Seasonal Prediction of Spring Drought over Northeast China

In recent decades, northeast China (NEC) has suffered increasingly severe spring droughts. Accurate short-term climate prediction for drought remains a great challenge. Using the standardized precipitation evapotranspiration index (SPEI), we develop a seasonal prediction model for spring drought in NEC during 1980–2022 by combining the year-to-year difference and EOF analysis methods. We construct two prediction models separately for the first two EOF modes. The preceding December SST in the subtropical Atlantic and autumn soil moisture in eastern Europe trigger significant Rossby wave trains propagating to northeast Asia, leading to a simultaneous distribution of drought anomalies in NEC through modulation of the Mongolian anticyclone. The early winter snow cover in East Asia, preceding September soil moisture in southwest China, and preceding December Beaufort Sea ice are associated with surface heat flux anomalies, vertical motion anomalies, and cyclonic anomalies over northeast Asia, contributing to the north–south dipole pattern of spring drought in NEC. Remarkable positive correlations between the predicted and observed SPEI are prevalent across NEC, with the maximum reaching 0.72. Additionally, the predicted and observed geographical SPEI distributions are highly consistent during eight extreme drought years, with pattern correlation coefficients exceeding 0.70 for six of these. In independent hindcasts, pattern correlation coefficients exceed 0.90 for two extreme drought years, 2017 and 2019. Seasonal drought predictions provide valuable insights to formulate strategies for disaster prevention and mitigation in NEC. Significance Statement Drought is a destructive meteorological phenomenon with a substantial impact on human lives. Northeast China, one of China’s largest agricultural regions for cereals, has recently experienced severe and frequent spring droughts. Based on the standardized precipitation evapotranspiration index (SPEI), we develop a statistical prediction model for spring drought over northeast China during 1980–2022 by combining the year-to-year difference and EOF analysis methods. We calculate significant positive correlations between the observed and predicted SPEI values, with a maximum of 0.72. In independent hindcasts, pattern correlation coefficients between the observations and predictions exceed 0.90 in two extreme drought years. Seasonal prediction of spring drought is considerably significant for disaster prevention and mitigation.