The influences of extratropics and circumnavigating waves on the MJO

The direct influences of extratropics and circumnavigating waves are found to be nonessential to the Madden-Julian Oscillation (MJO) using the Superparameterized Community Atmosphere Model. While some recent studies showed significant reduction of MJO variance when the extratropical influence is suppressed, the time-mean state in regions with high MJO activities tends to drift away from the controlled climatology in those studies, and evidence suggests that MJO activities are sensitive to the mean state. In the present study, we maintained the climatology in the tropics using time-invariant forcing to rule out the influence of mean state changes on the MJO.

Forced by perpetual February sea surface temperature, the control simulation is integrated for 12 years, and produces satisfactory climatology, except that MJO activities peak at around 15°S, a few degrees poleward compared to observations. An ensemble of 28 simulations are first conducted, in which extratropical eddies are suppressed by relaxing the prognostic variables north of 30°N and south of 40°S toward the mean state of the control simulation. Each of the ensemble members restarts from different points of the control simulation and lasts for 20 days. As the extratropical eddies are suppressed, the Hadley Circulation weakens in the ensemble simulations, and many aspects of the climatology drift away from the control simulation. Time-invariant forcing is calculated using the deviations from the control simulation at the end of the ensemble runs, and then applied in the next ensemble so that it drifts away from the controlled climatology more slowly. After a few iterations, the forcing can mostly compensate the effect of suppressed extratropical eddies on the mean state. With the constant forcing employed, the prognostic variables in the tropics are nudged back to the controlled climatology on an effective time scale longer than the intraseasonal timescale in the sensitivity experiment that is integrated for 12 years with extratropical eddies suppressed. Using the aforementioned methodology, the authors found that without extratropical influence, MJO activities are not weakened at all when the mean state is maintained the same as that in the control. When the method is repeated with prognostic variables relaxed toward the control north of 10°N and south of 30°S, the results do not change much, so the conclusion is not sensitive to the latitudes beyond which the eddies are suppressed. This study also seeks to constrain the influence of circumnavigating waves on the MJO. We further relax the prognostic variables toward the controlled climatology all over the globe except tropical Africa, Indian Ocean and Pacific Ocean (40°S-30°N, 0°E-90°W). The last experiment shows that MJO activities remain largely unchanged when the influences of both extratropics and circumnavigating waves are suppressed, and implies that the MJO is internal to the tropics, the equatorial Indian Ocean and Pacific Ocean in particular. The present results also illustrate the importance of maintaining mean state for MJO physics-denial studies.