Errors in forecasting social phenomena

In this paper, we examine the nature of forecasting errors associated with social phenomena. We introduce the notion of predictive likelihood and conclude a) that there is a need to separate the problem of parametric estimation and inference from the problem of forecasting, b) that all forecasts linked to decisions require at least an implicit structural model, c) that conventional cost functions have adversely affected the quality of social science forecasts, and d) that there is a need to develop forecasting procedures robust to different kinds of cost functions.     

Neuroid BP-type Model Applied to the Study of Monthly Rainfall Forecasting

A neuroid BP-type three-layer mapping model is used for monthly rainfall forecasting in terms of 1946-1985 Nanjing monthly precipitation records as basic sequences and the model has the form i × j ＝ 8 × 3, K = 1; by steadily modifying the weighing coefficient, long-range monthly forecasts for January to December, 1986 are constructed and 1986 month-to-month predictions are made based on, say, the January measurement for February rainfall and so on, with mean absolute error reaching 6,07 and 5,73 mm, respectively. Also, with a different monthly initial value for June through September, 1994, neuroid forecasting is done, indicating the same result of the drought in Nanjing dur-ing the summer, an outcome that is in sharp agreement with the observation.     

The Meta-Gaussian Bayesian Processor of Forecasts and Associated Preliminary Experiments

Public weather services are trending toward providing users with probabilistic weather forecasts, in place of traditional deterministic forecasts. Probabilistic forecasting techniques are continually being improved to optimize available forecasting information. The Bayesian Processor of Forecast (BPF), a new statistical method for probabilistic forecast, can transform a deterministic forecast into a probabilistic forecast according to the historical statistical relationship between observations and forecasts generated by that forecasting system. This technique accounts for the typical forecasting performance of a deterministic forecasting system in quantifying the forecast uncertainty. The meta-Gaussian likelihood model is suitable for a variety of stochastic dependence structures with monotone likelihood ratios. The meta-Gaussian BPF adopting this kind of likelihood model can therefore be applied across many fields, including meteorology and hydrology. The Bayes theorem with two continuous random variables and the normal-linear BPF are briefly introduced. The meta-Gaussian BPF for a continuous predictand using a single predictor is then presented and discussed. The performance of the meta-Gaussian BPF is tested in a preliminary experiment. Control forecasts of daily surface temperature at 0000 UTC at Changsha and Wuhan stations are used as the deterministic forecast data. These control forecasts are taken from ensemble predictions with a 96-h lead time generated by the National Meteorological Center of the China Meteorological Administration, the European Centre for Medium-Range Weather Forecasts, and the US National Centers for Environmental Prediction during January 2008. The results of the experiment show that the meta-Gaussian BPF can transform a deterministic control forecast of surface temperature from any one of the three ensemble predictions into a useful probabilistic forecast of surface temperature. These probabilistic forecasts quantify the uncertainty of the control forecast; accordingly, the performance of the probabilistic forecasts differs based on the source of the underlying deterministic control forecasts.     

A Regional Ensemble Forecast System for Stratiform Precipitation Events in Northern China. Part I: A Case Study

A single-model,short-range,ensemble forecasting system (Institute of Atmospheric Physics,Regional Ensemble Forecast System,IAP REFS) with 15-km grid spacing,configured with multiple initial conditions,multiple lateral boundary conditions,and multiple physics parameterizations with 11 ensemble members,was developed using the Weather and Research Forecasting Model Advanced Research modeling system for prediction of stratiform precipitation events in northern China.This is the first part of a broader research project to develop a novel cloud-seeding operational system in a probabilistic framework.The ensemble perturbations were extracted from selected members of the National Center for Environmental Prediction Global Ensemble Forecasting System (NCEP GEFS) forecasts,and an inflation factor of two was applied to compensate for the lack of spread in the GEFS forecasts over the research region.Experiments on an actual stratiform precipitation case that occurred on 5-7 June 2009 in northern China were conducted to validate the ensemble system.The IAP REFS system had reasonably good performance in predicting the observed stratiform precipitation system.The perturbation inflation enlarged the ensemble spread and alleviated the underdispersion caused by parent forecasts.Centering the extracted perturbations on higher-resolution NCEP Global Forecast System forecasts resulted in less ensemble mean root-mean-square error and better accuracy in probabilistic quantitative precipitation forecasts (PQPF).However,the perturbation inflation and recentering had less effect on near-surface-level variables compared to the mid-level variables,and its influence on PQPF resolution was limited as well.     

集合预报的现状和前景

综合论述了近年来已在国际上引起高度重视的新一代动力随机预报方法 ——— 集合预报。 随着计算机技术的迅猛发展和由于大气初值和数值模式中物理过程存在着不确定性的事实, 这一方法无疑代表了数值天气预报未来演变发展的方向。 未来的天气预报产品预计将从“决定论”的预报转变为“随机论”的预报来正确地表达气象科学中这一所谓“可预报性问题”, 以便更好地为用户服务。 文中扼要地叙述了集合预报的概念、基本问题及其最新的研究动态和发展, 包括(1)如何建立和评估一个集合预报系统;(2)如何正确地表征大气初值和模式物理过程的不确定性与随机性;(3)如何从集合预报中提炼有用的预报信息和合理地解释、检验集合预报的产品, 特别是概率预报。 除了直接在天气预报上的应用, 还提到集合预报在气象观测和资料同化方面应用的动态, 以引起有关研究人员的注意。     

Editor's note

Any opinions, findings, conclusions or recommendations expressed in this editorial are those of the author and not necessarily reflect the views of the National Science Foundation.     

Solar,volcanic, and CO2 forcing of recent climatic changes

The climate, as represented by the mean Northern Hemisphere temperature, has shown substantial changes within the past century. The temperature record is utilized as a means of elucidating the relative importance of anthropogenic CO₂ increase, volcanic aerosols, and possible solar insolation variations in externally forcing climate changes. Solar luminosity variations, suggested by observed solar radius variations on an ≈ 80 yr time scale, allow a self-consistent explanation of the hemispheric temperature trends. Evidence for solar influences on the climate is also found on the shorter 11 and 22 yr time scales present in solar activity cycles. The author is a staff scientist at the High Altitude Observatory, P.O. Box 3000, Boulder, CO 80307, of the National Center for Atmospheric Research. This work was completed while the author was a postdoctoral fellow in the Advanced Study Program of NCAR. Any opinions, findings and conclusions or recommendations expressed in this paper are those of the author and do not necessarily reflect the views of the National Science Foundation. The National Center for Atmospheric Research is sponsored by the National Science Foundation.     

美国天气预报的效果与预报估价

本文介绍美国天气预报的效果与对预报的估价。根据检验美国国家天气局从1967年到1985年逐日天气预报的效果表明，目前的预报水平明显高于15到20年前的预报水平。特别是客观预报效果明显的提高是值得引起注意的。预报的估价是一个很复杂的问题。本文主要介绍了美国的气象预报正确与否的得失估价及给经济带来获益或损失的经济评价工作。     

预报员在未来天气预报中的作用探讨

在没有数值天气预报的年代 ,天气预报是由预报员作出的 ,随着数值天气预报业务水平的不断提高 ,它首先取代了预报员的天气形势预报 ,目前正进入取代部分天气要素预报的时代。未来 1 0～ 2 0年中尺度数值天气预报和集合预报以及后处理技术将获得重大进展 ,客观预报有可能取代目前预报员的一些预报项目 ,但是预报员的作用仍然是不可缺少的 ,但需要按与时俱进的观点 ,认真分析客观预报与主观预报的最佳结合点 ,以不断提高天气预报水平     

北京地区中尺度数值业务预报的客观检验

对北京地区中尺度数值天气预报客观检验系统的框架和方法进行了详细的介绍，分析了北京地区冬季（2001年11月1日~2002年2月28日）主要预报要素的客观检验结果，在此基础上给出北京地区冬季中尺度数值预报业务的误差特征，并根据检验结果定性地分析了模式预报系统性和非系统性误差的可能来源。结果表明，模式地形的强迫作用是地面和对流层低层预报系统性误差的重要来源；测站高空预报要素的非系统性误差显著，初始时刻存在的误差在积分过程中被进一步放大；对北京 (54511) 单站地面要素预报的检验结果发现，2-m温度预报冬季整体偏低。在系统性误差为主要误差分量时对模式的平均误差进行订正能有效地提高温度预报的准确率。     

The social forecasting industry

This is an essay and report on the social forecasting industry-the collection of individuals and organizations, both profit and nonprofit, supplying forecasts of social futures in direct exchage for pay. It begins with a review of the industry's size, growth, and scope; contrary to initial impressions, the industry is by any comparative standard quite small, occupying a precarious niche in the much larger market for statistical services. Hallmark methodologies of the industry, including the Delphi method and cross-impact analysis, are discussed extensively. These methodologies are linked to a domain assumption of the industry—namely, that the complexities of the social world are best rendered tractable by forming consensuses of experts' judgments. It is argued that the key characteristic differentiating the social forecasting industry from forecasting in the social sciences proper—the existence of a direct, paying market for services—is an important element in maintaining the most salient methodological and product differences between the two arenas: The marketplace for forecasts is not the same as the marketplace for ideas.     

The Kansas and Minnesota experiments

Two experiments conducted almost 20 years ago have come to be regarded as important milestones in boundary-layer research. This paper recounts the motivation for those experiments, describes how they were conducted, and summarizes their results.The National Center for Atmospheric Research is sponsored by the National Science Foundation.     

Studies on Non-interpolating Semi-Lagrangian Scheme and Numerical Solution to KdV Equation

A new non-interpolating semi-Lagrangian scheme has been proposed, which can eliminate any interpolation, and consequently numerical smoothing of forecast fields. Here the new scheme is applied to KdV equation and its performance is assessed by comparing the numerical results with those produced by Ritchie’s scheme (1986). The comparison shows that the non-interpolating semi-Lagrangian scheme appears to have efficiency advantages.     

集合模式定量降水预报的统计后处理技术研究综述

集合数值模式预报已在定量降水预报业务中广泛应用,以获得预报不确定性、最可能预报结果以及极端天气预警。由于集合系统的数值模式不完善,且不能提供所有的不确定性信息,常表现出系统性偏差以及欠离散或过离散（如对于多模式集合）。为此,需要发展统计后处理技术,在尽量保持集合预报解析度的条件下,提高预报的技巧和可靠性。近年来,各种集合预报统计后处理技术得到快速发展。针对定量降水预报,依据技术方法的途径和成熟度将后处理研究归纳为3方面进行总结,包括:（1）不基于统计模型的非参数化后处理,包括集合定量降水预报偏差订正、多成员或模式信息集成以及基于空间分析的对流尺度模式后处理;（2）基于概率分布统计模型的参数化后处理,包括集合模式输出统计和贝叶斯模型平均两种方法框架;（3）考虑预报量的时间、空间和多变量间依赖关系或结构的处理方法,包括参数化和经验连接概率法。最后,讨论发展统计后处理技术需要关注的问题,包括考虑不同来源、不同尺度的多模式信息集成;提供高质量、高分辨率的降水分析资料;发展再预报技术扩充训练样本;基于不同的订正目的和应用场景来使用不同的后处理技术;发展面向海量预报数据、捕捉极端降水以及考虑预报量结构的新技术。      

Multischeme ensemble forecasting of surface temperature using neural network over Turkey

The ensemble method has long been used to reduce the errors that are caused by initial conditions and/or parameterizations of models in forecasting problems. In this study, neural network (NN) simulations are applied to ensemble weather forecasting. Temperature forecasts averaged over 2 weeks from four different forecasts are used to develop the NN model. Additionally, an ensemble mean of bias-corrected data is used as the control experiment. Overall, ensemble forecasts weighted by NN with feed forward backpropagation algorithm gave better root mean square error, mean absolute error, and same sign percent skills compared to those of the control experiment in most stations and produced more accurate weather forecasts.     

K 近邻非参数回归概率预报技术及其应用

针对参数回归技术制作概率预报存在拟合好、但预报结果不稳定的现象, 提出了用K近邻非参数回归技术制作概率预报的新途径。K 近邻非参数回归技术包括历史样本数据库、近邻子集生成和优化以及预报量估计4 个主要部分。利用该技术进行了单要素概率预报(主要包括云量和降水)和多维联合概率预报(降水、总云量、风速和气温)试验, 并对试验结果进行了检验。实例研究结果表明:该文所给出的计算方案预报稳定性好, 准确率较高,具有良好的业务应用价值。     

Quantifying the value of historical climate knowledge and climate forecasts using agricultural systems modelling

A well tested agricultural systems model was used together with 114 years of historical climate data to study the performance of a dryland wheat–fallow system as impacted by climate variations and nitrogen input levels in southeast Australia, and to investigate the value of: (1) historical climate knowledge, (2) a perfect climate forecast, and (3) various forecasts of targeted variables. The potential value of historical climate records increases exponentially with the number of years of data. In order to confidently quantify the long term optimal nitrogen application rate at the study site at least 30 years of climate data are required. For nitrogen management only, the potential value of a perfect climate forecast is about $54/ha/year with a reduction of excess nitrogen application of 20 kg N/ha/year. The value of an ENSO based forecast system is $2/ha/year. Perfect forecasting of three or six categories of growing season rainfall would have a value of $10–12/ha/year. Perfect forecasts of three or six categories of simulated crop yield would bring about $33–34/ha/year. Choosing integrated variables as a forecasting target, for example crop yield derived from agricultural modelling, has the potential to significantly increase the value of forecasts.     

Assessing the impact of climatic change on food production

Attempts to assess the impact of a hypothetical climatic change on food production have relied on the use of statistical models which predict crop yields using various climatic variables. It is emphasized that the coefficients of these models are not universal constants, but rather statistical estimates subject to several sources of error. Thus, any statement regarding the estimated impact of climatic change on food production must be qualified appropriately.The National Center for Atmospheric Research is sponsored by the National Science Foundation.     

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.