Assessing the accuracy of the Climate Trace global vehicular and power plant CO2 emissions

Accurate estimation of greenhouse gas (GHG) emissions at the infrastructure scale remains essential to climate science and policy applications. Powerplant and vehicle emissions often form the majority of fossil fuel CO2 (FFCO₂) emissions in much of the world at multiple scales. Climate Trace, co-founded by former U.S. Vice President Al Gore, is a new AI-based effort to estimate pointwise and roadway-scale GHG emissions, among other sectors. However, limited independent peer-reviewed assessment has been made of this dataset. Here, we update a previous analysis of Climate Trace powerplant FFCO₂ emissions in the U.S. and present a new analysis of Climate Trace urban on-road CO₂ emissions in U.S. urban areas. This is done through comparison to an atmospherically calibrated, multi-constraint estimates of powerplant and on-road CO₂ emissions from the Vulcan Project (version 4.0).

Across 260 urban areas in 2021, we find a mean relative difference (MRD) of 69.9% in urban inroad FFCO₂ emissions. Furthermore, differing versions of the Climate Trace on-road emissions releases shift from over to under-estimation in almost equal magnitudes. These large differences are driven by biases in Climate Trace’s machine learning model, fuel economy values, and fleet distribution values. An update to the powerplant FFCO₂ emissions analysis (from a 2024 paper) show both improved and degraded convergence of emissions. We continue to recommend that sub-national policy guidance or climate science applications using the GHG emissions estimates in these sectors made by Climate Trace should be done so with caution.