An Abstract of work to be considered for presentation at
the 23rd Conference on Hurricanes and Tropical Meteorology
Convective processes in the atmosphere play an extremely important role from both a dynamic and a thermodynamic perspective. Consequently, when numerical models of the atmosphere are configured with insufficient resolution to represent convective processes adequately, these processes must be introduced in parameterized form. Many different approaches to the parameterization problem have been implemented and varying degrees of success have been reported in their implementation. In this context, success has been indicated typically by 1.) minimizing the development of and strong dynamic response to saturated, conditionally unstable structures on well-resolved scales, 2.) producing precipitation in the absence of larger-scale saturation, and 3.) "good" agreement between simulated and observed structures on the well-resolved scales of the model.
As the resolution of operational numerical models has increased in recent years, Quantitative precipitation forecasts (QPFs) associated with convective systems have become another measure of success for convective parameterization schemes. Accurate QPFs demand that these schemes not only produce precipitation at roughly the right place and time, but also in the right amount. What this requirement means is that not only must a convective scheme modulate the rearrangements of mass that are necessary to keep a forecast model from making an inappropriate scale selection in its response to convective instability, it must do so in such a way that the vertically integrated heating and drying effects of parameterized convection are consistent with what occurs in nature. This additional requirement is very difficult to achieve because of the highly nonlinear interactions between parameterized convection and well-resolved processes in numerical models.
Quantitative estimates of parameterized precipitation rate (directly proportional to vertically-integrated parameterized heating) are typically by-products of other "closure" assumptions used in CPSs. For example, in Kuo-type schemes with moisture-convergence closures the precipitation rate is formulated as some fraction of the total amount of moisture converging into the column; in adjustment schemes such as the Betts-Miller scheme, the total precipitation rate is constrained by specified profiles of water vapor, again a vertically-integrated constraint; in mass-flux parameterizations precipitation rates strongly constrained by lower-tropospheric moisture values (i.e., the amount of moisture present in penetrative "bubbles"). In this study, these constraints on parameterized precipitation rate are compared and contrasted in the context of the QPF problem for the current generation of Numerical Weather Prediction models