An Absolutely Normal Risk Experiment: Broadcaster Use of Prototype Normalized SPC Outlook Products

One of the greatest strengths and challenges of the Storm Prediction Center (SPC) Probabilistic Outlook is its reliance on Absolute Probabilities of severe weather within 25 miles of a point. While presenting severe weather likelihoods in an absolute fashion can help users better interpret their personal risk from tornadoes, hail, and high winds, these absolute probabilities can also have subjectively low values (such as a 2 or 5% risk of tornadoes) and higher values can be climatologically restricted to tornado-prone regions of the country. To give forecasters and communicators another tool to highlight the context of rare severe weather risks, we have divided absolute tornado likelihood forecasts from the SPC by the daily climatological likelihood of tornadoes across the US to develop what we call Normalized Risk information. In this study we presented focus groups of broadcast meteorologists from across the US with absolute and normalized likelihood forecasts for four tornado outbreak days: the April 27^th, 2011 Super Outbreak; the June 1^st 2011 Springfield tornado; the December 10^th, 2021 Quad-State supercell outbreak; and finally the November 11^th, 2022 ArkLaTex outbreak. We then asked participants to describe how they thought they might be able to use such a product to communicate tornado risk to their viewers, when they thought this product might be useful, and what challenges the product may have to face before moving into an operational version. Participants generally found the Normalized Risk product to be of value for them, and preferred a graphical display of the product to numerical values presented for point locations. However, some highlighted concerns about the ability of the public to interpret and properly understand probability information, the lack of a “maximum” value of Normalized Risk for viewers to orient themselves to, and the difficulty of describing the Normalized Risk product to viewers with the limited time allotted to them in a newscast. Overall, results suggest that Normalized Risk information has value to broadcast meteorologists, although a better understanding of public interpretation of probability information and the Normalized Risk product in particular may increase the value of the product for broadcast partners. Future work should seek to identify public reactions to Normalized Risk information presented by broadcasters, as well as how to improve the visual presentation of Normalized Risk graphics for ease of consumption by users of all levels of weather knowledge.