Representing People in Severe Weather Warning Systems

Severe weather warning systems identify hazard risk and disseminate warnings so that individuals may keep themselves, loved ones, and property safe. An important barrier to integrating social and behavioral science research into operational severe weather warning systems is that current technology has limited ability to represent people as a data element in warning systems. There are data elements for threat identification, such as radar observations and forecast models that depict the spatial and temporal characteristics of the weather threats, indicate forward motion, intensity, and probabilities. There are data elements for warning, deterministic polygons that show the spatial extent of areas of risk for people, and contain qualitative information on hazard types, hazard arrival times by town, protective action recommendations, and confidence. Having threats and warnings as data elements allows NWS forecasters, emergency managers or software to rapidly link these elements. In contrast, individuals have not been not represented as data elements. Consequently, warning systems primarily broadcast general warning messages to the public at large, and it is up to individuals ensure they receive the warnings, interpret their level of personal risk, and determine if protective action is feasible and necessary. More recently, through services such as Wireless Emergency Alerts, people are represented by their current location, however, there is still no social and behavioral context associated with that person’s location, and warning messages are still general. Next generation warning systems should have the technical capability to disseminate personalized warnings to individuals. These warning systems will be able to represent individuals as a complex data element, so that dynamic weather risk can be rapidly linked to each individuals’ probable locations, perceptions, and vulnerability. Advances in high resolution weather sensing, the Internet of Things, smart phones, mobility enabled Information and Communications Technology (ICT), and cloud computing makes these individualized warnings possible.

CASA, a multidisciplinary center focused on high resolution warning systems, is pursuing research towards individualized warnings. CASA has created a research platform for next generation warning systems in the Dallas Fort Worth Metroplex where research can be conducted during live severe weather events with stakeholders and the general public. As part of this research platform, we have created a mobile phone app called CASA Alerts that delivers real-time, personalized weather alerts to individuals. The app also functions as a tool for conducting cross-sectional and longitudinal research on warning perception and behaviors through mobile phone surveys and collecting individual location data which can be linked in time and in space to weather data. Our research has identified that individual mobility patterns, derived from cell phone location data are a promising data element for representing people in operational warning systems. After describing the research platform, we will discuss the results of a pilot project based on warning people using their mobility patterns and individual contexts.