Assessing Anomalous Propagation of Convective Storms in Complex Terrain Using a Combined Dual-Doppler and Modeling Approach

Catalonia’s most vigorous convective activity (NE of Spain’s Iberian Peninsula) mainly occurs during summer and autumn, with most of the thunderstorms producing severe weather events and causing economical losses and even fatalities. In some cases, these severe thunderstorms have unexpected propagation characteristics, likely due to a combination of the complex topography or sea-land interaction and their own autopropagation mechanisms, the latter being related to internal dynamics and interactions with neighboring thunderstorms. Understanding the mechanisms that make these thunderstorms split, merge or change directions is a key factor in nowcasting, and consequently, the early warning system and decision-making chain for severe cases.

The Meteorological Service of Catalonia (SMC) nowcasting system is based on real-time radar reflectivity data, which is used to feed the centroid-type identification, tracking and nowcasting algorithm of the convective cells. This system is being updated (currently in internal pre-operative testing stage) with a new identification and tracking version to allow the obtention of warnings if a splitting thunderstorm is going to take place. This algorithm, along with the new operative Lightning Jump (LJ), tool has noticeably increased the capacity of the SMC to emit earlier severe thunderstorms warnings, especially those with hail and strong convective winds. However, a detailed understanding of the internal thunderstorm dynamics that trigger these anomalous propagations has not been thoroughly explored. Specifically, this work is the first time that a dual-Doppler analysis has been performed on the Iberian Peninsula with operational C-band weather radars. The analysis has been successfully performed over different complex cases using the Spline Analysis at Mesoscale Utilizing Radar and Aircraft Instrumentation (SAMURAI) dual-Doppler software within the open source Lidar Radar Open Software Environment (LROSE) project (Colorado State University and National Center for Atmospheric Research-Earth Observing Laboratory).

In this study, two splitting thunderstorms have been analyzed, including severe hailstorms that occurred on 10 July 2013 and 14 September 2014. Both of these events occurred over an area of 50 km²but displayed significantly different splitting dynamics. To support and extend the observational analysis, the cases are also simulated with the Weather and Forecast Research (WRFv4) model at 1 km horizontal grid spacing and ERA5 reanalysis initial and boundary conditions, thus allowing us to obtain a three-dimensional view of the pressure field and a complete description of the storm dynamics.

Given that an increase in heavy rains in the Northern Mediterranean is expected as a result of climate change, this improvement of the early warning system is considered a good measure of non-structural adaptation. That is why this contribution is presented within the Spanish project M-CostAdapt (CTM2017-83655-C2-2-R).