A forest fires They sow chaos in Spain. Over the past week they burned More than 30,000 haHaving received the total number of burned lands, very close to the annual record reached in 2022. The reasons are quite clear. Climate change led to an early and dry summer, which leaves the flame to expand freely. Unfortunately, this will be more and more, so it is important act with the maximum impulseIt would be perfectly perfect Predict forest firesField but it is possible?
In the past there was an unwritten rule that was used to predict forest fires or at least know what the most likely areas were. It was a 30-30-30 rule, which established that if the temperature is above 30 ºC, winds are more than 30 kilometers per hour and humidity below 30 %, the probability of a fire is quite high. Today, if we adhere to only this, we will have so many possible spotlights that it would be impossible to optimize available resources. It would be somewhat more precise.
Therefore, in recent years, thanks to Achievements of artificial intelligenceAlgorithms are introduced Predict forest fires Through different strategies. Some have already been tested with excellent results. It is necessary to polish them, but they are on the right track.
How do IA algorithms work to predict forest fires?
AI algorithms can predict forest fires using three different ways. On the one hand, through Analysis of satellite image. They can see factors associated with a higher risk of fire, such as the presence of dry vegetation or human activity.
Mainly, there are algorithms that are based on Historical dataAlthough they also take into account additional information, such as weather and vegetation. Finally, some algorithms do directly Modeling possible fires In different regions, establishing as risk, such as their behavior, and the best ways to stop them.
Two international cases
Canada was one of the countries that suffered most of the forest fires in recent years. Therefore, they have their own forecasting system, in which they are taken into account primarily Data on climatic conditions and atmospheric phenomena last yearFor example, it is important to establish when the snow layer is completely melted in each region, since from this moment on The risk of a forest fire It will be more. Thanks to AI, they can conclude when this will happen in different parts of the country.
On the other hand, of course, the presence of more or less dry vegetation, human activity and, above all, atmospheric phenomena that may affect. Phenomena such as Boy or girlSince it was noticed that both can increase the risk of forest fires.
But if Canada has introduced its own system for predicting forest fires in other parts of the world that they have chosen International cooperationFor example, at the Chile University, they have a system of AI called Cell2fire, which was achieved thanks to collaboration with scientists from other countries, including Spain.
In your case, Advanced simulation. Thanks to data as temperature, wind speed and many other related parameters, you can make modeling that allow you to install when and when the fires of the forest begin, as well as what the best way to stop them.
Great accuracy
Thanks to Cell2fire, for example, the best places for creating can be determined New firewalls Or the points at which the observation must be made. In the first tests with this model of AI, forest fires were already predicted using Accuracy 85 %This field is undoubtedly great news, although it is still impossible to predict absolutely all fires. Especially when they are called. And, unfortunately, in many cases there is no reason than it. To predict human evil is much more difficult, but you can still improve tools for predicting other factors related to factors.
Source: Hiper Textual
