Real-time MJO Identification and Subseasonal Predictions

There is evidence that the Madden-Julian Oscillation (MJO) can enhance predictability of sensible weather possibly out to lead times of ~40-50 days. This includes high impact events such as extreme surface air temperatures and extended periods of severe storms/excessive precipitation. However, identification of MJOs in real-time is challenging because the weather-climate system is dominated by noise. In fact, stochastic extratropical dynamics can organize large scale subtropical wind fields and envelopes of tropical rainfall that can be misrepresented as MJOs especially when using combined wind and outgoing longwave radiation indices (Wheeler and Hendon, 2004). These mixed global wind-tropical convective variations (Weickmann and Berry, 2009) partially reflect the Global Wind Oscillation (GWO), which is non-oscillatory and does not provide useful forecast information beyond ~15 days.

An analysis of outgoing longwave radiation (OLR) and global circulation data sets is performed for the 2006-07 through 2011-12 boreal cold seasons (which included the DYNAMO field experiment). During this roughly six year period nine MJOs were identified that had the potential to extend the range of skillful or useful prediction. The breakdown of these events and their interactions with ENSO are discussed. Examples of red noise dominated variations and predictability ramifications will also be given. These include the premature ending of the 2006-07 El-Niño, extratropical feedbacks during 2010-11 leading to a strong jet stream not consistent with La-Niña, and constructive interference of an MJO and La-Niña that contributed to the March 2012 Midwest-eastern USA “heat wave”. The criticality to distinguish in real time between “sustained, coherent MJOs” and other types of coherent or noisy tropical-extratropical variability is emphasized. The fast, moderate MJOs that initiated during DYNAMO might provide clues about MJO initiation and succession.