The Center for Data Innovation spoke with Dr. Arthur Lane, project coordinator of FireSat, a wildfire detection initiative sponsored by the National Aeronautics and Space Administration (NASA). Lane discussed how a network of 200 satellites could detect fires on Earth as quickly as 10 minutes after they begin, as well as how this technology could help monitor other hazards such as oil spills and volcanic eruptions.
This interview has been edited.
Joshua New: FireSat is a project based on technology developed by NASA’s Jet Propulsion Laboratory (JPL) to aid wildfire prevention around the world with hundreds of space-based sensors. Could you explain how FireSat will work? How is it feasible for space-based sensors to monitor the entire planet for minor changes that could indicate a fire?
Arthur Lane: The earliest conception of FireSat was from around 2011 or so and was the fusion of work and ideas of several groups at JPL. The use of infrared technologies and imaging systems for fire detection actually predates FireSat, albeit it was applied more locally through the use of sensor systems on helicopters and fixed-wing aircraft. There were also a few earth-orbital satellites active in the 2010 time frame, though their orbital configurations generally only allowed for views of the same piece of earth once or twice a day and their sensors were not typically devoted to fire detection.
What brought the FireSat concept forward, I believe, was three factors: The emergence of large and destructive wildfires in the western United States and Canada in 2012 and after; the development of advanced algorithms that can run on-board a satellite and rapidly process imaging data streams to recognize subtle anomalies, which vastly diminishes the need for downloading huge data streams to ground stations; and the feasibility of having a sufficiently large number of satellites in orbit to provide high-frequency repetitive observations, ranging from 10 to 20 minutes, of the same area.
As envisioned currently, FireSat would rely on almost identical sensor systems installed on over 200 communication satellites, with 8 to 12 satellites spaced along a single circular orbit, and with 20 or more orbits that would provide full coverage of Earth. These sensors can record a series infrared (IR) images when they detect the high IR emissions given off by a fire source. The satellites’ algorithms will assess thermal hotspots in real time and sophisticated JPL software would examine hotspot movement across the IR images to confirm a fire and assess its growth.
For this whole process, only a few hundred bytes of data would have to be downloaded to Earth each time a satellite passes. The next satellite following in the same orbit would see that same location 10 to 20 minutes later and if the hotspots were present in this second image set, additional data would also be transmitted to grounds stations where additional analysis could take place. The satellite array is expected to have a greater than seven year lifetime and critical replacements could be launched as needed in the future.
New: How does FireSat differ from existing space-based fire detection efforts? Surely we already have satellites that are capable of taking pictures of the Earth that could reveal a fire.
Lane: The superlative value in FireSat is not the fire detection, per se, but rather the ability to detect and monitor the growth and movement of a fire within the first hour or so of its onset. This is especially valuable for remote locations where multi-square mile destruction can occur within a short amount of time. There are satellites that take pictures of Earth and some have the ability to detect small-scale fires, but often times nobody examines this data until hours or days after it is acquired, which reduces its value in time-sensitive situations.
New: FireSat will monitor the entire globe, not just the United States. Will it be a challenge to share this data with other countries?
Lane: I am fairly certain that once the system is in place and functioning, it will be able to adequately provide information to other countries. FireSat sensors will be based on commercial products so there should be no technical obstacles to data sharing.
New: With so many sensors collecting so much data, how can FireSat or government agencies process this data quickly enough to respond to a wildfire before it grows out of control?
Lane: By having some of the analysis occur onboard the satellites, FireSat will not be sending huge quantities of data to ground stations, so it will be easy to quickly distribute and analyze this information. Confirming a fire after the first signs can take as little as 10 minutes, during which we can activate an alert network and provide information about the scene. At this point I am not able to provide details on these algorithms, but I know they will be tested in multiple capacities as the sensor package matures.
New: Can FireSat image data be useful for other purposes beyond fire detection? Will this data be publicly available for analysis?
Lane: Yes, depending upon the final technical specifications of the sensor systems, it would be possible to image other physical phenomenon. Based upon some preliminary published research, it appears possible to detect oil spills in bodies of water with this kind of technology. Also, satellites could monitor volcanic lava flows and track their movements to identify any safety risks. And it may even be possible to detect volcanic pre-eruptions. There are several other detection and monitoring capabilities that might be possible, but how well such concepts mature into practice will depend on hardware parameters and how the detection algorithm suite can be tweaked as we learn more about those possibilities.
