Enhancing Confidence in Volcanic Ash Forecasts: Approaches for Quantifying and Reducing Uncertainties

Generating accurate quantitative volcanic ash forecasts poses a considerable challenge, especially in remote volcanic locations where there are large uncertainties regarding the timing, quantity, and vertical distribution of ash released. Quantifying this uncertainty is crucial for providing timely and reliable volcanic hazard warnings. Our research presents methodologies aimed at quantifying the uncertainty associated with volcanic ash dispersion forecasts.  Ensemble forecasting techniques that account for input, parametric, and structural uncertainties result in volcanic ash forecasts with low confidence, making them challenging for decision-making. To enhance the utility of these forecasts, we introduce data assimilation methods that combine ensemble volcanic ash predictions with satellite retrievals, considering uncertainties from both data sources. This approach enables us to constrain emission estimates, subsequently increasing confidence in the forecasts. By applying a risk-based methodology to these refined dispersion simulations, we significantly reduce the high-risk areas, minimizing disruptions to flight operations. These findings offer insights for the design of ensemble methodologies, facilitating the transition from deterministic to probabilistic volcanic ash forecasting.