Evolution of model systematic errors in the Tropical Atlantic Basin from coupled climate hindcasts

Significant systematic errors in the tropical Atlantic Ocean are common in state-of-the-art coupled ocean–atmosphere general circulation models. In this study, a set of ensemble hindcasts from the NCEP coupled forecast system (CFS) is used to examine the initial growth of the coupled model bias. These CFS hindcasts are 9-month integrations starting from perturbed real-time oceanic and atmospheric analyses for 1981–2003. The large number of integrations from a variety of initial states covering all months provides a good opportunity to examine how the model systematic errors grow. The monthly climatologies of ensemble hindcasts from various initial months are compared with both observed and analyzed oceanic and atmospheric datasets. Our analyses show that two error patterns are dominant in the hindcasts. One is the warming of the sea surface temperature (SST) in the southeastern tropical Atlantic Ocean. This error grows faster in boreal summer and fall and peaks in November–December at round 2°C in the open ocean. It is caused by an excessive model surface shortwave radiative flux in this region, especially from boreal summer to fall. The excessive radiative forcing is in turn caused by the CFS inability to reproduce the observed amount of low cloud cover in the southeastern ocean and its seasonal increase. According to a comparison between the seasonal climatologies from the CFS hindcasts and a long-term simulation of the atmospheric model forced with observed SST, the CFS low cloud and radiation errors are inherent to its atmospheric component. On the other hand, the SST error in CFS is a major cause of the model’s southward bias of the intertropical convergence zone (ITCZ) in boreal winter and spring. An analysis of the SST errors of the 6-month ensemble hindcasts by seven coupled models in the Development of a European Multimodel Ensemble System for Seasonal-to-Interannual Prediction project shows that this SST error pattern is common in coupled climate hindcasts. The second error pattern is an excessive deepening of the model thermocline depth to the north of the equator from the western coast toward the central ocean. This error grows fastest in boreal summer. It is forced by an overly strong local anticyclonic surface wind stress curl and is in turn related to the weakened northeast trade winds in summer and fall. The thermocline error in the northwest delays the annual shoaling of the equatorial thermocline in the Gulf of Guinea remotely through the equatorial waveguide.     

混合误差协方差用于集合平方根滤波同化的试验

在集合卡尔曼滤波方法中, 根据预报集合统计提供的依流型而变的预报误差协方差对同化起到决定性的作用。但在集合样本容量不足及模式存在系统误差时, 由预报集合估计的预报误差协方差会出现明显偏差。既要减小这种估计偏差对同化产生的影响而又不增加计算量, 一种可供选择的方法是将定常或准定常的高斯型预报误差协方差和由预报集合估计的预报误差协方差加权平均用于集合卡尔曼滤波同化。利用浅水方程模式, 通过观测系统模拟试验检验在不同的模式误差、 集合成员数以及观测密度条件下, 将这种混合预报误差协方差矩阵用于在集合平方根滤波的效果。试验结果表明, 当预报集合成员数较多而模式又无误差时, 不必采用混合的预报误差协方差矩阵, 否则, 采用混合的预报误差协方差矩阵都有可能改进分析和预报。混合预报误差协方差的最优的权重系数与模式误差关系密切, 模式误差越大, 定常预报误差协方差的权重越大。最优的权重系数与集合成员数及观测密度也有一定关系。     

The role of ENSO in global ocean temperature changes during 1955–2011 simulated with a 1D climate model

Global average ocean temperature variations to 2,000 m depth during 1955–2011 are simulated with a 40 layer 1D forcing-feedback-mixing model for three forcing cases. The first case uses standard anthropogenic and volcanic external radiative forcings. The second adds non-radiative internal forcing (ocean mixing changes initiated in the top 200 m) proportional to the Multivariate ENSO Index (MEI) to represent an internal mode of natural variability. The third case further adds ENSO-related radiative forcing proportional to MEI as a possible natural cloud forcing mechanism associated with atmospheric circulation changes. The model adjustable parameters are net radiative feedback, effective diffusivities, and internal radiative (e.g., cloud) and non-radiative (ocean mixing) forcing coefficients at adjustable time lags. Model output is compared to Levitus ocean temperature changes in 50 m layers during 1955–2011 to 700 m depth, and to lag regression coefficients between satellite radiative flux variations and sea surface temperature between 2000 and 2010. A net feedback parameter of 1.7Wm^?2 K^?1 with only anthropogenic and volcanic forcings increases to 2.8Wm^?2 K^?1 when all ENSO forcings (which are one-third radiative) are included, along with better agreement between model and observations. The results suggest ENSO can influence multi-decadal temperature trends, and that internal radiative forcing of the climate system affects the diagnosis of feedbacks. Also, the relatively small differences in model ocean warming associated with the three cases suggests that the observed levels of ocean warming since the 1950s is not a very strong constraint on our estimates of climate sensitivity.     

Long-term climate response to stabilized and overshoot anthropogenic forcings beyond the twenty-first century

From multi-ensembles of climate simulations using the Community Climate System Model version 3, global climate changes have been investigated focusing on long-term responses to stabilized anthropogenic forcings. In addition to the standard forcing scenarios for the current international assessment, an overshoot scenario, where radiative forcings are decreased from one stabilized level to another, is also considered. The globally-averaged annual surface air temperature increases during the twenty-first century by 2.58 and 1.56°C for increased forcings under two future scenarios denoted by A1B and B1, respectively. These changes continue but at much slower rates in later centuries under forcings stabilized at year 2100. The overshoot scenario provides a different pathway to the lower B1 level by way of the greater A1B level. This scenario results in a surface climate similar to that in the B1 scenario within 100 years after the forcing reaches the B1 level. Contrasting to the surface changes, responses in the ocean are significantly delayed. It is estimated from the linear response theory that temperature changes under stabilized forcings to a final equilibrium state in the A1B (B1) scenario are factors of 0.3–0.4, 0.9, and 17 (0.3, 0.6, and 11) to changes during the twenty-first century, respectively, for three ocean layers of the surface to 100, 100–500, and 500 m to the bottom. Although responses in the lower ocean layers imply a nonlinear behavior, the ocean temperatures in the overshoot and B1 scenarios are likely to converge in their final equilibrium states.     

利用海洋强迫的大气主模态提高大气动力季节预报

数值模式的季节预报技巧主要与大气外强迫的变率密切相关。当前的大气环流模式（general circulation models,GCMs）通常不能准确地模拟出与大气外强迫有关的响应模态和响应强度,从而导致了预报误差的产生。本文给出了一种后处理方法,有助于降低模式的系统性误差,并提高季节预报技巧。     

Does temperature contain a stochastic trend? Evaluating conflicting statistical results

We evaluate the claim by Gay et al. (Clim Change 94:333–349, 2009) that “surface temperature can be better described as a trend stationary process with a one-time permanent shock” than efforts by Kaufmann et al. (Clim Change 77:249–278, 2006) to model surface temperature as a time series that contains a stochastic trend that is imparted by the time series for radiative forcing. We test this claim by comparing the in-sample forecast generated by the trend stationary model with a one-time permanent shock to the in-sample forecast generated by a cointegration/error correction model that is assumed to be stable over the 1870–2000 sample period. Results indicate that the in-sample forecast generated by the cointegration/error correction model is more accurate than the in-sample forecast generated by the trend stationary model with a one-time permanent shock. Furthermore, Monte Carlo simulations of the cointegration/error correction model generate time series for temperature that are consistent with the trend-stationary-with-a-break result generated by Gay et al. (Clim Change 94:333–349, 2009), while the time series for radiative forcing cannot be modeled as trend stationary with a one-time shock. Based on these results, we argue that modeling surface temperature as a time series that shares a stochastic trend with radiative forcing offers the possibility of greater insights regarding the potential causes of climate change and efforts to slow its progression.     

包含外强迫因子的平潭旬降水量预测

降水量是重要的预报要素之一,长期的降水预测更是能提前预测旱涝分布情况,为国民经济规划提供依据。但目前为止,长期的降水预测仍缺少客观的预报方法。为此,尝试利用非线性预测模型来预测旬降水量,并将该模型应用于福建平潭,分别用与原始数据的差值、与原始数据的相关系数、均方根误差,以及符号显著性检验方法,讨论了包含外强迫因子的平稳性模型与不包含外强迫因子的非线性模型的预测能力,结果表明:包含外强迫因子的模型第一步预测结果与原始观测数据的相关系数为0.73,不包含外强迫因子的模型第一步预测结果与原始观测数据的相关系数则为0.47。无论是从与原始数据的差值及相关系数,还是均方根误差等方面,外强迫模型都是优于平稳性模型,并且通过符号检验方法可看出两种模型存在差异性,这也说明加入外强迫因子可以有效地提高预测技巧,外强迫因子与状态变量在预测中扮演同等重要的角色。     

Which Features of the SST Forcing Error Most Likely Disturb the Simulated Intensity of Tropical Cyclones?

Among all of the sources of tropical cyclone(TC) intensity forecast errors, the uncertainty of sea surface temperature(SST) has been shown to play a significant role. In the present study, we determine the SST forcing error that causes the largest simulation error of TC intensity during the entire simulation period by using the WRF model with time-dependent SST forcing. The SST forcing error is represented through the application of a nonlinear forcing singular vector(NFSV)structure. For the selected 12 TC cases, the NFSV-type SST forcing errors have a nearly coherent structure with positive(or negative) SST anomalies located along the track of TCs but are especially concentrated in a particular region. This particular region tends to occur during the specific period of the TCs life cycle when the TCs present relatively strong intensity, but are still intensifying just prior to the mature phase, especially within a TC state exhibiting a strong secondary circulation and very high inertial stability. The SST forcing errors located along the TC track during this time period are verified to have the strongest disturbing effect on TC intensity simulation. Physically, the strong inertial stability of TCs during this time period induces a strong response of the secondary circulation from diabatic heating errors induced by the SST forcing error. Consequently, this significantly influences the subsidence within the warm core in the eye region, which,in turn, leads to significant errors in TC intensity. This physical mechanism explains the formation of NSFV-type SST forcing errors. According to the sensitivity of the NFSV-type SST forcing errors, if one increases the density of SST observations along the TC track and assimilates them to the SST forcing field, the skill of TC intensity simulation generated by the WRF model could be greatly improved. However, this adjustment is most advantageous in improving simulation skill during the time period when TCs become strong but are still intensifying just prior to reaching full maturity. In light of this, the region along the TC track but in the time period of TC movement when the NFSV-type SST forcing errors occur may represent the sensitive area for targeting observation for SST forcing field associated with TC intensity simulation.     

GFS模式地形高度偏差对地面2 m气温预报的影响

利用2016年1月1日—12月31日全球预报系统(GFS,Global Forecasting System)1～5 d的2 m气温预报资料,以及同期中国地面气象站2 m气温观测资料,研究模式地形高度偏差对地面2 m气温预报的影响。结果表明,较大模式地形高度偏差可严重影响2 m气温模式预报性能,导致较大预报误差。随着模式预报时效延长,2 m气温预报均方根误差也略有增加。比较模式地形高度偏差和预报时效对于模式预报性能的影响,发现模式地形高度偏差对于模式预报效果的影响更加显著。两种地形订正方案,即不做温度垂直订正的线性回归以及对温度进行垂直订正的线性回归都能显著减小2 m气温模式预报的误差,后者的订正效果更好。     

A Hybrid Coupled Model for the Pacific Ocean–Atmosphere System.Part I: Description and Basic Performance

A hybrid coupled model(HCM) is constructed for El Nino–Southern Oscillation(ENSO)-related modeling studies over almost the entire Pacific basin. An ocean general circulation model is coupled to a statistical atmospheric model for interannual wind stress anomalies to represent their dominant coupling with sea surface temperatures. In addition, various relevant forcing and feedback processes exist in the region and can affect ENSO in a significant way; their effects are simply represented using historical data and are incorporated into the HCM, including stochastic forcing of atmospheric winds, and feedbacks associated with freshwater flux, ocean biology-induced heating(OBH), and tropical instability waves(TIWs). In addition to its computational efficiency, the advantages of making use of such an HCM enable these related forcing and feedback processes to be represented individually or collectively, allowing their modulating effects on ENSO to be examined in a clean and clear way. In this paper, examples are given to illustrate the ability of the HCM to depict the mean ocean state, the circulation pathways connecting the subtropics and tropics in the western Pacific, and interannual variability associated with ENSO. As satellite data are taken to parameterize processes that are not explicitly represented in the HCM, this work also demonstrates an innovative method of using remotely sensed data for climate modeling. Further model applications related with ENSO modulations by extratropical influences and by various forcings and feedbacks will be presented in Part II of this study.     

Global climate models as forcing for regional ocean modeling: a sensitivity study in the Iberian Basin (Eastern North Atlantic)

This work evaluates the performance of several global climate models (GCMs) as forcing of a regional ocean model configuration centered in the Iberian Basin. The study is divided in two parts. First, the output of nine GCMs is analyzed based on the fields needed to force the ocean model (Regional Ocean Modelling System—ROMS). GCMs differ greatly between them and their performance depends on the field. In the second part, the two GCMs with the worst performances in both extremes of the ensemble are used as forcing for two ROMS simulations, with the purpose of assessing the range of uncertainty comprised in this set of GCMs. Two other ROMS runs are setup: one climatologically forced control run, and one forced with the average of all the nine GCMs—the ensemble mean. Results show that the tendency of overestimation/underestimation of the forcings is reflected in the modeled hydrography, both at the surface and deeper layers down to 500 m. Nevertheless, in terms of circulation, all four runs reproduce the Azores Current, as well as the coastal transition zone seasonality (winter poleward flow and summer upwelling-associated equatorward flow). The CGCMs output performance as forcing depends on the forcing variable: one performs well for one or more variables, but badly for others, and which field is well or badly reproduced varies for each CGCM. Therefore, there is not a single CGCM having the best forcing for all variables. Hence, our results indicate that the most adequate approach consists of using the ensemble mean as forcing rather than using an individual model. This is supported by the general low overall (i.e. for all forcing variables) errors of the ensemble mean regarding the control climatological dataset, and the good comparison of the ensemble-forced ROMS run with the control run.     

The influence of tropical Pacific forcing on the Arctic Oscillation

The potential role of tropical Pacific forcing in driving the seasonal variability of the Arctic Oscillation (AO) is explored using both observational data and a simple general circulation model (SGCM). A lead–lag regression technique is first applied to the monthly averaged sea surface temperature (SST) and the AO index. The AO maximum is found to be related to a negative SST anomaly over the tropical Pacific three months earlier. A singular value decomposition (SVD) analysis is then performed on the tropical Pacific SST and the sea level pressure (SLP) over the Northern Hemisphere. An AO-like atmospheric pattern and its associated SST appear as the second pair of SVD modes. Ensemble integrations are carried out with the SGCM to test the atmospheric response to different tropical Pacific forcings. The atmospheric response to the linear fit of the model’s empirical forcing associated with the SST variability in the second SVD modes strongly projects onto the AO. Idealized thermal forcings are then designed based on the regression of the seasonally averaged tropical Pacific precipitation against the AO index. Results indicate that forcing anomalies over the western tropical Pacific are more effective in generating an AO-like response while those over the eastern tropical Pacific tend to produce a Pacific-North American (PNA)-like response. The physical mechanisms responsible for the energy transport from the tropical Pacific to the extratropical North Atlantic are investigated using wave activity flux and vorticity forcing formalisms. The energy from the western tropical Pacific forcing tends to propagate zonally to the North Atlantic because of the jet stream waveguide effect while the transport of the energy from the eastern tropical Pacific forcing mostly concentrates over the PNA area. The linearized SGCM results show that nonlinear processes are involved in the generation of the forced AO-like pattern.     

Dynamic analogue initialization for ensemble forecasting

This paper introduces a new approach for the initialization of ensemble numerical forecasting: Dynamic Analogue Initialization (DAI). DAI assumes that the best model state trajectories for the past provide the initial conditions for the best forecasts in the future. As such, DAI performs the ensemble forecast using the best analogues from a full size ensemble. As a pilot study, the Lorenz63 and Lorenz96 models were used to test DAI’s effectiveness independently. Results showed that DAI can improve the forecast significantly. Especially in lower-dimensional systems, DAI can reduce the forecast RMSE by ～50% compared to the Monte Carlo forecast (MC). This improvement is because DAI is able to recognize the direction of the analysis error through the embedding process and therefore selects those good trajectories with reduced initial error. Meanwhile, a potential improvement of DAI is also proposed, and that is to find the optimal range of embedding time based on the error’s growing speed.     

A model study of the Little Ice Age and beyond: changes in ocean heat content,hydrography and circulation since 1500

The Earth System Climate Model from the University of Victoria is used to investigate changes in ocean properties such as heat content, temperature, salinity, density and circulation during 1500 to 2000, the time period which includes the Little Ice Age (LIA) (1500–1850) and the industrial era (1850–2000). We force the model with two different wind-stress fields which take into account the North Atlantic Oscillation. Furthermore, temporally varying radiative forcings due to volcanic activity, insolation changes and greenhouse gas changes are also implemented. We find that changes in the upper ocean (0–300 m) heat content are mainly driven by changes in radiative forcing, except in the polar regions where the varying wind-stress induces changes in ocean heat content. In the full ocean (0–3,000 m) the wind-driven effects tend to reduce, prior to 1700, the downward trend in the ocean heat content caused by the radiative forcing. Afterwards no dynamical effect is visible. The colder ocean temperatures in the top 600 m during the LIA are caused by changes in radiative forcing, while the cooling at the bottom is wind-driven. The changes in salinity are small except in the Arctic Ocean. The reduced salinity content in the subsurface Arctic Ocean during the LIA is a result from reduced wind-driven inflow of saline water from the North Atlantic. At the surface of the Arctic Ocean the changes in salinity are caused by changes in sea–ice thickness. The changes in density are a composite picture of the temperature and salinity changes. Furthermore, changes in the meridional overturning circulation (MOC) are caused mainly by a varying wind-stress forcing; the additional buoyancy driven changes due to the radiative forcings are small. The simulated MOC is reduced during the LIA as compared to the industrial era. On the other hand, the ventilation rate in the Southern Ocean is increased during the LIA.     

江苏—南黄海地区M≥6强震有序网络结构及其预测研究

基于TIGGE（THORPEX Interactive Grand Global Ensemble,全球交互式大集合）资料中欧洲中期天气预报中心（European Centre for Medium-Range Weather,ECMWF）、日本气象厅（Japan Meteorological Agency,JMA）、美国国家环境预报中心（National Centers for Environmental Prediction,NCEP）和英国气象局（United Kingdom Met Office,UKMO）4个中心的北半球地面2m气温集合平均预报资料,利用插值技术与回归分析,并引入了消除偏差集合平均（bias-removed ensemble mean,BREM）和多模式超级集合（superensemble,SUP）方法进行统计降尺度预报研究.结果表明,在2007年夏季3个月中,4个单中心的降尺度预报明显地改善了预报效果.引入SUP和BREM两种集成预报方法后,预报误差得到进一步减小.对比综合表现最好的单中心ECMWF的预报,1～7d的降尺度预报误差改进率均达20％以上.研究还发现,引入SUP方法的降尺度预报效果优于引入BREM方法的降尺度预报,利用双线性插值方法在上述两方案中的预报效果优于其他3种插值方法.     

基于TIGGE资料的北半球地面气温预报的统计降尺度研究

基于TIGGE(THORPEX Interactive Grand Global Ensemble,全球交互式大集合)资料中欧洲中期天气预报中心(European Centre for Medium-Range Weather,ECMWF)、日本气象厅(Japan Meteorological Agency,JMA)、美国国家环境预报中心(National Centers for Environmental Prediction,NCEP)和英国气象局(United Kingdom Met Office,UKMO)4个中心的北半球地面2 m气温集合平均预报资料,利用插值技术与回归分析,并引入了消除偏差集合平均(bias-removed ensemble mean,BREM)和多模式超级集合(superensemble,SUP)方法进行统计降尺度预报研究。结果表明,在2007年夏季3个月中,4个单中心的降尺度预报明显地改善了预报效果。引入SUP和BREM两种集成预报方法后,预报误差得到进一步减小。对比综合表现最好的单中心ECMWF的预报,1~7 d的降尺度预报误差改进率均达20%以上。研究还发现,引入SUP方法的降尺度预报效果优于引入BREM方法的降尺度预报,利用双线性插值方法在上述两方案中的预报效果优于其他3种插值方法。     

三种全球预报产品中国区近地面气温短期预报效果检验

全球气象预报产品是扩散模式、空气质量模式的重要基础资料和前提条件,其误差直接影响模拟结果的准确度。为考察不同气象预报产品的误差,选取2016年6月至2017年5月GFS、ECMWF、T639三种全球气象预报产品,利用中国2100个地面观测站数据,对预报产品中近地面气温进行了对比,并分析了其在不同季节、不同区域的特征。结果表明：在中国区域三种气象产品气温预报存在偏低预报的趋势,其均方根误差的年平均值为2.60—3.52℃,相关系数的年平均值为0.89—0.92,平均绝对误差的年平均值为1.87—2.67℃。整体而言,EC表现最佳,其余依次为GFS、T639。气温预报误差存在季节变化特征,三种产品均方根误差与平均绝对误差均表现为夏秋季优于春冬季,相关系数表现为秋冬季优于春夏季。气温预报误差存在明显的地域差异,三种气象预报产品的气温误差空间分布特征较为相似,在中国华东地区误差值表现较低,在西南地区误差较高。同时,其误差水平在中国沿海地区表现较低,在地形复杂地区表现较高。     

Evidence for the role of the Atlantic multidecadal oscillation and the ocean heat uptake in hiatus prediction

The recent hiatus in global temperature at the surface has been analysed by several studies, mainly using global climate models. The common accepted picture is that since the late 1990s, the increase in anthropogenic radiative forcings has been counterbalanced by other factors, e.g., a decrease in natural forcings, augmented ocean heat storage and negative phases of ocean–atmosphere-coupled oscillation patterns. Here, simple vector autoregressive models are used for forecasting the temperature hiatus in the period 2001–2014. This gives new insight into the problem of understanding the ocean contribution (in terms of heat uptake and atmosphere–ocean-coupled oscillations) to the appearance of this recent hiatus. In particular, considering data about the ocean heat content until a depth of 700 m and the Atlantic multidecadal oscillation is necessary for correctly forecasting the hiatus, so catching both trend and interannual variability. Our models also show that the ocean heat uptake is substantially driven by the natural component of the total radiative forcing at a decadal time scale, confining the importance of the anthropogenic influences to a longer range warming of the ocean.     

Efficient Anomalous Forcings for Linear Problems

For linear forcing problems, a method is developed to provide a set of forcing modes which form a complete orthonormal basis for the finite-time response to steady forcing in the energy inner product space. The forcing modes are found by calculating eigenvectors of a positive definite and symmetric matrix determined from given equations of motion. The amplitude of responses to forcing modes is given in terms of the associated eigenvalues. This method is used in a nondivergent barotropic model linearized about the 300 hPa zonally-varying climatological flow both for northern summertime and wintertime. The results show that the amplitude of response varies considerably with different forcing modes. Only a few of forcing modes associated with the leading eigenvalues, called efficient forcing mode, can excite significant response. The efficient forcing modes possess highly localized spatial structure with wavetrain appearance. Most of the efficient forcings are located to the south of regions of the jet cor