Advancing the use of satellites to constrain atmospheric methane fluxes

Abstract

CH4 is the second most potent anthropogenic greenhouse gas, after CO2, and is directly responsible for approximately 20% of the human-induced greenhouse effect. To improve our understanding of the global CH4budget, high quality measurements of its atmosphericmole fraction are needed with good resolution in space and time. They provide constraints to the so-called inverse models, which are used to convert the mole fraction gradients into surface fluxes. Unfortunately, it is difficult to take continuous measurements on the ground in many regions of the world due to political or geographical limitations, leaving the models incapable of estimating the fluxes in these regions. Measurements made by sensors onboard a satellite platform can be a suitable source of information as they provide near global coverage. However, satellite retrievals often suffer from errors due to scattering of light by aerosol and cirrus. Such errors, if unaccounted, can be wrongly attributed to the flux estimates of inverse models. One way to dodge this problem is to use the ratio between the retrievals of two tracers, so that the retrieval errors cancel out in the ratio. The traditional ‘proxy’ method achieves this goal by multiplying the ratio of satellite retrieval of CH4 and CO2 with numerically modeled CO2 mole fractions to obtain CH4 mole fraction estimates. The method assumes that the errors in modeled CO2 are negligible compared to residual errors in the ratio of satellite retrievals. However, the retrieval errors are becoming smaller with new and more advanced satellite data becoming available and CO2 model errors are becoming the performance limiting factor in inverse modelling. In this thesis, I present the ratio inversion method, which avoids the use of modeled CO2 by assimilating the ratio of satellite retrievals of CH4 and CO2 to constrain surface emissions. I tested the method with synthetic numerical experiments, where the satellite retrievals were simulated numerically with predefined ‘true’ fluxes. These experiments confirmed that the ratio inversion method is capable of reproducing the true emissions. Via a similar experiment, I also found that inversions using satellite retrievals perform better than inversions assimilating only ground-based measurements. I compared retrievals from the Greenhouse gases Observing SATellite (GOSAT) with in situ measurements from ground-based sensors of Total Carbon Column Observing Network (TCCON), to quantify the errors in CO2 and CH4 retrievals, in their ratio, and in modeled CO2 mole fractions. The errors in the ratio were found to be lower than the errors in CO2 and CH4, which confirms that a large fraction of the scattering related errors cancels out. The ratio measurements from GOSAT were then used to constrain the surface fluxes of 2009–2010, and the results were validated with independent aircraft measurements of CO2 and CH4. Further, I used GOSAT and ground-based measurements, along with carbon isotope ratio observations of CH4, to identify the impact of the 2011 La Niña on natural CH4 emissions. After removing the variability caused by meteorology, I found that emissions from wetlands in the Tropics were higher during this event. The enhancement in wetland emissions was primarily caused by an increase in the wetland extent driven by heavy precipitation during the La Niña.