Downscaling extreme winds to the building scale for resilient critical infrastructure

Extreme winds pose a growing threat to critical infrastructure under climate change; yet their impact at building scale remains poorly quantified because conventional atmospheric modelling approaches do not adequately represent small-scale processes. Here, we develop and apply a downscaling numerical methodology that transfers mesoscale atmospheric information to the building scale, enabling a physically consistent assessment of the impact of extreme winds. We examine a representative high-rise hospital situated in complex terrain and regularly exposed to severe windstorms (Bora). The wind dynamics and resulting loads are reproduced numerically, first by employing the idealised logarithmic boundary layer model for the incoming flow and subsequently by downscaling realistic outputs from a mesoscale meteorological model. The mesoscale-informed simulations capture terrain-induced wind structures, including upstream recirculation regions and a distinct low-level jet, which are not captured by the logarithmic approach. Incorporating mesoscale vertical velocity redistributes momentum and turbulence across height, increasing turbulent kinetic energy near the ground while reducing pressure coefficients along the windward façade. In contrast, simulations based on idealised inflow conditions underestimate both turbulence levels and windward pressure loads. It is shown, for the first time for severe windstorms, that considering realistic atmospheric conditions is key to more accurately reproducing wind dynamics and atmospheric load at building scale, and flow dynamics must be carefully represented in downscaling approaches to avoid biased results. Our findings support the incorporation of mesoscale information into engineering assessments of critical infrastructure exposed to extreme winds, thereby strengthening the physical basis upon which risk and resilience analyses are conducted.