Nonmeteorological factors in warning verification

In order to assess the effectiveness of its warnings, the National Weather Service (NWS) employs a warning verification process. The NWS uses three parameters to evaluate warning effectiveness and to establish national performance goals: probability of detection (POD), false alarm ratio (FAR), and critical success index (CSI). However, these goals do not take into account variables that change between County Warning Areas (CWAs) and could handicap individual NWS offices at different levels. This study evaluates whether time of day, population density, or distance from the radar impacts warning performance when obtaining storm reports for warning verification. This paper examines five CWAs to look for non-meteorological influences in warning verification: Amarillo (AMA), Lubbock (LUB), Norman (OUN), Tulsa (TSA) and Nashville (OHX). Warning verification (POD, FAR, and CSI) from the years 1998-2001 was broken down to the county level in each CWA. The population density and mean distance from the radar was obtained for each county and compared against the three parameters. For evaluating time of day, the POD, FAR, and CSI were calculated for all the day time events and warnings, and all the night time events and warnings for the CWA as a whole. Preliminary results indicate that population density and distance from the radar do not seem to have a strong effect on POD, FAR or CSI. Time of day has the strongest relation to warning performance of the three parameters tested. Events that happen at night tend to have slightly lower performance measures than those during the day. Regardless of the results, this study introduces a methodology to assist offices with identifying and potentially improving warning performance in specific areas.