Verification and Analysis of the Impact of Cold Wave Weather Process on the Numerical Prediction Skills of Wind/Photovoltaic Power Resource Elements

Based on the reanalysis data of ERA-5 and the 120-h prediction data of Global Forecast System (GFS), which is the NCEP prediction system, this paper examines the prediction skills of near-surface wind speed and net downward short-wave radiation flux of the numerical model during nine cold waves in East China from December 2020 to March 2021. The following test results are obtained: 1) The GFS prediction system can accurately predict the cold wave process (cooling range and minimum temperature) 1–4 days in advance, with an average prediction accuracy > 80%. 2) The near-surface wind speed increases significantly during the cold wave. Even though the Threat Score (TS) of the wind speed of grades 0–2 is significantly low, the TS score of the wind speed of grades 3–5 and >6 is higher than the general weather process. The relative error in the prediction of the net downward short-wave radiation flux is larger than the general weather processes, particularly on the day of the cold wave outbreak. 3) The prediction skills exhibit significant diurnal variation, during the cold wave. The prediction skill of the wind speed of grades 0–2 is the lowest in the afternoon, particularly on the strongest day of the cold wave. The TS score of the wind speed of grades 3–5 is generally the lowest at around 1800 LST and is extremely low at night on the strongest day of the cold wave. After the afternoon, particularly on the day of the cold wave outbreak, the error in the prediction of the net downward short-wave radiation flux increases significantly. 4) The prediction skills decrease and the prediction time increases during the cold wave. The 24-h prediction has the highest TS score and the minimum error, while the 72-h prediction has the lowest TS score and the maximum error.