Respective roles of shallow convection and stratiform rainfall on the simulation of Madden-Julian Oscillation

 
Poster PDF

Author

Joshua Xiouhua Fu — University of Hawaii

Category

Modeling

Description

Reducing systematic errors of state-of-the-art general circulation models (GCMs) will increase our confidence on the projection of future climate change and improve our forecast skills of present-day weather and climate variability. One systematic error of conventional GCMs is the failure to simulate a dominant tropical atmospheric variability: Madden-Julian Oscillation (MJO). Through downscale and upscale influences, MJO modulates the occurrence of extreme hydro-meteorological events (e.g., hurricane, flood, and drought, etc.) and the onset and development of ENSO. This common model error not only hinders the achievement of seamless forecasts, but also undermines our confidence on the projection of future climate change. In this proposed study, we aim to improve the representation of MJO in conventional GCMs. As a first step toward the proposed goal, two sets of sensitivity experiments with ECHAM GCM have been carried out to unravel the respective roles of shallow convection and stratiform rainfall on the simulation of the MJO. First, a series of retrospective forecast experiments, targeting a prominent MJO event observed during TOGA-COARE, has been conducted to assess the impact of shallow convection on the dynamical MJO forecast skill. It is found that the boundary-layer moistening from shallow convection plays a critical role on the eastward propagation of model MJO. On the other hand, a series of long-term free integrations reveal that an appropriate portion (> 30%) of stratiform rainfall (vs. total rainfall) is necessary to sustain a robust MJO in ECHAM GCM. Present findings highlight the need to improve the representations of both stratiform clouds and shallow convection in conventional GCMs. We are now developing a high-resolution dynamically consistent regional data set for the TWP-ICE field campaign period using advanced data assimilation techniques. We plan to use this data set to diagnose the interactions among shallow convection, deep convection, and stratiform clouds. The diagnostic results will be used to improve current cumulus parameterizations. Eventually, the improved parameterizations will be tested and implemented in ECHAM and other GCMs.