Simulated Madden-Julian Oscillation structure and sea-surface temperature dependence within an aquaplanet model

The response of simulated Madden-Julian Oscillations to changes in sea surface temperature climatologies, turbulent surface flux parameterizations, moisture time tendency parameterizations, and air-sea couplings is examined with the use of an aquaplanet model. The intent of this study is to determine the contribution of individual proposed mechanisms to the generation of variability at intraseasonal (30--60 days) timescales. Among the theoretical frameworks tested are Conditional Instability of the Second Kind (CISK) which is a class of scale interaction invoking a feedback between large-scale circulation and cumulus-scale convection, Wind Induced Surface Heat Exchange (WISHE) which asserts that condensational heating and adiabatic cooling in the atmosphere always balance which nullifies CISK so that the primary forcing for the Madden-Julian Oscillation is evaporative moisture fluxes induced by perturbations in easterly surface winds preceding the bulk of convection, and Air-Sea Convective Intraseasonal Interactions (ASCII) which consider the interactive couplings between surface wind magnitude and direction, ocean mixing layer depth, surface latent heat flux anomalies, and cloud shielding effects to create a zonally asymmetric boundary layer water vapor distribution induced by elevated sea surface temperatures to the east of the propagating convective mode and enhanced evaporative fluxes to the west which act to hinder rapid forward progression of the wave. A series of numerical experiments were conducted in which one or more of the model representations of these mechanism were suppressed and the resulting behavior of the simulated Madden Julian Oscillation was evaluated relative to its spatial and temporal structure. It was found that the suppression of WISHE lowered the variability of the simulated Madden Julian Oscillation by a factor of ∼80 while the suppression of the moisture feedback which is driven by difference in surface layer water vapor content had little impact on the model's ability to simulate the intraseasonal mode. Suppression of CISK through the adoption of climatological moisture time tendencies revealed little change in signal strength but a dramatic modulation in the spatial structure and the rate of propagation of the simulated Madden-Julian Oscillation was noted. Overall, the WISHE mechanism was found to be of great importance in maintaining oscillatory signal strength but it also tended to enhance phase speeds to rather unrealistic rates (≃20 ms-1) compared to the observed rates of progression obtained from the analysis of NCEP/NCAR Reanalysis OLR fields which served as the primary means of validation for the model results. Additional experiments employing a warm pool with small elevations in sea surface temperatures and air-sea interactive couplings provided for a more realistic representation of the tropical regime of the Indian and western Pacific Ocean basin associated with the bulk of observed Madden-Julian Oscillation variability. Simulated modes generated by the air-sea couplings were significantly more robust than those produced using a sea surface temperature distribution which were invariant in time. An ASCII-type response was only obtainable though the modulation of stability with respect to convection in the model's parameterization of moist processes which sent the simulation into a state of "superrotation" characterized by surface westerlies in the tropics instead of observed easterlies.