Towards a better understanding of the constraints and biases of atmospheric methane inversions

Atmospheric models, ranging from simple box models to advanced 3-D transport models, play a crucial role in interpreting observations related to atmospheric pollution and global warming. Their ubiquitous use has provided valuable insights, yet understanding the trade-offs and benefits of model complexity requires careful consideration, as the specific limitations and advantages depend on the application at hand. In an attempt to monitor atmospheric levels of methane with a 2-box inversion model, powered by global CAMS inventories for 5 major emission categories namely Agriculture, Wetlands, Pyrogenic, Fossils and Waste, sink specific lifetimes for troposphere, stratosphere and soil, hemispheric gradients and 40 years of polar observations of methane mole fraction and isotope composition from 10 stations, we identified several caveats of the methane budget. This work investigates the production and consumption of methane at source and sinks respectively, by the optimization of either CH4 emissions exclusively or both emissions and the isotopic signatures from the five emission categories. In addition, the significance of model parameters such as source isotopic composition, sink kinetic isotopic effects, errors associated with emissions and isotopic measurements, as well as model spin-up/spin-down criteria and the mutual controls of the tracers are evaluated to understand the dynamics of the atmospheric methane cycle. Incorporation of δ2H alongside methane mole fraction (χ(CH4)) and δ13C into inversion models has improved our understanding of the methane sources and sinks significantly, however the simplifications and assumptions need to be tested for model sensitivity to yield more accurate results as well as build more robust models.