Global seasonal climate predictability in a two tiered forecast system: part I: boreal summer and fall seasons

This paper shows demonstrable improvement in the global seasonal climate predictability of boreal summer (at zero lead) and fall (at one season lead) seasonal mean precipitation and surface temperature from a two-tiered seasonal hindcast forced with forecasted SST relative to two other contemporary operational coupled ocean–atmosphere climate models. The results from an extensive set of seasonal hindcasts are analyzed to come to this conclusion. This improvement is attributed to: (1) The multi-model bias corrected SST used to force the atmospheric model. (2) The global atmospheric model which is run at a relatively high resolution of 50 km grid resolution compared to the two other coupled ocean–atmosphere models. (3) The physics of the atmospheric model, especially that related to the convective parameterization scheme. The results of the seasonal hindcast are analyzed for both deterministic and probabilistic skill. The probabilistic skill analysis shows that significant forecast skill can be harvested from these seasonal hindcasts relative to the deterministic skill analysis. The paper concludes that the coupled ocean–atmosphere seasonal hindcasts have reached a reasonable fidelity to exploit their SST anomaly forecasts to force such relatively higher resolution two tier prediction experiments to glean further boreal summer and fall seasonal prediction skill.     

Probabilistic prediction of Indian summer monsoon rainfall using global climate models

Probabilistic seasonal predictions of rainfall that incorporate proper uncertainties are essential for climate risk management. In this study, three different multi-model ensemble (MME) approaches are used to generate probabilistic seasonal hindcasts of the Indian summer monsoon rainfall based on a set of eight global climate models for the 1982–2009 period. The three MME approaches differ in their calculation of spread of the forecast distribution, treated as a Gaussian, while all three use the simple multi-model subdivision average to define the mean of the forecast distribution. The first two approaches use the within-ensemble spread and error residuals of ensemble mean hindcasts, respectively, to compute the variance of the forecast distribution. The third approach makes use of the correlation between the ensemble mean hindcasts and the observations to define the spread using a signal-to-noise ratio. Hindcasts are verified against high-resolution gridded rainfall data from India Meteorological Department in terms of meteorological subdivision spatial averages. The use of correlation for calculating the spread provides better skill than the other two methods in terms of rank probability skill score. In order to further improve the skill, an additional method has been used to generate multi-model probabilistic predictions based on simple averaging of tercile category probabilities from individual models. It is also noted that when such a method is used, skill of probabilistic forecasts is improved as compared with using the multi-model ensemble mean to define the mean of the forecast distribution and then probabilities are estimated. However, skill of the probabilistic predictions of the Indian monsoon rainfall is too low.     

Evaluation of summer temperature and precipitation predictions from NCEP CFSv2 retrospective forecast over China

National Centers for Environmental Prediction recently upgraded its operational seasonal forecast system to the fully coupled climate modeling system referred to as CFSv2. CFSv2 has been used to make seasonal climate forecast retrospectively between 1982 and 2009 before it became operational. In this study, we evaluate the model’s ability to predict the summer temperature and precipitation over China using the 120 9-month reforecast runs initialized between January 1 and May 26 during each year of the reforecast period. These 120 reforecast runs are evaluated as an ensemble forecast using both deterministic and probabilistic metrics. The overall forecast skill for summer temperature is high while that for summer precipitation is much lower. The ensemble mean reforecasts have reduced spatial variability of the climatology. For temperature, the reforecast bias is lead time-dependent, i.e., reforecast JJA temperature become warmer when lead time is shorter. The lead time dependent bias suggests that the initial condition of temperature is somehow biased towards a warmer condition. CFSv2 is able to predict the summer temperature anomaly in China, although there is an obvious upward trend in both the observation and the reforecast. Forecasts of summer precipitation with dynamical models like CFSv2 at the seasonal time scale and a catchment scale still remain challenge, so it is necessary to improve the model physics and parameterizations for better prediction of Asian monsoon rainfall. The probabilistic skills of temperature and precipitation are quite limited. Only the spatially averaged quantities such as averaged summer temperature over the Northeast China of CFSv2 show higher forecast skill, of which is able to discriminate between event and non-event for three categorical forecasts. The potential forecast skill shows that the above and below normal events can be better forecasted than normal events. Although the shorter the forecast lead time is, the higher deterministic prediction skill appears, the probabilistic prediction skill does not increase with decreased lead time. The ensemble size does not play a significant role in affecting the overall probabilistic forecast skill although adding more members improves the probabilistic forecast skill slightly.     

2010年8月环流形势的多模式超级集合预报

利用多模式超级集合预报法,以欧洲中期天气预报中心、日本气象厅、德国气象局、中国气象局和中国空军气象中心共5个决定性7 d预报产品为集合成员,对2010年8月500 hPa高度场和850 hPa温度场分别进行固定训练期和滑动训练期超级集合预报。采用均方根误差和相关系数对超级集合预报、单一模式预报和简单集合平均预报进行对比检验,同时对各预报结果的均方根误差空间分布进行对比分析。结果表明:超级集合预报在所有预报结果中最佳,且滑动集合预报对8月后期时段预报要略好于固定集合预报,两者预报效果均好于参与集合预报的各模式,也好于集合平均预报。但随着预报时效的延长,集合平均预报的优势也随之提升。从预报结果均方根误差的空间分布可知,多模式超级集合预报相比于单一模式预报效果提高的区域,500 hPa位势高度场主要位于印度半岛、印度洋、青藏高原及以西地区,而850 hPa温度场则主要位于蒙古、青藏高原、中国新疆及以西地区。     

Ensemble hindcasts of ENSO events over the past 120 years using a large number of ensembles

Based on an intermediate coupled model (ICM), a probabilistic ensemble prediction system (EPS) has been developed. The ensemble Kalman filter (EnKF) data assimilation approach is used for generating the initial ensemble conditions, and a linear, first-order Markov-Chain SST anomaly error model is embedded into the EPS to provide model-error perturbations. In this study, we perform ENSO retrospective forecasts over the 120 year period 1886–2005 using the EPS with 100 ensemble members and with initial conditi...     

Probabilistic temperature forecasting with statistical calibration in Hungary

Weather forecasting is based on the outputs of deterministic numerical weather forecasting models. Multiple runs of these models with different initial conditions result in forecast ensembles which are used for estimating the distribution of future atmospheric variables. However, these ensembles are usually under-dispersive and uncalibrated, so post-processing is required. In the present work, Bayesian model averaging (BMA) is applied for calibrating ensembles of temperature forecasts produced by the operational limited area model ensemble prediction system of the Hungarian Meteorological Service (HMS). We describe two possible BMA models for temperature data of the HMS and show that BMA post-processing significantly improves calibration and probabilistic forecasts although the accuracy of point forecasts is rather unchanged.     

夏季亚欧中高纬度环流的集合预报效果检验

使用NCEP集合预报资料, 对亚洲中高纬地区2003年6—8月500 hPa高度场的集合预报效果进行了检验。环流预报效果检验结果表明:预报时效大于5 d时, 集合平均预报明显优于单一预报; 使用相同模式分辨率时, 集合平均能将可用预报时效延长12 h以上, 达到7.5 d; 通过集合预报可获得真正意义的概率预报结果, 取得较单一高分辨率预报好的预报效果。阻塞过程的个例分析也表明集合平均的预报效果明显优于单一确定性预报; 特征等值线可反映集合成员的不一致信息和少数集合成员的异常表现, 以此为基础, 可估计分析对象出现与否的概率, 达到提高预报效果的目的。     

区域集合预报系统2 m温度预报的校准技术

采用非齐次高斯回归 (NGR) 技术对国家气象中心区域集合预报系统的2 m温度预报结果开展了一阶偏差和二阶离散度的校准研究。对预报结果比较详尽的检验分析表明:校准后的2 m温度预报可靠性和预报技巧均显著提高,表现为校准后集合预报成员的均方根误差与离散度更为接近;原Talagrand直方图中的“L”形分布现象得到有效改善;Brier评分、最小连续分级概率评分 (CRPS) 明显减小,相对作用特征 (ROC) 面积增大,说明校准后的2 m温度预报表现出更好的预报技能。此外,NGR技术与自适应误差订正技术的对比试验表明,NGR在消除集合平均偏差和提高集合离散度两个方面均有优势。     

A theoretical relationship between probabilistic relative operating characteristic skill and deterministic correlation skill in dynamical seasonal climate prediction

Understanding the relationship between probabilistic and deterministic prediction skills is of important significance for the study of seasonal forecasting and verification. Based on the Brier skill score methodology, we have previously found a theoretical relationship between the probabilistic resolution skill and the deterministic correlation (i.e., anomaly correlation; AC) skill and a lack of necessary or consistent relationship between the probabilistic reliability skill and the deterministic skill in dynamical seasonal prediction. Here, we further theoretically investigate the relationship between the probabilistic relative operating characteristic (ROC) skill and the deterministic skill. The ROC measures the discrimination attribute of probabilistic forecast quality, another important attribute besides the resolution and reliability. With some simplified assumptions, we first derive theoretical expressions for the hit and false-alarm rates that are basic ingredients for the ROC curve, then demonstrate a sole dependence of the ROC curve on the AC, and finally analytically derive a relationship between the related ROC score and the AC. Such a theoretically derived ROC-AC relationship is further examined using dynamical models’ ensemble seasonal hindcasts, which is well verified. The finding here along with our previous findings implies that the discrimination and resolution attributes of probabilistic seasonal forecast skill are intrinsically equivalent to the corresponding deterministic skill, while the reliability appears to be the fundamental attribute of the probabilistic skill that differs from the deterministic skill, which constitutes an understanding of the fundamental similarities and difference between the two types of seasonal forecasting skills and predictability and can offer important implications for the study of seasonal forecasting and verification.

     

东亚地区冬季地面气温延伸期概率预报研究

利用TIGGE资料中的ECMWF、NCEP、UKMO三个中心集合预报系统以及由此构成的多中心集合预报系统所提供的地面2 m气温10~15 d延伸期集合预报产品,建立贝叶斯模式平均(Bayesian Model Averaging,BMA)概率预报模型,对东亚地区冬季地面气温进行延伸期概率预报研究。采用距平相关系数、均方根误差、布莱尔评分、等级概率评分等指标分别对BMA确定性结果与概率预报进行评估。结果表明,BMA方法明显地改进了原始集合预报结果,预报技巧优于原始集合预报,且多中心BMA预报优于单中心BMA预报,最佳滑动训练期取35 d。BMA预报为气温的延伸期概率预报提供了更合理的概率分布,定量描述了预报的不确定性。     

Downscaling of the global seasonal forecasts of hydrometcenter of Russia for North Eurasia

Proposed is a method of downscaling of the global ensemble seasonal forecasts of air temperature computed using the SLAV model of the Hydrometcenter of Russia. The method is based on the regression and suggests a probabilistic interpretation of forecasts based on the assessment of uncertainty associated with the regression and model forecast ensemble spread. The verification of the method for 70 weather stations of North Eurasia using the rank probability skill score RPSS showed a significant advantage of downscaled forecasts over the forecasts interpolated from the model grid points. It is concluded that the use of the downscaling method is reasonable for the long-range forecasting of the station air temperature for North Eurasia.     

Multimodel approach to seasonal prediction

Multimodel forecast fields of temperature at 850 hPa and seasonal precipitation are combined using a procedure of two-step averaging. It is shown that the resulting forecasts averaged over the multimodel ensemble outperform the forecasts of individual models. The verification of forecast production has been carried out on cross-validated hindcasts according to WMO requirements. The simulation of spatiotemporal variability of atmospheric variables is assessed. The results indicate that the combined models are rather skillful in the tropical oceans, while the accuracy in the extratropics is poor.     

Comparison of Statistical Post-Processing Methods for Probabilistic Wind Speed Forecasting

In this study, the statistical post-processing methods that include bias-corrected and probabilistic forecasts of wind speed measured in PyeongChang, which is scheduled to host the 2018 Winter Olympics, are compared and analyzed to provide more accurate weather information. The six post-processing methods used in this study are as follows: mean bias-corrected forecast, mean and variance bias-corrected forecast, decaying averaging forecast, mean absolute bias-corrected forecast, and the alternative implementations of ensemble model output statistics (EMOS) and Bayesian model averaging (BMA) models, which are EMOS and BMA exchangeable models by assuming exchangeable ensemble members and simplified version of EMOS and BMA models. Observations for wind speed were obtained from the 26 stations in PyeongChang and 51 ensemble member forecasts derived from the European Centre for Medium-Range Weather Forecasts (ECMWF Directorate, 2012) that were obtained between 1 May 2013 and 18 March 2016. Prior to applying the post-processing methods, reliability analysis was conducted by using rank histograms to identify the statistical consistency of ensemble forecast and corresponding observations. Based on the results of our study, we found that the prediction skills of probabilistic forecasts of EMOS and BMA models were superior to the biascorrected forecasts in terms of deterministic prediction, whereas in probabilistic prediction, BMA models showed better prediction skill than EMOS. Even though the simplified version of BMA model exhibited best prediction skill among the mentioned six methods, the results showed that the differences of prediction skills between the versions of EMOS and BMA were negligible.     

Sub-seasonal to Seasonal Hindcasts of Stratospheric Sudden Warming by BCC_CSM1.1(m):A Comparison with ECMWF

This study focuses on model predictive skill with respect to stratospheric sudden warming(SSW) events by comparing the hindcast results of BCC_CSM1.1(m) with those of the ECMWF's model under the sub-seasonal to seasonal prediction project of the World Weather Research Program and World Climate Research Program. When the hindcasts are initiated less than two weeks before SSW onset, BCC_CSM and ECMWF show comparable predictive skill in terms of the temporal evolution of the stratospheric circumpolar westerlies and polar temperature up to 30 days after SSW onset. However, with earlier hindcast initialization, the predictive skill of BCC_CSM gradually decreases, and the reproduced maximum circulation anomalies in the hindcasts initiated four weeks before SSW onset replicate only 10% of the circulation anomaly intensities in observations. The earliest successful prediction of the breakdown of the stratospheric polar vortex accompanying SSW onset for BCC_CSM(ECMWF) is the hindcast initiated two(three) weeks earlier. The predictive skills of both models during SSW winters are always higher than that during non-SSW winters, in relation to the successfully captured tropospheric precursors and the associated upward propagation of planetary waves by the model initializations. To narrow the gap in SSW predictive skill between BCC_CSM and ECMWF, ensemble forecasts and error corrections are performed with BCC_CSM. The SSW predictive skill in the ensemble hindcasts and the error corrections are improved compared with the previous control forecasts.     

Prospects of using Bayesian model averaging for the calibration of one-month forecasts of surface air temperature over South Korea

In this study, we investigated the prospect of calibrating probabilistic forecasts of surface air temperature (SAT) over South Korea by using Bayesian model averaging (BMA). We used 63 months of simulation results from four regional climate models (RCMs) with two boundary conditions (NCEP-DOE and ERA-interim) over the CORDEX East Asia. Rank histograms and residual quantile-quantile (R-Q-Q) plots showed that the simulation skills of the RCMs differ according to season and geographic location, but the RCMs show a systematic cold bias irrespective of season and geographic location. As a result, the BMA weights are clearly dependent on geographic location, season, and correlations among the models. The one-month equal weighted ensemble (EWE) outputs for the 59 stations over South Korea were calibrated using the BMA method for 48 monthly time periods based on BMA weights obtained from the previous 15 months of training data. The predictive density function was calibrated using BMA and the individual forecasts were weighted according to their performance. The raw ensemble forecasts were assessed using the flatness of the rank histogram and the R-Q-Q plot. The results showed that BMA improves the calibration of the EWE and the other weighted ensemble forecasts irrespective of season, simulation skill of the RCM, and geographic location. In addition, deterministic-style BMA forecasts usually perform better than the deterministic forecast of the single best member.     

Skill of seasonal hindcasts as a function of the ensemble size

 Forecast skill as a function of the ensemble size is examined in a 24-member ensemble of northern winter (DJF) hindcasts produced with the second generation general circulation model of the Canadian Centre for Climate Modelling and Analysis. These integrations are initialized from the NCEP reanalyses at 6 h intervals prior to the forecast season. The sea surface temperatures that are applied as lower boundary conditions are predicted by persisting the monthly mean anomaly observed prior to the forecast period. The potential predictability that is attributed to lower boundary forced variability is estimated. In lagged-average forecasting, the forecast skill in the first two weeks, which originates predominately from the initial conditions, is greatest for relatively small ensemble sizes. The forecast skill increases monotonically with the ensemble size in the rest of the season. The skill of DJF 500 hPa geopotential height hindcasts in the Northern Hemisphere and in the Pacific/North America sector improves substantially when the ensemble size increases from 6 to 24. A statistical skill improvement technique based on the singular value decomposition method is also more successful for larger ensembles. Received: 22 February 2000 / Accepted: 6 December 2000     

Ensemble Forecast： A New Approach to Uncertainty and Predictability

Ensemble techniques have been used to generate daily numerical weather forecasts since the 1990s in numerical centers around the world due to the increase in computation ability. One of the main purposes of numerical ensemble forecasts is to try to assimilate the initial uncertainty （initial error） and the forecast uncertainty （forecast error） by applying either the initial perturbation method or the multi-model/multiphysics method. In fact, the mean of an ensemble forecast offers a better forecast than a deterministic （or control） forecast after a short lead time （3-5 days） for global modelling applications. There is about a 1-2-day improvement in the forecast skill when using an ensemble mean instead of a single forecast for longer lead-time. The skillful forecast （65% and above of an anomaly correlation） could be extended to 8 days （or longer） by present-day ensemble forecast systems. Furthermore, ensemble forecasts can deliver a probabilistic forecast to the users, which is based on the probability density function （PDF） instead of a single-value forecast from a traditional deterministic system. It has long been recognized that the ensemble forecast not only improves our weather forecast predictability but also offers a remarkable forecast for the future uncertainty, such as the relative measure of predictability （RMOP） and probabilistic quantitative precipitation forecast （PQPF）. Not surprisingly, the success of the ensemble forecast and its wide application greatly increase the confidence of model developers and research communities.     

Indian summer monsoon rainfall variability in global coupled ocean-atmospheric models

Performance of seven fully coupled models in simulating Indian summer monsoon climatology as well as the inter-annual variability was assessed using multi member 1 month lead hindcasts made by several European climate groups as part of the program called Development of a European multi-model ensemble system for seasonal-to-inter-annual prediction (DEMETER). Dependency of the model simulated Indian summer monsoon rainfall and global sea surface temperatures on model formulation and initial conditions have been studied in detail using the nine ensemble member simulations of the seven different coupled ocean–atmosphere models participated in the DEMETER program. It was found that the skills of the monsoon predictions in these hindcasts are generally positive though they are very modest. Model simulations of India summer monsoon rainfall for the earlier period (1959–1979) are closer to the ‘perfect model’ (attainable) score but, large differences are observed between ‘actual’ skill and ‘perfect model’ skill in the recent period (1980–2001). Spread among the ensemble members are found to be large in simulations of India summer monsoon rainfall (ISMR) and Indian ocean dipole mode (IODM), indicating strong dependency of model simulated Indian summer monsoon on initial conditions. Multi-model ensemble performs better than the individual models in simulating ENSO indices, but does not perform better than the individual models in simulating ISMR and IODM. Decreased skill of multi-model ensemble over the region indicates amplification of errors due to existence of similar errors in the individual models. It appears that large biases in predicted SSTs over Indian Ocean region and the not so perfect ENSO-monsoon (IODM-monsoon) tele-connections are some of the possible reasons for such lower than expected skills in the recent period. The low skill of multi-model ensemble, large spread among the ensemble members of individual models and the not so perfect monsoon tele-connection with global SSTs points towards the importance of improving individual models for better simulation of the Indian monsoon.     

GRAPES-GEPS环流集合预报的分类释用方法研究与检验

集合预报在数值天气预报体系中具有重要地位,因此如何有效提取集合样本信息以提高集合预报技巧一直是一个重要课题.基于中国全球集合预报业务系统(GRAPES-GEPS)的500?hPa高度场集合资料开展对环流集合预报的分类释用方法研究,并对集合聚类预报结果进行了检验分析.通过在传统Ward聚类法中引入动态聚类的"手肘法"方案...     

Seasonal predictions based on two dynamical models

Abstract

Two dynamical models are used to perform a series of seasonal predictions. One model, referred to as GCM2, was designed as a general circulation model for climate studies, while the second one, SEF, was designed for numerical weather prediction. The seasonal predictions cover the 26‐year period 1969–1994. For each of the four seasons, ensembles of six forecasts are produced with each model, the six runs starting from initial conditions six hours apart. The sea surface temperature (SST) anomaly for the month prior to the start of the forecast is persisted through the three‐month prediction period, and added to a monthly‐varying climatological SST field.

The ensemble‐mean predictions for each of the models are verified independently, and the two ensembles are blended together in two different ways: as a simple average of the two models, denoted GCMSEF, and with weights statistically determined to minimize the mean‐square error (the Best Linear Unbiased Estimate (BLUE) method).

The GCMSEF winter and spring predictions show a Pacific/North American (PNA) response to a warm tropical SST anomaly. The temporal anomaly correlation between the zero‐lead GCMSEF mean‐seasonal predictions and observations of the 500‐hPa height field (Z₅₀₀) shows statistically significant forecast skill over parts of the PNA area for all seasons, but there is a notable seasonal variability in the distribution of the skill. The GCMSEF predictions are more skilful than those of either model in winter, and about as skilful as the better of the two models in the other seasons.

The zero‐lead surface air temperature GCMSEF forecasts over Canada are found to be skilful (a) over the west coast in all seasons except fall, (b) over most of Canada in summer, and (c) over Manitoba, Ontario and Quebec in the fall. In winter the skill of the BLUE forecasts is substantially better than that of the GCMSEF predictions, while for the other seasons the difference in skill is not statistically significant.

When the Z₅₀₀ forecasts are averaged over months two and three of the seasons (one‐month lead predictions), they show skill in winter over the north‐eastern Pacific, western Canada and eastern North America, a skill that comes from those years with strong SST anomalies of the El Niño/La Niña type. For the other seasons, predictions averaged over months two and three show little skill in Z₅₀₀ in the mid‐latitudes. In the tropics, predictive skill is found in Z₅₀₀ in all seasons when a strong SST anomaly of the El Niño/La Niña type is observed. In the absence of SST anomalies of this type, tropical forecast skill is still found over much of the tropics in months two and three of the northern hemisphere spring and summer, but not in winter and fall.