A Numerical Investigation of Several Factors Contributing to the Observed Variable Intensity of Deep Convection over South Florida

This study employs a revised version of the Colorado State University three-dimensional numerical cloud scale model to investigate the observed behavior of deep convection over South Florida on 17 July 1973. A brief description of recent model improvements is made. A combined balance and dynamics initialization procedure designed to introduce variable magnitudes and distributions of low-level wind convergence to the initial fields is described. Using radiosonde and PIBAL data collected by the NOAA/ERL Florida Area Cumulus Experiment (FACE) and the National Weather Service at Miami on 17 July 1973, composite wind, temperature, pressure and moisture profiles were constructed to depict the state of the atmosphere at the time of deep convection. Mesoscale convergence was estimated from results of a mesoscale model simulation of low-level sea breeze convergence made by Pielke (personal communication) for the same case study day. Several numerical simulations were performed using the sounding data as a basic state. The initial magnitude and distribution of low-level convergence was varied and the sensitivity of the model to some micro-physical parameters was examined. The results of the numerical experiments show that (i) the magnitude of surface convergence over a finite area has a pronounced influence on the simulated storm circulation, the eddy kinetic energy of the storm and the total rainfall of the storm system; (ii) the horizontal distribution of convergence has a relatively large effect on the rates of entrainment into the updraft below 5 km MSL resulting in significant modulations in predicted precipitation, but only moderate changes in storm kinetic energy; (iii) variations in terminal velocity of precipitation associated with the introduction of the ice phase has only a minor effect on precipitation and total kinetic energy of the storm; and (iv) increased rain evaporation rates result in a moderate increase in the kinetic energy of the simulated storm, but at the expense of surface precipitation. Pressure forces are also shown to play an important role in initiating downdrafts and in biasing the direction of downdraft-associated outflow. Implications of these results to the modification of convective clouds are discussed.