# Phone Mobility Data Tracks CO2 Emissions in Near Real Time Across Four Countries

- Published: 2026-11-07 03:07 UTC
- Section: Planet › Climate
- Tags: carbon emissions tracking, phone mobility data, The Innovation, intercity mobility networks, GRACED dataset, ODIAC inventory, climate governance
- URL: https://realtech.media/news/rt-86c1d563b8d6f5853e616de8_phone_mobility_data/

**Researchers used aggregated mobile network data to compute carbon emissions across 366 cities with high statistical precision.**

![Phone Mobility Data Tracks CO2 Emissions in Near Real Time Across Four Countries](https://realtech.media/media/2026/11/54494dd67d2dd9f7.jpg)
*Guido from Berlin, Wikimedia Commons, CC BY 2.0*

Accurate accounting of anthropogenic carbon dioxide emissions traditionally relies on statistical reports or satellite observations, which often lag by months or face resolution limits. In a study published in The Innovation, researchers from the National University of Defense Technology and the University of Hong Kong introduced a computational modeling framework that calculates carbon emissions using aggregated human mobility records [1]. The system uses movement as an observable proxy to assess multi-origin emissions without waiting for administrative inventories.

The researchers analyzed anonymized cellular signaling records from China Unicom, representing roughly 318 million active users, to construct daily intercity mobility networks during January and February 2020. The evaluation period covered the initial COVID-19 lockdowns in China, during which human mobility fell by 56.97% and national CO2 emissions dropped by 32.45% [1]. The correlation between national mobility and total carbon emissions reached r = 0.89. Structural network features, such as node centrality and clustering combined with local population attributes, served as inputs for machine learning models evaluated against the GRACED emission dataset.

## Computational Precision Across Global Regions

Across 366 cities over a 60-day window, the LightGBM model predicted multi-source, domestic aviation, and ground transportation emissions with R2 values of 1.00, 0.99, and 0.98, respectively [1]. The model retained consistency when applied to external datasets without altering the core network framework. When tested on data from Italy, the continental United States, and Mexico, the approach produced emission estimates with R2 values between 0.91 and 0.99 under the GRACED dataset, and between 0.90 and 0.99 against the independent OpenData Inventory for Anthropogenic Carbon (ODIAC).

## Intercity Networks and Operational Scope

The mobility networks revealed functional clusters of connected urban centers whose total emissions corresponded directly to their positional importance within the travel graphs [1]. The authors emphasized that mobility serves as a complementary indicator rather than a replacement for full physical inventories. The tested data focused primarily on intercity rather than intra-city movements over an extraordinary two-month window. Future iterations aim to integrate meteorological parameters, local energy structures, and remote sensing to monitor long-term shifts in transportation dynamics.

## Sources

1. [The Innovation](https://doi.org/10.1016/j.xinn.2026.101501)
2. [EurekAlert](https://www.eurekalert.org/news-releases/1150729)

---

Source: RealTech, realtech.media — https://realtech.media/news/rt-86c1d563b8d6f5853e616de8_phone_mobility_data/