VolcanEESM: Global volcanic sulphur dioxide (SO2) emissions database from 1850 to present - Version 1.0

The database is a combination of all available information possible from the wider literature with as many observations of the amount and location of SO2 emitted by each eruption. The database includes no information about the size, mass, distribution or optical depth of resulting aerosol. As such the database is model agnostic and it is up to each modeling group to make decisions about how to implement the emission file in their prognostic stratospheric aerosol scheme.

The dataset is divided into two parts based on the availability of satellite data. For the pre-satellite era, the necessary information about the emissions was gathered from the latest ice core records of sulphate deposition in combination historical accounts available in the wider literature (see references included in the database for specific citation for each record). In the satellite era, volcanic emissions were primarily derived from remotely sensed observations.

For the period 1850 CE to 1979 the project combined the most recent volcanic sulfate deposition datasets from ice cores with volcanological and, where applicable, petrological estimates of the SO2 mass emitted as well as historical records of large-magnitude volcanic eruptions. In detail, for the majority of eruptions between 1850 CE to 1979 , there are few direct measurement of SO2 emissions or quantitative observations of the plume height and very few measurements of the aerosol optical depth (AOD).

The project therefore focused on large-magnitude (VEI>4) eruptions, which was used to identify using the Smithsonian Global Volcanism Program database and ice-core data. The total amount of SO2 emitted by an eruption, including the spatial distribution of the volcanic injection, was determined by using parameters that based on analogous eruptions that occurred during the satellite era following the methodology of Stoffel et al. [2015] and Gao et al. [2008].

The project estimated lower and upper bounds of the SO2 emitted and constrained the season by an eruption [similar to Stoffel et al., 2015] and expect the uncertainties on our inventory to be of comparable magnitude to those prescribing the forcing. Before running long model integrations, they assessed the representativeness of the emission inventory and quantified uncertainties by comparing simulated sulphate deposition patterns for selected eruptions to ice-core records and by comparing simulated AOD to Crowley et al. [2008] and Sato et al. [1993]. Where applicable they also compared simulated optical properties to astronomical estimates of atmospheric transmission [Stothers 1996; 2001].