Human Factors Affecting Tornado Warning Decisions in National Weather Service Forecast Offices

The National Oceanic and Atmospheric Administration’s (NOAA) National Weather Service (NWS) is responsible for a myriad of weather warnings for the protection of life and property. One of the most important is the tornado warning, which is a notification of imminent danger. Tornado warnings are often the cause of interruptions on media outlets, as well as anxiety among many members of the public.

The link between the tornado science and the tornado warning is the NWS Weather Forecast Office (WFO) forecaster. It is the role of the forecaster to issue tornado warnings to provide the public and emergency managers with advance notice of tornadoes. Besides the obvious challenge of imperfect knowledge of tornadoes and their environments, some of the challenges for the forecaster’s tornado warning decision process include non- scientific influences. These are conditions that are not solved by improved remote sensing or increased physical understanding of tornadoes.

One example is weight of verification. Tornado warning verification is usually scored by four main metrics; Probability of Detection (POD), False Alarm Rate (FAR), Critical Success Index (CSI), and lead time (the time between the issuance of the tornado warning and the first damage cause by the tornado). However, verification statistics can be an inaccurate representation of skill due to the lack of ground truth reports in rural areas that are difficult to reach and have little population. Further, these statistics do not account for the forecaster decisions to not issue warnings, which can also be difficult due to limitations of the science and observations.

Another factor that can play an important role is the operational environment, including, but not limited to, working with others on the operational floor, and concern from danger to loved ones in the tornado path. There may also be additional stress in today’s WFOs from varying reliability of ground truth reports, new dissemination duties through social media outlets and chat rooms, the large volume of decision assistance tools available (which can conflict with each other), and the large number of phone calls that come in to the operations area during a tornado outbreak.

To date, there has been little research done to determine what factors within the WFO environment may affect the warning decision making process of NWS warning forecasters. This study will undertake that effort by surveying NWS personnel who are responsible for issuing tornado warnings.

The survey will cover a number of topics the authors deem as important to the tornado warning decision process. There will be questions on basic demographics, experience level of forecasters, knowledge among those forecasters of Government Performance and Results Act (GPRA) goals of the agency, and a host of human factors, such as the forecasters’ interaction with others on the operational floor during the tornado warning decision process, as well as their overall comfort with the task and responsibility when charged with making the tornado warning decision.

The ultimate goal of the survey is to develop best practices by comparing the behaviors of offices with higher GPRA scores with those of lower performing offices. The results are expected to help managers and lead forecasters arrange their human resources during a tornado warning situation as effectively as possible, and to also identify potential training options to improve tornado warning scores.