Abstract
Despite the years of climate change mitigation effort, per capita transportation emissions are on the rise. Reducing vehicle miles traveled, congestion mitigation and increasing vehicle efficiency are three strategies to reduce CO2 emissions from vehicles. Outcomes of these strategies may contradict each other considering their impacts on the road network and possible behavior changes within the transportation system. Though, models used in policy evaluations do not capture the interplay between vehicle characteristics, travel demand, and urban form. Understanding the spatial and temporal variations in vehicular emissions and the impact of each subsector requires collaboration between two seemingly separate fields: emissions modeling and urban science. This research combines state-of-the art methods from urban science and atmospheric chemistry to develop a Mobile Data Emission System (MODES), which is a portable framework for making fine grained vehicle emissions estimates using a large sample of mobile phone data for the Bay Area, Location Based Service Data from SafeGraph, and Uber Movement Speeds data. The MODES results were validated with two different sensor-based emission estimates, including the direct CO2 measurements of the Berkeley Air quality and CO2 Network (BEACO2N).
