Abstract
Wind wave energy balance equation has a capability to predict time development of a spatial distribution of wind wave induced by strong wind fields. While it has been expected that the equation is suitable for the wind wave estimations in the lakes because the equation can take account of topographic effects caused by spatial changes of the lake's depth and shoreline, huge computational cost is obstructing the utilization of numerical simulations of the equation in daily wind wave predictions. In this work, to overcome the difficulty, GPGPU technology was applied and its efficiency of acceleration of numerical solutions was investigated. By using decoupled model as a source term of wind wave generation, efficient acceleration was realized. The developed GPGPU solver was applied to Kasumigaura Lake. The simulation results for the Typhoon show that the solver can provide highly rapid solutions.
