Evaluation of SPC Outlooks using the Symmetric Extremal Dependency Index

A new verification measure devised by Ferro and Stephenson, the extremal dependency index converges to a meaningful limit for forecasts of rare events, while other common verification measures converge to zero or infinity. An alternative version of this measure, the symmetric extremal dependency index, shares all of the same benefits as the extremal dependency index, such as having a non-degenerate limit, a fixed range, a meaningful origin, and being non-trivial to hedge, while also being complement symmetric. This measure is applied to convective outlooks and similar forecasts of severe weather, and the results are compared to those obtained from other common verification measures. One notable benefit of the symmetric extremal dependency index is its reduced variance across a wide range of spatial scales when compared to other measures, making it a candidate for further use when evaluating severe weather forecasts.