Downscaling and projection of precipitation from general circulation model predictors in an equatorial climate region by the automated regression-based statistical method

The authors have applied an automated regression-based statistical method, namely, the automated statistical downscaling (ASD) model, to downscale and project the precipitation climatology in an equatorial climate region (Peninsular Malaysia). Five precipitation indices are, principally, downscaled and projected: mean monthly values of precipitation (Mean), standard deviation (STD), 90th percentile of rain day amount, percentage of wet days (Wet-day), and maximum number of consecutive dry days (CDD). The predictors, National Centers for Environmental Prediction (NCEP) products, are taken from the daily series reanalysis data, while the global climate model (GCM) outputs are from the Hadley Centre Coupled Model, version 3 (HadCM3) in A2/B2 emission scenarios and Third-Generation Coupled Global Climate Model (CGCM3) in A2 emission scenario. Meanwhile, the predictand data are taken from the arithmetically averaged rain gauge information and used as a baseline data for the evaluation. The results reveal, from the calibration and validation periods spanning a period of 40 years (1961–2000), the ASD model is capable to downscale the precipitation with reasonable accuracy. Overall, during the validation period, the model simulations with the NCEP predictors produce mean monthly precipitation of 6.18–6.20 mm/day (root mean squared error 0.78 and 0.82 mm/day), interpolated, respectively, on HadCM3 and CGCM3 grids, in contrast to 6.00 mm/day as observation. Nevertheless, the model suffers to perform reasonably well at the time of extreme precipitation and summer time, more specifically to generate the CDD and STD indices. The future projections of precipitation (2011–2099) exhibit that there would be an increase in the precipitation amount and frequency in most of the months. Taking the 1961–2000 timeline as the base period, overall, the annual mean precipitation would indicate a surplus projection by nearly 14~18 % under both GCM output cases (HadCM3 A2/B2 scenarios and CGCM3 A2 scenario). According to the model simulation, the September–November periods might be the more significant months projecting the increment of the precipitation amount around over 50 %, while the precipitation deficit would be seen in March–May periods.     

Non-stationary extreme value analysis in a changing climate

This paper introduces a framework for estimating stationary and non-stationary return levels, return periods, and risks of climatic extremes using Bayesian inference. This framework is implemented in the Non-stationary Extreme Value Analysis (NEVA) software package, explicitly designed to facilitate analysis of extremes in the geosciences. In a Bayesian approach, NEVA estimates the extreme value parameters with a Differential Evolution Markov Chain (DE-MC) approach for global optimization over the parameter space. NEVA includes posterior probability intervals (uncertainty bounds) of estimated return levels through Bayesian inference, with its inherent advantages in uncertainty quantification. The software presents the results of non-stationary extreme value analysis using various exceedance probability methods. We evaluate both stationary and non-stationary components of the package for a case study consisting of annual temperature maxima for a gridded global temperature dataset. The results show that NEVA can reliably describe extremes and their return levels.     

西南地区极端降水变化趋势

利用西南地区90个气象台站1970-2010年逐日降水量资料,依据世界气象组织(WMO)定义的连续5d最大降水量、总降水量、强降水比等6种极端降水指数,采用F检验、11a滑动平均等统计方法,研究了西南地区极端强降水变化趋势的时空变化特征。在时间上,西南地区近41年来冬、春、夏季连续5d最大降水量缓慢波动上升,秋季连续5d最大降水量呈下降趋势;强降水、降水强度及强降水比呈上升趋势,但总降水量和最长持续无降水日数呈减少趋势;另外,各极端降水指数还存在明显的年际、年代际变化。在空间上,西南地区极端降水变化趋势具有显著的地域差异,呈东西或西北东南向梯度变化特征。其中冬季连续5d最大降水量、降水强度、强降水比及最长持续无降水日数,在西南大部分地区呈增加趋势。秋季连续5d最大降水量与总降水量在西南大部分地区呈减少趋势。而春、夏季连续5d最大降水量和强降水的增减区域大致相当。     

7个IPCC AR4模式对中国地区极端降水指数模拟能力的评估及其未来情景预估

利用中国区域550个站点1961～2000年日降水量资料, 考察参与政府间气候变化委员会(IPCC)第四次评估报告的7个新一代全球模式及多模式集合对现代气候情景下(20C3M)5个极端降水指数的模拟能力, 同时进行中国区域未来不同排放情形下极端降水事件变化的预估, 结果表明: 最新全球模式能较好地模拟出极端降水指数气候场的空间分布及其中国区域的线性趋势, 且模式集合模拟能力优于大部分单个模式, 但在青藏高原东侧、 高原南部存在虚假的极端降水高值区, 模拟的东部季风区的极端降水强度系统性偏低, 区域平均序列年际变率的模拟能力也较低。中国地区21世纪与降水有关的事件都有趋于极端化的趋势, 极端降水强度可能增强, 干旱也将加重, 且变化幅度与排放强度成正比。     

IAP AGCM 4.1对淮河流域夏季降水的预报技巧评估

基于中国科学院大气物理研究所新一代大气环流模式IAP AGCM 4.1共30 a(1981—2010年)的集合回报试验结果,评估了模式对淮河流域夏季降水的预报技巧。分析结果表明,模式总体上可以较好地再现出淮河流域夏季平均降水南多北少的空间分布特征,其中模式模拟的6月降水量与观测值的空间相关可达0.93。但降水强度与观测相比具有系统性的偏差,且模式模拟的降水年际变率显著偏弱。基于降水距平相关系数的确定性预报技巧分析表明,模式对流域西南部夏季降水的预测技巧较高,达到0.2以上,且模式对6月降水异常的预测能力相对最好,7月次之。针对淮河不同子流域的预报技巧分析表明,IAP AGCM 4. 1对蚌埠、鲁台子、王家坝水文控制站以上集水面积的夏季面雨量异常具有一定的预报技巧,30 a集合回报的时间相关系数分别为0. 11、0. 13、0. 16。基于降水等级的概率预报技巧评估表明,模式对7月淮河流域南部少雨事件具有很好的预报能力,同时对6月流域中部多雨事件的预报技巧也较高。     

RCPs情景下汉江流域未来极端降水的模拟与预估

采用应用于跨行业影响模式比较计划(ISIMIP)的5个CMIP5全球气候模式模拟的历史和未来RCP排放情景下的逐日降水数据,在评估模式对汉江流域1961—2005年极端降水变化特征模拟能力的基础上,进一步计算了RCP2.6、RCP4.5和RCP8.5排放情景下汉江流域未来2016—2060年极端降水总量(R95p)、极端降水贡献率(PEP)、连续5 d最大降水(RX5d)和降水强度(SDII),结果表明:RCP4.5情景下的极端降水指数上升最明显,R95p和RX5d分别较基准期增加12.5%和8.2%,PEP增加3.2个百分点,SDII微弱上升。在不同排放情景下,PEP均有一定的增幅,以流域西北和东南部增幅较大;R95p在流域绝大部分区域表现出一定的增加,且流域东南部和北部是增幅高值区;RX5d在RCP2.6和RCP4.5情景下整体表现为增加的特征,但在RCP8.5情景下整体表现为减少的特征。对极端降水预估的不确定性中,SDII的不确定性最小,RX5d的不确定性最大;不确定性大值区主要位于流域东部、东南部和西北部部分区域。     

Impact of projected climate change on hydrologic regime of the Upper Paraguay River basin

We present an assessment of climate change impacts on the hydrologic regime of the 600,000 km² Upper Paraguay River basin, located in central South America based on predictions of 20 Atmospheric/Ocean General Circulation Models (AOGCMs). We considered two climate change scenarios from the Intergovernmental Panel on Climate Change (IPCC) and two 30-years time intervals centered at 2030 and 2070. Projected temperature and precipitation anomalies estimated by the AOGCMs for the study site are spatially downscaled. Time series of projected temperature and precipitation were estimated using the delta change approach. These time series were used as input to a detailed coupled hydrologic-hydraulic model aiming to estimate projected streamflow in climate change scenarios at several control points in the basin. Results show that impacts on streamflow are highly dependent on the AOGCM used to obtain the climate predictions. Patterns of temperature increase persist over the entire year for almost all AOGCMs resulting in an increase in the evapotranspiration rate of the hydrological model. The precipitation anomalies show large dispersion, being projected as either an increase or decrease in precipitation rates. Based on these inputs, results from the coupled hydrologic-hydraulic model show nearly one half of projections as increasing river discharge, and other half as decreasing river discharge. If the mean or median of the predictions is considered, no discernible change in river discharge should be expected, despite the dispersion among results of the AOGCMs that reached +/?10 % in the short horizon and +/? 20 % in the long horizon, at several control points.     

未来气候变化对西南地区地质灾害的可能影响

本研究利用区域气候模式RegCM4提供的RCP8.5(高排放)和RCP4.5(中排放)情景下的逐日平均气温和降水量,计算并分析了西南地区21世纪不同阶段平均气温、平均降水、连续干旱日数(CDD)、> 20mm的降水日数(R20mm)、连续5天最大降水量(Rx5day)和单日降水强度指数(SDII)相对于参照期(1986～2005年)的变化特征,进而定性地给出未来气候变化对西南地区地质灾害的可能影响。结果表明:未来西南地区因平均气温升高、平均降水量变化、持续干旱变化、强降水变化、降水集中程度变化和单日降水强度变化将导致地质灾害发生的可能性增大,但诱因不同、影响区域有差异;另外温室气体浓度越高,平均气温、平均降水以及相关极端指数相对于基准期变幅基本上都越大,相关地质灾害风险增加的可能性也越大。      

Probability Distribution of Precipitation Extremes over the Yangtze River Basin

 Based on the daily observational precipitation data of 147 stations in the Yangtze River basin for 1960-2005, and the projected daily data of 79 grids from ECHAM5/MPI-OM in the 20th century, time series of precipitation extremes which contain annual maximum (AM) and Munger index (MI) were constructed. The distribution feature of precipitation extremes was analyzed based on the two index series. Research results show that (1) the intensity and probability of extreme heavy precipitation are higher in the middle Mintuo River sub-catchment, the Dongting Lake area, the mid-lower main stream section of the Yangtze River, and the southeastern Poyang Lake sub-catchment; whereas, the intensity and probability of drought events are higher in the mid-lower Jinsha River sub-catchment and the Jialing River sub-catchment; (2) compared with observational data, the averaged value of AM is higher but the deviation coefficient is lower in projected data, and the center of precipitation extremes moves northwards; (3) in spite of certain differences in the spatial distributions of observed and projected precipitation extremes, by applying General Extreme Value (GEV) and Wakeby (WAK) models with the method of L-Moment Estimator (LME) to the precipitation extremes, it is proved that WAK can simulate the probability distribution of precipitation extremes calculated from both observed and projected data quite well. The WAK could be an important function for estimating the precipitation extreme events in the Yangtze River basin under future climatic scenarios.     

1960-2005年长江流域降水极值概率分布特征

Based on the daily observational precipitation data of 147 stations in the Yangtze River basin for 1960-2005,and the projected daily data of 79 grids from ECHAM5/MPI-OM in the 20th century,time series of precipitation extremes which contain annual maximum(AM)and Munger index(MI)were constructed.The distribution feature of precipitation extremes was analyzed based on the two index series.Research results show that(1)the intensity and probability of extreme heavy precipitation are higher in the middle Mintuo River sub-catchment,the Dongting Lake area,the mid-lower main stream section of the Yangtze River,and the southeastern Poyang Lake sub-catchment;whereas,the intensity and probability of drought events are higher in the mid-lower Jinsha River sub-catchment and the Jialing River sub-catchment;(2)compared with observational data,the averaged value of AM is higher but the deviation coefficient is lower in projected data,and the center of precipitation extremes moves northwards;(3)in spite of certain differences in the spatial distributions of observed and projected precipitation extremes,by applying General Extreme Value(GEV)and Wakeby(WAK)models with the method of L-Moment Estimator(LME)to the precipitation extremes,it is proved that WAK can simulate the probability distribution of precipitation extremes calculated from both observed and projected data quite well.The WAK could be an important function for estimating the precipitation extreme events in the Yangtze River basin under future climatic scenarios.     

Projections of temperature and precipitation extremes in the North Western Mediterranean Basin by dynamical downscaling of climate scenarios at high resolution (1971–2050)

The North Western Mediterranean basin (NWMB) is characterised by a highly complex topography and an important variability of temperature and precipitation patterns. Downscaling techniques are required to capture these features, identify the most vulnerable areas to extreme changes and help decision makers to design strategies of mitigation and adaptation to climate change. A Regional Climate Model, WRF-ARW, is used to downscale the IPCC-AR4 ECHAM5/MPI-OM General Circulation Model results with high resolution (10 km), considering three different emissions scenarios (B1, A1B and A2) for 2001–2050. Model skills to reproduce observed extremes are assessed for a control period, 1971–2000, using the ERA40 reanalysis to drive the WRF-ARW simulations. A representative set of indices for temperature and precipitation extremes is projected. The modelling system correctly reproduces amplitude and frequency of extremes and provides a high degree of detail on variability over neighbouring areas. However, it tends to overestimate the persistence of wet events and consequently slightly underestimate the length of dry periods. Drier and hotter conditions are generally projected for the NWMB, with significant increases in the duration of droughts and the occurrence of heavy precipitation events. The projected increase in the number of tropical nights and extreme temperatures could have a negative effect on human health and comfort conditions. Simulations allow defining specifically vulnerable areas, such as the Ebro Valley or the Pyrenees, and foreseeing impacts on socio-economic activities in the region.     

Simulation of extreme precipitation indices in the Yangtze River basin by using statistical downscaling method (SDSM)

In this study, the applicability of the statistical downscaling model (SDSM) in modeling five extreme precipitation indices including R10 (no. of days with precipitation ≥10?mm?day^?1), SDI (simple daily intensity), CDD (maximum number of consecutive dry days), R1d (maximum 1-day precipitation total) and R5d (maximum 5-day precipitation total) in the Yangtze River basin, China was investigated. The investigation mainly includes the calibration and validation of SDSM model on downscaling daily precipitation, the validation of modeling extreme precipitation indices using independent period of the NCEP reanalysis data, and the projection of future regional scenarios of extreme precipitation indices. The results showed that: (1) there existed good relationship between the observed and simulated extreme precipitation indices during validation period of 1991–2000, the amount and the change pattern of extreme precipitation indices could be reasonably simulated by SDSM. (2) Under both scenarios A2 and B2, during the projection period of 2010–2099, the changes of annual mean extreme precipitation indices in the Yangtze River basin would be not obvious in 2020s; while slightly increase in the 2050s; and significant increase in the 2080s as compared to the mean values of the base period. The summer might be the more distinct season with more projected increase of each extreme precipitation indices than in other seasons. And (3) there would be distinctive spatial distribution differences for the change of annual mean extreme precipitation indices in the river basin, but the most of Yangtze River basin would be dominated by the increasing trend.     

6.25 km高分辨率降尺度数据对雄安新区及整个京津冀地区未来极端气候事件的预估

基于RCP4.5情景下6.25 km高分辨率统计降尺度数据,使用国际上通用的极端气候事件指数,分析雄安新区及整个京津冀地区未来极端气候事件的可能变化。首先对当代模拟结果进行评估,结果表明,集合平均模拟可以较好地再现大部分极端气候事件指数的分布,且对与气温有关的极端气候事件指数模拟效果较好。但也存在一定偏差,特别是对连续干旱日数(CDD)的模拟效果相对较差。集合平均的预估结果表明,未来在全球变暖背景下,雄安新区及整个京津冀地区均表现为极端暖事件增多,极端冷事件减少,连续干旱日数减少,极端强降水事件增多。具体来看,到21世纪末期,日最高气温最高值(TXx)和日最低气温最低值(TNn)在整个区域上都是增加的,大部分地区增加值分别超过2.4℃和3.2℃;夏季日数(SU)和热带夜数(TR)也都表现为增加,但两者的变化分布基本相反,其中SU在山区增加幅度较大,平原地区增加幅度较小,而TR在平原地区的增加值较山区更显著,两个指数未来增加值分别为20~40 d和5~40 d;霜冻日数(FD)和冰冻日数(ID)都表现为减少,减少值分别超过10 d和5 d;与降水有关的极端气候事件指数,CDD、降雨日数(R_1mm)和中雨日数(R_10mm)的变化均以减少为主,但数值较小,一般都在?10%~0之间;最大5 d降水量(R_X5day)、降水强度(SDII)和大雨日数(R_20mm)主要表现为增加,增加值一般在0~25%之间。从区域平均的变化来看,与气温有关的极端气候事件指数的变化趋势较为显著,与降水有关的极端气候事件指数变化趋势较小。两个区域对比来看,雄安新区模式间的不确定性更大,反映出模式对较小区域模拟的不足。     

NEX-GDDP降尺度数据对中国极端降水指数模拟能力的评估

利用1986—2005年中国地面气象台站观测的格点化逐日降水数据(CN05.1)评估了NASA高分辨率降尺度逐日数据集NEX-GDDP中21个全球气候模式在0.25?(约25 km×25 km)分辨率下对中国极端降水的模拟能力.选取年最大日降水量(RX1D)、年最大5 d降水量(RX5D)、湿日总降水量(PRCPTOT...     

Thirty-two-year ocean–atmosphere coupled downscaling of global reanalysis over the Intra-American Seas

This study examines the oceanic and atmospheric variability over the Intra-American Seas (IAS) from a 32-year integration of a 15-km coupled regional climate model consisting of the Regional Spectral Model (RSM) for the atmosphere and the Regional Ocean Modeling System (ROMS) for the ocean. It is forced at the lateral boundaries by National Centers for Environmental Prediction-Department of Energy (NCEP-DOE R-2) atmospheric global reanalysis and Simplified Ocean Data Assimilation global oceanic reanalysis. This coupled downscaling integration is a free run without any heat flux correction and is referred as the Regional Ocean–Atmosphere coupled downscaling of global Reanalysis over the Intra-American Seas (ROARS). The paper examines the fidelity of ROARS with respect to independent observations that are both satellite based and in situ. In order to provide a perspective on the fidelity of the ROARS simulation, we also compare it with the Climate Forecast System Reanalysis (CFSR), a modern global ocean–atmosphere reanalysis product. Our analysis reveals that ROARS exhibits reasonable climatology and interannual variability over the IAS region, with climatological SST errors less than 1 °C except along the coastlines. The anomaly correlation of the monthly SST and precipitation anomalies in ROARS are well over 0.5 over the Gulf of Mexico, Caribbean Sea, Western Atlantic and Eastern Pacific Oceans. A highlight of the ROARS simulation is its resolution of the loop current and the episodic eddy events off of it. This is rather poorly simulated in the CFSR. This is also reflected in the simulated, albeit, higher variance of the sea surface height in ROARS and the lack of any variability in the sea surface height of the CFSR over the IAS. However the anomaly correlations of the monthly heat content anomalies of ROARS are comparatively lower, especially over the Gulf of Mexico and the Caribbean Sea. This is a result of ROARS exhibiting a bias of underestimation (overestimation) of high (low) clouds. ROARS like CFSR is also able to capture the Caribbean Low Level Jet and its seasonal variability reasonably well.     

CMIP6全球气候模式对中亚极端降水模拟能力的评估

本文基于NOAA再分析逐日降水数据和22个CMIP6模式的降水模拟数据,选取了6个极端降水指数,从气候态和相对变率两个角度对CMIP6模式在中亚地区极端降水方面的模拟能力开展了评估。结果表明,在气候态方面,中亚地区降水的空间分布表现为由西南向东北递增,其东南部山地迎风侧降水偏多;多模式集合对SDII(简单降水强度)和CDD(最大无雨期)模拟的平均误差分别为-5.43%和0.45%,对PRCPTOT(年总降水量)、R1mm(有雨日数)、Rx5day(最大连续五日降水)和CWD(最大雨期)的模拟结果存在明显高估,且在中亚东南部高海拔地区误差偏高。在相对变率方面,多模式集合模拟的中亚极端降水的相对变率偏小,其中对CWD的模拟效果相对较好,平均误差为-4.78%;对R1mm的模拟效果最差,平均误差为-36.16%。模式间进行比较,TaiESM1、EC-Earth3-Veg-LR和GFDL-ESM为22个CMIP6模式中模拟能力最好的前3个模式。     

Precipitation extremes in the Yangtze River Basin,China: regional frequency and spatial–temporal patterns

Regional frequency analysis and spatial–temporal patterns of precipitation extremes are investigated based on daily precipitation data covering 1960–2009 using the index-flood L-moments method together with some advanced statistical tests and spatial analysis techniques. The results indicate that: (1) the entire Yangtze River basin can be divided into six homogeneous regions in terms of extreme daily precipitation index. Goodness-of-fit test indicates that Pearson type III (PE3, three parameters), general extreme-value (GEV, three parameters), and general normal (GNO, three parameters) perform well in fitting regional precipitation extremes; (2) the regional growth curves for each homogeneous region with 99 % error bands show that the quantile estimates are reliable enough and can be used when return periods are less than 100 years, and the results indicate that extreme precipitation events are highly probable to occur in regions V and VI, and hence higher risk of floods and droughts; and (3) spatial patterns of annual extreme daily precipitation with return period of 20 years indicate that precipitation amount increases gradually from the upper to the lower Yangtze River basin, showing higher risks of floods and droughts in the middle and lower Yangtze River basin, and this result is in good agreement with those derived from regional growth curves.     

Variability and trends in daily precipitation extremes on the northern and southern slopes of the central Himalaya

This study focuses on the precipitation extremes recorded on the northern and southern slopes of the central Himalaya, especially those documented at higher altitudes. Daily precipitation data recorded over a 35-year period at nine meteorological stations in the region were studied. We used the precipitation extreme indices delineated by the Expert Team on Climate Change Detection and Indices (ETCCDI). The spatial and temporal variations in these precipitation extremes were calculated. When regional patterns were investigated to detect any anomalies, only 1 of the 10 precipitation extreme indices from the southern slopes of the central Himalaya showed a statistically significant trend; none from the northern slopes of the central Himalaya highlighted a statistically significant trend. On the southern slopes, all indices increased, apart from the maximum 1-day precipitation (RX1) and simple daily precipitation intensity (SDII) indices. Indices such as the consecutive dry days (CDDs) and RX1 indices exhibited similar increases on both the northern and southern slopes of the central Himalaya. These results suggest that increases in precipitation have been accompanied by an increasing frequency of extremes over the southern central Himalaya. Nonetheless, no relation could be established between the precipitation extreme indices and circulation indices for higher altitudes.

     

Simulation of regional climate change under the IPCC A2 scenario in southeast China

A variable-grid atmospheric general circulation model, LMDZ, with a local zoom over southeast China is used to investigate regional climate changes in terms of both means and extremes. Two time slices of 30?years are chosen to represent, respectively, the end of the 20th century and the middle of the 21st century. The lower-boundary conditions (sea-surface temperature and sea-ice extension) are taken from the outputs of three global coupled climate models: Institut Pierre-Simon Laplace (IPSL), Centre National de Recherches Météorologiques (CNRM) and Geophysical Fluid Dynamics Laboratory (GFDL). Results from a two-way nesting system between LMDZ-global and LMDZ-regional are also presented. The evaluation of simulated temperature and precipitation for the current climate shows that LMDZ reproduces generally well the spatial distribution of mean climate and extreme climate events in southeast China, but the model has systematic cold biases in temperature and tends to overestimate the extreme precipitation. The two-way nesting model can reduce the ??cold bias?? to some extent compared to the one-way nesting model. Results with greenhouse gas forcing from the SRES-A2 emission scenario show that there is a significant increase for mean, daily-maximum and minimum temperature in the entire region, associated with a decrease in the number of frost days and an increase in the heat wave duration. The annual frost days are projected to significantly decrease by 12?C19?days while the heat wave duration to increase by about 7?days. A warming environment gives rise to changes in extreme precipitation events. Except two simulations (LMDZ/GFDL and LMDZ/IPSL2) that project a decrease in maximum 5-day precipitation (R5d) for winter, other precipitation extremes are projected to increase over most of southeast China in all seasons, and among the three global scenarios. The domain-averaged values for annual simple daily intensity index (SDII), R5d and fraction of total rainfall from extreme events (R95t) are projected to increase by 6?C7, 10?C13 and 11?C14%, respectively, relative to their present-day values. However, it is clear that more research will be needed to assess the uncertainties on the projection in future of climate extremes at local scale.     

The eastward shift of the Walker Circulation in response to global warming and its relationship to ENSO variability

This study investigates the global warming response of the Walker Circulation and the other zonal circulation cells (represented by the zonal stream function), in CMIP3 and CMIP5 climate models. The changes in the mean state are presented as well as the changes in the modes of variability. The mean zonal circulation weakens in the multi model ensembles nearly everywhere along the equator under both the RCP4.5 and SRES A1B scenarios. Over the Pacific the Walker Circulation also shows a significant eastward shift. These changes in the mean circulation are very similar to the leading mode of interannual variability in the tropical zonal circulation cells, which is dominated by El Niño Southern Oscillation variability. During an El Niño event the circulation weakens and the rising branch over the Maritime Continent shifts to the east in comparison to neutral conditions (vice versa for a La Niña event). Two-thirds of the global warming forced trend of the Walker Circulation can be explained by a long-term trend in this interannual variability pattern, i.e. a shift towards more El Niño-like conditions in the multi-model mean under global warming. Further, interannual variability in the zonal circulation exhibits an asymmetry between El Niño and La Niña events. El Niño anomalies are located more to the east compared with La Niña anomalies. Consistent with this asymmetry we find a shift to the east of the dominant mode of variability of zonal stream function under global warming. All these results vary among the individual models, but the multi model ensembles of CMIP3 and CMIP5 show in nearly all aspects very similar results, which underline the robustness of these results. The observed data (ERA Interim reanalysis) from 1979 to 2012 shows a westward shift and strengthening of the Walker Circulation. This is opposite to what the results in the CMIP models reveal. However, 75 % of the trend of the Walker Circulation can again be explained by a shift of the dominant mode of variability, but here towards more La Niña-like conditions. Thus in both climate change projections and observations the long-term trends of the Walker Circulation seem to follow to a large part the pre-existing dominant mode of internal variability.