Predictive modeling of coastal hydrophysical processes in multiple-processor systems based on explicit schemes

Abstract

The computational efficiency of implicit and explicit regularized schemes is compared for the suspension transportation and storm surge, which are two actual hydrophysics problems reduced to nonlinear systems of diffusion-convection equations. Regularized schemes are applied to the reformation of the bottom surface caused by the suspension sedimentation on the bottom of the water body under the dumping soil and for modeling the surging storm in the Taganrog Gulf of the Azov Sea on September 24–25, 2014 when the water level exceeded the normal level by 420 cm under the effect of a hurricane that lasted 30 h with a wind speed of 40 m/s. The computational experiments demonstrate that the explicit scheme obtained on a 512-core of a supercomputing system of the Southern Federal University in Taganrog with a computational grid with 5001 × 5001 × 101 nodes is faster by a factor of 71.5 than the implicit scheme