With the convective season now upon us, an all-to-familiar challenge revisits the meteorologist:  forecasting thunderstorms.  We are all well acquainted with classic thunderstorm weather – developing cumulus clouds, increasing winds, the subsequent rainfall, and the occasional thunder and lightning.  During the convective season, meteorologists expect thunderstorms, but can we go one step further and forecast them? 

First, what do customers require from a thunderstorm forecast?   Such specific information as exact precipitation amounts, precise timing, and specific location of thunderstorms, is not possible, as the current state of meteorological science cannot provide such specific information in a thunderstorm forecast.  However, a general forecast consisting of a broad range of precipitation amounts, and approximate timing and location of thunderstorms, is possible using current meteorological tools, such as Numerical Weather Prediction (NWP) models. 

The limiting factor in current operational NWP models is their grid resolution, which is similar to the dimensions of one television pixel (which can only show one shade of color). This resolution is not small enough for models to resolve particular thunderstorms.

Figure 2 illustrates what the model can depict given its coarse resolution. Note that the first frame of Figure 2 (the left-most frame) is identical to Figure 1. The second and third frames (from the left) differ markedly from the middle frame of Figure 1. The second frame (Building Convection) shows an updraft that completely fills the grid column, which is similar to the aforementioned television pixel.  Note that this grid-scale updraft is considerably weaker, generating a shallow rain cloud. The third frame (Mature Phase of Convection) shows a downdraft that completely fills the grid (again, similar to one television pixel).
 

At present, the operational NWP models get around this problem by invoking the convective parameterization (CP) scheme (a type of numerical band-aid) to each grid column. In a nutshell, the CP scheme re-arranges heat and moisture across a vertical profile to offset the model’s tendency to create a grid-scale thunderstorm (see Fig. 2). This re-arrangement to a vertical profile determines the amount of precipitation that will occur, and the amount of stabilization necessary needed to prevent the aforementioned “infinite loop” of convection (or grid-scale thunderstorm).  While this sounds confusing, this problem with model grid-scale thunderstorms can create unusually high and false precipitation amounts!

In summary, specific, detailed forecasts for thunderstorms are not yet a reality. Next generation NWP models are currently in the works, reportedly with much finer resolution that will be able to explicitly forecast thunderstorms.  Until then, forecasters will base potential thunderstorm forecasts off the strengths and limitations of current models, including large-scale forcing, instability, and moisture.