中期气温预报评估和校正的新框架

Increasing the skill of short-term wind speed ensemble forecasts combining forecasts and observations via a new dynamic calibration

All numerical weather prediction models used for the wind industry need to produce their forecasts starting from the main synoptic hours 00, 06, 12, and 18 UTC, once the analysis becomes available. The six-hour latency time between two consecutive model runs calls for strategies to fill the gap by providing new accurate predictions having, at least, hourly frequency. This is done to accommodate the request of frequent, accurate and fresh information from traders and system regulators to continuously adapt their work strategies. Here, we propose a strategy where quasi-real time observed wind sp…

超越技能分数：通过法国前一个月能源预测案例研究探索次季节预测价值

Beyond skill scores: exploring sub-seasonal forecast value through a case study of French month-ahead energy prediction

我们量化了真实世界预测问题的亚季节预测的价值：法国未来一个月能源需求的预测。使用地面温度作为预测因子，我们构建了一个交易策略，并根据实际能源需求和价格数据评估了使用气象预测的财务价值。我们表明，预测的提前时间大于2周，可以有价值的应用程序，无论是在自己的短期预测，特别是在北方冬季。我们认为这个例子的基础上的成本/损失框架，并表明，虽然它捕捉的短期预测的性能，它错过了在长期预测的边际价值。我们还将我们对预测值的评估与传统技能评分进行了对比，我们发现如果单独使用，可能会产生误导。我们强调的重要性，基于最终用户实际使用的变量的预测技能的评估。

一种新的低成本技术改善了全球天气预报

A new low-cost technique improves weather forecasts across the world

计算机生成的预测将地球表面划分为网格，每个网格现在约为伦敦大小的25%，每个网格预测一个值。如果天气在一个网格框内变化明显，那么对特定地点的预报不可避免地会失败。一个全新的统计后处理方法，使用集合预报作为输入，预计两个gridbox天气依赖的因素：在每个gridbox的变化程度，并在gridbox规模的偏见。在全球范围内，技能大大提高;例如，对于极端降雨，有用的预报提前5天。如果没有后处理，此限制小于1天。相对于历史预测的进步，这是一个突破性的进步。在校准过程中纳入的关键驱动因素是对气象的了解，并放弃了只使用当地数据的传统观念。相反，我们简单地认识到“阵雨就是阵雨，无论它们发生在世界各地”，这使得校准数据集的大小大幅增加。许多多方面的应用包括改进的山洪预警，对模型弱点的物理相关见解和全球逐点重新分析。

关于评估观测在对流可分辨数值天气预报中的影响与作用的方法

On methods for assessment of the influence and impact of observations in convection-permitting numerical weather prediction

本文比较评估对流可分辨NWP中观测价值的方法，聚焦集合预报系统适用性及业务考量，并指出未来研究方向。

温度的概率预报：对贝叶斯模型平均法的评论

Probabilistic forecasting of temperature: comments on the Bayesian Model Averaging approach

最近提出了一个具体的实施贝叶斯模型平均作为一种方法，集合温度预报的校准。我们指出这种新方法和早期的方法称为内核回归之间的相似之处。我们还认为，贝叶斯模型的平均方法（如应用）有一些缺陷，将导致次优校准的平均值和不确定性的预测。

利用确定性天气预报通过深度学习进行原位概率温度预测

Leveraging deterministic weather forecasts for in-situ probabilistic temperature predictions via deep learning

提出一种神经网络方法，通过为多预报时效训练单一模型，从确定性数值天气预报生成概率预报，将CRPS降低15%并改善分布校准，无需模拟多轨迹即可为任意网格预报增加不确定性信息。

概率温度预报：四种扩散-回归模型的比较

Probabilistic temperature forecasting: a comparison of four spread-regression models

扩散回归是线性回归的扩展，允许包含包含方差信息的预测因子。它可用于从天气预报集合中获取信息，并对未来温度进行概率预测。有许多方法可以详细地制定扩散回归。我们进行实证比较的四个最明显的方法，适用于校准一年的ECMWF温度预测伦敦希思罗机场。

Novel strategies of Ensemble Model Output Statistics (EMOS) for calibrating wind speed/power forecasts

The issue of the accuracy of wind speed/power forecasts is becoming more and more important as wind power production continues to increase year after year. Having accurate forecasts for the energy market clashes with intrinsic difficulties of wind forecasts due to, e.g., the coarse resolution of Numerical Weather Prediction models. Here, we propose a novel Ensemble Model Output Statistics (EMOS) which accounts for nonlinear relationships between predictands and both predictors and other weather observables used as conditioning variables. The strategy is computationally cheap and easy-to-implem…

概率中期温度预报需要考虑非正态性吗？

Do probabilistic medium-range temperature forecasts need to allow for non-normality?

高斯扩散回归模型用于特定站点集合温度预报的校准依赖于温度预报的不确定性呈正态分布这一明显的限制性假设。我们使用核密度来概括模型，以允许更灵活的分布形状。然而，我们没有发现任何有意义的改善，在由此产生的概率预测时，使用基于可能性的分数进行评估。我们得出结论，不确定性的分布要么非常接近正态分布，要么不接近正态分布，那么我们测试的集合预报就不会预测到非正态分布。

Systematic assessment of the effects of space averaging and time averaging on weather forecast skill

Intuitively, one would expect a more skillful forecast if predicting weather averaged over one week instead of the weather averaged over one day, and similarly for different spatial averaging areas. However, there are few systematic studies of averaging and forecast skill with modern forecasts, and it is therefore not clear how much improvement in forecast performance is produced via averaging. Here we present a direct investigation of averaging effects, based on data from operational numerical weather forecasts. Data is analyzed for precipitation and surface temperature, for lead times of rou…

Uniform reliability tests for forecasting systems with small lead time

A long noted difficulty when assessing the reliability (or calibration) of forecasting systems is that reliability, in general, is a hypothesis not about a finite dimensional parameter but about an entire functional relationship. A calibrated probability forecast for binary events for instance should equal the conditional probability of the event given the forecast for {\em any} value of the forecast. Attempts to estimate deviations from calibration at a specific forecast value meet with the difficulty that the probability of the forecast assuming that value is typically zero. Considering the …

基于对数正态分布的概率风速预测EMOS模型

Log-normal distribution based EMOS models for probabilistic wind speed forecasting

预报集合是从具有不同初始条件的数值天气预报模型的多次运行中获得的，通常用于解释预报的不确定性。然而，预报集合中经常出现偏差和离散误差，它们通常是欠离散和未校准的，需要统计后处理。我们提出了一个包围模型输出统计（EMOS）方法校准风速预测的对数正态（LN）分布的基础上，我们还显示了一个政权切换扩展的模型，结合了以前研究的截断正态（TN）分布与LN。这两个模型被应用到风速预测的8名成员的华盛顿大学中尺度合奏，50名成员的ECMWF合奏和11名成员的ALADIN-HUNEPS合奏的匈牙利气象服务，和他们的预测性能进行比较的TN和一般极值（GEV）分布为基础的EMOS方法和TN-GEV混合模型。结果表明，改进校准的概率和准确性的点预测相比，原始集合和气候预测。此外，TN-LN混合模型优于传统的TN方法，其预测性能能够跟上利用GEV分布的模型，而无需将质量分配给负值。

MAUSAM：南亚季风期间全球AI天气预报模型的以观测为中心的评估

MAUSAM: An Observations-focused assessment of Global AI Weather Prediction Models During the South Asian Monsoon

本文提出MAUSAM评估框架，利用地面站、雨量计和卫星观测检验七个AI天气模型在南亚季风期的表现，发现其技巧显著但精细尺度误差仍存，且再分析基准会高估性能。