Fine‐scale analysis of a severe hailstorm using crowd‐sourced and conventional observations

ABSTRACT On 1 July 2015, severe hailstorms developed over northern England. One storm tracked across an area with a dense network of privately owned (i.e. ‘home’) automatic weather stations (AWSs), permitting analysis of surface wind, pressure and temperature fields on the storm scale. The home AWS data were filtered and corrected by comparison with data from the nearest United Kingdom Met Office AWS, where measurements are made with calibrated sensors having known error characteristics. A time‐compositing technique was applied to the corrected home AWS data, before interpolation onto a 1 km grid using Delaunay triangulation. The resulting analyses were compared with radar data to assess their quality and provide insights into storm evolution and structure. Surface analyses resolved a pressure anomaly couplet on the right, rear flank of the storm, gust fronts, and regions of inflow and outflow. The pressure couplet was closely collocated with the radar‐observed mid‐level updraught position, with the mesolow (mesohigh) situated underneath the downshear (upshear) flank. The mesolow was also collocated with the strongest inflow winds. Structural features in the surface analyses (e.g. forward‐ and rear‐flank gust fronts and a prominent inflow notch) compared well with radar‐observed structures, and conformed closely to established conceptual models of supercell storm morphology. Further insight into storm structure was gained by synthesis of the surface analyses, Doppler radar data, crowd‐sourced hail reports and eyewitness photographs. Collectively, the results demonstrate that the gridded home AWS data may be of sufficient quality for use in post‐event studies of severe thunderstorms and, potentially, in the operational forecasting environment. Data from privately owned automatic weather stations (AWSs) are used to construct high‐resolution, gridded‐surface analyses near a severe hailstorm. Data were filtered and corrected by comparison with nearby Met Office AWS data. Following the application of a time‐compositing technique, corrected data were interpolated onto a regular grid. The gridded analyses resolved surface wind and pressure features on the storm scale. The realism of features was assessed by comparison with radar data and with reference to conceptual models of severe thunderstorms.