Fossil fuel CO2 emission signatures over India captured by OCO-2 satellite measurements

Recent developments in space-based measurements provide new possibilities for monitoring greenhouse gas (GHG) emissions at all scales, from global to local. We traced CO₂ emission sources such as power plants over India using OCO-2 satellite measurements from 2014 to 2021. India is the third-largest CO₂ emitter in the world, with coal accounting for nearly 60% of total fossil fuel combustion.

The Gaussian plume model was used to assess the power plant emissions. Cross-sectional (c/s) CO₂ emission flux is estimated to validate the results. In 13 out of 26 cases, the estimated power plant CO₂ emissions agreed within ± 25% of the emissions reported in the Carbon Brief (CB) database, and in 21 cases, the estimated emissions are within ± 50%. There is one case where the CB database significantly overestimated the CO₂ emission for power plants. Further, in two cases, in which the Gaussian plume model gives much higher estimated CO₂ emissions, there are emission sources other than the power plants in the vicinity. The c/s emission flux and emission inventories can be used to confirm such cases. 

In addition, the c/s emission flux method was employed to assess the emissions reported in the EDGAR and ODIAC CO₂ emission inventories. Our study demonstrated the capability of OCO-2 to uncover missing or underestimated CO₂ emission sources in emission inventories. Our approach could be a vital tool to conduct an initial estimate of missing or underestimated sources in the databases and emission inventories at various scales, as c/s emission flux estimation and Gaussian plume model require less computation than other modeling approaches. More sophisticated methods, such as Bayesian inversion combined with extensive transport modeling, can then be used to derive emissions with less uncertainty. Future satellites with high spatio-temporal resolution and coverage (e.g., Microcarb, Tansat2, CO2M) will enhance the possibilities of continuous monitoring of point sources as well as detecting missing or underestimated emission sources.