The National Institute for Computational Sciences

Tracking Pollutants

Research May Explain How Air Pollution Can be Detrimental Even When Concentrations Drop

When we see smog or experience the respiratory consequences of compromised air, we may be curious as to the origins and nature of the pollutants. But researchers in the realm of fluid mechanics seek answers by performing some seriously scientific detective work. Theirs is the task of tracking the trajectories of the molecules of a pollutant backwards in time, starting from its current position.

"In order to do so, and to obtain results of sufficient accuracy, we need to have a complete record of the position coordinates of many millions of such molecules spanning a fairly long time period," says Georgia Tech Professor P. K. Yeung. "More than likely, the wind that carries the pollutants is turbulent, which brings us to the science of studying turbulence via numerical simulations capable of providing tremendous detail."

Yeung and his students, along with Yeung's collaborator, Brian Sawford of Monash University in Australia, have been advancing the fundamental understanding required to follow the complex path of molecules, thanks to a competitive allocation of advanced high-performance computing resources from the National Science Foundation's Extreme Science and Engineering Discovery Environment (XSEDE)—specifically, the supercomputers Kraken at the National Institute for Computational Sciences (NICS) and Stampede at the Texas Advanced Computing Center (TACC) (under XSEDE project number TG-MCA99S022).

Yeung explains that his group has conducted simulations in which the positions of millions of particles are tracked forwards in time and saved for analysis at a thousand steps or more.

Without running into conflict with the fact that the equations of motion are not reversible in time, the group has been able to obtain backwards path information through a post-processing calculation that counts many billions of pairs of molecules from existing data sets, he says.

"This research has shown that dispersion happens faster going backwards than forwards, which means that, under a suitable combination of conditions, turbulence brings different fluid parcels together faster than it moves them apart," Yeung explains. "This observation may be important in explaining how chance accumulations of pollutant material may occur locally with detrimental consequences even when the overall average concentration drops. The actual details may, of course, also depend on how the action of turbulent transport is coupled with molecular diffusion, which may be weak or strong depending on the nature of the pollutants involved."

Yeung notes that because of the size of the problems involved in this research, the generation and analysis of the necessary data are inherently resource intensive, which led to his group's use of the NICS-managed Kraken, before it was decommissioned on April 30, and Stampede at TACC.

Each of the two resources provided the necessary computing performance to investigate turbulent transport and molecular diffusion, including direct simulations using up to 20483 grid points, Yeung says.

But he explains that working with the two supercomputers not only allowed the group to pursue its efforts at promoting a greater understanding of the physics but also to have a working exposure to the challenges of optimizing a given code to obtain the best performance on multiple architectures with machine-dependent characteristics. Experience gained in developing flexible codes will be very helpful as new kinds of computing platforms arrive in the multi-petaflop and exaflop eras to come in the future, Yeung adds.

The challenge involved in data generation for Yeung's project are the subject of a Science Track Talk at the XSEDE14 conference, which takes place in Atlanta, July 13–18.

Scott Gibson, science writer, NICS

Article posting date: 13 June 2014

About JICS and NICS: The Joint Institute for Computational Sciences (JICS) was established by the University of Tennessee and Oak Ridge National Laboratory (ORNL) to advance scientific discovery and state-of-the-art engineering, and to further knowledge of computational modeling and simulation. JICS realizes its vision by taking full advantage of petascale-and-beyond computers housed at ORNL and by educating a new generation of scientists and engineers well versed in the application of computational modeling and simulation for solving the most challenging scientific and engineering problems. JICS runs the National Institute for Computational Sciences (NICS), which had the distinction of deploying and managing the Kraken supercomputer. NICS is a leading academic supercomputing center and a major partner in the National Science Foundation's eXtreme Science and Engineering Discovery Environment, known as XSEDE. In November 2012, JICS sited the Beacon system, which set a record for power efficiency and captured the number one position on the Green500 list of the most energy-efficient computers.