Recent improvements of 1D models of volcanic plume rise: Implications for characterizing sulfur emissions by explosive volcanic eruptions and the subsequent sulfate aerosol forcing.

Abstract:

Explosive volcanic eruptions can inject sulfur gases directly into the stratosphere and are a major source of stratospheric sulfate aerosols. Predicting the spatio-temporal evolution of the aerosol burden and subsequent forcing requires to characterize the vertical distribution of sulfur inputs into the stratosphere. This necessitates the use of model of volcanic plume rise for reconstructing past eruptions or designing future eruption scenarios. Such models are also commonly used during volcanic crisis to estimate the rate of emission of gases and ash.
Several classes of model of volcanic plume rise exist, but 1D "integral" models are widely used because they are computationally efficient and facilitate sensitivity studies or near real-time predictions of eruption source parameters. In these models, both turbulent entrainment of atmosphere into the volcanic plume and condensation of entrained water vapor contribute significantly to the strength of a plume (i.e., its buoyancy flux) and must be parameterized. Various models exist for both processes and uncertainties on the parameters on which they rely are large. Despite a recent “intercomparison” of model results for conditions representative of two eruptions, no extensive evaluation and comparison using a large database of eruptive parameters yet exists.

Here, we use a new database of 94 eruptive phases with independently constrained plume heights and source conditions to evaluate four integral models of volcanic plumes. Using a Monte Carlo approach to account for uncertainties in eruption parameters, we evaluate the best set of entrainment and condensation parameters for each model, and compare their performance. We further test the models using a recent set of analogue laboratory experiments. Our work provides guidance on the best integral plume model to use, and places constraints on each entrainment or condensation parameter, as well as the associated uncertainties. We show the implications of our results for predictions for the vertical distribution of SO2 in the atmosphere, and in particular for characterizing eruptive conditions leading to significant stratospheric injection of sulfur.