AAAR 34th Annual Conference
October 12 - October 16, 2015
Hyatt Regency
Minneapolis, Minnesota, USA
Abstract View
Application of Big Data Technologies for Aerosol Modeling: A Perspective
SATISH VUTUKURU, Independent Researcher
Abstract Number: 707 Working Group: Urban Aerosols
Abstract Computational modeling has become indispensable to study and help solve our environment, energy, and climate challenges. The first generation computer models for air quality were developed in 1980s and 1990s. This period also saw great advances in large-scale parallel computing especially for scientific computing applications. Consequently, computational modeling, most notably air quality and aerosol modeling, benefited from such advances. Most of the models used today still use parallel computing paradigms developed in 1990s. In recent years, large-scale computing has been widely adopted by data-driven consumer internet companies. This has led to the development of a suite of technologies, most notably the Apache Hadoop ecosystem, to handle large amounts of data and perform robust parallel computing using commodity hardware. In this paper we propose research directions to develop next-generation models that take advantage of these web-scale technologies. We primarily focus on the following three major themes: (a) Application of the map-reduce paradigm of parallel computing using Hadoop and associated technologies, (b) Large scale data management, analysis, and visualization tools, and (c) Mobile and cloud computing technologies. We discuss how these new methodologies can be used for air quality, aerosol, and exposure modeling.