The relationships that account for the observed anomalies in precipitation and severe weather over and downwind of St Louis, Missouri are still not well understood, in spite of good field studies and data analyses. Similar uncertainties exist for the city of Houston, Texas and other urban areas. One of the leading hypotheses explaining these anomalies is the so-called “glaciation” mechanism. This mechanism is related to air pollutants emanating from an urban area which are generally rich in cloud condensation nuclei (CCN). It is well known that enhanced CCN concentrations can result in narrower droplet spectra and thereby suppress warm rain processes. On the other hand, some urban areas, like St. Louis can also be sources of giant CCN (GCCN) or ultra-giant particles which can enhance warm rain processes. The presence of supercooled raindrops greatly increases the rate of glaciation of cumuli as the supercooled droplets readily collect ice crystals and freeze, and the Hallett-Mossop rime-splintering process is also enhanced.

A series of numerical model simulations using the Regional Atmospheric Modeling System developed at Colorado State University (RAMS@CSU) has been designed to investigate the hypothesized glaciation mechanism on convective storm evolution and precipitation over and downwind of St. Louis, MO. The sophisticated "Town Energy Budget"(TEB) model of the urban land surface, and the two-moment bulk microphysics which allows for the prognosis of CCN and GCCN, are both being utilized. Forty vertical levels and three two-way interactive nested model grids centered over St. Louis, with horizontal grid spacing of 37.5 km, 7.5 km and 1.5 km for grids 1 through 3 respectively, are employed. The fine grid spacing on grid 3 resolves convection explicitly. RAMS is initialized with vertical profiles of aerosol concentrations that were generated using data acquired during the Metropolitan Meteorological Experiment (METROMEX). Aerosol source functions are employed throughout the simulations to represent the supply of these aerosols over urban and rural regions. In the CONTROL simulation, aerosol source functions and vertical profiles of CCN and GCCN concentrations over rural areas were utilized. For the sensitivity tests, the rural and / or the urban source functions and vertical profiles for CCN and / or GCCN are being used. A factor separation analysis is being performed on the data to determine the contributions and forcing associated with each of these aerosol species individually, as well as those due to the interactions between the species. In this way, the impacts of variations in aerosol concentrations associated with urban regions, and the possible role that CCN and GCCN may play in the glaciation mechanism can be assessed. The results of these sensitivity tests will be presented.