Seasonal prediction of extreme high-temperature days over the Yangtze River basin

Abstract

Extreme high temperatures occur frequently over the densely populated Yangtze River basin (YRB) in China during summer, significantly impacting the local economic development and ecological system. However, accurate prediction of extreme high-temperature days in this region remains a challenge. Unfortunately, the Climate Forecast System Version 2 (CFSv2) exhibits poor performance in this regard. Thus, based on the interannual increment approach, we develop a hybrid seasonal prediction model over the YRB (HM_YRB) to improve the prediction of extreme high-temperature days in summer.The HM_YRB relies on the following four predictors: the observed preceding April–May snowmelt in north western Europe; the snow depth in March over the central Siberian Plateau; the CFSv2-forecasted concurrent summer sea surface temperatures around the Maritime Continent; and the 200-hPa geopotential height over the Tibetan Plateau. The HM_YRB indicates good capabilities in predicting the interannual variability and trend of extreme high-temperature days, with a markable correlation coefficient of 0.58 and a percentage of the same sign (PSS) of 76% during 1983–2015 in the one-year-out cross-validation. Additionally, the HM_YRB maintains high PSS skill (86%) and robustness in the independent prediction period (2016–2022). Furthermore, the HM_YRB shows a good performance for years with high occurrence of extreme high-temperature days, with a hit ratio of 40%. These predictors used in HM_YRB are beneficial in terms of the prediction skill for the average daily maximum temperature in summer over the YRB, albeit with biases existing in the magnitude. Our study provides promising insights into the prediction of 2022-like hot extremes over the YRB in China.