![Sample our Environment & Agriculture journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5934/TOC-Subject-Sample-2021-Environment-Agriculture.jpg"})
Abstract
The simulations and potential forecasting of dust storms are of significant interest to public health and environment sciences. Dust storms have interannual variabilities and are typical disruptive events. The computing platform for a dust storm forecasting operational system should support a disruptive fashion by scaling up to enable high-resolution forecasting and massive public access when dust storms come and scaling down when no dust storm events occur to save energy and costs. With the capability of providing a large, elastic, and virtualized pool of computational resources, cloud computing becomes a new and advantageous computing paradigm to resolve scientific problems traditionally requiring a large-scale and high-performance cluster. This paper examines the viability for cloud computing to support dust storm forecasting. Through a holistic study by systematically comparing cloud computing using Amazon EC2 to traditional high performance computing (HPC) cluster, we find that cloud computing is emerging as a credible solution for (1) supporting dust storm forecasting in spinning off a large group of computing resources in a few minutes to satisfy the disruptive computing requirements of dust storm forecasting, (2) performing high-resolution dust storm forecasting when required, (3) supporting concurrent computing requirements, (4) supporting real dust storm event forecasting for a large geographic domain by using recent dust storm event in Phoniex, 05 July 2011 as example, and (5) reducing cost by maintaining low computing support when there is no dust storm events while invoking a large amount of computing resource to perform high-resolution forecasting and responding to large amount of concurrent public accesses.
Acknowledgements
Research reported is supported by NSF (CSR-1117300 and IIP-1160979), NASA (NNX07AD99G), and Microsoft Research.