Exploring multi-model atmospheric GCM ensembles with ANOVA

Analysis of variance (ANOVA) is a powerful statistical technique for making inferences about experiments that are influenced by multiple factors. Whilst common in many other scientific fields, its use within the climate community has been limited to date. Here we review the basis for ANOVA and how, in particular, it can be applied to partition the variance in a multi-model ensemble of Atmospheric General Circulation Model simulations. We examine an ensemble of four AGCMs forced with observed twentieth century sea surface temperatures (SST). We show that the dominant contributions to the total variance of seasonal mean sea level pressure arise from between-model differences (the bias term) and internal noise (the noise term). However, which term is most important varies from region to region. Of particular interest is the interaction term, which describes differences between the models in their responses to common SST forcing. The interaction term is found to be largest over the Indian Ocean (in all seasons), and over the subtropical Northwest Pacific in boreal summer. The differences between the model responses in these regions suggest differences in their simulation of atmospheric teleconnections, with potentially important implications, e.g. for seasonal predictions of the South and East Asian Monsoons. Examination of these differences may lead to an understanding of the reasons why models respond differently to common forcing, and ultimately to improvements in the performance of climate models.     

大气环流模式与春玉米生长模拟模式联接的影响评估模拟试验

根据春玉米田间试验资料和历史气候资料,对春玉米生长模拟模式进行了验证与灵敏性分析,在此基础上,运用逐步订正法将当前气候前景和大气环流模式输出资料结合历史气候资料生成的未来气候情景订正到１ｏ×１ｏ网格点上,与春玉米生长模拟模式相联接,就未来气候变化对我国东北地区春玉米生长、发育和最终产量的可能影响进行了网格化定量模拟,并对一些适应性对策的效果进行了定性或定量的分析。结果表明,在ＤＫＲＺＯＰＹＣ模拟的未来情景下,若保持当前作物品种和生产技术措施不变,研究区域除北部将平均增产７０％外,其余地区都将有不同程度的减产,幅度在－１０％～－５０％之间,而在ＮＣＡＲ模拟的情景下,中西部地区将增产,其它地区可维持当前产量水平。适应性对策将对开发利用未来可能的气候资源,减缓未来气候变化的负效应,充分发挥其正效应起到积极作用,进而绝大部分区域将受益于未来水热条件的改变。     

Evaluation of HadCM2 and Direct Use of Daily GCM Data in Impact Assessment Studies

This paper investigates two important aspects of methods used to explore possible effects of climatic changes on agricultural productivity on regional spatial scales. First, an evaluation of precipitation and near surface air temperature in two successive versions of the Hadley Centre General Circulation Model (GCM) has been performed to consider to what extent GCMs are capable of simulating the mean and variability of local climates. This is explored by comparing the output of an individual GCM grid box with three station observations. Several ancillary issues associated with the comparisons of observations of daily precipitation and model output that affect the statistical results are also discussed. Finally, daily data from the control and sulphate runs of the latest Hadley Centre GCM (HadCM2) have been used directly as input to the CERES-Wheat model, and the modelled yield distribution is compared to that produced with the historical data series. Our results imply that for this particular grid box covering the study region in central France, the daily raw data from HadCM2 experiment can be used directly to assess the potential impact of the greenhouse gas and sulphate aerosol radiative induced forcings and the associated climatic change on average regional winter wheat production. On the other hand, less confidence should be placed on their use regarding the estimation of future agricultural risk and variability assessment. Furthermore, a possibly more severe methodological problem that has arisen from our study is the inability of CERES-Wheat to simulate the waterlogging effects of excessive soil water on crop growth and development. Finally, we assess the potential impact of changing climate on regional winter wheat production by using the daily data from the sulphate integration up to the end of the 21st century.     

Impact of global warming on the interannual and interdecadal climate modes in a coupled GCM

In this study, we investigated the impact of global warming on the variabilities of large-scale interannual and interdecadal climate modes and teleconnection patterns with two long-term integrations of the coupled general circulation model of ECHAM4/OPYC3 at the Max-Planck-Institute for Meteorology, Hamburg. One is the control (CTRL) run with fixed present-day concentrations of greenhouse gases. The other experiment is a simulation of transient greenhouse warming, named GHG run. In the GHG run the averaged geopotential height at 500?hPa is increased significantly, and a negative phase of the Pacific/North American (PNA) teleconnection-like distribution pattern is intensified. The standard deviation over the tropics (high latitudes) is enhanced (reduced) on the interdecadal time scales and reduced (enhanced) on the interannual time scales in the GHG run. Except for an interdecadal mode related to the Southern Oscillation (SO) in the GHG run, the spatial variation patterns are similar for different (interannual?+?interdecadal, interannual, and interdecadal) time scales in the GHG and CTRL runs. Spatial distributions of the teleconnection patterns on the interannual and interdecadal time scales in the GHG run are also similar to those in the CTRL run. But some teleconnection patterns show linear trends and changes of variances and frequencies in the GHG run. Apart from the positive linear trend of the SO, the interdecadal modulation to the El Niño/SO cycle is enhanced during the GHG 2040?～?2099. This is the result of an enhancement of the Walker circulation during that period. La Niña events intensify and El Niño events relatively weaken during the GHG 2070?～?2090. It is interesting to note that with increasing greenhouse gas concentrations the relation between the SO and the PNA pattern is reversed significantly from a negative to a positive correlation on the interdecadal time scales and weakened on the interannual time scales. This suggests that the increase of the greenhouse gas concentrations will trigger the nonstationary correlation between the SO and the PNA pattern both on the interdecadal and interannual time scales.     

Tropical disturbances in a GCM

We have analyzed the tropical disturbances in a 11-layer atmospheric general circulation model (GCM) on a 2.5° × 3.75° horizontal grid coupled to a 50 m-mixed layer ocean. Due to the coarse resolution, the GCM is unable to resolve adequately tropical cyclones. The tropical disturbances simulated by the GCM are much weaker and have a much larger horizontal extent. However, they still display much of the essential physics of tropical cyclones, including low-level convergence of mass and moisture, upper tropospheric outflow and a warm core. For most ocean basins the spatial and temporal distribution of the simulated tropical disturbances compares well with the observed tropical cyclones. On doubling the CO₂ concentration, the number of simulated tropical disturbances increases by about 50%. There is a relative increase in the number of more intense tropical disturbances, whose maximum windspeed increases by about 20%. This agrees with the theoretical estimate of Emanuel. However, because the low-resolution of the GCM severely restricts their maximum possible intensity, simulated changes in tropical disturbance intensity should be interpreted cautiously.     

北京地区冠心病发病率的气象评估模型

基于1984～2000年北京地区自然人群的长期、持续、跟踪监测的旬冠心病发病资料和同期日气象资料,通过分析揭示了冠心病发病率的季节变化规律和年际变化特点;通过发病率和气象因子的相关分析,确定了全年及各季、月可能诱发冠心病发病率的主要气象因子。结果表明,气压、风速、气温和水汽压等因子与发病率关系显著,从各个季节来说,气象因子对发病率的影响又各有侧重并具有不同的表现形式。建立了各代表月冠心病发病率的气象评估模型,历史回代和预测结果表明,所建方程具有较高的评估能力。     

A GCM study of Antarctic glaciation

An atmospheric general circulation model, the NCAR CCM, has been used to investigate the possible effects of two specific tectonic mechanisms on Antarctic glaciation. These two mechanisms are: (1) closing the Drake Passage (connecting South America with Antarctica), which is assumed to effectively represent an increased meridional heat transport by the ocean; and (2) changing the elevation of Antarctica. Perpetual season (summer and winter) and seasonal cycle simulations with warmer sea-surface temperatures and no sea ice prescribed for mid- to high-latitude southern oceans have been made with both present-day (high) Antarctic elevations and with low Antarctic elevations (all points 200 m). The results suggest a relatively minor role for oceanic heat transport in the formation/elimination of Antarctic glaciation. That is, under the warmer conditions inferred to have prevailed prior to the opening of the Drake Passage, conditions would still have been favorable for the maintenance of an Antarctic ice-sheet. If anything, a moderate ocean warming would promote glaciation, by increasing snowfall. Lowering the elevation of Antarctica has a larger effect on the model simulations, reducing the likelihood of glacial conditions. In the absence of snowcover, summer temperatures over Antarctica can warm considerably, leading to a monsoon-like circulation. However, it may be difficult to achieve such snow-free conditions, even with greatly increased atmospheric carbon dioxide. A tundra-like climate is the closest the model has come to representing a non-glacial climate, even when both seasurface temperatures and elevations are maximally varied.     

Regional climate simulation with a high resolution GCM: surface hydrology

Aspects of the surface hydrology of high resolution (T106) versions of the ECHAM3 and ECHAM4 general circulation models are analysed over the European region and compared with available observations. The focus is on evaporation, and surface measurements are shown to be useful for the identification of systematic deficiencies in the regional-scale performance of climate models on an annual and seasonal basis, such as the excessive summer dryness over continents. The annual mean evaporation at the available European observation sites is overestimated by 4 mm/month by the ECHAM3 T106, quantitatively consistent with an overestimated surface net radiation of 4 Wm^–2 over Europe. In winter, ECHAM3 shows an overestimated evaporation which compensates for an overestimated downward sensible heat flux. This is primarily related to a too strong zonalisation of the large-scale flow and associated overestimated warm air advection and windspeed. Inaccurate local land surface parameters (e.g. leaf area index, roughness length) are minor contributors to the overestimation. In early summer, the excessive solar radiation at the surface calculated with the ECHAM3 radiation scheme generates a too large evaporation and an excessive depletion of the soil moisture reservoirs. This favours the subsequent excessive summer dryness over Europe with too low values of evaporation, convective precipitation and soil moisture content, leading to a too high surface temperature. In the ECHAM4 T106 simulation, the problem of the European summer dryness is largely reduced, and the simulated evaporation as well as convective precipitation, cloud amount and soil moisture content during summer are substantially improved. The new ECHAM4 radiation scheme appears to be an important factor for this improvement, since it calculates smaller insolation values in better agreement with observations and subsequently may avoid an excessive drying of the soil. Received: 20 September 1995 / Accepted: 10 May 1996     

臭氧对作物影响的模型研究概述

近年来，在大量试验研究的基础上，有关臭氧(O3)对作物影响的研究重点为如何建立有效的模型，以评估和预测O3对作物产量和农业造成的损失。文中介绍了有关臭氧对作物影响模型的研究概况，将模型分为3类：统计模型、光化学模型和机理模型；在分析各类模型局限性的基础上，指出O3对作物影响的模型研究将向模块化、普遍化方向发展，形成综合、全面的模型，并讨论了我国相关研究的发展方向。     

用Vis5D软件包在PC机上实现模式预报输出结果的可视化

介绍了Ｖｉｓ５Ｄ系统的运行环境、安装要点及主要功能;给出了Ｖｉｓ５Ｄ制作的图象实例;着重介绍了把模式输出结果转换为ｖ５ｄ格式文件的方法,以及几种常用网格数据文件转换为ｖ５ｄ格式文件的方法;简要说明了在Ｖｉｎｄｏｗｓ环境下使用Ｖｉｓ５Ｄ制作的气象图像及进行动画播放的过程。     

Reproducibility of seasonal ensemble integrations with ECMWF GCM and its association with ENSO

Summary This study evaluates seasonal climate potential predictability with a dataset of nine-member ensemble seasonal integrations produced by the ECMWF GCM for the ERA-15 period (1979–1993). The methodology used here is the measure of the ensemble reproducibility for a particular season defined by Yang et al (1998). High reproducibility reflects the dominant role of the lower boundary forcing in seasonal climate anomaly, indicating good potential predictability. Spatial patterns of the reproducibility for selected variables are documented, which exhibit obvious regionality and seasonality. Such variables are always highly reproducible over most of the tropical regions. Over the northern extratropics, primary reproducible information, taking the 500hPa geopotential height for example, is found over the PNA region during winter while over most of Asia during summer. Winter has the largest reproducible area; autumn has the lowest, while summer and spring are in between. Association of the reproducibility with ENSO events was examined regionally. Internal variances due to ensemble spread were broken down for individual years, and the reproducibility was computed for four categories: El Niño, La Niña, ENSO, and non-ENSO years. The reproducibility during winter especially over the PNA region is insensitive to ENSO events, with exceptions over the tropical western Pacific, central Siberia and Western Europe. Contrarily, ENSO events have significant impacts on the reproducibility over the southwest USA and most of Asian monsoon region during summer. These results suggest that ENSO events may not be helpful to the seasonal climate predictability over the PNA region during winter, but they may increase predictable information over many regions of the northern extratropical continents during summer.     

Evaluating the Ozone Valley over the Tibetan Plateau in CMIP6 Models

Total column ozone (TCO) over the Tibetan Plateau (TP) is lower than that over other regions at the same latitude, particularly in summer. This feature is known as the “TP ozone valley”. This study evaluates long-term changes in TCO and the ozone valley over the TP from 1984 to 2100 using Coupled Model Intercomparison Project Phase 6 (CMIP6). The TP ozone valley consists of two low centers, one is located in the upper troposphere and lower stratosphere (UTLS), and the other is in the middle and upper stratosphere. Overall, the CMIP6 models simulate the low ozone center in the UTLS well and capture the spatial characteristics and seasonal cycle of the TP ozone valley, with spatial correlation coefficients between the modeled TCO and the Multi Sensor Reanalysis version 2 (MSR2) TCO observations greater than 0.8 for all CMIP6 models. Further analysis reveals that models which use fully coupled and online stratospheric chemistry schemes simulate the anticorrelation between the 150 hPa geopotential height and zonal anomaly of TCO over the TP better than models without interactive chemistry schemes. This suggests that coupled chemical-radiative-dynamical processes play a key role in the simulation of the TP ozone valley. Most CMIP6 models underestimate the low center in the middle and upper stratosphere when compared with the Microwave Limb Sounder (MLS) observations. However, the bias in the middle and upper stratospheric ozone simulations has a marginal effect on the simulation of the TP ozone valley. Most CMIP6 models predict the TP ozone valley in summer will deepen in the future.     

Northern Hemisphere midlatitude cyclone variability in GCM simulations with different ocean representations

The impact of different ocean models or sea surface temperature (SST) and sea-ice concentrations on cyclone tracks in the Northern Hemisphere midlatitudes is determined within a hierarchy of model simulations. A reference simulation with the coupled atmosphere ocean circulation model ECHAM/HOPE is compared with simulations using ECHAM and three simplified ocean and sea-ice representations: (1) a variable depth mixed layer (ML) ocean, (2) forcing by varying SST and sea-ice, and (3) with climatological SST and sea-ice; the latter two are from the coupled ECHAM/HOPE integration. The reference simulation reproduces the observed cyclone tracks. The cyclones are tracked automatically by a standard routine and the variability of individual cyclone trajectories within the storm tracks is determined by a cluster approach. In the forced simulation with varying SST, the geographical distribution and the statistics of the cyclones are not altered compared to the coupled reference simulation. In the ML- and the climatological simulation, deviations of the mean cyclone distribution are found which occur mainly in the North Pacific, and can partially be traced back to missing El Niño/Southern Oscillation (ENSO) variability. The climatological experiment is superior to the ML-experiment. The variability of the individual cyclone trajectories, as determined by the cluster analysis, reveals the same types and frequencies of propagation directions for all four representations of the lower boundary. The largest discrepancies for the cluster occupations are found for the climatological and the ML-simulation.     

Regional climate simulation with a high resolution GCM: surface radiative fluxes

The ability of a high resolution (T106) version of the ECHAM3 general circulation model to simulate regional scale surface radiative fluxes has been assessed using observations from a new compilation of worldwide instrumentally-measured surface fluxes (Global Energy Balance Archive, GEBA). The focus is on the European region where the highest density of observations is found, and their use for the validation of global and regional climate models is demonstrated. The available data allow a separate assessment of the simulated fluxes of surface shortwave, longwave, and net radiation for this region. In summer, the incoming shortwave radiation calculated by the ECHAM3/T106 model is overestimated by 45 W m^–2 over most of Europe, which implies a largely unrealistic forcing on the model surface scheme and excessive surface temperatures. In winter, too little incoming shortwave radiation reaches the model surface. Similar tendencies are found over large areas of the mid-latitudes. These biases are consistent with deficiencies in the simulation of cloud amount, relative humidity and clear sky radiative transfer. The incoming longwave radiation is underestimated at the European GEBA stations predominantly in summer. This largely compensates for the excessive shortwave flux, leading to annual mean net radiation values over Europe close to observations due to error cancellation, a feature already noted in the simulated global mean values in an earlier study. Furthermore, the annual cycle of the simulated surface net radiation is strongly affected by the deficiencies in the simulated incoming shortwave radiation. The high horizontal resolution of the GCM allows an assessment of orographically induced flux gradients based on observations from the European Alps. Although the model-calculated and observed flux fields substantially differ in their absolute values, several aspects of their gradients are realistically captured. The deficiencies identified in the model fields are generally consistent at most stations, indicating a high degree of representativeness of the measurements for their larger scale setting.     

一个全球耦合模式的ENSO后报试验

A group of seasonal hindcast experiments are conducted using a coupled model known as the Flexible Global Ocean-Atmosphere-Land System Model-gamil1.11 (FGOALS-g1.11) developed at the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG). Two steps are included in our El Niño-Southern Oscillation (ENSO) hindcast experiments. The first step is to integrate the coupled GCM with the Sea Surface Temperature (SST) strongly nudged towards the observation from 1971 to 2006. The second step is to remove the SST nudging term. We carried out a one-year hindcast by adopting the initial values from SST nudging experiments from the first step on January 1st, April 1st, July 1st, and October 1st from 1982 to 2005. In the SST nudging experiment, the model can reproduce the observed equatorial thermocline anomalies and zonal wind stress anomalies in the Pacific, which demonstrates that the SST nudging approach can provide realistic atmospheric and oceanic initial conditions for seasonal prediction experiments. The model also demonstrates a high Anomaly Correlation Coefficient (ACC) score for SST in most of the tropical Pacific, Atlantic Ocean, and some Indian Ocean regions with a 3-month lead. Compared with the persistence ACC score, this model shows much higher ACC scores for the Nino3.4 index for a 9-month lead.     

Tropical Atlantic Variability in a Coupled GCM

Summary:Tropical Atlantic Variability (TAV) is simulated in a coupled GCM. The TAV seems to be consistent with a dipole mode that involves both surface and subsurface oceanic dynamics. The poor correlation of the tropical North and South Atlantic SST is suggested to be distorted by the presence of a symmetric tropical Atlantic mode.     

GCM studies of the African summer monsoon

A 37-year simulation of global climate by a 9-level GCM on an 8°×10° grid showed realistic interannual variation of the computed precipitation over the African Sahel. The model includes an interactive ocean so that interannual variations of sea-surface temperature (SST) also occur. Comparison of an ensemble of five summers that were rainy over the Sahel with five summers of simulated drought showed that insufficient ambient moisture was the immediate cause of the lack of moist convection. The drier conditions are shown to result from weaker moisture advection over the southeast Atlantic Ocean. Weaker southerly winds there and lower sea-level pressure gradients seemed to result from anomalously warm SST. Such SST anomalies have been linked to Sahelian drought in previous observational studies. These regional circulations that were conducive to lower rainfall rates during the north African summer monsoon were not manifestations of the more generalized zonal mean circulation.     

A GCM AND ITS MONTHLY SIMULATION EXPERIMENT

A global general circulation model (AMS-GCM) including many physical processes is presented in this paper.And a month's simulation has been made by use of the globa/climate data in January.The results seem to be in agreement with the observations.     

热带大气季节内振荡的一个数值模拟研究

文中分析了中国科学院大气物理研究所全球气候谱模式ALGCM (R4 2L9) 12a(1978～ 1989年 )积分的逐日输出结果 ,并与 1978～ 1989年的逐日NCEP资料对照 ,以此对热带季节内振荡 (30～ 6 0d振荡 )进行数值模拟研究。分析表明 ,该模式在热带地区可以模拟出明显的季节内振荡 (ISO)的准周期信号 ,并抓住了热带ISO的基本传播特征 ,能较好地再现东、西半球传播速度的差异 ,同时模式模拟存在东传要好于西传 ,冬、春季的模拟要好于夏、秋季的现象。该模式模拟的热带ISO的强度较许多大气模式明显提高 ,尤其是对 2 0 0hPa上ISO动能强度的模拟。模式基本模拟出了ISO低层辐合、高层辐散的水平风场特征。模式较好地再现了热带ISO纬向风的垂直结构。此外 ,观测资料表明热带ISO在冬、春强 ,而夏、秋弱的季节性倾向与ISO的年际变化相联系 ,模拟的ISO在季节性倾向偏差上表现为冬、夏相对强 ,而春、秋相对弱。垂直速度、散度、水汽等物理量的配置同NCEP资料的结构特征仍有明显差异 ,模拟的ISO空间分布也不太理想 ,表明要很好模拟ISO结构和空间分布特征 ,还须做不少工作。