An Analysis of Doppler Radar Coverage and its Effect on Tornado Warning Lead Time in Southeastern Ohio

Southeastern Ohio is a region not commonly prone to tornadoes, as evident by only 29 recorded tornadoes from 2009 to 2022. In Southeastern Ohio, the nearest National Weather Service Stations are located in Wilmington, OH (KILN), and Charleston, WV. (KRLX). This leaves many counties further from the beam stationed between 70-90 miles from a radar, and the beam height at 6-10 KFT above the surface. For Southeastern Ohio, this distance from both radar stations in turn could make detection of tornadic rotation, a lower-level atmospheric process, more difficult to detect and therefore warn. Notably, the warnings issued by the National Weather Service for these rarer events exhibit varying degrees of quality. Despite a number of satisfactory warnings, several examined unsatisfactory warnings range from too little lead time or late warnings, to failure to warn entirely.

Using the Storm Events Database created by NOAA, all tornado reports from 2009 to 2022 were downloaded for 12 counties which fall into the region of Southeastern Ohio. A timeframe from 2009 to 2022 was selected, accounting for the integration of Dual-Polarization within NWS radars. Using publicly-available NEXRAD data, the corresponding radar scans for all reports for KILN and KRLX were analyzed using GR2-Analyst. All data were then reported to include the maximum incoming and outgoing velocities accounting for radar folding within the supercellular velocity couplet, and the latitude and longitude, pixel difference, radar beam height, and recorded time of those maximums, for both stations. To create a control sample of data for radar detection of tornadoes, warnings from 2009-2022 were analyzed under the same conditions, but instead in Ohio counties within 30 mile radius of the station KILN, to keep geographic conditions somewhat similar, and only low-level data were accounted for.

With a set of established criteria, it was determined whether or not each radar was able to properly “observe” the tornado. For NWS-issued warnings, a catalog of severe thunderstorm and tornado warnings was created for each event. From the storm report data, a begin and end time is always provided for the recorded tornado, and therefore, lead times were derived by comparing the time of the issued polygon with theNWS-provided begin-time. All lead time data was sorted into categories of 5 minute intervals, including a “late” category, based on the disparity between the storm’s beginning from the radar signature and the warning from the NWS. Additionally, warning-determination data was classified by “warned,” “unwarned,” and “late.” The purpose of this analysis is to determine whether or not farther proximity to radar stations has indeed contributed to a number of unsatisfactory tornado warnings for Southeastern Ohio compared to reports and warnings closer in radius to a station, as well as to analyze the quality in lead time for issued warnings. With the findings presented, this data can hopefully be utilized in the future in an attempt to remedy this problem with other radar solutions and prevent unnecessary loss of life in Southeastern Ohio.