Sleep efficiency ran 3 points lower above 12 µg/m³; Sacramento's burn ban starts at 31.

In bedrooms where fine particles ran above 12 micrograms per cubic meter, sleep efficiency was 3.27 percentage points lower than in the cleanest bedrooms. The 95% confidence interval runs from 1.03 to 5.51 points. Sleep efficiency is the share of time in bed spent asleep. The figure comes from a 2023 study in Sleep Health of 62 adults in Louisville, Kentucky, who wore a wrist motion sensor for 14 nights while monitors in their bedrooms logged PM2.5, the particles smaller than 2.5 micrometers.

Two cautions come with that number. The comparison group breathed air at 0.1 micrograms or less, so this is dirty against clean, not 15 against 35. And the paper calls the drop dose-dependent, while its table gives an overall p-value for PM2.5 of 0.063 in the fully adjusted model, short of the usual 0.05 line. The safe reading is that the top band slept worse than the bottom band, not that sleep fell by a step at every level.

Deep sleep, scored by a Fitbit

A 2026 paper in Scientific Reports asked about deep sleep directly. Lin and colleagues put air monitors in the bedrooms of 163 undergraduates in Shanghai, one night each, and scored sleep with a Fitbit Alta HR. Bedroom PM2.5 averaged 16 micrograms per cubic meter. A higher level went with a smaller share of the night in deep sleep: a standardized coefficient of -0.19, 95% confidence interval -0.36 to -0.02, per one standard deviation of PM2.5.

In minutes, by my arithmetic: deep sleep averaged 20.0% of the night, with a standard deviation of 5.1 points. One standard deviation of bedroom PM2.5 therefore goes with about one point less deep sleep, somewhere between 0.1 and 1.8 points. On that study's 409-minute average night, one point is about four minutes. The low end of the interval is almost zero, and each student was measured for one night.

"Deep sleep" here is an estimate. The Fitbit infers stages from movement and heart rate, not from brain waves, and the authors write that this can blur stages compared with a clinical sleep lab. The mechanism they offer (inflammation, oxidative stress, a nervous system pushed toward its alert side) comes from earlier work. Neither study tested it. Basner's team also cites a 20-person Fitbit study that found better sleep with more PM2.5, at low levels.

Sacramento's winter days land in that range

I pulled the Environmental Protection Agency's daily PM2.5 file for Sacramento County. Five monitors report: Del Paso Manor, 1309 T Street, Bercut Drive, Folsom and Sloughhouse. I took each monitor's 24-hour average and counted a day when any of them read above a line. The air district's wood-burning season runs November through February, 120 days.

In 2024-25, a monitor read above 12 on 49 of those days, above 15 on 37, above 25 on 8 and above 31 on 1: December 7, 2024, when the T Street monitor read 35.6. In 2025-26 it read above 12 on 53 days, above 15 on 38, above 25 on 8 and above 31 on none. Fifteen is the World Health Organization's 24-hour guideline. January and February 2026 are the newest values EPA has posted and may still be revised.

Where you live moves the count. In 2025-26 the Del Paso Manor monitor read above 12 on 47 days and the Folsom monitor on 25, about half as many, in the same county and the same weeks.

Outdoor readings, indoor studies

Those are outdoor numbers. Both studies measured bedroom air, and Lin's team writes that it cannot say how much of the PM2.5 came from outside, from indoor sources or from particular chemicals in the mix. I have no bedroom measurement for Sacramento, so I cannot tell you how many local nights cost sleep. What the monitors do show is that dozens of winter days sit in the range where these studies found a difference.

The law acts at the tail

Rule 421 of the Sacramento Metropolitan Air Quality Management District bans burning in fireplaces, wood stoves, inserts and pellet stoves when the district forecasts a 24-hour PM2.5 average above 31 micrograms per cubic meter (Stage 1) or above 35 (Stage 2). Between 25 and 31 it asks for voluntary restraint. Those cut points are in the text the district posts, adopted in 2007 and amended on September 24, 2009. The two sleep studies came out in 2023 and 2026. The rule covers unincorporated Sacramento County and the cities of Citrus Heights, Elk Grove, Folsom, Galt, Isleton, Rancho Cordova and Sacramento.

Across the 240 days of the last two seasons, a monitor read above 31 once. On 86 of the 102 days a monitor read above 12, none reached 25, where the voluntary tier begins. The rule runs on forecasts, so declared burn days can differ from readings. The district's season tally loads by script and I could not read it; the district's rule contact is [email protected].

The burn season opens Sunday, November 1. The district posts same-day and next-day burn status at airquality.org and on 1-877-NO-BURN-5. A first violation draws a $50 penalty, or a passing grade on the district's wood smoke exam. Burning on a prohibited day can be reported at 279-207-1122.

A sleeper sees none of this. The night looks like every other night, and the morning after a 15 feels like the morning after a 25, if it feels like anything. The studies describe groups, the monitors describe the street, and the bedroom is the one place in this chain where I have no number.