Phone Mobility Data Tracks Carbon Emissions in Real Time
Researchers estimated daily carbon output across four countries using anonymized cell phone signals from hundreds of millions of users.
Updated on Nov. 10, 2026 in Environmental

Researchers used anonymized signaling data from 318 million Chinese mobile users to estimate carbon emissions across China, Italy, the continental United States, and Mexico. Their study, published in The Innovation, found mobility fell 71.3% and emissions fell 41.2% during the January and February 2020 lockdowns.
Why it matters
Conventional emissions inventories lag reality by months, so a method that tracks movement could offer faster estimates for policymakers. The approach suggests physical activity such as commuting and deliveries may serve as a practical proxy for carbon output.
The mobility-based model reached R² values of 0.91 for Italy, 0.95 for the continental United States and 0.99 for Mexico, with the correlation between national mobility and CO2 emissions in China at r = 0.89. Whether the approach holds for regions with different energy mixes remains untested in the study.
The details
Researchers converted daily China Unicom signaling data into intercity networks in which each city is a node and each movement of people between cities is an edge. Structural features of these networks, combined with population and geographic data, trained machine learning models that were then validated against the GRACED dataset, which tracks daily CO2 emissions across seven sectors.
Timeline
January and February 2020: Study window covering COVID-19 lockdowns, when mobility fell 71.3%.
Early 2020: COVID-19 lockdowns caused carbon emissions to drop sharply.
The Big Picture
The study positions human mobility as an alternative to satellite atmospheric inversions, which trade spatial resolution against coverage. If the framework scales, it could offer emissions estimates that update far more frequently than current inventories allow.
If the mobility-based framework is adopted, emissions estimates could be refreshed far more often than current inventories allow, which could inform faster policy responses. Readers should note the study has so far been validated only in four countries and needs further testing before it becomes a routine tool.
The takeaway
Human movement may act as a practical proxy for carbon output, giving researchers a faster way to track emissions than traditional inventories. The approach still requires validation beyond the four countries tested before it can be relied on widely.
Further reading
For broader context on emissions tracking methods, explore the Environmental section.
Source note: This article includes information reported by Earth.



