Navigating toward sustainable traffic safety and transportation resilience in response to extreme hydro-meteorological events (1/3)

Taiwan's roads sit at the intersection of three pressures: intensifying extreme weather, PM2.5 levels among the region's highest (annual averages ~2.7× the WHO guideline), and a motorcycle-dominated traffic mix in which riders are directly exposed to both pollution and crash risk. This NSTC integrated program — Navigating toward Sustainable Traffic Safety and Transportation Resilience in Response to Extreme Hydro-Meteorological Events — spans four sub-projects that address these pressures together. The E3 Center is responsible for Sub-Project 4, Impacts of Vehicle Electrification on Traffic Safety and Public Health, with Prof. I-Yun Lisa Hsieh as co-PI.

Sub-Project 4 asks a deceptively simple question: can vehicle electrification improve air quality, traffic safety, and urban resilience at the same time? It works along four linked threads — air pollution → accident risk (causal identification with AI), air pollution → visibility (physics-informed prediction), electrification → public health (air-quality and health modelling), and electrification → urban resilience (spatial vulnerability and adaptive capacity) — to deliver a data-driven, policy-relevant framework rather than a single model.

Methodologically it combines causal machine learning (double machine learning with instrumental variables) over hundreds of thousands of accident, air-quality, and weather records; a physics-informed visibility model coupling neural networks with the IMPROVE extinction formula; the group's TW-FLEET fleet-turnover model feeding CMAQ–WRF air-quality simulation and a health-impact assessment; and GIS-based resilience mapping across electrification scenarios for 2024, 2030, and 2050.

At its mid-term stage the results point in a consistent direction: reductions in PM2.5 carry measurable traffic-safety co-benefits — strongest for motorcyclists, the most exposed road users; visibility can be predicted from routine monitoring data, opening a path to early warning; and vehicle electrification improves air-pollution resilience without erasing the underlying spatial disparities, so targeted interventions remain necessary. These threads converge on the sub-project's headline deliverable: an integrated real-time platform — traffic-accident risk maps, visibility forecasts, and air-pollution resilience maps — that turns the research into actionable insight for traffic-safety, public-health, and urban-planning decisions.