Subseasonal-to-seasonal (S2S) forecasting of the South China Sea (SCS) summer monsoon (SCSSM) onset is critical for agricultural and water resource management in East Asia. While state-of-the-art S2S forecasting systems offer skillful predictions of SCSSM onset up to two pentads in advance, longer-lead-time forecasts remain challenging due to the complex interactions across multiple timescales. Based on reforecasting experiments of the ECMWF S2S forecasting system, we demonstrate that forecast lead times can break through this limitation under two distinct regimes driven by air-land-sea interactions in the Afro-Asian monsoon region. The first regime, associated with decaying El Niño events, is characterized by increased soil moisture in the Horn of Africa two months prior to SCSSM onset. This intensifies the zonal land-sea thermal gradient between East Africa and the Indian Ocean, thereby promoting the eastward propagation of the Madden-Julian Oscillation. The second regime, linked to a cold Indian Ocean and a warm subtropical North Pacific Ocean, leads to a colder Tibetan Plateau and an anomalous meridional circulation around the SCS three pentads before monsoon onset. The intrusion of cold air into the northern SCS reduces atmospheric stability to trigger the monsoon onset convection. Subseasonal predictability of SCSSM onset is enhanced as the forecasting system can predict these regimes more than two pentads in advance. These findings suggest the critical role of air-land-sea interactions in determining regime-dependent subseasonal forecast skill for SCSSM onset, which improves the application of S2S forecasts in decision-making processes.
 

subseasonal prediction skill,South China Sea summer monsoon