Abstract
This study establishes a regional inverse framework to refine methane (CH4) emission inventories for Melbourne, Australia. Methane is a long-lived greenhouse gas and the second most significant contributor to radiative forcing from greenhouse gases after carbon dioxide. Improved understanding of methane emissions from different sectors in Australia is necessary to focus and prioritise mitigation efforts and to track progress towards emissions reduction; however, methane emissions are uncertain, especially at fine resolution (urban and regional scales) needed for mitigation. Moreover, improving predictions of atmospheric methane mole fractions requires precise and accurate emission estimates; However, previous studies indicate a mismatch between current emission estimates and atmospheric observations.Here, we use a combination of surface atmospheric measurements of methane and an inversion approach based on Bayes’ theorem to improve urban-scale methane emission estimates for Melbourne, Australia. Our inversion system is a Python-based four-dimensional variational (Py4DVar) data assimilation system. Due to lack of local methane inventories, prior emission estimates for Melbourne are compiled from globally-accessible datasets, including (1) anthropogenic emissions from the Emissions Database for Global Atmospheric Research (EDGAR), (2) fire emissions from the Global Fire Assimilation System (GFAS) dataset and (3) biogenic emissions from the Model of Emissions of Gases and Aerosols from Nature (MEGAN). Boundary condition adjustments are made using Kennaook/Cape Grim continuous in-situ CH4 mole fraction measurements and the Whole Atmosphere Community Climate Model (WACCM) dataset. The boundary condition adjustments are necessary to develop the efficiency of the regional inversion. The main goal of our inversion system is to provide more precise estimates of regional methane emissions. Independent satellite measurement comparisons are used to assess the system.The comparison with assimilated data shows improvements in modelling methane mole fraction at the suburban Aspendale site with a bias reduction from ~70 ppb (prior) to ~3 ppb (posterior). Our detailed investigations indicate that although the prior results in a reasonable match of modelled mole fraction with observations, the EDGAR dataset does not provide a realistic spatial pattern for the main anthropogenic sources (enteric fermentation and landfills) around Melbourne. The possibility of improving the spatial distribution of the prior emissions has been tested using available local/global datasets, including national maps of livestock and landfills. Eventually, to obtain more comprehensive improved emission inventories in Melbourne, more CH4 mole fraction observational data are required in this area. The results of this study are being used to expand the methane monitoring network for Melbourne.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have