Authors: Zoe Schroder*, Florida State University, James B. Elsner, Florida State University
Topics: Climatology and Meteorology, Hazards, Risks, and Disasters
Keywords: Tornadoes, Atmospheric Environments, Tornado Clusters, Bayesian Inference
Session Type: Paper
Presentation File: No File Uploaded
Environmental factors are consistently used in forecasting when and where an outbreak of tornadoes is likely to occur. Factors such as convective available potential energy (CAPE), convective inhibition, helicity, and bulk shear provide information about the potential development of tornadoes on a given day. However, more work is needed to quantify how the number of tornadoes generated on a convective day varies with specific environmental factors. The objective of this research is to quantify the relationship between the number of tornadoes and the environmental factors on a big day. A big day is defined as a 24-hour period starting at 6 AM (convective day) with ten or more tornadoes that are clustered in both space and time. On average, big days produce twenty-two tornadoes and occur most frequently in April, May, and June. Here we develop a statistical model that can be used to predict the number of tornadoes while controlling for environmental factors on these big days. The model uses a truncated Poisson likelihood and is fit within a Bayesian framework using functions from the “brms” package in R. The model contains a random effect term for month and predicts the number of tornadoes for unit changes in CAPE, helicity, bulk shear, CIN, latitude, and population.