Evaluation of the Forecast Performance for Extreme Cold Events in East Asia With Subseasonal‐to‐Seasonal Data Sets From ECMWF

Abstract

is often subjected to the influence of extreme cold events during the boreal winter. Observations show that the frequency of winter extreme cold events has increased during the past two decades and East Asia has experienced several severe extreme cold events in the past few winters (Cohen et al., 2020; Johnson et al., 2018). For example, an unprecedented low-temperature and freezing disaster occurred in southern China in early January 2008, which caused hundreds of human deaths and great economic loss (Ding et al., 2008; Tao & Wei, 2008). Record-breaking blizzards and low temperatures frequently affected many areas in East Asia during the winter of 2010-2011 (Gong et al., 2014). In January 2016, a strong cold event occurred in East Asia and snowfall was observed for the first time during past 115 years in Amami-Oshima (Qian et al., 2018; Yamaguchi et al., 2019). Moreover, the winter of 2017-2018 was extremely cold in East Asia and many countries, including China, Japan, and Korea, experienced record-breaking low temperatures (Tachibana et al., 2019). All of these extreme cold events have caused great damage to human lives, agriculture, transportation, and power infrastructure in East Asia. Skillful forecasts of these extremes may have benefits for effective hazard preparedness and risk management. The lead time for skillful forecasts of atmospheric motion is limited owing to its chaotic features. The study of atmospheric predictability was first recognized by Lorenz (1963). After that, he investigated the behavior of error growth and found that motions in small scales have a shorter saturation time by utilizing a two-dimensional barotropic vorticity equation (Lorenz, 1969). His results suggest that the intrinsic predictability of the atmosphere is about 16.8 days. He also pointed out that it seems possible to forecast the instantaneous weather patterns with a lead time of 2 weeks by investigating the operational products from the European Centre for Medium-Range Weather Forecasts (ECMWF) (Lorenz, 1982). However, recent studies have shown that skillful forecast lead times may exceed 2 weeks (Buizza & Leutbecher, 2015; Xiang et al., 2018), which indicates that it may be possible to achieve long lead times for the skillful forecast of extreme events. The predictability of weather is mainly based on the atmospheric conditions, and weather forecasts on short timescales usually depend on the accuracy of the atmospheric initial conditions. But for timescales longer than one season, the boundary conditions play the principal role in atmospheric circulation prediction.

The unprecedented Pacific Northwest heatwave of June 2021

“…Forecasts and reforecasts are initialized twice a week. Previous evaluations of ECMWF indicate good performance up to lead times of 7-10 days for extreme cold events 93 , up to 11 days for daily precipitation 94 , up to 15 days for heatwaves 95,96 , and up to 20 days for soil moisture 97 .…”

Influence of Arctic Sea Ice Concentration on Extended‐Range Prediction of Strong and Long‐Lasting Ural Blocking Events in Winter

“…Dai, Mu, Li et al. (2021) proposed that certain long and persistent extreme cold events may exhibit a skillful forecast lead time longer than 2 weeks. Besides, Liu et al.…”

“…However, Buizza and Leutbecher (2015) pointed out that forecast skills longer than 2 weeks are achievable owing to major advances in data assimilation, modeling and ensemble techniques. Dai, Mu, Li et al (2021) proposed that certain long and persistent extreme cold events may exhibit a skillful forecast lead time longer than 2 weeks. Besides, Liu et al (2016) indicated that boundary conditions are important for the forecast at extended-range timescales.…”

Extent of the Impact of Arctic Atmospheric Uncertainty on Extended‐Range Forecasting of Cold Events in East Asia

“…Previous research suggests that long‐lasting phenomena may have a long predictability that can reach four pentads or even beyond (Buizza & Leutbecher, 2015; Dai et al., 2021; Vitart & Robertson, 2018). Dai et al.…”

“…minus one standard deviation was about 12 days (Dai et al, 2021). In the winter of 2020/2021, three successive cold surges adversely affected East Asia, in which the daily mean temperature in many regions of China, Korea and Japan dropped below the 10th percentile of the period 1979-2019 (Dai et al, 2022;X.…”