MJO prediction in GCMs: Is an accurate heating distribution critical?

Description

What if an accurate distribution of heating in the tropics was all that was needed in order to simulate a realistic Madden-Julian Oscillation (MJO) in large-scale models? If we could simplify the problem to this extreme, we could focus our efforts on improving the simulation of diabatic heating in large-scale models. This study assesses whether or not taking an approach like this is a viable way to proceed. In this study, heating distributions are added to the atmospheric component of CCSM4.0 (CAM4) using a newly developed technique. The heating distributions added range from idealized blobs representing top-heavy, bottom-heavy, and middle-heavy heating to realistic distributions of heating derived from observations. All vertical heating profiles input to CCSM4 are Gaussian in the horizontal. The heating input is shifted among latitude and longitude points in a manner that is consistent with the eight phases of the MJO. The CCSM4 is then run for 15 years, and the resulting MJO that is produced by all distributions is compared to that of a control run done with no additional heating. While the runs done with an idealized heating distribution simulate a more realistic MJO than that of the control run, these runs are missing some key features of the MJO. Many of these missing features are captured by the run done with observed heating, suggesting that realistic variations in the height of maximum heating are necessary to simulate the MJO.