Authors: Bradley Baker*, University of Memphis, Dorian Burnette, University of Memphis
Topics: Climatology and Meteorology
Keywords: severe thunderstorms, climatology, storm environments, climate variability
Session Type: Paper
Start / End Time: 3:05 PM / 4:45 PM
Room: Buchanan, Marriott, Mezzanine Level
Presentation File: No File Uploaded
The Southeast U.S. is vulnerable to severe thunderstorms, but severe thunderstorm variability is not well understood in the region prior to the 1950s. Here we compute the significant severe parameter from version 2c of the 20th Century Reanalysis (20CR) dataset and use it as a proxy to develop a climatology of severe thunderstorm days for the Southeast from 1925 to 2014. We test the sensitivity of the significant severe parameter to different thresholds (10,000 and 20,000) by reconstructing and analyzing two separate time series. Although 20CR is embraced for its long-term record, inhomogeneities have been detected in the dataset prior to 1950 due to the changing density of the observations. Thus, we evaluate our reconstructed time series by sampling and verifying severe thunderstorm days before and after 1950 using archived newspapers and severe weather datasets. We fit an Autoregressive Conditional Poisson (ACP) model to the time series and calibrate it using severe thunderstorm days after 1950. Our results indicate an overcount bias in severe thunderstorm days before 1950. We conclude that applying a monthly, proportionally-based bias correction factor to the time series before 1950 shows promise in minimizing non-climatic biases.