The Canadian Severe Storms Laboratory (CSSL) has been operating for a year now. It’s a nationwide project based at Western University, where data from scientists and engineers across the country are used to study extreme weather and help communities better prepare for future events. 

London Morning host Andrew Brown checked in with Western’s Northern Tornadoes Project director David Sills to find out what they’ve learned so far. 

The following has been edited for length and clarity.

Andrew Brown: I want to start with this lab. What is it? 

David Sills: We started the Northern Tornadoes project back in 2017, and that’s been a big success. We’re detecting more tornadoes than ever now. 

In 2022, we started the Northern Hail Project, and it’s mostly busy in Alberta, where the biggest hail problem is. And they’ve been out every year chasing storms, collecting the hail behind them before it melts, and learning lots about hail. 

Then in 2024, we started a new project called the Northern Mesonet Project. That looks at regional weather networks, collecting data to support all of the work that we’re doing across the country. Actually, there’s a network of 3,000 live stations across the country that we’re pulling together. And all of this is now under the umbrella of the(CSSL). 

What’s unique about the CSSL is there’s an American version which is called the National Severe Storms Lab and there’s a European version called the European Severe Storms Lab. So we’re joining the club. 

AB: Are you guys learning from each other and working together? 

DS: Definitely. We’re very plugged in with our partners in the US and in Europe. They helped us to develop how this would work in Canada. 

AB: Is there anything that a tornado guy in Ontario can learn from a hail guy in Alberta?

DS: Once we started the CSSL, we started to kind of merge our capabilities in the field. We used to have separate teams that would investigate hail, another team for hail and tornadoes, and so on. Now we’re training all of our interns and staff to be able to do any kind of surveys because we’re doing our hail survey now in Ontario, and we’re doing tornado surveys in Alberta. So all the teams need to know how to do the surveys for any kind of hazard. And so it’s a multi-hazard approach now. 

AB: There were 116 confirmed tornadoes in Canada in 2025. Every time I say that, it feels like such a big number. What does that number tell you?

DS: Before we started this program, the average in Canada was about 62 tornadoes per year. That’s what we knew about, and we knew there was a lot more going on away from populated areas. And certainly looking at satellite imagery and getting into what’s happening in the forest just north of our populated areas has really told us a lot more about how many tornadoes we get. 

In fact, last year seemed to be a really quiet year across the country. In southern Ontario, there were only five tornadoes and they were just mostly weak tornadoes. And then we started looking more closely at the satellite imagery. There were just a whole bunch of tornadoes that occurred in the north last year that we didn’t know about until we reviewed the satellite imagery and suddenly the numbers went back up to above normal. 

AB: How do you use satellites to figure out that tornadoes happened in the past?

DS: There are satellite constellations out there that give us high-resolution global daily imagery. If we know a storm has occurred, we can look at the days before, the days after and compare the imagery. And if you kind of just swipe back and forth between the images, the damage really pops out. If it’s a long, narrow path, it’s generally a tornado. If it’s more of a kind of a circular or a wide and long path, it’s more likely to be an outburst. Sometimes we have to investigate with teams or with aircraft to see which one it actually was. But because of all of that satellite imagery, we’re learning a lot more.

In the past, at the end of the season, if we missed anything, we’d review all the satellite imagery across Canada to see if anything turned up. And we’d always find a dozen or so tornadoes. But that involves about 300 hours of staff time pouring over satellite imagery. 

Last year, for the first time, we used an AI tool that would go across it and see if it could find stuff. And it found far more than I think we’d normally get. And so we’ve been spending all of the time basically verifying if what the AI tool picked up was actually wind damage, then trying to characterize whether there was a tornado or downburst. 

AB: Do you expect to be leaning on AI a bunch going forward with this?

DS: Only where it makes sense. I mean, we’re using it to save staff time. All you’re doing is these tedious tasks so they can focus on doing things like writing papers and more productive tasks. So I think it’s good there. We’ve got another AI tool that looks at the way trees have fallen after a tornado has gone by. And these tree fall patterns can actually tell you the intensity of the tornado. So it’s actually helping us rate the tornado on the EF scale more accurately. 

AB: When we talk about the Canadian severe storms lab and all this technology you’re using, I’m picturing you sitting in this really high-tech laboratory doing all of this research. Like, what does it look like where you work?

DS: Basically a desk with a computer. The lab next door to me, which is a bit bigger, that’s where a lot of the analysis work gets done. The university is building a new engineering building next to Alumni Hall, and our project is going to get the top floor of that building. We’re planning to have much more sophisticated-looking offices.