Electric vehicles (EVs) constitute just a fraction of the current U.S. transportation fleet; however, EV market-share is surging. EV adoption reduces on-road transportation greenhouse gas emissions by decoupling transportation services from petroleum, but impacts on air quality and public health depend on the nature and location of vehicle usage and electricity generation. Here, we use a regulatory-grade chemical transport model and an electricity dispatch algorithm to characterize neighborhood-scale (~1 km) air quality and public health benefits and tradeoffs associated with a multi-modal EV transition. We focus on a Chicago-centric regional domain wherein 30% of the on-road transportation fleet is instantaneously electrified and changes in on-road, refueling, and power plant emissions are considered. We find decreases in annual population-weighted domain mean NO2 (-11.84%) and PM2.5 (-2.56%) with concentration reductions of up to-5.1 ppb and-0.97 µg m-3 in urban cores. Conversely, annual population-weighted domain mean MDA8O3 concentrations increase +0.65%, with notable intra-urban changes of up to +2.3 ppb. Despite mixed pollutant concentration outcomes, we find overall positive public health outcomes, largely driven by NO2 decreases that produce mortality reductions that are ~5 times greater in census tracts with disproportionately large non-white populations.

To determine added electricity generation unit (EGU) emissions in the 30% electric all transport (eAT) scenario, we employ an electricity dispatch algorithm that provides a first-order approximation of altered EGU demand. We begin by determining the vehicle miles traveled (VMTs) for each vehicle type (k) in each U.S. county from the EPA's National Emissions Inventory 1 . For our 30% EV adoption scenario, eAT, we convert 30% of VMTs in each category k to electric VMTs (eVMTs) using Eq.1:where 𝑒𝑉𝑀𝑇 𝑐,𝑘 is a proportion of a county's VMTs that will demand electricity for battery charging for vehicle type (k), 𝑉𝑀𝑇 𝑐,𝑘 is a county's total VMTs for vehicle type (k), and 𝑓 𝐸𝑉 is the fractional EV adoption rate (0.3 in the eAT scenario). Given the variability in driving habits among vehicle owners, we note that simulating the electrification of 30% of VMTs is likely to be different from 30% of vehicles.Newly converted eVMTs are then translated into increased electricity demand from each CONUS county via Eq.