model. J. Climate, 17, 930 951. Emori, S., and S. Brown, 2005: Dynamic and thermodynamic changes in mean and extreme precipitation under changed climate. Geophys. Res. Lett., 32, doi: 10.1029/2005GL023272. Frich, P., L. Alexander, P. Della-Marta, B. Gleason, M. Haylock, A. M. G. K. Tank, and T. Peterson, 2002: Observed coherent changes in climatic extremes during the second half of the twentieth century. Climate Res., 19, 193 212. Feng, L., T. Zhou, B. Wu, T. Li, and J. Luo, 2010: Projection of future precipitation change over China with a high-resolution global atmospheric model. Adv. Atmos. Sci., doi: 10.1007/s00376-010-1016-x. Gao, X., J. Pal, and F. Giorgi, 2006: Projected changes in mean and extreme precipitation over the Mediterranean region from a high resolution double nested RCM simulation. Geophys. Res. Lett., 33, L03706, doi: 10.1029/2005GL024954. Gao, X., Z. Zhao, Y. Ding, R. Huang, and F. Giorgi, 2001: Climate change due to greenhouse effects in China as simulated by a regional climate model. Adv. Atmos. Sci., 18, 1225 1230. Gao, X., Z. Zhao, and F. Giorgi, 2002: Changes of extreme events in regional climate simulations over East Asia. Adv. Atmos. Sci., 19, 927 942. Gao, X., Y. Shi, R. Song, F. Giorgi, Y. Wang, and D. Zhang, 2008: Reduction of future mosoon precipitation over China: Comparison between a high resolution RCM simulation and the driving GCM. Meteor. Atmos. Phys., 100, 73 86. Kharin, V., and F. Zwiers, 2000: Changes in the extremes in an ensemble of transient climate simulations with a coupled atmosphere-ocean GCM. J. Climate, 13, 3760 3788. Kharin, V., and F. Zwiers, 2005: Estimating extremes in transient climate change simulations. J. Climate, 18, 1156 1173. Kimoto, M., 2005: Simulated change of the East Asian circulation under global warming scenario. Geophys. Res. Lett., 32, L16701, doi: 10.1029/2005GL023383. Li, B., and T. Zhou, 2010: The potential climate change over China under IPCC A1B scenario: Multi-model ensemble and uncertainties. Advances in Climate Change Studies, 6, 270 276. (in Chinese) Li, H., T. Zhou, and R. Yu, 2008: Analysis of July August daily precipitation characteristics change over East China during 1958 2000. Chinese J. Atmos. Sci., 32, 358 370. (in Chinese) Li, H., A. Dai, T. Zhou, and J. Lu, 2010: Responses of East Asian summer monsoon to historical SST and atmospheric forcing during 1950 2000. Climate Dyn., 34, 501 514, doi: 10.1007/s00382-008-0482-7. Meehl, G., and J. Arblaster, 2003: Mechanisms for projected future changes in south Asian monsoon precipitation. Climate Dyn., 21, 659 675, doi 10.1007/s00382-003-0343-3. Meehl, G., J. Arblaster, and C. Tebaldi, 2005: Understanding future patterns of increased precipitation intensity in climate model simulations. Geophys. Res. Lett., 32, doi: 10.1029/2005GL023680.

NO. 2 LI ET AL. 447 Min, S., S. Legutke, A. Hense, U. Cubasch, W. Kwon, J. Oh, and U. Schlese, 2006: East Asian climate change in the 21st century as simulated by the coupled climate model ECHO-G under IPCC SRES scenarios. J. Meteor. Soc. Japan, 84, 1 26. Semenov, V., and L. Bengtsson, 2002: Secular trends in daily precipitation characteristics: Greenhouse gas simulation with a coupled AOGCM. Climate Dyn., 19, 123 140. Sun, Y., S. Solomon, A. Dai, and R. Portmann, 2006: How often does it rain? J. Climate, 19, 916 934. Taylor, K., 2001: Summarizing multiple aspects of model performance in a single diagram. J. Geophys. Res., 106, 7183 7192. Tebaldi, C., K. Hayhoe, J. Arblaster, and G. Meehl, 2006: Going to the extremes: An intercomparison of modelsimulated historical and future changes in extreme events. Climatic Change, 79, 185 211. Trenberth, K., 1998: Atmospheric moisture residence times and cycling: Implications for rainfall rates and climate change. Climatic Change, 39, 667 694. Trenberth, K., 1999: Conceptual framework for changes of extremes of the hydrological cycle with climate change. Climatic Change, 42, 327 339. Turner, A., and J. Slingo, 2009a: Uncertainties in future projections of extreme precipitation in the Indian monsoon region. Atmos. Sci. Letts., 10, 152 158. Turner, A., and J. Slingo, 2009b. Subseasonal extremes of precipitation and active-break cycles of the Indian summer monsoon in a climate change scenario. Quart. J. Roy. Meteor. Soc., 135, 549 567, doi: 10.1002/qj.401. Wang, B., J. Liu, J. Yang, T. Zhou, and Z. Wu, 2009: Distinct principal modes of early and late summer rainfall anomalies in East Asia. J. Climate, 22, 3864 3875. Wilby, R., and T. Wigley, 2002: Future changes in the distribution of daily precipitation totals across North America. Geophys. Res. Lett., 29, 1135 1139, doi: 10.1029/2001GL013048. Xu, Q., 2001: Abrupt change of the mid-summer climate in central east China by the influence of atmospheric pollution. Atmos. Environ., 35, 5029 5040. Xu, Y., Y. Zhang, E. Lin, W. Lin, W. Dong, R. Jones, D. Hassell, and S. Wilson, 2006: Analyses on the climate change responses over China under SRES B2 scenario using PRECIS. Chinese Science Bulletin, 51, 2260 2267. (in Chinese) Xu, Y., C. Xu, X. Gao, and Y. Luo, 2009: Projected changes in temperature and precipitation extremes over the Yangtze river basin of China in the 21st century. Quaternary International, 208, 44 52. Yu, R., and T. Zhou, 2007: Seasonality and threedimensional structure of the interdecadal change in East Asian monsoon. J. Climate, 20, 5344 5355. Yu, R., B. Wang, and T. Zhou, 2004: Tropospheric cooling and summer monsoon weakening trend over East Asia. Geophys. Res. Lett., 31, L22212, doi: 10.1029/2004GL021270. Zhai, P., X. Zhang, H. Wan, and X. Pan, 2005: Trends in total precipitation and frequency of daily precipitation extremes over China. J. Climate, 18, 1096 1108. Zhang, Y., Y. Xu, W. Dong, L. Cao, and M. Sparrow, 2006: A future climate scenario of regional changes in extreme climate events over China using the PRE- CIS climate model. Geophys. Res. Lett., 33, 1 6. Zhou, T., and Z. Li, 2002: Simulation of the East Asian summer monsoon by using a variable resolution atmospheric GCM. Climate Dyn., 19, 167 180. Zhou, T., and R. Yu, 2005: Atmospheric water vapor transport associated with typical anomalous summer rainfall patterns in China. J. Geophys. Res., 110, D08104, doi: 10.1029/2004JD005413. Zhou, T., and R. Yu, 2006: Twentieth-century surface air temperature over China and the globe simulated by coupled climate models. J. Climate, 19, 5843 5858. Zhou, T., B. Wu, and B. Wang, 2009a: How well do atmospheric general circulation models capture the leading modes of the interannual variability of Asian Australian monsoon? J. Climate., 22, 1159 1173. Zhou, T., and Coauthors, 2009b: The CLIVAR C20C Project: Which components of the Asian-Australian monsoon circulation variations are forced and reproducible? Climate. Dyn., 33, 1051 1068. Zhou, T., D. Gong, J. Li, and B. Li, 2009c: Detecting and understanding the multi-decadal variability of the East Asian Summer Monsoon Recent progress and state of affairs. Meteorologische Zeitschrift, 18 (4), 455 467.