Planning the electric vehicle transition by integrating spatial information and social networks

Abstract

The transition from gasoline-powered vehicles to plug-in electric vehicles (PEVs) offers a promising pathway for reducing greenhouse gas emissions. Spatial forecasts of PEV adoption are essential to support power grid adaptation, yet forecasting is hindered by limited data at this early stage of adoption. While different model calibrations can replicate current trends, they often yield divergent forecasts. Using empirical data from states with the highest levels of adoption in the United States, this study shows that accounting for spatial and social networks among potential PEV adopters produces forecasts that are only one-third of benchmark predictions for 2050. Results further demonstrate that incorporating spatial social networks improves the ability to capture the spatial autocorrelation observed in the empirical diffusion process. This study also evaluates the potential impact of various PEV marketing campaigns under prevailing uncertainties, highlighting the importance of tailoring strategies to network dynamics for effective PEV promotion.