Impact of input parameter uncertainty on tephra fallout deposition forecasts: insights from Mount Ruapehu, Mount Etna, and Villarrica

Tephra fallout poses a significant hazard to communities and infrastructure, yet forecast accuracy is highly sensitive to uncertainty in eruption source parameters. A common approach is to inform these input parameters using prior knowledge of a volcano’s magma composition(s) or past eruptions. However, this effect of such informed input parameter values on forecast accuracy has not been systematically tested. In this study, we test how different input parameter choices influence forecasts of tephra mass deposition, using the 17 June 1996 eruption of Mount Ruapehu, Aotearoa New Zealand, the 28 February 2021 eruption of Mount Etna, Italy, and the 3 March 2015 eruption of Villarrica, Chile, as case studies. Forecasts were evaluated in two ways: (i) mass deposition forecasts, comparing simulated and observed tephra mass at specific locations, and (ii) impact-based forecasts, assessing whether simulated tephra mass exceeded defined impact thresholds. We show that forecast accuracy was generally low, with only 2.8% of the simulations achieving ≥ ≥ 50% impact-based agreement. Moreover, forecast performance did not differ systematically between reference and informed input parameter distributions, highlighting that reliance on past eruption data does not necessarily improve, and can sometimes bias, forecast outcomes.