The Extreme Forecast Index (EFI) calculations are performed using the ECMWF 2-m air temperature forecasts produced in the framework of the Subseasonal to Seasonal (S2S) Prediction Project. Four computation schemes are implemented using empirical and theoretical distributions of heat wave characteristics as well as the one-dimensional test statistics of histograms and linear interpolation formulas. Case studies (for different initial dates and regions) characterized by the significant air temperature anomalies in Northern Eurasia are performed using traditional forecast skill scores and the spatial verification methods to evaluate the efficiency of the proposed schemes for different threshold values of EFI. It is shown that the forecast quality can be considered satisfactory in most cases. The dependence of forecast skill on the intensity, spatial scales, and duration of temperature anomalies is revealed. Further studies should be carried out using larger samples based on several hydrodynamic models and the multimodel approach.

中文翻译：

---

关于肢体指数EFI的评估

极端预报指数（EFI）的计算使用在次季节到季节性（S2S）预测项目框架内生成的ECMWF 2-m气温预测。利用热波特性的经验和理论分布以及直方图和线性插值公式的一维检验统计量，实现了四种计算方案。使用传统的预报技能得分和空间验证方法，对以欧亚大陆北部明显气温异常为特征的案例研究（针对不同的初始日期和地区）进行评估，以评估针对不同的EFI阈值提出的方案的效率。结果表明，在大多数情况下，预报质量可以认为是令人满意的。预测技能对强度的依赖，揭示了空间尺度和温度异常的持续时间。应基于几个流体动力学模型和多模型方法，使用较大的样本进行进一步的研究。