Characterization of Storms in and around Urban Areas of the Central Plains

The urbanization process and its influences on severe weather have been understudied to date, particularly for small to mid-sized cities. To go beyond analyses of storm reports, which are inherently biased towards population centers, NEXRAD radar data is used to provide a nearly continuous evolution of storm severity, from convective initiation through storm dissipation across multiple cities in the central plains of the United States. An initial exploration of radar data surrounding Oklahoma City was performed to determine the best methodology for radar data quality control, storm tracking, and processing of multiple severe weather proxy algorithms (including hail size and rotational velocity estimation). Following this initial processing to tune the methodology, NEXRAD radar data covering 10 years (2002-2012) was processed for Minneapolis-St. Paul (MSP), Dallas-Ft. Worth (DFW), Omaha, and Oklahoma City (OKC). Individual events were initially chosen based on the occurrence of any type of severe report (wind, hail, tornado) in or around the urban domain (i.e., within the city borders or any of the surrounding counties). WSR-88D reflectivity data from the radar in closest proximity to a given city (e.g., KTLX – Oklahoma City) was then examined to judge the suitability of the storm event. This initial evaluation focused on storm mode, interaction with the city, and complexity of the meteorological environment; if an event was deemed worthwhile (e.g., a storm crossed city boundaries and remained relatively isolated), it was included in the study. The Warning Decision Support System – Integrated Information (WDSS-II) was used to complete all the radar data visualization and processing. Processing steps included: (1) radar merger (produces a three-dimensional grid of radar data incorporating the near storm environmental information - removes issues due to radar coverage holes and data artifacts due to storm movement to or away from an individual radar), (2) severe weather algorithms (including hail-size estimation, vertically integrated liquid, azimuthal shear and mesocyclone detection), and (3) storm identification and tracking (to extract storm attributes with time). In total, over 141 (OKC), 167 (DFW), 101 (Omaha), and 126 (MSP) unique storm clusters were tracked over the urban areas. Results will examine the differences in storm size, track, and severity prior, during and after crossing each of the different urban domains. We will also examine differences that arise due to diurnal effects, seasonality, and varying the storm type and size.