A Mechanism on Parallel Processing to Numerical Weather Prediction for Weather Routing-Ⅱ-Accuracy Evaluation and Performance Benchmark of the Sea Surface Wind Prediction

摘要

Regarding the development of a land support system based on weather routing to calculate the optimal route of the vessel, which is required to use the high accuracy datasets of predicted ocean environments. However, recent weather routing services are not possible to forecast the environmental datasets by themselves and use limited datasets by forecast agencies with high-cost in the limited sea area. In this paper, we show the forecast results of the surface wind velocity at a local sea area, which is based on our cluster system. In addition, we evaluate the forecast accuracy of the surface wind velocity between the WRF (Weather Research and Forecasting) and the GSM (Global Spectral Model) models. Finally, we measure the peak performance of the cluster system, the elapsed time of the forecast simulation based on GCC and Intel compilers, and the network performance of the bandwidth and the round trip time.