Inventory data is a key component of scientific and regulatory efforts focused on air pollution, climate and global change and also a critical compliment for observational emission efforts. The Community Emissions Data System (CEDS) project aims to provide consistent estimates of historical anthropogenic emissions using an open-source data system. The first product from this system was anthropogenic emissions over 1750-2014 of reactive gases, aerosols, and carbon dioxide, for use in CMIP6. These data are annually resolved, have monthly seasonality, were estimated at a moderately detailed level of ~50+ sectors and 8 fuel types, and were mapped to spatial grids.

CEDS combines bottom-up default emissions estimates that are calibrated to country-level inventories where these are deemed reliable. Outside of years where inventories are available, driver data and emission factors are extended using user-defined rules. The system is designed to facilitate annual updates (so the most recent inventory data is available). The software and most input data are being released as open source software in order to provide access to assumptions, improve emission estimates, and allow access to fundamental emissions data for research purposes.

We report on our efforts to expand the spatial resolution by estimating emission trends by state/province for large countries. This will allow spatial shifts in emissions over time to be better represented and make the data more useful for research such as that discussed in this session. As part of these improvements we will add support for use of regionally-specific emission proxies and point sources.

A key focus of ongoing research is better quantification of emissions uncertainty. Our goal is consistent estimation of uncertainty over time, sector, and country. We will also report on results estimating the additional uncertainty associated with extending emissions data over recent years.