In a new tactic in Beijing’s growing battle on choking smog, sensors and analytics will pinpoint the source and trajectory of polluting particles and forecast levels three days in advance down to the resolution of individual streets.
IBM, which is working with Beijing officials, says its cognitive computing systems will “analyze and learn from streams of real-time data” from air-monitoring stations and satellites. IBM hopes to have built an analytics platform by early next year based on the inputs of existing sensors in the Chinese capital. After that, IBM will install additional sensors and develop analytics programs at the direction of the Beijing government.
The Beijing government has said that it will spend $160 billion to reduce the density of so-called PM2.5—fine smog particles that are 2.5 micrometers or smaller in diameter—by 25 percent by 2017. The particles come from a variety of sources, including coal burning and industrial and vehicle emissions.
The problem is so bad that the Shanghai Academy of Social Sciences recently said that Beijing is “almost unfavorable for human living.” Beijing’s 2013 PM2.5 concentrations averaged 89.5 micrograms per cubic meter last year, and surged above 600 on the worst days this past January. The World Health Organization says levels must remain below 25 to be safe.
Tao Wang, resident scholar at the Carnegie-Tsinghua Center for Global Policy, a Beijing-based think-tank, says better data will help. “They know roughly the sources, but not the proportion from the different sources, or the interaction between pollutants,” Wang said. “The capability to monitor is lacking and they need to improve that. A lot of what this is about is getting this down to a lot more micro areas of the city, and getting more discrete measurements.”