Abstract 096

Abstract ID: 096

The Indo-Pacific Maritime Continent Barrier Effect on MJO Prediction in Subseasonal-to-Seasonal Prediction Models

Lead Author: S. Abhik
School of Earth, Atmosphere, and Environment, Monash University, Australia, Australia

Keywords: Maritime Continent, Madden–Julian oscillation, S2S Forecast verification/skill, Tropical subseasonal variability, Atmosphere–ocean interaction

Abstract: The eastward propagating Madden-Julian Oscillation (MJO) is influenced by the barrier effect across the Indo-Pacific Maritime Continent (MC) islands: MJO convection weakens or sometimes completely dissipates upon encountering the archipelago. This MC barrier effect is generally overestimated in most of the current general circulation models. Here, we examine the impact of the MC barrier effect on the MJO prediction during the austral summer season using retrospective forecasts from 13 subseasonal-to-seasonal (S2S) prediction systems. The overall MJO prediction skill for December-March start dates of all S2S models is evaluated and the top prediction systems are identified based on bivariate correlation, root-mean-square-error (RMSE), and real-time multivariate MJO (RMM) phase error. This skill assessment is extended to the prediction of blocked and continuously propagating MJO events over the MC islands. Consistent with our earlier study, the prediction skill of the continuously propagating MJOs is significantly higher than the blocked MJOs in the S2S models. The higher prediction skill for the continuous MJO events is mostly insensitive to the higher initial RMM amplitude of the continuous MJO events. Rather, the greater prediction skill is shown to be related to the more persistent behavior of the continuously propagating MJOs to the eastern flank of the MC islands. Our forecast skill assessment suggests that the lifetime of the predicted MJOs initialized over the Indian Ocean is similar to the MJOs initialized over the western Pacific. The forecast skill difference between blocked and continuous MJO events is also evaluated in a multi-model ensemble by utilizing all the ensemble members of the ‘top models’. The ‘top model’ forecasts reasonably capture the observed enhanced subsidence over the colder sea-surface temperature in the central Pacific that is associated with the blocked MJO events. The role of a large-scale environment in governing the MJO propagation across the MC islands is discussed, and the shortcomings of S2S forecast systems are qualitatively identified.

Chidong Zhang (NOAA/Pacific Marine Environmental Laboratory, Seattle, Washington)
Harry H. Hendon (School of Earth, Atmosphere and Environment, Monash University, Australia)