Severe thunderstorm warning bias

One of my master's students Megan White recently completed her thesis on the relationship between the issuance of severe thunderstorm warnings and underlying demographic variables. The map above shows the distribution of National Weather Service severe thunderstorm warnings. Each grid cell represents the number of severe thunderstorms issued over the five years of the study. Megan's thesis can be accessed here.

In this map of local R2 derived from geographically weighted regression, one can see the varying relationship between the number of severe thunderstorm warnings and the properties of the underlying demographic template. Boundary lines are National Weather Service forecast zones.  White and red shades indicate areas where there was a stronger association between demographic variables (population count, percent white, and median income) and severe thunderstorm warning issuance.  Several cities stood out for their propensity to issue warnings that more frequently coincided with increasingly populated, whiter,  higher income areas.  Forecasters in Washington DC, Nashville, the Atlanta I-75 corridor, and St. Louis were more likely to  to take into account these underlying demographic variables when alerting the public about severe weather.