Integrated Sakurajima Volcanic Ash Fall Distribution Under Climate Change An Analysis based on JRA-55 Data using Neural Network 1. Background • A volcanic eruption is one of the events which emit several dangerous pollutants that led to catastrophes. • Volcanic ashes, apparently hampering not only the resident who lived near the volcano but also the other citizens in the farther area that can get indirect impact from that. • The volcanic ash released by large-scale eruptions may cause serious damage to critical infrastructures, buildings, and cause health problem. PUFF Model DPRI, Kyoto University 3. Methodology • Integrated volcanic ashfall risk assessment in large-scale eruption. a. Systematically estimate the risk of volcanic ashfall from probable large- scale volcanic eruption in Sakurajima. b. Historical weather data is used for risk estimation to validate and understand ashfall risk. c. Get critical characteristic from historical eruption data and past climate data for verifying implemented policies obtained from the prior result. • Confirmation of risk assessment between historical weather data and predictive weather data that will enhance the decision support system for volcanic disaster risk management for the development of early warning system. a. Compare daily ash fallout map between results obtained from both predictive data and historical data on general occasion. b. Compare daily ash fallout map between results obtained from both predictive data and historical data on specific occasion. c. Thorough analysis by comparing similarities and anomalies detection between both results. • The confirmed ash fallout distribution data, added with population data, will be extended to the development of integrated evacuation planning that considered evacuation willingness of the potential endangered resident. 2. Objectives 4. Results on Annual and Monthly Observation 5. Result under Typhoons No. 24 (Sep 28 th -30 th ) 6. Conclusions • Ash fallout distribution maps produced from PUFF model with JRA-55 data give risk assessment coverage confirming the risk threshold on different risk components. • In typhoon approaching situation, the alterations in ashfall dispersion course is detected by using JRA-55 data. • Lastly, by conducting ashfall risk-assessment on large-scale eruption scenario for long period and confirming the development of ashfall early warning system, we can build more comprehensive evacuation planning based on people behavior towards the ashfall hazard. 7. Future Works :Position vector of the −th particle at time :Initial locations of the particles as a source term :Local wind velocity to transport the particle :Diffusion velocity :Gravitational fallout velocity ∆:Time step (300) Probability Distribution Eruption Scenario Risk Map for 0.1 mm thick ash fall in 2018 The risk threshold for ash fall hazard Risk Map for 5 cm thick ash fall in 2017 Risk Map for 5 cm thick ash fall in January 2017 Risk Map for 5 cm thick ash fall in August 2017 Risk Map for 50 cm thick ash fall in August 2017 Risk Map for 50 cm thick ash fall in October 2017 Typhoon Trajectory Risk Map on Sep 28 th , 2018 Risk Map on Sep 29 th , 2018 Risk Map on Sep 30 th , 2018 Haris Rahadianto Master Student Graduate School of Informatics - Kyoto University [email protected] Kyoto - JAPAN Haris Rahadianto・Subhajyoti Samaddar・Hirokazu Tatano