Smoke maps became a normal part of planning an outdoor day sometime in the last decade, and most people learned to read them the way you learn to read a weather app: by looking at the color and deciding.
The color is real information. It is also the output of several assumptions, some of which fail in exactly the conditions you most want the map for.
What the number is
The air quality index is an index, not a measurement. It runs from 0 to 500 and it is computed from measured pollutant concentrations through a set of published breakpoints, so that the same number means roughly the same thing about health regardless of which pollutant produced it.
The categories and their colors are standardized. Green is good, yellow is moderate, orange is unhealthy for sensitive groups, red is unhealthy for everyone, purple is very unhealthy and maroon is hazardous. Those bands carry the agency's own guidance about what to do at each level, and that guidance is the right place to take health advice from rather than anywhere else.
The first thing worth knowing is that the value shown for a location is the worst single pollutant at that moment, not a blend. During a smoke event that is almost always fine particulate matter, the very small particles that make up wood smoke and travel enormous distances. But on a hot still day it might be ozone instead, which means a yellow reading in August is not automatically smoke and the map may be telling you about a different problem entirely.
How to Read the AQIThe number is smoothed on purpose
Here is the technical detail that matters most for someone standing outside deciding whether to go.
The particulate index displayed on the national map is not the concentration in the last five minutes. It is a value derived from recent hours, weighted so that recent readings count for more, which produces a figure that is stable and comparable rather than jumping around minute to minute.
That is a sensible design for a public health index and a poor one for a fast-moving plume. When smoke arrives quickly, the displayed number lags behind the air, and the lag runs in the dangerous direction: the map is still showing you the average of the last few hours, which were better than now.
The practical response is to look at the trend rather than the value. Most of these maps let you see a monitor's recent history. A monitor that has gone from 40 to 90 to 160 over three hours is telling you something the current headline number is actively understating. A stable 160 is a different situation from a rising one.
And there is a simpler check that costs nothing. Look outside. If visibility has dropped noticeably since you last looked, the map is behind.
Where the dots come from
Not all the points on a smoke map are the same kind of thing, and the map usually distinguishes them if you look.
Permanent regulatory monitors are the reference instruments. They are accurate, maintained, and sparse. Large parts of the rural West have very few, which is unfortunate given that is where most of the smoke is.
Temporary monitors get deployed during incidents, which fills gaps exactly when it matters and produces a network whose shape changes from week to week.
Low-cost sensors are the newer and much denser layer, largely privately owned and voluntarily shared. They are the reason a smoke map can show you neighborhood-level detail at all. They are also less accurate, and specifically they tend to read high in wood smoke, so the national map applies a correction to them before display. A raw uncorrected reading from a consumer sensor in heavy smoke will typically be worse than the truth.
Individual sensors also fail. A sensor sitting next to a barbecue, a sensor indoors, a sensor with a dying component, all produce numbers that are wrong in ways nothing about the display reveals. If one dot disagrees dramatically with every dot around it, believe the neighborhood rather than the outlier.
AirNow Fire and Smoke Map: Lesson 1, BasicsThe gap between the dots is drawn, not measured
Any smooth colored surface on a smoke map is interpolation. The map has values at monitor locations and it is guessing at everything in between using an algorithm.
In a city with many sensors, the guess is well constrained and the surface is close to reality. In a rural county with one monitor forty miles away, the surface is essentially decorative. It will still be drawn in a confident color.
This is the same lesson as everywhere else in map reading, and it is the same one we made about boundaries in why map boundaries are wrong: the rendering conveys a confidence the underlying data does not have. Zoom in and look for the actual monitor locations before trusting a color over a valley where there is no instrument.
Smoke in the sky is not smoke in your lungs
This is the single most consequential misreading of these maps and it deserves to be blunt.
Many smoke maps carry a plume overlay derived from satellite imagery, showing where analysts have identified smoke in the atmosphere. It is genuinely useful for understanding where smoke is going.
It says nothing directly about ground level. Satellites see the whole atmospheric column from above, so a plume drawn over your town may be smoke at twenty thousand feet moving across the continent while the air you are breathing is clean. The reverse also happens: air can be foul at the surface under a sky the satellite reads as clear, particularly at night when smoke settles.
The overlay usually indicates density in broad classes, and those classes describe the column rather than the surface. Use the plume layer to understand the weather and the monitors to understand the air.
Smoke has a daily rhythm
Smoke behavior in complex terrain is strongly diurnal and knowing the pattern is worth more than refreshing the app.
The common mountain valley pattern is that overnight cooling creates a stable layer, cold air drains downhill, and smoke settles into the valley bottoms. Mornings are frequently the worst air of the day at valley elevations even though nothing new has burned. Then daytime heating breaks the inversion, the layer mixes upward, and the surface concentration improves through the middle of the day before settling again in the evening.
Which means two things practically. Elevation is a variable: climbing out of a valley can put you above the worst of it. And time of day is a variable: an afternoon plan may be entirely reasonable on a morning that looked impossible.
Fires themselves also have a daily cycle, generally more active in the afternoon as humidity drops and winds pick up, which is the same set of conditions described in red flag warning explained.
The fire layers are a different beast
Smoke maps usually sit alongside fire layers, and those come with their own reading rules that are worth separating out.
Satellite hotspot detections are the red dots. They are thermal detections from orbiting sensors, and three things about them get misread constantly. They are detections of heat, so an industrial flare, a steel mill or a large lit gas facility can produce one. They have a footprint measured in hundreds of meters, so a dot is a pixel rather than a point and the fire is somewhere within it. And they are snapshots from the moments a satellite happened to be overhead, so a gap in the dots may mean the fire stopped or may mean nothing passed over.
Fire perimeters are the polygons, and they are made by people. They are typically flown and mapped periodically rather than continuously, frequently at night when infrared works best, which means a perimeter you are looking at in the afternoon may describe the situation as of some hours ago. On a fast-moving day that gap is the whole story.
Neither layer tells you about closures. Whether a road, trail or area is closed is an administrative decision made by the managing agency, published by that agency, and not derivable from a fire's shape on a map. Same principle as everywhere else in this category: the agency is authoritative and the layer is derived. The evacuation side of that is covered in wildfire evacuation levels explained.
What these maps do not cover at all
Worth naming the gaps, because they are silent.
Particulate is what gets measured because it is what these networks were built for. Wildfire smoke also carries gases, and near an active fire the mix is considerably more complicated than a single number. The index is a public health tool for regional air, not a hazard assessment for standing close to something burning.
Ash fall is not measured. A day can have a moderate reading and still be depositing ash on everything, which is a different nuisance and a different problem for filters and gear.
Blowing dust is not smoke, and in the arid West a particulate reading can spike for reasons that have nothing to do with fire. The index reports the particles, not their origin.
And indoor air is not on the map at all. Concentrations inside a building can be dramatically better or worse than outside depending on the building and what you have done to it, which is the whole subject of the clean air room approach.
Forecast versus observation
Keep the two straight, because they answer different questions.
Observations are monitors reporting what is happening. Forecasts are model output predicting where smoke will be, and smoke forecast models have improved a great deal while remaining models. They are excellent at large-scale transport and considerably less good at what happens in a specific drainage.
Use the forecast to decide whether to go at all and where to point the car. Use the observations, plus your eyes, to decide once you are close.
What to actually do about it
That is a separate subject and we have it covered rather than repeated here.
For deciding whether to be out at all and how hard to work, exercising in wildfire smoke covers the exertion question, which matters because harder breathing means more of everything getting deeper into your lungs.
For a day already underway, hiking in wildfire smoke deals with the field decisions.
For protection, N95 masks and wildfire smoke covers what actually filters fine particulate and what does not.
And for base camp, a vehicle or a rental, clean air room for wildfire smoke is the one that makes a bad week survivable.
The gear
The national fire and smoke map, which combines regulatory monitors, temporary monitors and corrected low-cost sensors in one place and is the best single source available.
A monitor's history view, not just its current value, because the trend carries information the headline number is smoothing away.
A satellite plume layer, used for understanding transport and never for estimating what you are breathing.
A smoke forecast model, for the day-before decision about whether the trip happens at all.
Your own sensor, if you spend a lot of time in smoke country, with the understanding that consumer sensors read high in wood smoke and that a single sensor is a data point rather than a network.
Well-fitted N95 or P100 respirators, which are the only thing on this list that changes your actual exposure rather than your knowledge of it.
And a plan with elevation in it, because moving up out of a valley inversion is frequently more effective than any equipment.
The short version
The color is an index, not a measurement, and it shows the worst pollutant rather than smoke specifically. The particulate number is deliberately smoothed, so during a fast-arriving plume the air is worse than the map. The smooth surface between dots is interpolation and it is confident wherever there are no instruments. And a satellite plume tells you about the sky, not about your lungs.
Read the trend, look at the actual monitors, and check the window. The map is a good tool that is easy to over-trust, which is a description that fits almost everything in this category.
It is worth saying that none of this is an argument against the maps. Twenty years ago a person in a smoky valley had no way to know whether the air was bad enough to matter, no way to see it coming, and no way to compare today with last week. Now that information is free, updated continuously, and detailed enough to plan around at the scale of a single afternoon. The failure modes described here are the price of a system built quickly out of whatever instruments were available, and knowing them is what lets you use it for the decisions it is actually good for.
The habit worth building is small: before a trip in smoke season, look at the trend rather than the number, find where the nearest real monitor is, and have an elevation plan. Three things, thirty seconds, and they convert a colored map into an actual decision.
