LPDM footprint dataset - Brazil 2017

This is based on the Zenodo dataset at DOI https://doi.org/10.5281/zenodo.16748754, which does not include the 2017 data. Sample dataset of LPDM footprints for GOSAT measurements over and near Brazil, generated with the Met Office's NAME III v7.2. The footprints were generated for GOSAT measurements for 2017 in the [35.8° S - 7.3° N, 76.0° W - 32.8° W] domain. Dispersion was modelled in a domain that covers the whole South American continent (domain [61.0° S - 22.3° N, 91.3° W - 24.8° W]). Created for Tunnicliffe et al. (2020) and published as part of the journal paper "Enabling fast greenhouse gas emissions inference from satellites with GATES: a Graph-Neural-Network Atmospheric Transport Emulation System" (Fillola et al., Geoscientific Model Development 19.5 (2026): 1893-1915, https://doi.org/10.5194/gmd-19-1893-2026; code hosted on GitHub at https://github.com/elenafillo/GATES_LPDM_emulator). Additionally, it is published as part of the preprint "GATESBackground: Emulating background greenhouse gas mole fractions for regional atmospheric inverse modelling with graph neural networks" (Keshtmand et al., preprint egusphere-2026-3361). See Ganesan et al. (2017) for details on how they were generated. The datasets are provided as monthly netCDF arrays, which can be opened and explored in python using the xarray library. Each file contains backwards dispersion plumes (under variable "fp"), with the coordinates of the satellite measurement stored in the "release_lat" and "release_lon" variables. Additionally, there are boundary condition sensitivities (under variable "particle_locations_x", where x is designated as n for north, s for south, e for east, or w for west) which span 20 vertical levels from 500 metres to 19,500 metres at 1000 metre intervals. The north/south boundary sensitivities have a horizontal resolution of 0.352°, whilst east/west have a horizontal resolution of 0.234°. These boundary condition sensitivities are used in conjunction with CAMS reanalysis products, described in the Zenodo dataset https://doi.org/10.5281/zenodo.21035682, to calculate the background mole fraction. Fillola, E. and Tunnicliffe, R.: LPDM footprint dataset – Brazil, Zenodo [data set], https://doi.org/10.5281/zenodo.16748754, 2025. Fillola, Elena, et al.: Enabling fast greenhouse gas emissions inference from satellites with GATES: a Graph-Neural-Network Atmospheric Transport Emulation System, Geoscientific Model Development, 19.5 (2026): 1893-1915, https://doi.org/10.5194/gmd-19-1893-2026 Tunnicliffe et al. (2020): Quantifying sources of Brazil's CH4 emissions between 2010 and 2018 from satellite data, Atmospheric Chemistry and Physics, 20(21), 13041–13067, https://doi.org/10.5194/acp-20-13041-2020 Keshtmand et al. (2026): GATESBackground: Emulating background greenhouse gas mole fractions for regional atmospheric inverse modelling with graph neural networks, preprint egusphere-2026-3361 Ganesan et al. (2017): Atmospheric observations show accurate reporting and little growth in India's methane emissions, Nature Communications, 8, 836, https://doi.org/10.1038/s41467-017-00994-7 Tunnicliffe, R. and Keshtmand, N.: CAMS background methane mole fraction curtains - South America, Zenodo [data set], https://doi.org/10.5281/zenodo.21035682, 2026.