18 µg/m³ measured at Green Bay East High, 2 km away, at Aug 06, 11:00AM. The NowCast smoke average is 18 µg/m³ — Moderate on the US EPA scale.
Unusually sensitive people should consider reducing prolonged outdoor exertion.
Forecast concentrations are below current levels and expected to ease over the next 48 hours.
No significant smoke is expected in the next 48 hours.
● Observed · ● FireSmoke · ● FireWork · ● HRRR
Background bands (US EPA AQI, 24-h basis): ■ Good · ■ Moderate
| Model | Peak PM2.5 | Peak around |
|---|---|---|
| HRRR-Smoke (NOAA, 3 km) | 6 µg/m³ | Aug 06, 6:00PM |
| FireWork (ECCC) | 12 µg/m³ | Aug 07, 5:00AM |
| FireSmoke (UBC BlueSky) | 3 µg/m³ | Aug 07, 12:00PM |
Model values are for the ~10 km grid cell nearest Green Bay; the two models are independent and often disagree — treat the higher value as the cautious planning number.
Observations and NowCast are total PM2.5 — wildfire smoke plus every other source — as regulatory monitors measure it, and FireWork models that same total. FireSmoke is a smoke-dispersion model: it estimates the wildfire contribution only, so it reads lower than the monitors whenever other sources matter. The map's optional plume layer is likewise fires-only.
Little smoke at the surface or overhead.
● At the ground · ● Overhead (if mixed down)
Overhead smoke dims the sun but is not what you breathe. The dashed line shows the surface level if the whole column mixed through the boundary layer — a worst case, not a forecast · Data: NOAA HRRR-Smoke
1-day record · 2 nearby monitors
| Model | Typical error | Tendency |
|---|---|---|
| HRRR-Smoke | 11.5 µg/m³ | reads low 11.5 |
| FireSmoke | 16.3 µg/m³ | reads low 16.3 |
Mean absolute error and average tendency in µg/m³, from forecasts verified 0–6 hours ahead at monitors within 50 km. A short-range record, not a measure of how far ahead a model stays reliable. FireSmoke and HRRR model wildfire smoke only, so they read below the monitors whenever other sources of PM2.5 matter.
Open the live chart for Green Bay East High — hourly observations, NowCast and both model forecasts.