A U.S. climatology of mesoscale convective systems

This study explores the utility of an objective, radar-based algorithm to identify mesoscale convective systems (MCSs) for the purpose of generating a long-term climatology of these events for the conterminous U.S. NEXRAD composite reflectivity data are employed to extract spatiotemporal attributes of long-lived (≥ 3 hours) convective (≥ 40 dBZ) clusters that spanned 100 km or greater in at least one dimension. Initially, radar images were thresholded by reflectivity intensity and contiguous convective clusters were identified and labeled. The major axis of each cluster, calculated using a bounding ellipse, was used to determine which clusters met the spatial requirement for MCSs. The centroids of each qualifying cluster were then associated between successive images using a nearest neighbor approach. Each track continued until no centroids were reasonably close (in space and time) to the last known position. The objective algorithm is assessed for accuracy; manual tests suggest that object-oriented approach is robust at detecting MCSs.