How Satellites Connect the Dots to Measure Methane in the Permian?

Methane is the most potent short‑term greenhouse gas, and tracking it is vital to fight climate change. In the past decade, several satellites have been launched to watch methane from orbit. Some, like TROPOMI, map emissions over large areas, while others, such as Carbon Mapper, pinpoint short‑lived "super‑emitters" in near real time.

In this study researchers compared two approaches over the Permian Basin, a US region rich in oil and gas. The first method, using TROPOMI divergence data, estimated 1.83 ± 0.96 Tg of methane per year, spreading emissions across the whole basin. The second, an event‑based analysis of methane plumes detected by Carbon Mapper, gave 1.26 Tg (range 0.78‑2.02 Tg) and highlighted the most intense, brief sources.

The two maps align well (R²≈0.64) where they overlap. By replacing overlapping cells with the more precise event‑based values, the combined total rises to 2.68 Tg (range 1.88‑3.54 Tg), matching previous estimates. Moreover, data show a slight decline in emissions over the past five years. The outcome? A hybrid model that leverages the broad coverage of regional satellites and the pinpoint accuracy of point‑source observations, delivering a fuller picture of atmospheric methane.

This strategy could become a key tool for quickly monitoring and cutting greenhouse‑gas emissions in other parts of the world.

The divergence method leverages mass‑conservation: it computes the net methane flux across a closed surface using vertical concentration profiles from TROPOMI. This approach captures diffuse, widespread emissions but may miss localized spikes due to its ~7 km resolution. In contrast, Carbon Mapper’s event‑based technique employs higher‑resolution spectrometers (~1 km) to detect sudden column‑density increases typical of pipeline leaks or compressor stations. Each plume is quantified via radiative‑transfer inversion, converting spectral absorption into column concentration, then into emission rates using atmospheric dispersion models.

Data‑fusion merges the two datasets: cells with consistent values are retained, while discrepant cells are replaced by the more reliable estimate (usually the event‑based one). This reduces overall uncertainty because the combined variance is lower than that of either method alone. The observed R²≈0.64 indicates that about 64 % of the variance is explained by the overlap, suggesting substantial agreement. By substituting overlapping cells with event‑based values, the integrated estimate reaches 2.68 Tg (range 1.88‑3.54 Tg), aligning with previous basin‑scale studies. The hybrid framework thus offers a robust, scalable pathway for near‑real‑time methane monitoring worldwide.