A new study from Northern Arizona University (NAU) suggests that the world’s major greenhouse gas emissions databases may be significantly underestimating carbon dioxide emissions from urban vehicles. Researchers say the Climate TRACE database, developed by the Climate TRACE consortium co-founded by former Vice President Al Gore, underestimates urban vehicle CO2 emissions by an average of 70%.
The survey results are announced today. environmental research letters By Kevin Gurney, professor in the NAU School of Informatics, Computing, and Cyber Systems (SICCS). The study focused on carbon dioxide (CO2) emissions from cars and trucks as reported in the recently released Climate TRACE database.
Gurney said the results, along with previous research that identified similar problems in Climate TRACE estimates for power plants, raise concerns about the reliability of emissions data used to guide climate policy and decision-making.
“Given the importance of vehicle CO2 emissions in cities, we carefully examined the Climate TRACE data, which relied on a promising new artificial intelligence-based approach,” Gurney said. “Combined with previous research on Climate TRACE power plant CO2 emissions, our results suggest that Climate TRACE data significantly underestimates more than half of fossil fuel-based CO2 emissions in U.S. cities.”
Comparison of Climate TRACE and Vulcan emissions databases
To evaluate Climate TRACE’s estimates, Gurney and his colleagues compared them to data from Vulcan, an “on-the-road” emissions database developed by his lab. The Vulcan system is calibrated using official traffic records and energy consumption data, providing an independent benchmark for measuring vehicle emissions.
Researchers compared vehicle CO2 emissions data from 260 U.S. cities between two databases.
“Vulcan’s on-road data is not perfect, with an uncertainty of about 14%, but this is much lower than the discrepancy found when comparing the CO2 emissions of 260 U.S. city vehicles to the Climate TRACE database,” said Bilal Aslam, SICCS postdoctoral fellow and co-investigator on the study. “Climate TRACE CO2 emissions were on average 70% lower than the same emissions in the Vulcan on-road CO2 emissions database.”
The differences were even greater in some places, the researchers said.
“Individual cities like Indianapolis and Nashville were over 90% lower,” added Pauluk Das, a SICCS researcher who contributed to the study.
The authors believe this underestimation could extend beyond the United States and impact climate TRACE data around the world. They also expressed concerns about other aspects of the database that may require additional scrutiny.
Concerns about AI-based emissions tracking
The researchers emphasized that artificial intelligence has great potential to monitor environmental conditions and generate emissions estimates. However, they argue that strong scientific standards remain essential.
In their view, ensuring emissions data is accurate and reliable requires transparency, expert review, and rigorous scientific methodology. Reliable greenhouse gas measurements are a fundamental part of developing effective climate policies and assessing progress towards reducing emissions.
This document also outlines several recommendations aimed at strengthening and improving Climate TRACE efforts, helping policymakers and budget planners make better-informed decisions about reducing greenhouse gas emissions.
“While we will never estimate emissions with perfect accuracy, we must ensure that the data shared with policymakers and the public is unbiased and meets best practices and the most rigorous scientific standards available,” Gurney said. “Without this, we could mislead decision-makers and lose public confidence in our ability to tackle climate change.”
Kevin Gurney’s research on greenhouse gas emissions
Mr. Gurney’s expertise spans atmospheric science, ecology, and public policy, and he has spent more than 20 years developing a standardized approach to measuring greenhouse gas emissions across the United States.
His Vulcan and Hestia projects, with support from multiple federal agencies, quantify and visualize greenhouse gas emissions across the country, from individual power plants and roads to neighborhoods. These systems can help identify emissions “hotspots” and support more targeted strategies to reduce pollution. Gurney’s emissions estimates also show strong agreement with direct air monitoring measurements.
Throughout his career, Gurney wrote more than 180 scientific papers that received more than 20,000 citations. His work includes contributions to the recent U.S. National Academies report, Greenhouse Gas Emissions for Decision Making.
He has also been involved in the United Nations Framework Convention on Climate Change and Kyoto Protocol processes for more than 25 years, and is a lead author on the Intergovernmental Panel on Climate Change (IPCC).
