Exploring Pathways for Improving MJO Predictions

“The Madden-Julian Oscillation (MJO) is the dominant mode of tropical convection variability on the intraseasonal time scale. The recurrent nature of the MJO with a period of 30-60 days offers an opportunity to bridge the gap between weather forecasting and seasonal prediction. Over the recent years, significant improvements have been made in MJO prediction skill in operational forecasts. However, the MJO prediction performance still differs greatly among the operational systems and efforts are continually underway to seek further improvements. Furthermore, a general guidance on what would be most beneficial developmental pathways to improve MJO simulation and prediction skill remains unclear. Addressing the issues for improving MJO forecasts requires quantifying the sensitivity of MJO predictions to different factors that affect the prediction, and quantifying the relative importance of individual factors. Even though the existing subseasonal to seasonal (S2S) prediction data sets provide an opportunity to evaluate the current capability in predicting atmospheric and oceanic variability at S2S time scale, isolating influences of individual factors based on the existing datasets alone is not sufficient, because various models that were used to produce these databases have different resolutions, initialization procedures and model physics. The objective of this project is to study the influence of different aspects of forecast configurations and their relative importance on the MJO prediction based on a perfect model framework with the CFSv2.

To achieve the objective, the following steps will be taken: 1) A set of long-term control simulations with different model configurations for the atmospheric resolution, convection, and ocean component will first be performed. From them, a control configuration that produces the most realistic MJO simulation will be selected. 2) Potential predictability of the MJO will be assessed for the selected control configuration based on prediction runs with slight perturbations to atmospheric initial conditions taken from its long-term simulation. 3) A suite of forecast experiments will be done with changes to various aspects to the control configuration to determine (and understand) the sensitivity of the MJO predictions to those changes. Our focus will be on the changes in model resolution, convective parameterization, and representation of air-sea coupling. We will work toward isolating the most important factors governing MJO prediction performance.

We anticipate that the proposed research will identify the beneficial development pathways to further improve MJO predictions. It will provide guidance for improving the next generation CFS and other coupled forecast system models in the climate community. Results from this project will also help improve our understanding of the processes related to the MJO dynamics and how to properly represent them in the coupled climate models.”