Arguments
Dangerous and historic wildfire smoke pollution event engulfs the U.S. and Canada
Posted on 21 July 2026 by Guest Author
This is a re-post from Yale Climate Connections by Jeff Masters
As climate change bakes forests across North America, dense smoke from dozens of out-of-control wildfires burning in northern Minnesota and adjacent portions of Ontario is blanketing tens of millions of people with hazardous pollution.
The fires are bringing the worst air quality on record to much of the Great Lakes, mid-Atlantic, and Northeast United States. Pollution from small particles called PM2.5 — the fine particles less than 2.5 microns in diameter are the primary air pollution killers — has been far into the “Hazardous” range across five states since Wednesday morning.
The award for worst air in the nation on July 16 went to the city that has in the past billed itself as a climate haven: Duluth, Minnesota. The city’s 24-hour air quality index, or AQI, for PM2.5 particle pollution hit 934, over three times the threshold for “Hazardous” pollution. This shattered Duluth’s previous all-time AQI record of 159 set July 20, 2021. EPA pollution records go back to 1999.
The award for worst air in the nation on July 16 went to the city that has in the past billed itself as a climate haven: Duluth, Minnesota. The city’s 24-hour air quality index, or AQI, for PM2.5 particle pollution hit 934, over three times the threshold for “Hazardous” pollution. This shattered Duluth’s previous all-time AQI record of 159 set July 20, 2021. EPA PM 2.5 pollution records go back to 1999.
Record 24-hr PM 2.5 air quality on July 16, 2026
Duluth, MN: 934 AQI for PM2.5 (Old record: 159, July 20, 2021)
Toledo, OH (5 monitors): 624 (Old record: 190, June 28, 2023)
Chicago. IL (42 monitors): 511 (Old record: 246, June 28, 2023)
Detroit, MI: 490 (Old record: 226, June 28, 2023)
Cleveland, OH: 297 (Old record: 285, June 28, 2023)
Milwaukee, WI: 414 (Old record: 270, June 27, 2023)
Flint, MI: 343 (Old record: 178, June 27, 2023)
Minneapolis, MN (39 monitors): 251 (Old record: 193, July 29, 2021)
Green Bay, WI (5 monitors): 372 (Old record: 179, June 29, 2023)
Grand Rapids, MI (4 monitors): 482 (Old record: 227, June 27, 2023)
Lansing, MI (2 monitors): 408 (Old record: 194, June 29, 2023)
Buffalo, NY (11 monitors): 206 (Old Record: 176, June 7, 2023)
Record 24-hr PM 2.5 air quality on July 17, 2026
Washington D.C. (21 monitors): 246 AQI at Ashburn, VA, and 234 at Springfield, VA (Old record: 222, June 8, 2023, at Franconia Park, VA)
Pittsburg, PA (25 monitors): 261 (Old record: 237, June 29, 2023)
Dover, DE (2 monitors): 223 (Old record: 207, June 8, 2023)
Columbus, OH (8 monitors): 272 (Old record: 210, June 28, 2023)
Because of the huge number of people affected, and since this is occurring at the same time as a severe humid heat wave, this extreme and widespread pollution event — which will be followed by many months of repeated wildfire smoke incursion into the U.S. — will undoubtedly cause hundreds and perhaps thousands of premature deaths. The only comparable wildfire smoke event affecting this portion of North America occurred in 2023; a 2025 study blamed that event for 33,000 premature deaths in the United States, 8,300 in Canada, and 23,000 in Europe. According to the EPA, a premature air pollution death is one that occurs on average 14 years before a person would have otherwise died.
The climate change connection to the wildfires
As the climate warms, fire danger increases, mostly because the atmosphere gets “thirstier” – more water vapor can evaporate into warmer air. This results in more water vapor evaporating from plants, which dries them out and creates an increased risk of large and intense fires that can generate huge smoke plumes. Most of the fires grew out of control under extreme heat conditions made up to five times more likely by climate change (Fig. 1). According to Climate Central, the heat that helped fuel these fires would have been “highly unlikely” to have occurred in a world without climate change.
Figure 1. Climate Shift Index for Monday, July 13, 2026, showing the factor increase in high temperatures because of human-caused climate change. (Image credit: Climate Central)
Five states recorded “Hazardous” air quality on Thursday
On Thursday, July 16, portions of Ontario and five states — Minnesota, Wisconsin, Michigan, Illinois, and Ohio — experienced 24-hour levels of PM2.5 with an air quality index in the “Hazardous” (brown) range. According to the U.S. Environmental Protection Agency, these conditions necessitate health warnings of emergency conditions, with the entire population more likely to be affected. Purple “Very Unhealthy” air was observed in three other states — New York, Pennsylvania, and Indiana. This level of pollution triggers a health alert, and everyone – not just people with vulnerabilities – may experience more serious health effects.
Figure 2. Observed 24-hour air quality index (AQI) for PM2.5 pollution for Jul. 16, 2026. Fifty monitors in five states had “Hazardous” air, plus an additional nine monitors in Ontario. (Image credit: EPA)
According to air pollution scientist Ryan Stauffer, yesterday’s air pollution event blows away the previous most extreme wildfire smoke event in this region — in June 2023 — for extremity. During the 2023 event, only about three EPA monitors, all in Pennsylvania, recorded a 24-hour AQI in the hazardous range. But on Thursday, 50 official EPA monitors recorded a 24-hour AQI in the “Hazardous” range (Fig. 2), plus an additional nine monitors in Ontario. According to rankings at iqair.com, Detroit was the most polluted major city worldwide for most of Thursday, with Chicago bumping Detroit out on Thursday night. On Friday morning, the top five most polluted cities in the world were all in North America: Detroit, Chicago, Washington D.C., Toronto, and New York City.
On an hourly scale, some truly extreme AQI readings above 1,000 were recorded Thursday in Minnesota, Wisconsin, and Michigan. The “Hazardous” (brown) range is for an AQI in above 300, so these readings were more than three times beyond the “Hazardous” threshold. The most extreme readings occurred in northern Minnesota downwind of the fires burning in the Boundary Waters park, where an AQI of 3,567 was measured by a purpleair.com sensor.
How the trouble started: record heat and a record-strong high-pressure system
The wildfire event began on Monday, when the strongest upper-level ridge of high pressure ever observed in the north-central U.S. baked the region. All-time record heat exceeding 100 degrees Fahrenheit (37.8°C) was observed, worsening existing moderate to severe drought conditions. Thunder Bay, Ontario, on the north shore of frigid Lake Superior, hit 39.5 degrees Celsius (103.1°F), smashing its all-time heat record by over 2°C. To the north, Armstrong hit 40.7°C (105.3°F), the hottest temperature observed in all of Ontario since the great Dust Bowl heat wave of July 1936.
Strong winds moved in with the heat, fanning multiple wildfires that feasted on the dry fuels, which featured plenty of dead trees from a spruce budworm infestation and a 1999 derecho event that felled thousands of trees. Extreme fires with a rapid rate of spread resulted and created at least two massive pyrocumulus clouds — giant thunderstorms spawned by the heat of intense fires that reached the stratosphere. Tomer Burg has an excellent thread explaining the meteorology that led up to the wildfire event:
How to interpret PM 2.5 readings at airnow.gov and purpleair.com
You can access EPA’s real-time PM2.5 NowCast AQI numbers at https://www.airnow.gov (note that you can’t just type in “airnow.gov” to get to the website, since they haven’t configured it to allow that). The AQI numbers are updated once per hour (shortly after the top of the hour). Because the EPA standard is based on a 24-hour average, the raw hourly data is not reported. Instead, EPA’s NowCast AQI numbers are computed using a 12-hour weighted average from the past 12 hours of data.
When air quality is highly variable (e.g., during a wildfire or a sudden wind shift), the algorithm heavily weights the past 1-3 hours. Otherwise, the past 12 hours are weighted nearly equally, and the AQI number is a 12-hour average. At the end of a day, the official daily AQI is calculated using a true arithmetic 24-hour average.
Don’t miss out: If you sign up by August 1 to support our extreme weather coverage with a monthly donation, you’ll be invited to an exclusive conversation with our meteorologists later this season. Become a monthly supporter of Jeff Masters, Bob Henson, and Irene Sans’s extreme weather coverage today. Donate
In contrast, the AQI numbers at purpleair.com update every few seconds, so you get a much better idea of rapid changes in air quality. Clicking on a dot will bring up a graph of the past few days of data; the default averaging time for this graph is 10 minutes, but you can change this to a longer time span. Important: you need to choose “US EPA” in the “Apply Conversion” setting, or else the PM 2.5 readings will be too high.
The current situation: over 20 large fires out of control
As of Friday, there were 193 active fires covering 1.7 million acres (673,000 ha) in Ontario, with the vast majority of these considered to be “out of control,” according to the Canadian Interagency Forest Fire Centre. Of the 22 large fires in over 2,000 acres burning nearest the U.S. border, 18 were receiving a full fire-fighting response, and four were not being fought. In the U.S., the National Interagency Fire Center reported six large fires in northern Minnesota, all 0% contained, covering 60,000 acres.
Figure 3. Monthly fire assessment for North America for July 2026 (left) and August 2026 (right). Red shading indicates areas where conditions would favor increased fire activity. Green shading indicates areas where conditions would favor decreased fire activity. (image credit: National Interagency Fire Center).
The forecast: relief coming by Saturday for the worst-affected states
Thunderstorms moved over the Minnesota/Ontario fire area Friday morning, bringing up to two inches of rain. This water will help firefighting efforts, but the fire danger index in Ontario is still high to very high, and the thunderstorms may have sparked additional fires.
The cold front accompanying this storm will bring westerly winds that should flush the worst of the smoke out of Chicago and Wisconsin by Friday afternoon, and out of Michigan by early Saturday morning. However, a renewed invasion of smoke is predicted for Michigan and Wisconsin Saturday night into Sunday, bringing more “Hazardous” AQI conditions. Another pulse of smoke is predicted to move deep into the central U.S. on Wednesday, perhaps reaching Arkansas and Missouri.
We can anticipate that some of the major smoke-emitting fires in Minnesota/Ontario will continue to burn for an extended period, with some lasting until the first snows come in October. The long-range fire assessment (Fig. 3) calls for above-average fire risk over much of the forested areas of northern Canada and the western U.S. this summer, and we should anticipate frequent bouts of poor air quality from wildfire smoke across much of North America. July and August are usually the peak months of fire season, which typically extends well into September. A potent North American Monsoon is now bringing heavy rains to parts of the southwest U.S., so the overall fire risk in that region may decrease to average levels by August.
Many lightning-caused fires occurred in the Pacific Northwest yesterday. With hot, windy weather expected to move in during the coming week, some of these fires may well become significant smoke producers.
Climate change predicted to further worsen wildfires
The number of people in the U.S. who experienced at least one day each year with smoke-related fine particle pollution levels at three times over the EPA standard has increased 27-fold over the last decade, and we can expect climate change to significantly worsen wildfire smoke problems in North America in the coming years.
For example, the frequency and magnitude of extreme wildfires around the globe have doubled in the past 21 years because of climate change, according to a study published last year in the journal Nature Ecology & Evolution. Rising temperatures have ushered in an era of hotter and drier weather, lending the right conditions for wildfires to erupt, the researchers found.
Here are some additional resources on climate change and wildfires:
- Explore how climate change has worsened wildfire smoke.
- Explore how fire weather has changed in your area on Climate Central’s interactive map.
- Visit their More U.S. Fire Weather analysis for science background and methodology.
- Explore our guide to staying safe from wildfires and smoke.
Here is another example of the environment affecting the people in a very damaging way. Along with increasing horrific algae blooms, fresh water declines, weather patten anomalies, ad nauseam.
Many people understand the existential threat the co2 increase in the atmosphere is creating day by day, yet till it affects them meanfully, personally and constantly, this is what most people in the world really worry about, www.gallup.com/analytics/701519/worlds-most-important-problem-report.aspx?utm_source
We can come together every 4 yrs for sport or to save our ozone layer yet I'm told constantly not to scare people with the truth of an increasingly malevolent weather from our modern industrial practises and lifestyles.
Alas, the "you need to manage your forests" deflection point was used in the US a few years ago, when California was suffering from severe fires. The usual suspects were in "climate change can't be a factor"" mode, and looking for any possible alternative explanation - ABC (Anything But Carbon dioxide). It's the standard playbook on climate change contrarianism - look for any other factor, and try to deflect attention towards that squirrel.
Trump has threatened (and now announced) tariffs as a result of the smoke.
The Canadian boreal forest covers an area of some 550 million hectares. We have about 41 million people. That averages out to about 13 hectares each. We'd have to get each person cleaning 13 hectares to cover the forest - and much of it is not accessible by roads. I can just imagine the bulk email:
The boreal forest is dependent on fire for much of its ecology. Forest management does play a role in fire frequency and intensity, but claiming forest fires will stop with proper management is loony.
Bob Loblaw,
It is my understanding that the great majority of the Canadian boreal forest is not managed at all, and has not been managed in the past. The large fires of the last few years, which did not occur in the past, cannot be caused by mismanagement.
While a lot of forrest in the USA, and Canada, is managed, there are large areas of both that have never been managed. Increases in fire in unmanaged forest are directly caused by climate change. The deniers are simply looking for a simple excuse that sounds good, not an actual explanation.
But you already knew that. Hopefully it helps other readers.
Michael:
Yes, much of the forest is unmanaged. Much of it has no road access, and is not harvested. What management happens is managing harvesting and reforestation. The forestry companies that are given the rights to an area will build roads as needed - because they want to be able to take the wood out.
In British Columbia, historically there have been locations with huge trees where they have used large helicopters for extraction of highly-valuable timber, but that is in coastal forests, not the boreal forest. Much of the boreal forest timber is used for basic dimensional lumber (the proverbial 2x4 for house construction), fence posts, pulp for paper, etc.
Fire-fighting priorities have changed somewhat over the decades, but much of it is geared towards protecting human habitation, areas of active timber harvesting, etc. They do try to hit small fires early, to prevent large ones from developing. They also try to leave less slash around after harvesting, but again that will only be in the harvest areas.
In the boreal forest, a lot of the wood is deadwood on the ground, rather than in living trees. That's the natural state, long before anyone could imagine "managing" the forest. The largest carbon storage in the boreal forest is actually in soil carbon.
Lets be clear, 25% of the CO2 in the globes atmosphere has been spewed out by one country to last for centuries. This country willingly elects known climate change deniers as their leaders and role models. May the smoke from these new normal fires settle over their land and force the consequences of climate change inaction into their thoughts.
The Republican charge that the boreal forests are being mismanaged and the appropriate push back that this is driven by climate change does overlook another important factor for the future: active suppression of cultural fires set by indigenous nations who used to manage these boreal forests much more effectively than they are currently managed. Starting in the 1600s, indigenous fire management of the landscapes in question were actively suppressed across much of the boreal forest region, and yet those fire management traditions are still alive in many of the 600 predominantly indigenous communities across the north woods, who ironically are the most heavily impacted by the current mega-fires. Reflecting a recognition of the importance of aboriginal fire management techniques in Australia, there has been a quiet revolution that has been a long overdue recognition that indigenous burning practices, small in scale and well timed earlier in the year, can head off much of the large scale dangers that are inherent in boreal forests that are not logged or otherwise accessed by western lumber/agricultural interests.
It is a misnomer to say that these boreal forests were not actively managed just because they weren't being logged. For an excellent research paper outlining these practices and their results on biodiversity, forest composition, wildfire management and human's relationship to these forests, I recommend reading "Centering Indigenous Voices: The Role of Fire in the Boreal Forest of North America" in the journal Fire Science and Management: https://link.springer.com/article/10.1007/s40725-022-00168-9
[BL] Link activated.
Wild @ 6:
It may be that you hadn't read Zeke Hausfather's reposted article from today that covers several aspects of the Canadian boreal forest management question in detail. Among other things, he specifically talks about the large areas of Canada's boreal forest that have never seen any form of management.
Regarding your link to the indigenous forest management practices, I think there is a matter of scale to consider. A quick read of that article gives me the impression that many of the practices they talk about are largely local. These can be highly important in helping protect a settlement and managing local land use, but do not represent a methodology that can be applied to managing the boreal forest writ large.
Their figure 1 shows the scattering of current indigenous land (reservations) within the boreal forest. In the past, many indigenous cultures were somewhat nomadic, which would increase the likelihood that the use of fire (or mitigation of fire risks) is largely local. The map shows large areas of the boreal forest with little or no indigenous presence today. (The figure cannot be displayed here directly. This link will take you to it, if you don't want to follow wild's link to the entire article.)
https://link.springer.com/article/10.1007/s40725-022-00168-9/figures/1
Going back to the 1800s, the indigenous population in those areas was probably of the order of 100,000 (although very hard to estimate, I would guess). This is one source I could find with some actual numbers. (I ignored what Google AI thinks it can intelligently espouse.) As I mentioned above, the boreal forest covers some 550 million hectares. That would be 5500 hectares per individual for the indigenous population to "manage".
Bob @ 7:
Actually I had read Zeke's article from his Climate Brink posts, and I also suggested to him that he read over the article I linked to. The very same figure you linked from that article shows that the native populations were well placed to indeed manage the boreal forests through the use of local, small scale burns that created buffers in much the same way as is being documented in Australian aboriginal fire management practices, most vividly described in Victor Steffensen's excellent book "Fire Country." The goal is not to burn over the entire boreal forest as you suggested in your calculations of hectares/person, but rather to create and mantain a mosaic of burned areas that stimulates biodiversity, breaks up the potential for megafires, protects villages, etc. The nomadic nature of the communities, particularly the seasonal hunting and gathering cycle, is conducive to and consistent with such management practices and there is evidence that such practices covered more land than you might expect. All of this is not to say that the megafire problem exacerbated by climate change is not a huge problem, but as in Australia, the indigenous fire management practices that were practiced for thousands of years, then suppressed, offers an excellent tool for the future in regards to optimizing biodiversity, strengthening the links of human communities to the surrounding boreal forests, and helping mitigate some of the threats of megafires to those communities. Nothing short of reducing our fossil fuel emissions will truly mitigate the mess we're in, but it's a both/and situation that is possible here!
wild @ 8:
I disagree that native populations are well-distributed to manage the boreal forest writ large. The figure from the paper you cited shows vast areas of the boreal forest that are very sparsely covered. (Keeping in mind that locations of reserves, as provided in that paper, are only partly representative of the historic range of the indigenous population.)
Compare that figure to the following figure, which shows today's fire distribution in Canada, taken from the Canadian Wildland Fire Information System. (I am grabbing today's display as a local file, as the image will change over time.)
Note that many of the fires are in the northern portions of the boreal forest, where there would have been relatively little human activity historically. These areas are often still very difficult for travel.
The same applies to today's smoke patterns (same source):
Again, the origins of much of the smoke is in sparsely settled areas.
I would agree that many historical indigenous practices would be of benefit for protecting settled areas, and that our current forestry, building and development habits have made for very difficult times in recent years. But nothing done on local or regional scales surrounding current settlements and forest harvest areas will do much for the large areas that are currently unpopulated and burning.