Dnipropetrovsk. Measured by low-cost sensors, too few to rank; the model estimate below is the cross-check.
24-hour average, not ranked: 77US AQI
1–2 sensors, not rankedModerate
PM2.5 22.9 µg/m³, 1.5× the WHO 24-hour guideline, from the one low-cost sensor in Kryvyy Rih. One cheap unit can read well off, so Kryvyy Rih isn't ranked.
7-day average: 86 US AQI (27.8 µg/m³)
30-day average: 78 US AQI (23.5 µg/m³)
Latest hour: 69 US AQI (19.0 µg/m³), local time
The model's estimate for the same 7 days is about 10 µg/m³: the city's reading is 2.8× that.
Measured, but by one low-cost sensor and no other monitor. It passes the same checks as every sensor we count (stuck, frozen, spikes, far from the model), yet one unit can still be off. A third sensor here would rank Kryvyy Rih. How sure each figure is. How accurate low-cost sensors are, from our own measurements. Its sensors are Sensor.Community's (Sensor.Community, ODbL); the values they set are published under the ODbL on our sources page.
Hour by hour
Hourly values over the last 4 days, local time. Dashed line: WHO 24-hour guideline (15 µg/m³).
Last 30 days
One cell a day, coloured by its band: 21 of 30 days had 18 hours or more of readings. Grey: not enough.
Indoors: in a study of 1,274 buildings, most of them homes, the typical home had 42% of the outdoor PM2.5 inside on ordinary days, and half of homes between 35% and 56%. At the last 24 hours of Kryvyy Rih's unranked sensors, 77 US AQI (22.9 µg/m³), that is about 9.6 µg/m³ inside from outdoors (8.0 to 13), before cooking, candles and cleaning add their own. Only a monitor indoors tells a home's own level. How much outdoor air gets inside
The model's estimate
7-day average, estimated: ≈ 10µg/m³
Estimate, not measuredLikely good to moderate
Likely between 5 and 20 µg/m³ (28–71 US AQI): the model's 7-day average to 15 Nov 2026, about 0.7× the WHO 24-hour guideline.
30-day average, estimated: ≈ 10 µg/m³
Model estimate from the Copernicus Atmosphere Monitoring Service (CAMS). No city within 300 km was measured by reference monitors on most days of the last week, and adjusting it takes 3 (low-cost sensors aren't used), so this is the raw model value: fewer than half of its days land within ×1.5 of a monitor, and one in ten is off by ×2.8 or more. Not ranked on SmogRank's boards.
Last 30 days, estimated
Daily means, UTC days. Dashed line: WHO 24-hour guideline (15 µg/m³).
Estimates contain modified Copernicus Atmosphere Monitoring Service information 2026. Neither the European Commission nor ECMWF is responsible for any use that may be made of the Copernicus information or data it contains. How estimates are made.
Use this page
SmogRank, estimated PM2.5 daily means (CAMS model estimate, not measured), Kryvyy Rih, Ukraine, 17 Oct 2026 – 15 Nov 2026, smogrank.com/ua/kryvyy-rih (accessed 17 Nov 2026). Contains modified Copernicus Atmosphere Monitoring Service information 2026.
You're welcome to reuse a single chart, image or figure with this citation and a link. Terms. Raw monitor readings in bulk are published by OpenAQ, under each provider's licence.