Fire. Crédito: Christopher Burns/Unsplash

The new trick against wildfires: supercomputers

Devices can assess a big number of variables simultaneously, such as weather, land, and vegetation, predicting the fire destination and helping prevent and fight against great wildfires

Fire is one of the biggest forces of nature, as seen in the recent California disaster, in which more than 70 people died. In events of such magnitude, the flame may be as tall as a tree, and destroy forests, houses, and kill. They may appear suddenly and spread fast, becoming rapidly uncontrollable. Luckily, a new tool can help to control this destruction power - supercomputers and artificial intelligence.

Though it may seem unpredictable, the fire follows certain rules when some variables like wind, temperature, and fuel are observed. It means that, a priori, if its moves could be tracked, and these data sent to a supercomputer, its dynamics could be predicted.

Learn more:
Sirius: Brazil launches the most up-to-date electron accelerator
Japanese-made radar monitors the Guanabara Bay

However, some challenges must be overcome. One of them is to develop the most appropriate model to collect and process data, which would provide information about the blaze patterns. Even with some difficulties, there are success stories. According to Wired magazine, the atmospheric scientist Alexandra Jonko, from Los Alamos National Laboratory, is working on a supercomputer and a system called FIRETEC to model fires. It simulates, among other things, air density and temperature, as well as the properties of the grass or leaves in an area.

Jonko runs a bunch of simulations with different wind speeds, typically on the scale of 40 acres. She says that it takes about four hours to simulate between 10 and 20 minutes of a fire spreading. FIRETEC produces physics-based data on fire dynamics. Wildfire agencies generally know the ideal conditions - low winds, for instance -, but this type of modeling could help give even more granular insight.

How can supercomputers help?

The first step is to collect information. One of the data resources comes from a scanning laser installed on an airplane, which lets researchers visualize trees in 3D and detail the vegetation underneath them.

Another way of collecting data is from a ground-based station. On mountaintops in San Diego, CA, lookout stations are loaded with sensors like high-definition cameras and wind and moisture detectors that work uninterruptedly. It’s the ALERTWildfire. If the camera catches a fire breaking out, the sensor data is immediately sent to the supercomputer, which does real-time modeling of the blaze for fire agencies. This way, the information provided to the firefighters becomes more accurate, and evacuation warnings can be issued.

Only a supercomputer can process such complex data in such a short time. That’s why the equipment is so necessary. Wind-driven fires move quickly, and the bigger a fire gets, the more data the device produces.

Currently, 85 cameras are deployed in San Diego, but the researchers hope to expand to over 1,000 across California. Also, human eyes must watch the camera feeds to detect fires, though the idea is to get AI to do that in the future.

Causes and proportions

According to experts, the main causes of wildfires are lightning, fire on the pasture to clean or promote new leaf growth, campfire or hunt, arson, and acts of negligence, such as burning cigarettes or matches discarded on the ground.

In 2018, Greece was strongly hit by wildfire, which spread to many parts of the country, killing at least 74. The fire destroyed houses, blocking important routes and leading people to flee the region. It started in Mati, a village located in the Attica region. Authorities suspect arson was provoked to eliminate houses in the region.

California also suffered from the same problem, as mentioned above. In 2018, the state was hit by the most destructive wildfire in California history. The Camp Fire broke out in Northern California and covered an area of around 125,000 acres, and affected 7,600 structures, 6,453 of which were family houses.


Published 16 January 2018