End-to-end simulations for greenhouse gas monitoring from space with a spectrometer - lidar - camera sensor triplet

The greenhouse gases (GHG) methane (CH4) and carbon dioxide (CO2) have been emitted at an increasing rate since the Industrial Revolution, leading to amplified global warming. The Paris agreement, signed by 175 nations, represents the world’s first sound political framework to regulate GHG emissions. It entails a need to quantify GHG fluxes, ideally with global coverage. 

Since the pioneering missions able to detect and quantify trace gases in the Troposphere, green gas monitoring instrument (GMI) and scanning imaging absorption spectrometer for atmospheric cartography (SCIAMACHY) almost 30 years ago, a number of satellite missions that provide global coverage have been launched and are used to serve that need. There is, however, a significant discrepancy between bottom-up GHG emission estimates from inventories and top-down estimates using a combination of space-borne GHG concentration measurements and atmospheric dispersion modeling. Over the last 12 years or so, a new generation of satellites-borne imaging spectrometers emerged with sub-kilometre pixel resolution, able to map trace gas plumes and thus able to quantify GHG fluxes directly at the source, contributing to improved GHG inventories. Among those are the first commercial Earth observation missions to monitor GHG sources.

The commercial AIRMO mission aims to quantify GHG fluxes, notably CH4, in the planetary boundary layer using a combination of push-broom spectrometer, pulsed micro-lidar and visible camera. Raw lidar and spectrometer data (level-0 data) are processed (level-1b) to retrieve satellite images of CH4  and CO2 (level-2), from which images of column averaged enhancements are retrieved (level-3) and mass flux from point sources (level-4) are derived. This work will report on results of a sensitivity analyses aimed to identify the variables with the highest impact on level-4 error, including instrumental parameters (e.g., signal-to-noise ratio SNR, integration time, detector smile) and environmental variables (e.g., wind speed, surface albedo, aerosols). The approach used here is a combination of top-down and bottom-up analysis. The top-down analysis starts from the level-4 requirement of 100 kgCH4 /h and estimates from this the required precision and accuracy (bias). The bottom-up analysis simulates, end to end (from level-0 to level-4) or between sub-levels, the error propagation, validating the top-down approach. From the sensitivity study an error budget is established.