Optimizing mass eruption rate estimates by combining simple plume models

Abstract

Tephra injected into the atmosphere by volcanic ash plumes poses one of the key hazards in explosive eruptions. Forecasting the atmospheric dispersal of volcanic ash requires good knowledge of the current eruption source parameters, in particular of the mass eruption rate (MER), which quantifies the mass flow rate of gas and tephra at the vent. Since this parameter cannot be directly measured in real-time, monitoring efforts aim to assess the MER indirectly, for example, by applying plume models that link the (relatively easily detectable) plume height with the mass flux at the vent. By comparing the model estimates with independently acquired fallout measurements from the 130 eruptions listed in the Independent Volcanic Eruption Source Parameter Archive (Aubry et al., J. Volcanol. Geotherm. Res., 2021, 417), we tested the success rates of six 0D plume models along with four different modelling approaches with the aim to optimize MER prediction. According to our findings, instead of simply relying on the application of one plume model for all situations, the accuracy of MER forecast can be increased by mixing the plume models via model weight factors when these factors are appropriately selected. The optimal choice of model weight factors depends on the availability and type of volcanological and meteorological information for the eruption monitored. A decision tree is presented that assists the reader in finding the optimal modelling strategy to ascertain highest MER forecast accuracy.