Host: Baoqiang Xiang
Subseasonal prediction (between 10 days and one season) has garnered great interest in recent decades as a focused effort to bridge the gap between the weather and seasonal forecasts. As the dominant intraseasonal mode over the tropics, the Madden-Julian Oscillation (MJO) is one of the major predictability sources for S2S predictions. Recently, we have developed a subseasonal forecast system using the GFDL SPEAR model, and the model shows encouraging MJO prediction skills. Different from many previous studies focusing on assessing the prediction of MJO propagation features, here we propose to investigate the development of errors of MJO-related convections over the Indian Ocean-Western Pacific sector and try to track the origins of the forecast errors within the modeling system. A variety of statistical methods and machine learning techniques (such as K-means cluster analysis) will be used to study the proposed issue. We hope this project will lead to an improved understanding of the so-called maritime continent prediction barrier issue as well as the root causes of the model biases limiting the MJO prediction. It is also expected to provide some guidance for the further development of the GFDL subseasonal forecast system.