Feature - Seeing particles with VPM
Before we can make use of data, we need to make sense of it. But with complex concepts such as particulate air pollution, you could just as easily drown in the data.
And that’s exactly what was happening when NASA first approached Uma Shankar, an atmospheric scientist at the Institute for the Environment at UNC Chapel Hill, to ask what sort of advanced visualizations the particulate matter research community needed.
After some thought, Shankar suggested an application to visualize particulate matter across the range of sizes in which it occurs.
“Particulate matter has such important impacts on a variety of air quality issues, especially human health” Shankar explained. “Now we better understand the connection between particulate matter and climate, so its importance is even greater than we originally understood.”
Despite this better understanding, the existing visualization tools were not up to the task of seeing how PM evolves over the entire range of particle sizes, as time passes and the PM moves with the wind, according to Shankar.
“We didn’t want to have to sift through multiple images to get the information,” Shankar said.
“I wanted some efficient software development that RENCI is just aces at,” Shankar said. “I knew that from looking at some of their visualizations for other models.”
The result of that collaboration, Visualize Particulate Matter, is unusual in several ways. According to Shankar, existing software for visualizing PM is not interactive, and only a handful of private applications will superimpose model data of how PM varies over its size range over the equivalent observed data.
“There are a number of different types of particles often interpreted as pollutants,” said RENCI senior visualization researcher David Borland. “These individual [data] files can have over a hundred different kinds of pollutants and variables that we’re trying to visualize.”
The solution Borland and RENCI senior software developer Steve Chall came up with is a big improvement, according to Shankar.
In the left-hand panel, three separate graphs plot particle size against the number of particles, their surface area, and volume of particles in a given volume of air. Each graph may have several curves of various colors, each corresponding to a location marker of the same color on the map in the right-hand panel. Those locations are selected by clicking on the map on the right-hand side, or by manually entering the location’s coordinates; the data in question can be observational, or generated using a simulation. When the visualization is running, the graphs and map vary over time, as well as through the atmospheric layers, for each of the selected locations.
VPM can also show the simulation data for a trajectory, a valuable feature when trying to find out where PM pollution (which can travel long distances) came from, where it will go, and how it changes along the way.
Currently, VPM is in the form of a desktop application for Linux. To use it to view simulation data, the user has to start by running a simulation. (At the Institute for the Environment, they run the simulation annually on the Emerald Beowulf Linux cluster, which consists of 850 processors). Next, they take the data output from their simulation and plug it into VPM.
Soon, however, the VPM collaborators hope to be running a version of the application on a larger web-based platform, the Visibility Information Exchange Web System. The VIEWS version will access existing online data sets, rather than allowing users to upload their own.
“One of the big challenges was actually architecting it in such a way that it would be easy to write multiple interfaces to it,” Borland said. “The core code is written in C++, so that required figuring out how to wrap that in Java to be able to call it from a website.”
“I think over the next few months I’ll be looking at what capabilities can be added,” Shankar said. “There’s very much a need to plan for the longevity of this tool, so we are going to be trying to make this sustainable long term.”
—Miriam Boon, iSGTW