BreatheCheck Get alerts

The official monitor and the sensors near you disagree

In most US metros the regulatory monitor and the median of nearby outdoor sensors differ by 15 to 47 AQI points at the same hour. The monitor is one instrument, sited for regulatory coverage; the sensor network is dozens to hundreds of devices sited where people live. Neither is wrong. They measure different things, and the difference is what a resident actually notices.

The gap right now

Official AQI from EPA AirNow. Sensor AQI is the median of outdoor PurpleAir devices within roughly 17 miles, corrected with the published US EPA formula for PA-II units (0.524 × PA − 0.0862 × RH + 5.75) — raw PurpleAir readings run high, and shipping them uncorrected would be the same error we criticise elsewhere.

MetroOfficialSensorsGapDevices
Salt Lake City, UT 44
Los Angeles, CA 35
New York, NY 48
Chicago, IL 30
Houston, TX 63
Denver, CO 61
Atlanta, GA 85
Phoenix, AZ 42
Seattle, WA 18
Minneapolis, MN 27
Bakersfield, CA 58
Spokane, WA 67
Sacramento, CA 19
Fresno, CA 55
Pittsburgh, PA 45
Eugene, OR 32
Boise, ID 38
Fort Collins, CO 57
Reno, NV 45
Albuquerque, NM 44
Madison, WI 32
Tucson, AZ 38
Colorado Springs, CO 43
St. Louis, MO 49
Irvine, CA
Oakland, CA 52
Long Beach, CA 38
Milwaukee, WI 36

Why they disagree

Four mechanisms account for most of it, and they are not errors. Averaging window: the official figure is often a 24-hour average, so a sharp afternoon plume is diluted by a clean morning. Siting: one monitor covers a whole metro, and it sits where regulation requires, not where you stand. Height: smoke can travel aloft, above the sensor layer. Instrument: low-cost sensors read high in humidity, which is why the correction above exists.

A resident in Seattle put the experience plainly on a day the official number read 30: “I can smell the smokey smell, and the sun is bright red. Am I missing something?” They were not. Both numbers were describing something real.

How to use this

When the two numbers are close, either one is fine to act on. When they diverge by more than about 20 points, treat the higher one as the better guide for the next hour and the lower one as the better guide for the day — the monitor is slower and smoother, the sensors are faster and noisier.

Method and reuse

Both series are refreshed hourly and stored from day one. Sources: EPA AirNow (regulatory monitors) and PurpleAir (outdoor devices only, `location_type=0`). Metros with fewer than three reporting devices are left blank rather than averaged into a number we cannot stand behind.

If you want to cite or reuse this: the live figures are at /current as JSON, and per-metro history at /history/<slug>. Attribution to breathecheck.com is the only ask.

Updated