In July 2026, the skies over the eastern United States turned orange. On July 16 and 17, the New York City skyline and Washington, D.C.'s Washington Monument and Lincoln Memorial were swallowed in thick haze [source: PBS NewsHour, 2026]. The fires were burning not in the United States but across the border, in Canada's Ontario (roughly 190 blazes in the north alone) and the Boundary Waters of northern Minnesota [source: NPR, 2026]. The smoke did not stop at the border checkpoint.
That is the heart of what this article is about. Fine particulate matter (PM2.5), the main component of wildfire smoke, can travel thousands of kilometers and settle into the lungs of people in cities far from any flame. This piece separates two things. One is what is already well established — what PM2.5 is and why it endangers the respiratory and cardiovascular systems. The other is where research is still under way. And between the two, it lays out what you can actually do.
One note on method before the numbers begin. Three kinds of evidence appear below, each labeled where it appears. Some figures are measured — a concentration recorded by a monitor, a fire detected by a satellite. Some are modeled — an estimate produced by running observations through a model, which is true of every death figure here. Some are reported — a number that circulated in coverage, useful for scale but not precision. Two units also do different jobs and are never interchangeable: micrograms per cubic meter (µg/m³) is a concentration, the mass of particles in a volume of air, while the Air Quality Index (AQI) is a scaled communication number derived from concentrations. Nothing here converts one into the other.
In this article
- The summer of 2026, in numbers
- What PM2.5 is and why it is dangerous
- What happens in the body — the settled and the still-emerging
- Smoke has no borders
- Who is most at risk, and how to protect them
- Conclusion — what to watch
The summer of 2026, in numbers
Start with the facts of what happened. Between July 14 and 20, 2026, smoke from wildfires in Canada and Minnesota poured into the Great Lakes, Northeast and Mid-Atlantic regions of the United States [source: NASA, 2026]. According to satellite analysis from NASA, the smoke from Ontario reached the densely populated Northeast within about two days, having traveled a total of thousands of miles [source: NASA, 2026].
What the satellites actually saw
How that trajectory was established explains what kind of statement it is. NASA's Scientific Visualization Studio combined two products. Active fires were detected by VIIRS, the imaging instrument on the NOAA-20 and SUOMI-NPP satellites — a direct observation of where things were burning. The smoke itself was tracked through brown carbon aerosol optical depth from the GEOS-CAM model, aerosol optical depth being a measure of how much light airborne particles block [source: NASA, 2026]. So the fire locations are observations and the plume is a modeled field. On the U.S. side those fires burned in Minnesota's Boundary Waters Canoe Area Wilderness, roadless lake country far from the cities that ended up breathing the result [source: NPR, 2026].
The numbers reveal the weight of the situation. On July 17, PM2.5 concentrations across the Baltimore–Washington, D.C. area spiked to roughly 200 µg/m³, triggering a "Code Purple" (very unhealthy) alert [source: NASA, 2026]. In the U.S. Air Quality Index (AQI) system, the purple category means the "risk of health effects is increased for everyone" [source: AirNow, 2026]. The population within the affected zone ran into the tens of millions from the Great Lakes to the Atlantic coast, and by some tallies exceeded 100 million [source: PBS NewsHour, 2026].
Concentration and index are not the same number
Those two figures describe the same air but are not the same kind of statement, and almost every public misreading of smoke data starts here. The 200 µg/m³ is a measured concentration; Code Purple is the label on the 201–300 band of a derived index. The full EPA scale runs Good 0–50 (green), Moderate 51–100 (yellow), Unhealthy for Sensitive Groups 101–150 (orange), Unhealthy 151–200 (red), Very Unhealthy 201–300 (purple) and Hazardous at 301 and above (maroon), with an AQI of 100 corresponding roughly to the short-term national air-quality standard [source: AirNow, 2026]. The index is built so the same number means the same thing whichever pollutant drives it — which is why it cannot be read back as a concentration.
Closer to the fires, local coverage during the event reported AQI values far higher: 1008 in Marquette, Michigan, and 682 in Duluth, Minnesota, against a hazardous band beginning at 301 [source: local news reports, 2026]. One tally likewise put the exposed population above 124 million [source: local news reports, 2026]. Both belong to the reported tier — cited for gradient and scale, not as the anchor, which stays the Baltimore–Washington measurement of roughly 200 µg/m³.
The official response
The scale of the official response tells its own story. New York, New Jersey and Pennsylvania handed out free N95/KN95 masks, cancelled outdoor events, and opened libraries and other public buildings as shelter spaces [source: PBS NewsHour, 2026]. New York City's Emergency Management activated an air-quality emergency plan on July 15 and kept it in force until the skies cleared on the 19th [source: NYC Emergency Management, 2026]. A Major League Baseball game was postponed and local events were called off.
As a sequence, that is a useful record of how such an episode gets managed: activated on the 15th, extended as the smoke persisted, concluded on the 19th [source: NYC Emergency Management, 2026]. The measures divide into three kinds — reduce the dose (masks), remove the exposure (cancelled outdoor programming, including a Cleveland Guardians game and a soapbox derby), and offer cleaner air to go to (public buildings as shelter) [source: PBS NewsHour, 2026].
Not the first time
If the scene feels familiar, it is because of a memory from three years earlier. On June 7, 2023, New York City briefly had the worst air quality of any major city in the world. Its daily mean PM2.5 concentration that day was 203.9 µg/m³ — about five times the U.S. daily standard of 35 µg/m³ [source: Annals of Emergency Medicine, 2024]. The orange sky of 2026 is less an exceptional event than one scene in a recurring pattern.
That 203.9 µg/m³ comes from a peer-reviewed analysis of emergency medical services use in New York City during the record particulate levels of June 2023 — the episode was studied as a load on the emergency system, not only as a weather story [source: Annals of Emergency Medicine, 2024]. Contemporaneous reporting judged 2023 the worst U.S. smoke event since at least 2006, against Canada's worst fire season on record with more than 400 active fires burning at once [source: news reports, 2023]. That ranking is background from coverage, not an official measurement.
What PM2.5 is and why it is dangerous
At the center of all of this is a particle you can barely see. PM2.5 refers to fine particulate matter with a diameter of 2.5 micrometers (µm) or less. If that is hard to picture, think of it this way: a human hair is about 70 µm across, which makes a PM2.5 particle roughly 28 times thinner than a hair [source: EPA, 2026].
How small 2.5 micrometers really is
That comparison is worth holding onto, because it makes everything else intuitive. The World Health Organization uses the same image with rounder arithmetic, calling PM2.5 about 30 times finer than a human hair [source: WHO, 2021]. Either way, this is not dust you brush off a sleeve — particles that small stay suspended and travel with the air, which is why smoke crosses a continent.
Why the size is the whole problem
Its small size is precisely the reason it is dangerous. Coarse dust is mostly filtered out in the nose or throat, but PM2.5 slips past those defenses, penetrates into the deepest reaches of the lungs, and in part crosses into the bloodstream, where it can affect the heart and other organs [source: EPA, 2026]. This is why wildfire smoke is a health problem and not merely "air that smells bad."
Where the guideline line sits
A baseline makes the numbers easier to read. In its 2021 updated air-quality guidelines, the World Health Organization (WHO) recommends that PM2.5 stay at or below 5 µg/m³ as an annual average and 15 µg/m³ over 24 hours [source: WHO, 2021]. That gives a sense of just how far Washington, D.C.'s roughly 200 µg/m³, or New York's 203.9 µg/m³ in 2023, sat from the recommended line.
One caution about those lines: they are different instruments with different averaging periods. The WHO figures are health-based guidelines stated separately for an annual and a 24-hour average, and the two are not interchangeable [source: WHO, 2021]. The U.S. daily standard of 35 µg/m³ is a regulatory threshold — the yardstick against which New York's 203.9 µg/m³ daily mean was called roughly five times the standard [source: Annals of Emergency Medicine, 2024]. Setting a peak reading beside an annual guideline would exaggerate the gap; a 24-hour mean against a 24-hour standard is the fair comparison.
What happens in the body — the settled and the still-emerging
From here, the layers have to be kept carefully apart. Some health effects of wildfire smoke are established by a large body of research; others are still being investigated. Blurring the two produces exaggeration.
Respiratory — the clearest signal
The best-supported area is the respiratory system. The U.S. Environmental Protection Agency (EPA) notes that hospitalizations and emergency-department visits for asthma and respiratory disease rise during wildfire smoke events, and that this signal is the most consistent across studies [source: EPA, 2026]. Symptoms range from persistent coughing, phlegm, wheezing and difficulty breathing to pulmonary inflammation and transient reductions in lung function [source: EPA, 2026].
One point worth noting is the observation that PM2.5 from wildfire smoke may be more harmful to the respiratory system than PM2.5 from other sources. In a peer-reviewed study of Southern California, a 10 µg/m³ rise in wildfire-specific PM2.5 was associated with increases in respiratory hospitalizations of 1.3% up to as much as 10%, whereas the same rise in ordinary, non-wildfire PM2.5 was associated with an increase of only 0.67% to 1.3% [source: Nature Communications, 2021]. This should be read, however, as an observational study of one particular region.
Those percentages need one more piece of care. They are relative increases — how much the rate of respiratory hospitalizations rose against the rate otherwise expected, per 10 µg/m³ of additional PM2.5. They are not absolute risks. A 10% relative increase is not a 10% chance of being hospitalized; where a baseline rate is small, a large relative increase can still be a small absolute change. The study supports a comparison between two sources of the same pollutant in one region over one period [source: Nature Communications, 2021]. The broader idea that wildfire particles are more toxic per unit mass than ordinary urban particles remains a research-stage hypothesis rather than settled.
Cardiovascular and beyond
The cardiovascular side points the same way, but the evidence is not as dense. The EPA finds that as smoke grows thicker, the risk of both cardiovascular- and respiratory-related emergency visits rises [source: EPA, 2026]. Associations with heart attacks and strokes have been reported too, but the EPA is explicit that the cardiovascular field has fewer studies than the respiratory one and that the results are somewhat more mixed [source: EPA, 2026]. In other words, the association is observed, but the strength of the causal case is not as firm as for the respiratory system.
"More mixed" is easy to hear as either dismissal or confirmation, and it is neither. It means the studies do not all point the same way with the same strength, and that there are fewer of them than in the respiratory literature [source: EPA, 2026]. That describes the maturity of the evidence, not a verdict that the effect is absent.
What "emerging research" means here
Beyond that lies a set of still-emerging, not-yet-settled research areas. Outcomes such as preterm birth, effects on cognitive performance, and increased susceptibility to infection are treated as emerging research topics, not as established causal facts [source: EPA, 2026]. Here it matters to separate "a possibility has been raised" from "it has been proven."
The infection question shows what that category contains. Some studies have reported increases in COVID-19 during wildfire smoke periods, and the EPA files this beside preterm birth and cognitive effects under emerging research rather than established effects [source: EPA, 2026]. An increase observed during a smoke period is a signal worth pursuing, not a demonstrated causal chain.
The mortality estimates, read carefully
The same caution applies to the mortality burden. One study using causal-inference modeling estimated roughly 24,100 all-cause deaths per year attributable to wildfire smoke PM2.5 in the contiguous United States, and suggested that wildfire-derived PM2.5 may be about five times as toxic as ordinary PM2.5 [source: Science Advances, 2025]. A separate long-term exposure study estimated more than 10,000 non-accidental deaths per year [source: PNAS, 2024]. Both figures are model-based estimates rather than an actual count of the dead, and they vary considerably with method and time period.
These are the numbers most likely to be quoted out of context, so it is worth spelling out what they are. No death certificate says "wildfire smoke." These are statistical attributions: a model estimates how many deaths would not have occurred at a lower level of exposure, given an assumed relationship between exposure and mortality. Change that assumption, the exposure window or the study period, and the number moves. Roughly 24,100 a year from a causal-inference approach and more than 10,000 a year from a long-term exposure approach are therefore neither a contradiction nor two independent confirmations [source: Science Advances, 2025; PNAS, 2024].
The toxicity multiplier works the same way: the estimate that wildfire PM2.5 may be about five times as toxic as ordinary PM2.5 is that study's modeled parameter, not an independently measured property of the particles [source: Science Advances, 2025]. And a last point of arithmetic — short-term studies of hospitalizations during smoke episodes and long-term studies of mortality across years measure different exposures over different windows, so their conclusions cannot be added into one total burden.
Smoke has no borders
What makes the 2026 case distinctive is that where the fire burned and who fell ill were events in different countries. Wildfire smoke travels thousands of kilometers, crossing borders, oceans and continents [source: WMO, 2025]. Intense fires can generate their own thunderstorms — pyrocumulonimbus — that loft smoke into the upper troposphere or even the stratosphere, and particles carried that high can linger for days to weeks and be transported far afield [source: WMO, 2025].
How smoke gets that far
Altitude is the mechanism. A pyrocumulonimbus is a thunderstorm generated by the fire's own heat, and it can inject smoke more than five miles up, into the upper troposphere and sometimes the stratosphere [source: WMO, 2025]. Height decides what happens next: near the surface, rain washes particles out within days, but there is little weather that high to remove them, so smoke can stay aloft for days to weeks and ride high-altitude winds far from its origin [source: WMO, 2025]. That is also why satellite instruments and modeled aerosol fields are the practical tools for following a plume no ground monitor can chase across a continent [source: NASA, 2026].
The 2023 crossing of the Atlantic
That this is not an abstraction was shown recently. Smoke from Canada's 2023 wildfires crossed the Atlantic and raised PM2.5 concentrations in Europe [source: Nature, 2025]. A fire that began in one country's forests left a measurable trace on the air quality of another continent.
Note the tier: that comes from a peer-reviewed study of long-range PM2.5 pollution and health impacts from the 2023 Canadian wildfires, so the transatlantic transport was measured and analyzed, not inferred from a map [source: Nature, 2025]. It also reframes the problem: if a fire in one jurisdiction raises particle levels in another, forecasting stops being a purely domestic function.
The background trend, stated carefully
A longer trend sits in the background. The rise in wildfire activity in the western United States is correlated with shifts in temperature and precipitation patterns associated with climate change, and is regarded as a significant departure from historic fire regimes. But here, too, correlation and causation must not be blended. No single fire can be declared, on its own, the result of climate change. The more accurate statement is that climate conditions are widening the backdrop against which fires grow larger and smoke lingers longer. That background trend makes it likely that the problem of borderless smoke will keep recurring.
That claim gets flattened in both directions, so state it precisely. The defensible version works at the level of conditions: fire activity has shifted away from historic regimes and correlates with climate-linked changes in temperature and precipitation. What does not follow is that any particular fire — including those of July 2026 — was caused by climate change; nor does the opposite error, treating each smoke episode as an isolated accident with no trend behind it.
Who is most at risk, and how to protect them
Smoke is not equally dangerous to everyone. The EPA names as especially vulnerable groups adults aged 65 and older, children, people with asthma or lung disease, those with heart disease, pregnant women, and outdoor workers [source: EPA, 2026]. For them, the same concentration of smoke carries a heavier burden.
The people the same air hits hardest
One group stands out within that list: the EPA notes that adults aged 65 and older show the highest risk of emergency-department visits across all the health outcomes examined [source: EPA, 2026]. The other categories follow different logics. People with asthma, lung disease or heart disease start from a compromised baseline, so a given increment of exposure costs more. Children breathe more air relative to their body size. Outdoor workers are exposed not because they are more susceptible but because their exposure is longer and harder to avoid — a workplace problem as much as a medical one [source: EPA, 2026].
Step one: check the number
Fortunately, the ways to reduce exposure are clear and doable. The first step is to check the measurement. The EPA recommends using AirNow (airnow.gov) to check your local AQI; sensitive groups should cut back on outdoor activity once the AQI passes 100, and everyone should reduce exposure above 150 [source: EPA, 2026]. The AQI color categories — from green (good) through purple (very unhealthy) to maroon (hazardous) — are a traffic light for reading the day's risk at a glance [source: AirNow, 2026].
Those thresholds sit on the scale in a way that makes them easy to remember. AQI 100 is roughly the short-term national standard and the top of the yellow moderate band; the orange band above it is named Unhealthy for Sensitive Groups. AQI 150 is the top of that orange band, so above it the scale turns red and the warning stops being group-specific [source: AirNow, 2026].
Step two: make the indoor air cleaner
Indoors, a few simple measures make a large difference. When smoke is heavy, stay inside and keep windows closed, and set heating or cooling systems to a recirculate mode that pulls in less outside air. Run a portable HEPA air purifier, and hold off on activities that add to indoor pollution, such as burning candles, frying, smoking or vacuuming [source: EPA, 2026]. If you must go out, a snug-fitting, NIOSH-approved N95 respirator reduces exposure to fine particles. Children under the age of two, however, should not be made to wear a respirator [source: EPA, 2026].
Each instruction does one of three jobs. Closed windows and recirculated air reduce what comes in; a HEPA purifier removes what is already inside; skipping candles, frying, smoking and vacuuming avoids adding a second particle source while the first is being kept out. The N95 covers the case none of those can — air you cannot control — and it works by sealing against the face, which is also why it is unsuitable for children under two [source: EPA, 2026].
Conclusion — what to watch
To sum up: the orange skies over the eastern United States in the summer of 2026 came from fires across the border, and PM2.5, the main component of that smoke, penetrates deep into the lungs and poses an established risk, above all to respiratory health. Cardiovascular effects show an observed association but a comparatively mixed evidence base, and outcomes such as preterm birth or cognition remain areas of ongoing research. Estimates of the mortality burden should be read as model-based values.
What should you watch from here? First, how the research converges on whether wildfire-smoke PM2.5 is genuinely more harmful than ordinary PM2.5. Second, how tightly cross-border air-quality cooperation and forecasting — through tools such as AirNow and satellite observation — are woven together. Third, how far habits at the individual level, such as checking the AQI, filtering indoor air and using an N95, take hold. Smoke may have no borders, but the preparation we can take against it has clear lines and clear levers.
One last habit, for this smoke story and the next. Ask which tier a number belongs to — measured, modeled or reported. Ask which scale it sits on, a concentration in µg/m³ or an index value on the AQI. Ask whether a percentage is a relative change or an absolute risk, and whether the study behind it looked at a few smoky days or at years of exposure.