Table1_Optimizing mass eruption rate estimates by combining simple plume models.XLSX

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.

火山灰柱注入大气中的火山碎屑（tephra）是爆炸性火山喷发的核心灾害之一。对火山灰大气扩散过程进行预测，需要精准掌握当前的喷发源参数，尤其是喷发质量速率（mass eruption rate，MER）——该参数用于量化火山口处气体与火山碎屑的质量流动速率。由于该参数无法实时直接测量，相关监测工作旨在间接评估MER，例如通过构建羽柱模型，将相对易于观测的喷发柱高度与火山口处的质量通量建立关联。本研究将模型估算结果与独立火山喷发源参数数据库（Independent Volcanic Eruption Source Parameter Archive, Aubry等, J. Volcanol. Geotherm. Res., 2021, 417）中收录的130次火山喷发的独立沉降测量数据进行对比，测试了6个零维（0D）羽柱模型以及4种不同建模方法的预测成功率，以优化MER的预测精度。研究结果表明，切勿仅针对所有场景单一使用某一款羽柱模型；若合理选取模型权重因子对各羽柱模型进行加权组合，可提升MER预测的准确性。模型权重因子的最优选取，取决于待监测火山喷发可获取的火山学与气象学信息的类型及完备程度。本研究构建了一套决策树，可辅助研究人员选取最优建模策略，以实现最高精度的MER预测。