Signal versus noise: Propagation modeling applied to the detection of the tornadoes produced in Northern Alabama during the passage of a storm front in March, 2018

Abstract

On March 19, 2018, a storm front that produced at least eight confirmed tornadoes passed across northern Alabama. During that period, we had a network of infrasound arrays deployed throughout the region. Infrasonic signals were detected from each of the confirmed tornadoes; however, detections were not made on all of the arrays simultaneously. To understand why certain tornadoes were, or were not, detected on a given array, an infrasound signal propagation analysis was undertaken. A parabolic equation (PE) propagation algorithm was employed with atmospheric temperature and wind profiles input from the weather research and forecasting (WRF) model. To complement the propagation studies, a signal verses noise analysis was performed for each array. Taking the noise analysis into account, the correlation between detection versus non-detection and the predictions of the propagation model was quite good.