Multi-timescale exploration of teleconnection/tornado activity relationships in the Southeastern United States

Abstract

Abstract Teleconnections like the El Niño/Southern Oscillation affect climate and weather conditions across the globe, including conditions that modulate tornado activity. Early studies of teleconnection/tornado activity relationships provided evidence of links between one teleconnection and tornado activity. Later attempts introduced multivariate approaches by analyzing bivariate distributions and integrating multiple teleconnections in statistical models to predict variability in tornado activity. However, little attention has been given to teleconnection interactions and the role of these interactions in modulating tornado activity. Here, we employ a data-driven, multiple logistic regression modelling approach to explore the interactions between the El Niño/Southern Oscillation, North Atlantic Oscillation, Artic Oscillation, and Pacific North American pattern and their ability to predict the odds of an active tornado period in the southeastern United States. We develop models at the annual, seasonal, and monthly scales and, in doing so, illustrate that the teleconnections and teleconnection interactions that best predict the odds of an active tornado period differ across timescales and that the relationships exhibit clear seasonality. We also show climate conditions associated with select interactions that help explain the elevated tornado activity, namely anomalously high near-surface air temperature and humidity steered by an anomalously strong subtropical high.