Climate change has increased the unpredictability of weather patterns, and as residents of the Western U.S. know, that means a jump in the number of forest fires. In the past year alone, 76,000 individual fires have consumed 5.8 million acres, according to the U.S. National Interagency Fire Center. We might not be able to stop fires from starting, but researchers at the University of Maryland, Baltimore County (UMBC), think they can at least provide us with better information once fires have been ignited.
At the moment, smoke pattern analysis is limited to front-line reports, low-resolution satellite images, and weather data that gets updated every six hours. But by using IBM InfoSphere Streams (technology that analyzes information from a stream of sources), researchers hope to cull data from already-existing surface, aerial, and satellite sensors to identify the progression of wildfires, to model fire and smoke behavior forecasts, and to issue real-time forecasts for firefighters. UMBC researchers will also use servers to process massive amounts of smoke plume models for predictions on where and when fires will spread. Such information could be invaluable for anyone living close to the fires — more accurate smoke and safety alerts could help people get out of the way faster.
It will presumably take time for the UMBC researchers to figure out whether their research is effective, but if it is, residents of fire-prone areas can at least cut down on smoke inhalation, and maybe even on loss of property.