Speaker 3
I'm based in Seattle and I'm in London today. Yes.
Speaker 1
Now, we've got together to talk about wildfires, hugely problematic, hugely in the news this summer. We're seeing more and more of them and the effects can be so catastrophic. And traditionally, it's been difficult, hasn't it, to predict not only where, but also when they might break out? The
Speaker 3
process of forest fires is completely stochastic, meaning that it follows a random distribution, which means that it's impossible to take when specifically which date or when or where are they going to start that we know for a fact. No matter how much technology we have, there will be a lot of randomness as part of the problem. But what we can do is we can use technology to help us understand the probabilities of forest fires from both, like from a calendar perspective, but also location. Technology can definitely play a role there.
Speaker 1
And how accurate is that at the moment? So, from
Speaker 3
a probability perspective, there's still room for improvement, but it's quite accurate to understand again from a probability distribution, which means that you can map, let's say we're talking about Canada, for example, you can map Canada and using historical data, you can build them all that tells you for the next forest fire seasons, what are the probabilities of forest fires and which locations can have higher chances for fires? That can be actually quite accurate. Of course, that's not going to tell you where or when, but it's going to tell you this area here has higher chances than another area there
Speaker 1
basically. So, in theory, that would be a good place for the emergency services to perhaps prioritize.
