TrailSmoke publishes a live smoke and air-quality page for 17 US cities and 8 national parks and outdoor destinations, each built on the same forecast: NOAA HRRR-Smoke near-surface PM2.5 for the next 48 hours, active fires and perimeters from NIFC, and ground readings from EPA AirNow monitors. Every one is linked below.
Cities differ less in how much smoke they get than in what it takes to get rid of it. Seattle and Missoula can record the same PM2.5 on the same afternoon and be in completely different situations, because Seattle is one wind shift from clean air and Missoula is not. Four mechanisms cover the cities on this list, and knowing which one a place runs on tells you whether to wait out an episode, change your plans for the week, or leave.
Flushed by marine air
Typical clearing time: Hours
Clean ocean air is never far away, and when it returns it scours the basin in a single afternoon. Episodes here start abruptly and end just as abruptly, so the useful question is when the wind turns onshore, not when the fire is contained. The trap is that smoke can ride above a shallow marine layer while surface monitors still read clean, which is why the sky and the numbers disagree so often on this coast.
One mechanism ends it: onshore flow. When the marine push returns through the Strait of Juan de Fuca and over the Chehalis Gap, clean Pacific air scours the lowland, often within a single afternoon. Until that happens, no amount of local wind helps, because the surrounding terrain keeps recirculating the same air. Watch the forecast for a westerly shift rather than for the fire to be contained.
The Gorge east wind is usually broken by returning onshore flow off the Pacific, which arrives through the same Coast Range gaps that bring the marine layer. Once westerly flow is established the valley flushes quickly. The valley also pools smoke overnight on its own, so mornings routinely read worse than the previous afternoon even with no new smoke arriving.
Onshore flow clears the surface quickly, often in hours. The complication is that smoke can ride above the marine layer while surface air stays clean, so the sky can look apocalyptic while monitors read moderate. The near-surface forecast and the visible sky routinely disagree here, and the near-surface number is the one that matches what you breathe.
The Delta breeze is the mechanism. Marine air pushes through the Carquinez Strait from San Francisco Bay most summer afternoons and ventilates the valley from the southwest. When a ridge shuts the Delta breeze off, smoke has nowhere to go and concentrations climb for days without any new fire growth. Watching for the breeze to return is more predictive than watching containment numbers.
Held in a basin
Typical clearing time: Days to weeks
Air drains downhill into these places overnight and pools there, so smoke that arrives once tends to stay. Concentrations follow a daily rhythm rather than a trend: afternoon heating mixes the layer upward and readings improve, then the next night undoes it. Waiting for the number to fall for good usually means waiting for a front, and in the deepest valleys that can take weeks.
The valley inversion is the problem. Smoke that arrives during the day settles overnight and is still there at dawn, so a single calm night can undo an afternoon of improvement. Real clearing usually waits for a frontal passage rather than a sea breeze, which means Spokane episodes tend to last longer than coastal ones.
The valley is a sink, so clearing depends on wind rather than time. Daytime heating mixes the layer upward and improves readings by afternoon; the following night's drainage flow undoes much of it. Episodes here often show a strong daily rhythm rather than a steady level, and the hourly forecast is more useful than a daily average.
Slowly. The valley holds what settles into it, and a light afternoon breeze is often not enough to turn the air over. Missoula episodes are measured in weeks rather than days, and the practical question is usually which hours of the day are tolerable rather than when the smoke will be gone.
The bowl needs a real wind event to flush. Lake-driven breezes shuffle air within the valley without removing much of it, so a smoke layer can persist through several days of otherwise pleasant weather. Afternoon mixing gives the daily minimum concentration, and overnight settling gives the maximum.
Downslope flow off the Sierra can scour the basin quickly, but the overnight inversion in the Meadows will re-form and re-concentrate whatever is left. Reno tends to have sharper daily swings than Tahoe just over the hill, because the basin is shallower and the daytime mixing is stronger.
Daytime heating usually does it. Once the surface warms and the mixed layer deepens, drainage smoke that pooled overnight is diluted through a very deep column and surface readings drop sharply. The pattern is a morning peak and an afternoon minimum, the reverse of Denver.
Mixed down from above
Typical clearing time: Days
Smoke arrives overhead first and reaches the ground when the afternoon mixed layer deepens enough to connect the two. That inverts the usual advice: the hours that clear most cities are the hours these cities get worse, and readings peak in mid-afternoon rather than at dawn. In the largest of them the sea breeze relocates smoke inland instead of removing it, so the worst air is well away from the fire.
The afternoon mixing that clears most cities can make Denver worse, because it connects the surface to a smoke layer that was overhead all morning. Readings frequently peak in mid-afternoon rather than at dawn, which is the opposite of the valley cities. Clearing normally waits for the upper-level flow to change.
The sea breeze does not remove smoke from the basin so much as relocate it. Coastal neighbourhoods clear first, and the San Gabriel Valley, Pasadena, and the Inland Empire read worst in the afternoon even when the fire is on the coastal side. Real clearing needs either sustained offshore flow or a trough deep enough to ventilate the whole basin.
Daytime mixing handles the imported layers within a day or two of the source easing. Spring Mountains smoke behaves differently and follows a drainage cycle, worst overnight and into early morning.
Carried in from far away
Typical clearing time: Days, steady
Almost no wildfire of consequence starts near these cities. The smoke is Canadian or western, a thousand miles old by arrival, and a wind shift is the only thing that ends it. Because the source is distant and the plume is broad, concentrations hold steadier hour to hour than in a valley city, which makes the multi-day forecast more trustworthy here than anywhere else on this list.
Entirely synoptic. A wind shift or a frontal passage ends the episode and nothing else does. Because the source is a thousand miles away and the transport is broad, concentrations tend to be steadier hour to hour than in a valley city, which makes the multi-day outlook more useful than the hourly one.
A wind shift, and nothing else. These episodes typically arrive and depart with the synoptic pattern over a period of days rather than hours, and the lake breeze modulates where within the metro the worst readings show up rather than whether they occur.
A shift in the synoptic flow, typically the passage of a front or the movement of the upper-level trough steering the plume. Once the wind turns onshore off the Atlantic the improvement is fast, often within a few hours.
The regional smoke leaves with the synoptic pattern, but the valley inversions release more slowly than the surrounding plateau does, so low-lying neighbourhoods lag the general improvement by hours. Readings across the metro can differ substantially by elevation during the same episode.
National parks and outdoor destinations
Parks invert the problem. Nobody needs to know whether a park is smoky right now, because nobody is standing in it wondering; they need to know whether a trip booked for next weekend is worth taking. That pushes the useful forecast horizon past what any smoke model can honestly offer, so the practical approach is to watch the fires upwind rather than the plume itself, and to treat the 48-hour forecast as the last check before driving rather than the basis for the decision.
Elevation also does more work here than it does in a city. Smoke pools in valleys and drains off ridges overnight, so in most of these parks the difference between a clear morning and an unbreathable one is a few hundred feet of camp elevation.
Valley and high country behave like different parks. Smoke pools on the valley floor overnight and can leave Tuolumne Meadows clear on the same morning Yosemite Valley is unhealthy, so the elevation you sleep at matters more than the one you hike to.
A closed basin at 6,200 feet ringed by Sierra forest. Smoke that gets in over the crest circulates rather than leaves, and the 2021 Caldor Fire showed how fast an approaching fire can turn a smoke problem into an evacuation.
The canyon is the mechanism. Smoke sinks below the rim overnight and sits in the inner gorge, so a rim viewpoint can be clear while a corridor trail below is not, and prescribed burns on the Kaibab add planned smoke in spring.
A high plateau that ventilates better than the valleys around it, with the Lamar and Hayden basins as the exceptions that pool overnight. Late August and September are the months to plan around.
Jackson Hole is an enclosed valley with a river running through it, which is a recipe for overnight pooling. Smoke settles into the hole and lifts through the morning, so an early start is often the smokiest part of the day here.
Northern Rockies timber plus imported British Columbia smoke, at the late end of the season. The North Fork and Lake McDonald valleys hold smoke long after the passes have cleared.
Most of its smoke is imported and arrives aloft, so the park frequently reads cleaner at the Alpine Visitor Center than Denver does at the same hour. Afternoon mixing is what brings it to the surface.
High, exposed, and close to the Cascade and Klamath fires that supply it. The rim ventilates well in wind and holds a layer in calm, and the visibility that the park exists for is the first thing to go.
Somewhere not on this list
Any town in North America works on the main map: search it and the forecast centres there. These pages exist for the places people search by name, and the state directory covers the same ground one level up. For how to read the forecast itself, including why the sky and the monitors disagree so regularly, start here.