Intraseasonal convective initiations – precursors and predictability

Abstract

The equatorial Indian Ocean is widely recognized as a key region for Madden–Julian Oscillation (MJO) convective initiation, but it does not behave as a single, uniform initiation basin. Instead, strong zonal and seasonal contrasts in sea surface temperature, moisture, and circulation create distinct environments in the western (WEIO) and eastern (EEIO) equatorial Indian Ocean. This thesis identifies the locations where MJO convective initiations occur within the basin and examines the physical processes that lead to intraseasonal convective initiations in these subregions, with a particular focus on their thermodynamic and dynamical precursors, as well as the implications for subseasonal prediction. Using the OLR MJO Index (OMI) over the period 1979–2021, 83 MJO convective initiation events are identified, with 41 occurring during boreal winter and 42 during summer. Two preferred initiation regions are identified: one over the WEIO, where initiations are more frequent and displaced off the equator —to the south in winter and north in summer —and the other over EEIO, where initiations are less frequent but remain near the equator and are associated with relatively strong convection. Winter initiations are preceded by a quiescent MJO convective state, while summer initiations follow vigorous earlier MJO activity. During boreal summer, the Indian Ocean Dipole (IOD) modulates where initiation occurs, with positive (negative) IOD phases favoring WEIO (EEIO) initiations. In winter, initiations do not exhibit any such preference, suggesting a minimal role of IOD influence, indicating a significant role by local intraseasonal processes. Examining the thermodynamic features, with a focus on winter initiations, reveals two distinct initiation regimes. Over the EEIO, column instability increases rapidly (within ~9 days) before initiation, whereas it is gradual over WEIO (within ~2 weeks). Column instability buildup in both regions is driven by moistening in the planetary boundary layer (PBL) and deep cooling above with both factors having a relatively large magnitude over EEIO. Because PBL moistening is a key contributor to the buildup of tropospheric instability in both regions, with a more rapid increase over the EEIO than the WEIO, we examine the processes responsible for this moistening using a moisture budget analysis. This analysis shows that the strong PBL moistening over EEIO is primarily driven by horizontal moisture advection associated with strong intraseasonal easterlies, which arise as an anticyclonic Rossby response to suppressed convection over the Maritime Continent acting on a strong background zonal moisture gradient. In contrast, WEIO moistening is controlled primarily by the evaporation of shallow clouds and precipitation in a subsiding pre-initiation environment, with a secondary contribution from horizontal moisture advection by weaker intraseasonal easterlies acting on a weaker background moisture gradient; together, these processes lead to a relatively gradual moistening. This thesis also examines the role of precursor Mixed Rossby–Gravity (MRG) waves in MJO convective initiations and their predictability. During boreal winter, about 40% of WEIO and 35% of EEIO initiations are preceded by low-level MRG activity, compared with only around 10–12% in summer. Further focusing on winter WEIO initiations, it is found that most MRG-preceded cases exhibit a clear top–down eddy kinetic energy dispersion pathway proposed by Takasuka et al., 2021, in which upper-tropospheric MRG amplification during the suppressed phase is followed by the development of lower-tropospheric MRG wave packets leading to MJO convective initiation and following eastward propagation. Reforecasts from two S2S models (NCEP and CMA) show that the impact of MRG precursors on initiation skill depends strongly on how the MRG-convection coupling is in each model. Overall, this work shows that MJO convective initiation over the Indian Ocean is characterized by regionally varying regimes, majorly controlled by background moisture gradients, intraseasonal Rossby responses to suppressed convection, and MRG waves dynamics. The results highlight how diverse the initiation processes are within the basin and checking whether forecast models can reproduce this diversity is a useful way to test how realistically they represent MJO dynamics.