The Impact of wind-driven ocean circulation variability on the Madden-Julian Oscillation and Kelvin Waves

The Madden-Julian Oscillation (MJO) and convectively coupled equatorial waves (CCEWs) have a significant impact on the size, location, variability, and strength of convection in the tropics. The behavior of the MJO and CCEWs is, in turn, modulated by moisture gradients and sea surface temperature (SST). In the tropics, the primary sources of SST variability are dynamically coupled ocean modes such as El Niño-Southern Oscillation (ENSO), the Indian Ocean Dipole, and Atlantic Niño, with a secondary influence from thermodynamically driven variability. The preponderance of these dynamically coupled modes in the tropical oceans obscures the variability unique to the MJO and CCEWs in observed datasets and coupled models. This leads to a gap in our understanding of to what extent and through which physical processes dynamically coupled ocean modes modulate the characteristics and evolution of the MJO and CCEWs. To address this, we examine how wind-driven ocean circulation variability impacts the MJO and CCEWs by comparing two versions of the Community Earth System Model version 2 (CESM2). One is a fully coupled (FC) control version in which dynamically coupled ocean modes can influence the MJO and CCEWs, like in the real world, and one in which they cannot. The latter is referred to as the Mechanically Decoupled (MD) simulation, where the ocean component of the model is forced by seasonally varying climatological values of wind stress. Thus, the MD simulation removes dynamically coupled ocean modes, like ENSO, while retaining the thermodynamic coupling crucial for the MJO and CCEWs and having more realistic ocean-damping processes.

To investigate the behavior of the MJO and CCEWs with and without influence from dynamically coupled modes, we compare total precipitation and net longwave flux at the top of the model from the two simulations. By comparing the wavenumber-frequency spectra of precipitation and net longwave flux at the top of the model, we find that, compared to the FC simulation, the MD simulation has stronger Kelvin waves and a weaker MJO. The difference in MJO activity is most notable over the Indian Ocean and for Kelvin waves over the Western and Central Pacific (Figure 1). The total unfiltered variance in the tropics for total precipitation and net longwave flux at the top of the model is less in the MD simulation compared to the FC simulation. Unsurprisingly, therefore, the MJO is weaker in the MD simulation compared to the FC simulation in absolute quantities. However, it is also proportionally weaker as a percentage of the total variance in the MD simulation versus the FC simulation. Consequently, we hypothesize a relationship between the MJO and Kelvin waves that is dependent on the mean state of the tropics. We also consider how wind stress-driven ocean circulation could modulate this relationship by how it impacts the distribution and convergence of low-level moisture.

Finally, we introduce a third simulation to test these hypotheses and more closely examine the impact ENSO has on CCEWs, given that ENSO is the strongest dynamic ocean mode. In this simulation, named NoENSO, anomalous wind stress is only absent over the equatorial Pacific Ocean. This eliminates ENSO while preserving other dynamic ocean modes, most notably the Indian Ocean Dipole, which has been known to impact the strength of the MJO. By further isolating ENSO, we hope to better ascertain the impacts of specific dynamic ocean modes on CCEW behavior.