An Adaptive Dynamic Downscaling Strategy to Obtain Real-time Multi-resolution Wind-field during Evolving Tropical Cyclones to Support Real-Time Risk Forecast of Powerlines

Abstract

Tropical cyclone (TC) induced high winds often cause significant damages to powerline systems and lead to widespread power failure and socio-economic losses. Meteorological numerical weather prediction during an evolving TC has the potential to enable real-time risk forecast of power systems. However, operational weather forecast only simulates wind-field at mesoscale (i.e.,1~10 km resolution) which is too coarse to be used directly to assess risks of powerlines located in complex terrains. Moreover, micro-scale (10~100 m) wind forecasts realized by dynamic downscaling from mesoscale predictions can only be achieved in existing studies for small domains (<~100 km2) due to its overwhelming computational demand. This study proposes an adaptive nested dynamic downscaling (ANDD) strategy, which is designed to obtain multi-resolution wind-field forecast in real-time to facilitate risk forecast of powerlines distributed over large region across complex terrains. The obtained wind-field can i) capture critical microscale wind dynamics caused by terrain features by performing dynamic downscaling only to the selected critical domains containing powerline segments suffering high winds; and ii) adapt to the evolution of a TC in real time by proactively and continuously searching for those critical domains based on the latest mesoscale forecasts reflecting the most recent TC evolution. The ANDD strategy is illustrated through the powerline system in Zhejiang Province (105,500 km2), China, during Typhoon Lekima.