Accurate weekly rainfall predictions are essential for effective disaster prevention and sustainable development. However, the reliability of current forecasting systems declines sharply beyond the short-term horizon, posing challenges for their application in long-term planning and preparedness. To address this limitation, we propose the Windows of Opportunity Identification (WOI) approach, a unified framework that enhances prediction reliability by identifying physical signals as precursors for more precise forecasts within target models. For instance, within the European Centre for Medium-Range Weather Forecasts (ECMWF) model, the WOI method objectively identifies the tropical Madden-Julian Oscillation (MJO) and mid-latitude waves as key precursors for accurately forecasting weekly East Asian Summer Monsoon (EASM) rainfall. When strong WOI signals are detected, prediction accuracy can surpass 0.5 for forecasts extending from week 2 to week 4. This approach illustrates how leveraging physical signals and the capabilities of existing models can significantly enhance medium-range predictive performance.
 

Sub-seasonal prediction, windows of opportunity identification, sustainable development