A Model Review of the Historic West Virginia Historic and Devastating Floods of 23 June 2016

A large scale frontal system with flow over a strong subtropical ridge set up a nearly classic Maddox frontal rainfall event over the Ohio Valley on 23 June 2016. The larger scale pattern was relatively well predicted in the NCEP global forecast systems and thus these models showed a signal for a potential rain event. However, lacking the ability to produce convection these models failed to produce the precipitation amounts relative to observed extreme rainfall. These models also exhibited large spatial and temporal issues. Several Convection Allowing Models (CAMS) produced significantly higher QPF amounts in closer proximity to the general locations where the heavier rainfall was observed. The probability matched mean products produced from the CAMS models showed great promise in forecasting future extreme rainfall events.  

The large scale pattern favoring heavy rainfall is presented along with forecasts from traditional NCEP models and ensemble forecast systems.  Short range forecasts from CAMS are also presented. In this event, the CAMS provided insights into the potential for an extreme rainfall event over the portions of Kentucky and West Virginia. The High Resolution Rapid Refresh Model (HRRR) also created a cold pool as the event progressed and thus indicated a reduced threat of additional flash flooding during the late morning and evening hours of 24 June 2016.

This paper will show how CAMS and emergent CAM ensembles should improve 0-24 hour forecasts of convective driven flash flood events with strong synoptic forcing. The ability to correctly leverage the power of rapid updating CAMS and CAM ensembles should be a key training focus to improve the forecasting and warning of heavy rainfall and flash flooding in the near future. These data, if effectively employed, have significant potential to improve short-term forecasting during the warm season relative to models that employ convective parameterization schemes.