Modeling Study of Aerosol Effects on a Tropical Cyclone with Bulk Microphysics
The anthropogenic aerosols in an urban area are mostly coated with sulfates and possess high hygroscopicity. These hygroscopic aerosols tend to become cloud condensation nuclei at the initiation of cloud particles and also participate in some cloud microphysical processes, hereafter referred as aerosol physics. When a severe storm approaches the polluted region, the convection, as well as storm intensity and track, can be modified through the cloud microphysics by the ingestion of the urban aerosols. This particularly concerns the local weather agency. Understanding the possible effects of these hygroscopic aerosols on the oncoming storm is greatly valuable to TC forecasts. This work presented two set of simulations, respectively, at 9-km and 4-km horizontal grid spacing by the Advanced Research core of Weather Research and Forecasting (ARW-WRF) model with the state-of-the-art aerosol-aware Thompson bulk microphysics scheme. These are the resolutions used for Numerical Weather Prediction (NWP) nowadays. Bulk microphysics schemes have been used widely in both research and operational communities. Incorporating aerosol physics into a bulk scheme appears to make subtle changes to the cloud physics but, in fact, turning a new page of our numerical modeling by combining the sub-micron particulate matter into the cloud and convective systems. The first set of simulations modeled a super typhoon Chan-Hom (2015) at 9-km resolution in Northwestern Pacific region. Typhoon Chan-Hom (2015) is an unprecedented case as it persistently ingests the ambient aerosols during its steady intensification over the ocean for three days before landfalling at Zhejiang. A three-dimensional monthly mean of aerosol number concentration from the Goddard Chemistry Aerosol Radiation and Transport (GOCART) model is implemented for the initial condition of the Polluted experiment to contrast with one-dimensional scheme-default maritime aerosol number concentration in the Clean experiment. The third simulation used the original non-aerosol version, double-moment Thompson, as a reference for the previous two aerosol-aware simulations. The results showed that aerosol number concentration influenced the properties of liquid cloud hydrometeors more than those of the ice ones. The particle sizes of cloud droplets and raindrops, as well as horizontal precipitation pattern, were substantially affected by hygroscopic aerosols. In addition, there were also some subtle but feasible modifications in a TC's secondary circulation. It was particularly surprising because the dynamical adjustments were assumed to be achieved by the limited convective dynamics resolved at such resolution. Lastly, the eyewall's microphysical and kinematic structures, as well as minimum SLP, also varied between aerosol-aware and non-aerosol-aware simulations; this reveals that the aerosol physics can make substantial differences to the development of TCs via microphysical parameterization. To better understand the significance of TC's dynamical responses to aerosol physics, another set of simulations are conducted at the 4-km resolution to resolve convective systems in the size of a few tens kilometers with several different orders of hygroscopic aerosol number concentration in the initial conditions. Three simulations with pristine, scheme-default and extreme polluted aerosol number concentrations are conducted in an idealized environment to avoid land-surface interference. The results show that hygroscopic aerosols not only substantially change liquid cloud properties but also the storm structures, including eyewall, anvil, and rainbands. A significant amount of deep convection is displaced from eyewall to outer rainband region in the polluted simulation, which reveals that a competition of convective available energy between eyewall and rainbands. The rainbands become more active in polluted condition, which broadens the strong tangential wind field, increases the storm size, and have nonlinear effects on storm intensity. The storm track is consistently shifted with storm size, with compact storm drifting to the east and large storm to the west. In conclusion, clouds, convective systems, and storm structures are sensitive to hygroscopic aerosol number concentration when they are explicitly resolved by the aerosol-aware Thompson bulk microphysical parameterization at this resolution. Therefore, scheme's parameters, such as the hygroscopicity, number concentration, and surface emission of aerosols need to be well defined with caution. This finding suggests that a thorough uncertainty quantification study over the bulk microphysics parameter space is needed sooner than later as bulk schemes gain more realistic but sophisticated features in the future.
Tung, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our