By: Luke Jones, Published on March 13, 2018 10:04 AM, Last Update on March 13, 2018 07:07 AM
Wildfires have been a major problem in recent years. In 2016, Fort McMurray was swept by fires, causing 90,000 people to be misplaced for a month, while British Columbia suffered a record wildfire season in 2017. However, research out of the University of Calgary’s Schulich School might be able to prevent wildfires.
The research team has created an artificial intelligence that leverages satellite data to predict where lightning is likely to strike and cause wildfire.
“That will give us a more precise description about the patterns happening in the lightning and the wildfire hazard,” said Dr. Xin Wang, one the researchers in the team. “It also can be used for the future predictions about those hazards.
“Those events don’t just randomly happen. They also have spatial and temporal patterns.”
Lightning is a leading cause of wildfire, especially in Western Canada where fires have taken a heavy toll, stretching fire authorities. The new study uses machine learning so the AI can understand historical data from 2010 to 2016. With the data, the data can be examined to see how lightning affects land, soil, vegetation, and more.
The study published in Sensors open access journal was written by Wang, Jeong Woo Kim and DongHwan Cha.
“We analyzed a number of different types of data over a number of years so we can pinpoint the location that has a high chance of wildfire. Statistically I would say it is more than 90 per cent accurate.”
Kim says the model could be an important tool for fire authorities as they seek to better manage fire-fighting resources.
“If they use this method, they can probably monitor those areas more closely and also they can build more fire stations and so on. They should be able to avoid any pipelines or power lines so they can reduce the hazard.”