In-situ ensemble visualization for analyzing pollutant dispersion

Yano, Midori  ; Kawamura, Takuma ; Hasegawa, Yuta   ; Idomura, Yasuhiro   

Uncertainty quantification in numerical simulations is important for improving the reliability of simulation results. In-Situ visualization techniques that visualize simulations at runtime are available to evaluate large-scale, high-speed, and cutting-edge ensemble simulations using supercomputers. This study proposes an in-situ visualization method for ensemble simulations and applies it to analyze statistical properties in ensemble calculations of real-time plume dispersion analysis.