A Diagnostic and Predictability Case Study: Was the Atlanta 2014 Snowmaggedon Rare?

A Diagnostic and Predictability Case Study: Was the Atlanta 2014 Snowmaggedon Rare?

Robert Haynes¹, Violeta Toma¹, James Belanger¹ and Laura Belanger²

¹ Georgia Institute of Technology, School of Earth and Atmospheric Sciences

² National Weather Service, Peachtree City, GA

High impact snowfall events in North Georgia are thought to be uncommon for the region, but their presence is not negligible.  For example, the 28–29 January 2014 snow event had severe consequences for Metro-Atlanta area residents, leaving Georgia highways gridlocked for hours and causing schools and businesses to close for several days after approximately 2.6” fell. In this study, we first investigate the uniqueness of this snowfall event.  Return periods are calculated using both annual maximum series and partial duration series of 24-hr snowfall accumulation using historically long weather station records in North and Central Georgia.  Preliminary findings suggest the 24-hr accumulation of 2–3'' is not uncommon, with a recurrence interval of four-to-ten years, depending on assumptions.  The January 2014 winter event proved to be challenging to forecast in regards to lead-time, positioning of snow gradients, and total accumulations.  This raises a broader question regarding how predictable snowfall events are in North Georgia.  The original ensemble runs from the ECMWF Ensemble Forecast System for the event demonstrate the struggle to determine, at sufficient lead-time, the correct snowfall accumulation for Snowmaggedon.  To place the forecasts into proper context, we examine the predictability of snowfall events with a similar return period ranking that occurred within the 20-yr ECMWF hindcast period.  The analysis considers the hit-rate, false-alarm ratio, and reliability of snowfall in the reforecasts, to provide a recommendation on how to interpret future real-time ensemble snowfall forecasts for North Georgia.