Numerical modeling studies of deep convection: 1-km ensemble forecasting, secondary ice production, and aerosol-cloud interactions

Abstract

This dissertation presents numerical modeling studies of deep convection triggered by sea breezes, which remains a challenge for operational numerical weather prediction (NWP) models due to the fine scale of sea breeze circulation, uncertainties in large-scale forcings, model uncertainties innate to physics parameterizations, and varying aerosol conditions. This research is organized into three parts. Part I describes a 24-hour, 1-km, 48-member convection-allowing ensemble (CAE) produced with the Weather Research and Forecasting (WRF) Model during the Experiment of Sea Breeze Convection, Aerosols, Precipitation and Environment (ESCAPE) field campaign near Houston, Texas, in June 2022. This CAE incorporates three initial/boundary conditions (ICs/BCs), four microphysics schemes, two planetary boundary layer (PBL) schemes, and two aerosol loadings. Ensemble precipitation forecasts are evaluated against observations for eight SBC cases. It is shown that ensemble diversity, especially through varied ICs/BCs, improves forecasts of light-to-moderate precipitation, though rare and extreme rainfall remains difficult to capture. An optimal single-physics configuration (Thompson microphysics with the Mellor–Yamada–Nakanishi–Niino PBL) achieves ensemble skill scores but retains precipitation biases, emphasizing the need for continued model physics parameterization development. Part II expands the Thompson microphysics scheme to include three secondary ice production (SIP) mechanisms, which exert strong influences on convection and precipitation but have remain underrepresented in NWP models for decades. The three SIP mechanisms are: Hallett-Mossop (HM), ice–ice collisional breakup (IICOL), and fragmentation of freezing raindrops (FFD). Idealized simulations of a continental supercell and a tropical mesoscale convective system are performed to investigate the impacts of SIP on precipitation in deep convection. Results show that SIP systematically reduces precipitation by redistributing condensate from liquid to ice and from low to high levels, primarily suppressing the warm-rain pathway. Detailed analyses of SIP impacts on cloud microphysics, thermodynamics, dynamics, and convection intensity, as well as interactions among different SIP mechanisms, are also presented. Part III investigates the combined effects of SIP and aerosols on deep convection and precipitation using the SIP-included, aerosol-aware Thompson scheme in the WRF Model and a cloud tracking technique. An SBC case observed during ESCAPE is reproduced. Results show divergent, and even opposite outcomes when SIP is included, compared to previous studies. Overall, the findings highlight the dual importance of SIP and aerosols in shaping convective cloud microphysics, thermodynamics, and precipitation, underscoring the need to incorporate SIP and realistic aerosol treatments in NWP models to improve precipitation forecasts.