Performance of WRF cloud resolving simulations with data assimilation on public cloud and HPC environments

Abstract

The Weather Research and Forecasting (WRF) model is a numerical weather prediction system commonly used for atmospheric research and operational forecasting. Given the great amount of computing resources needed by this model, a HPC or cloud computing infrastructure is needed. In this paper, high resolution simulations (1 km) with data assimilation have been ran on different configurations in Amazon Web Services Cloud computing environment, comparing the performance obtained with different computing sizes and storage technologies. A comparison between performance obtained with a Cloud computing setup and a HPC cluster has also been made.