The Decibel - The challenge with preparing for wildfires in B.C.

Episode Date: March 18, 2024

Canada went through its most destructive wildfire season in 2023. Wildfire services rely on data, forestry photography and mapping as a way to proactively control forest fires. However, a recent study in B.C. has found that the data being used is inaccurate and insufficient.Jen Baron, lead author of the study and PhD candidate at UBC’s Department of Conservation and Forestry, explains the inaccuracies in the data, the problems it creates and the ways it could be improved.Questions? Comments? Ideas? Email us at [e-mail]

2023 was the most destructive wildfire season on record in Canada. More than 180,000 square kilometers burned by the end of September. And fire seasons are expected to get worse. Wildfire services across the country are trying to predict and manage what could happen in upcoming fires. But those efforts, in BC in particular, could be hampered by inaccurate data, which means that resources might not be going to where they're actually needed. Jen Barron is a PhD candidate and a lecturer at UBC in the Department of Forest and Conservation Sciences. She led the study looking into this data issue. Today, Jen explains the problems in understanding

wildfire risks, especially in BC, and how to better prepare for wildfires in the future. I'm Maina Karaman-Wilms, and this is The Decibel from The Globe and Mail. Jen, thank you for being here. Thanks for having me. So the study that you worked on focused on the data around something called forest fuel load. I think we just need to explain that. So what exactly does that mean? Yeah, so when we're talking about wildfire, fuels are essentially anything in the natural environment that can burn, but most of the time we're talking about vegetation. So that can include the trees, obviously, and, you know,

the canopy of the trees, the needles or the leaves, as well as the trunks and the branches, but also anything else in a forest or in an ecosystem that might burn. So that includes things like the grasses and vegetation, moss, pine cones, needles on the forest floor, as well as everything that's kind of in that understory layer. So all of those things put together make up the fuel load. Okay, so understory, I imagine, is like closer to the ground, canopies up top? Yeah, so in wildfire, we often separate things into layers

to try to simplify things. So we often talk about things in terms of the overstory or canopy fuels, like what's above our heads, the surface or ground fuels, the things on or below our feet. And then this kind of intermediate layer, which is actually really important because it modifies or mediates that transition from when fire goes from the surface to the canopy. So the more stuff there is in this intermediate layer, the easier it is for fires to transition up to the ground. Gotcha. Okay. How is this organized? Like, can you give me an example of how this is organized? Yeah, sure. So fuels are very important. We care about them a lot because the type of fire behavior that we see is dependent on a couple of things. It's dependent on the weather conditions that a fire burns under, the topography, the lay of the land, and then the fuel, what's there to burn.

And so in locations that have more fuel or fuel different characteristics, for example, pine needles that have a lot of resin in them, that can have a more intense fire, more energy output, more severe consequences than, for example, a grass fire. So they're very important, but they're also very complicated because like we just talked about, there's so many things in the fire environment that can burn on essentially infinite combinations of all of these different things put together. So for that reason, we've developed internationally and in Canada description systems to simplify that problem of characterizing fuels. And so in Canada, we use what's called the fire behavior prediction system, which is a

national system that was developed by the Canadian Forest Service in the mid 20th century, primarily to represent boreal forests. So we have 16 categories that we use to describe these different fuel types across Canada. So I'll give an example. So one of the fuel types that we use is the C3 fuel type. And the name for that fuel type is mature jack or lodgepole pine. So one of the fuel types that we use is the C3 fuel type. And the name for that fuel type is mature jack or lodgepole pine. So what that fuel type describes is not only the overstory, which is dense, mature jack or lodgepole pine, maybe, you know, over a thousand trees per hectare, high density of trees with a closed canopy. So no light penetration in the overstory, but it also kind of indirectly

describes a bunch of other attributes. So it tells us that there's a high distance from the forest floor up to where the canopy starts. There's not very much fuel in the understory, maybe some deciduous shrubs. And then it has a moderately deep and compacted moss in the surface. So even though the name is only describing the overstory, the fuel type describes a whole bunch of other attributes that would also be related to how a fire might burn in that forest. Obviously, 16 fuel types will never be able to describe the diversity of conditions that we have across Canada.

And these were primarily developed to be used kind of at the stand level for fire suppression operations. So essentially, the way these fuel models work is that if I can assign a fuel type to the stand and then I'm, you know, I'm there trying to decide, is it safe to put firefighters on the ground? Do we need to issue an evacuation order? These types of things that I look at, here are my weather conditions. Here's my fuel type. Here's where, how the fire might behave if it moves into the stand. And so then I make some decisions based on that.

The challenge is that we often use these fuel types now beyond that stand level and beyond that operations and management situation. So we have fuel type maps at provincial and national scales that we use to make a lot of proactive management decisions. And so what our study did is it looked at the accuracy and applicability of those maps and whether they could appropriately inform management. Okay. Okay. So there's these 16 categories and they kind of describe specific conditions then in the forest. But Jen, there's so much forest in Canada, right? So I guess, how do we know what kind of fuel is in those forests? Like where does the data come from? Yeah. So both at a national level and at a provincial level in

British Columbia, those fuel type maps come from forest inventory data. So those are data sets that are managed by the province. And that data is collected to describe conditions in our forests and primarily related to how we would make decisions for forest timber management. So it describes primarily merchantable timber, timber that we could sell at market. And so one of the challenges with this approach and using this data set is obviously there's a lot more trees on the landscape than those that are viable at market. So how are they actually like physically doing this though, Jen? Like, do you have researchers, I guess, going out in helicopters, looking at forests? How does this work? Essentially, yeah. So most of the forest

inventory that happens in BC is based on air photo interpretation. So that's essentially what it sounds like, taking pictures out of aircrafts. And then someone in an office will look at those images and try to delineate different stand conditions. So this one is a lodgepole pine stand. This one is an aspen stand. This one is maybe 100 years old. This one's 50 years old. And we've been doing it this way for a long time. So that technology actually came out of the Second World War, the air photos. So it's a few decades old at this point. It's a few decades old.

But the reason we do it that way is because, A, air photos are relatively cheap. So, I mean, other provinces do this as well. But British Columbia is very large. They're also very high resolution. And so one of the challenges with that is that when we have closed canopy forests, those images don't show us any of the understory. And that's where a lot of the fuels that we're concerned about are. So it's not only the canopy fuels, but the understory fuels that matter for fire, and those get obscured with this approach. Yeah, that's a really important point,

I would imagine, because you're only getting a picture of the top canopy, not of the stuff that's underneath. How reliant is BC Wildfire Service on this data right now? Yeah, so I would say at present we worked with a number of people, experts from BC Wildfire on this study. And we found that they have pretty vast systems of internal knowledge and expertise to the extent that they do not rely directly on the maps. So instead of using the maps to decide the fuel type, if they were, for example, on a large fire complex, they want to do some modeling to try to predict where the fire might go, they will go out, watch the fire behave, and assign the fuel type that is most consistent with the fire behavior, as opposed to the forest conditions. So they don't rely heavily on these maps to make operational decisions, although they kind of universally articulated that having better

information without having to do that field validation, especially to watch it burn, would be particularly helpful to make more informed decisions about management. Okay. Okay. So it's not that BC Wildfire Service is using the data while fires are happening. So I guess, yeah, can you give me an example of how we would actually use this data? Yeah. So the way that these data sets are often used is to inform proactive management approaches, as well as for research, which is how we were interested in using this data. So unlike putting fire suppression during the fire season, a lot of the proactive management we do that's related to

forest restoration and conservation, and especially fuels management to protect communities from wildfires. So fuel reduction treatments like forest thinning and prescribed burning, we have to decide where to put those treatments before a fire happens. So management we might do outside of the fire season to reduce that risk. it's really important to have accurate data on fuels to make those decisions. We'll be back in a moment. So Jen, let's talk about your study. In really broad strokes, what were you looking at? Yeah, so we were interested in understanding how well the system, the fire

behavior prediction system, was able to represent the conditions on the ground. So essentially, is there a fuel type that exists that represents what we see? And then we assigned fuel types, the closest match fuel type on the ground, and we compared that to what was in the map. So in cases where there's discrepancies, where are those discrepancies coming from, and how might we be able to improve them? So we went to a number of plots, and this was specifically in southeastern British Columbia and kind of dry or lower elevation forests, similar to what you'd find in the Okanagan as well. And so we went to each forest stand, and we first looked at the system, and we just assigned a fuel type, which is how the system's designed to be used. So essentially, you get a picture and a paragraph and you say, which one does it match most closely with?

So what does the canopy look like? What does the understory look like? What's on the ground? Which one it matches with? Yeah. Which one does it match? And then we collected a bunch of measurements so that we could also try to assign a fuel type based on the process that the province was using. So we said, what ecosystem type is this? What's the tree density? What's the crown closure? What's the leading species? All of this information so that we could try to figure out not only what the mismatch was, if there was a disagreement, but also where it might be coming from. Is it the accuracy of the forest inventory data? Is it because no fuel type exists to represent the stand? Is it a combination of both of those factors? Okay, so then how common was it that the data you had didn't match with what you actually saw on the ground? Yeah, so we found that at 26% of the plots that we looked at, the fuel type we saw on the ground matched with what was in the provincial map. And so that means

that 74% of the time they disagreed with each other. So only 26% of the time it actually matched. Correct. Yeah. More often than not, that was often because we found something in the field that had more fuel or was denser than what the map would have assigned. And more fuel or denser would mean more prone to burning, I would imagine. Potentially higher risk. Yeah. Can you, I guess, can you give me an example of what you saw when we're talking about someplace, so 26% match, but that's 74% that didn't. So yeah, what's an example of what you would actually see in the field that, yeah, where that would be different?

The most common instance or most common category of disagreement we found is in kind of dry interior forests that have been impacted by a history of management and disturbance. So there's only one fuel type within the system, C7, which is like a dry open forest that's designed to represent these conditions. But because we have been harvesting and suppressing fire and grazing and urbanizing in these ecosystems for so long, there's so many types of departures from those historical conditions. And in particular, what we found is that the fuel type map would describe it as this kind of classic open ponderosa pine Douglas fir forest that you might have seen in the past. It

was frequently maintained by fire. Today, because fire has been excluded from these systems for so long, there's a lot more fuel in those stands. And so they're kind of along these gradients of tree encroachment, really suppress small trees with a lot of fuel in the stands that aren't represented well by that fuel type. And is a part of that because we tried to prevent forest fires for so long and that hasn't burned then in a way that it would naturally if we didn't do that? Exactly. Yeah. So these forest types in interior BC, especially these dry forests, are adapted to burn at low severity very frequently. And these dry systems often also occur at valley bottoms where we've built most of our communities. And so those are the places both where we're that used to burn the most in the past, both from lightning ignitions and indigenous fire stewardship, fire adapted systems. And they've gone the longest without fire today because they're where we're most effective at putting fires out right by our communities. And in the absence of fire, which would have taken out a bunch of smaller trees

and seedlings, all of those trees have regenerated and are in competition with each other. So they're very dense and very suppressed. And so that's allowed a lot of fuel to accumulate as well. And then we also found that for 58% of the plots, there was no fuel type within the system to represent what we found there. So that essentially means that there was never really a fair chance for them to match because what exists on the ground does not exist in the system. Oh, so there was no chance for it to match because your system basically didn't have those options. Correct. Okay. So what are the implications of this then, Jen? Like, how would this inaccurate data be affecting the way that we today manage and prevent forest fires? Yeah, so particularly in fire-adapted systems, especially in southern British Columbia and in

interior British Columbia, there has been a lot of conversation, especially in the past couple years, about the need to move towards a more proactive management approach that's not necessarily removing the need for fire suppression systems those will always have a role but a need to also invest in kind of more strategic management outside of the fire season to prevent these catastrophic fires during extreme fire weather when we get these long fire seasons with droughts and so there's a particular need to do that also not just near our communities, but across the landscape. So one of the implications of this is that it's very hard to identify priority areas for treatment, to scale up our approaches and to make good use of our limited resources if we don't know where the fuels are. Right. And so when we're talking about

identifying dense fuel structures and trying to prevent, you know, bigger forest fires, then is this, I mean, we talk about prescribed burns. Is this the kind of thing that could help here? Exactly. Yeah. So part of what those interventions can look like include prescribed fires. A lot of these stands have gone without fire for so long that it's not safe to do a prescribed burn as is. So we first do a thinning treatment. So we go in and remove some of those smaller trees first, take them out of the system, and then we burn the understory. That type of information is also especially important for designing prescribed fire prescriptions. So being able to understand what the current forest structure is and which areas might be most at risk, for example, of a fire spreading into a community is incredibly important.

Could you give me an example of something like this, like a place that, you know, really needed accurate data here and just, I guess, didn't really have access to the stuff they needed? Yeah. So one of the communities that I work with, Acom First Nation, which is a Tanaha community in southeastern British Columbia, they recently did a large fuels treatment last year, 1,200 hectares, where they did first forest thinning over a large portion of that land and then prescribed fire. And in order to make planning decisions for the prescribed fire over that large area on the reserve, we had to reclassify these maps. So the fuel types all described kind of open forest conditions across the reserve,

when in reality, there was a lot more dense forest. And that was something we had to manage for during this fuels treatment. And so having more accurate data about where those fuel structures are on and off reserve would be helpful in designing some of these prescriptions and also being able to prioritize. So we have limited resources. These treatments are still expensive and we have very limited capacity. And so being able to decide where treatments are going to be most effective is very important. So Jen, what are some of the ways to fix the problems with this data that we have in BC? Yeah, so we identified two approaches or two ways that we could improve this challenge. So the first is the forest inventory data really was not designed to look at fuels and it's not doing

a particularly good job of it. But of course, that wasn't what it was built to do. So there is an opportunity using more modern remote sensing technology, LIDAR, which uses lasers, to get some more information about forest structure. And instead of measuring kind of forest structural attributes for harvesting, measure the fuels themselves. So how much biomass is there, the measurements that we would use to

inform fire behavior prediction. So LIDAR, so instead of just taking a photograph of the canopy, you're actually having a laser come down and like look at the density throughout the layers of the forest. Exactly. Yeah. So you're shooting lasers down into the forest. And so you get a little bit better penetration into the understory, still not great in a dense forest. But what happens is that you get a ton more information about forest structure. And so we can tell things, especially when we have a lot of ground plots to create models with, we can measure things much more accurately. So fundamentally, we identified that, you know, better forest inventory data would help, and we can make decisions based on that data.

But ultimately, if we're thinking about fuel type maps, better data would not be able to be ingested by the current system. So we would have a lot of data and the system wouldn't really be able to do much better with it. And so fundamentally, what we identified as a significant need was the ability of adapted or new systems that do a better job of describing the conditions that we have in BC. And in particular, on the predicting fire behavior side, the data in the current system is from the 60s, 70s and 80s has a very large bias towards the boreal forest and eastern Canada. And so BC is just not well represented. And even in those conditions, even in those stands, the climate conditions, the fire weather conditions during

the mid 20th century are very different than they are today. And so simplistically, we can't expect to predict things in systems that we don't have data for. And so there's a significant need to collect new data under the fuels that we have, especially in B.C., and also the fire weather conditions of today to make more accurate predictions about how fire might behave and how we might want to manage it. So, I mean, just lastly here, it sounds like it would be a lot of work to do all this, but what are the consequences if we don't do it? I'm not super pessimistic about the future of fire, but I am realistic. We know fires are going to continue to get worse. On average, the frequency of extreme fire seasons will increase.

And so we have to make some strategic decisions about how we want to manage our priorities. Essentially, these forests are going to burn and we have the ability to control how they might burn, whether they might have the negative consequences that we saw last fire season or whether we might be able to reintroduce fire in a more beneficial way.

Jen, it was great to hear about this research. Thank you so much for being here today. Yeah, thanks for having me. That's it for today. I'm Maina Karaman-Wilms. Our intern is Manjot Singh. Our producers are Madeline White, Cheryl Sutherland, and Rachel Levy-McLaughlin. David Crosby edits the show. Adrienne Chung is our senior producer, and Angela Pachenza is our executive editor. Thanks so much for listening, and I'll talk to you tomorrow.