Predictability of Hurricane Irma's (2017) Genesis: The Impact of Initial Moisture and Cloud Radiative Feedbacks

Recent research has highlighted the importance of both moisture content within the incipient disturbances and diurnal cycle in influencing the formation of tropical cyclones. Despite this recognition, the precise impact of these two factors on the predictability of tropical cyclogenesis remains uncertain. This study aims to elucidate the impacts of initial moisture and diurnal cycle on the ensemble forecast uncertainty of Hurricane Irma (2017). We employ a series of ensemble forecasts and sensitivity experiments, using initial conditions derived from an ensemble-based data assimilation system assimilating satellite all-sky infrared and microwave radiances. All ensemble forecasts were initialized with the ensemble mean values of each state variable; however, the ensemble perturbations in moisture within the pre-Irma disturbance were retained.

Using the control ensemble forecasts initialized with different initial moisture uncertainties, we will show the substantial sensitivity of Irma’s formation to the moisture content within the pre-Irma disturbance. Also, we will explain how that sensitivity translates to considerable forecast uncertainty in Irma genesis timing. Through sensitivity experiments assuming either perpetual day or perpetual night, we will show the impacts of diurnal cycle on convective development and further the ensemble forecast uncertainty of genesis timing. Relative to the control ensemble where convection was enhanced during the nighttime hours and suppressed during the daytime hours, the perpetual night ensemble exhibits prolific convection that was unsuppressed during the day. This enhanced convection is shown to significantly increase the forecast uncertainty of the genesis timing. This is because the initially moister members produce stronger and more widespread convection that leads to the much earlier spin-up of the low-level vortex that becomes Irma. Conversely, the uncertainty of genesis timing forecast is significantly reduced in the perpetual day experiment. Additional cloud-radiative-feedback experiments are conducted to investigate the relative roles of longwave and shortwave radiation in the ensemble forecast uncertainty of Irma genesis timing.