Effect of complex orography on numerical simulations of a downburst event in Spain

TL;DR

This study compares nested and single-domain simulations of a severe supercell event in Spain using the HARMONIE-AROME model, finding similar accuracy but different computational costs, informing better forecasting strategies.

Contribution

It evaluates the performance of nested versus single-domain configurations of the HARMONIE-AROME model in simulating a real supercell outbreak, highlighting efficiency and accuracy trade-offs.

Findings

Both configurations captured the supercell structure effectively.

Nested approach provided marginally better reflectivity and thermodynamic detail.

Single-domain simulation was more computationally efficient with comparable results.

Abstract

Supercells, characterized by their deep, persistent mesocyclones, are among the most severe convective weather systems, often associated with large hail, damaging winds, and significant societal impacts. Accurately simulating these events is essential for improving forecasting capabilities. This study evaluates the performance of two different nested strategies of the HARMONIE-AROME model in simulating a severe supercell outbreak over eastern Iberia on July 31, 2015, where at least six confirmed supercells were reported, producing large hail and significant damage. The configurations include H500_NESTED, a two-step one-way nested approach with a parent domain at 2.5 km resolution and a nested domain at 500 m resolution, and H500, a single-domain simulation at 500 m resolution. The simulated reflectivity is validated against OPERA radar data, while precipitation and temperature outputs are compared to observations from weather stations. Key convective parameters, derived from the Murcia radiosonde, are analyzed to assess the pre-convective environment. Results show that both configurations adequately capture the overall structure of the supercell event, with H500_NESTED offering marginally improved representations of reflectivity and thermodynamic variables. However, these improvements do not justify the 30% higher computational cost of the nested approach. The H500 simulation emerges as a more efficient option, providing comparable results at a lower cost.