Specific to the Houston anomaly, our findings indicate that the spatial excess in CG flash density over Houston is a robust feature, even if days with flash densities exceeding the upper quartile of the highly skewed cumulative flash density distribution for our gridded domain are removed (this subset comprises less than 10% of the total daily sample, but produces 75-85% of the total lightning in the Houston area). This is also somewhat true of the surrounding coastal region, pointing to the importance of spatial preference in storm frequency in the determination of regional CG lightning anomalies.
By comparison, when examining inland regions (e.g., toward Dallas) we found that even fewer large events produced 80-85% of the total flash density. When only “lightning-days” were averaged to produce a conditional mean, we found that the CG anomaly still existed in the Houston area but that the conditional mean flash density became even larger moving inland into central Texas. The combined findings suggest that although the Houston area sees an increased frequency of lightning producing storms (including more large flash density events), storms occurring further inland (e.g., vicinity of Dallas), may actually produce more lightning on an event basis.