Estimating Global Nitrous Oxide Emissions with a Multi-Network Atmospheric Inversion

Abstract: Nitrous oxide (N₂O) is a major long-lived greenhouse gas and the dominant stratospheric ozone-depleting substance emitted today. Current N₂O mitigation strategies rely largely on its emission estimates by bottom-up inventories and process-based methods. Independent verification of these estimates using atmospheric observations and top-down inverse modeling would help ensure that mitigation efforts are accurately targeted and evaluated.

Here, we present a computationally efficient framework for estimating global N₂O emissions. We developed an offline N₂O forward simulation based on GEOS-Chem, in which stratospheric loss are prescribed. The offline framework substantially reduces the computational cost of the forward model and enables the large ensembles required for atmospheric inversion. It is coupled to a Bayesian inversion system developed at MIT to estimate spatially and temporally resolved emissions, using EDGAR anthropogenic emissions together with natural soil and ocean sources as the a priori emissions. The inversion assimilates surface in situ observations from surface networks, such as AGAGE, together additional available observations. To improve regional constraints without increasing computational cost, we design the inversion grid to focus on the regions of greatest scientific and mitigation relevance. This framework provides an independent approach for verifying global N₂O emissions and identifying priority regions for effective mitigation.