Outdoor Air Quality in Cities: Where We Stand in 2026

Real problems. Real data. Real solutions.

Sensorbee AB  •  February 2026  •  sensorbee.com

As a manufacturer of environmental monitoring systems, Sensorbee works daily with the organisations that actually deal with urban air pollution — construction companies managing dust on site, municipalities tracking traffic emissions, and environmental consultants producing compliance reports. This article summarises the current state of outdoor air quality in cities and explains how monitoring technology, combined with artificial intelligence, is turning raw sensor data into decisions that measurably improve the air people breathe.

The problem in numbers

In 2021, the Health Effects Institute reported that air pollution contributed to 8.1 million deaths globally, making it the second-leading risk factor for death worldwide. Over 90 percent of those deaths were linked to fine particulate matter (PM2.5). The World Health Organisation puts ambient outdoor air pollution alone at 4.2 million premature deaths per year, with 89 percent occurring in low- and middle-income countries. These are not abstract statistics. They translate into shortened lives from heart disease, stroke, chronic obstructive pulmonary disease, and lung cancer.

In 2024, only seven countries worldwide met the WHO’s recommended annual average PM2.5 guideline of 5 µg/m³. Ninety-nine percent of the global population breathes air that exceeds WHO guideline limits. The air quality monitoring software market alone is projected to grow from USD 0.72 billion in 2025 to USD 1.11 billion by 2031 — a clear sign that governments and industries are waking up to the fact that you cannot manage what you do not measure.

The five pollutants of greatest concern remain particulate matter (PM2.5 and PM10), nitrogen dioxide (NO₂), ozone (O₃), sulphur dioxide (SO₂), and carbon monoxide (CO). To these, volatile organic compounds (VOCs) and hydrogen sulphide (H₂S) are increasingly added in monitoring programmes, particularly near construction sites, industrial zones, and wastewater treatment facilities.

Why cities still struggle

Urban population continues to grow. The United Nations projects 68 percent of the world’s population will live in cities by 2050. More people means more vehicles, more construction, more energy demand, and more pollution sources packed into smaller areas. Urban road canyons — streets flanked by tall buildings — trap emissions and create pollution hotspots where concentrations far exceed what a regional monitoring station would report.

Traditional reference monitoring stations remain essential for enforcement and regulatory compliance. They produce legally defensible data that holds up in court. But they are expensive to build and maintain, difficult to site, and they provide data from only one fixed point. A city with five reference stations and 500 km² of area has one data point per 100 km² — and yet air quality can vary by a factor of ten within a single city block.

This spatial data gap has historically meant that urban air quality maps rely heavily on modelling with very few calibration points. That is no longer acceptable when regulations are tightening, citizens are checking air quality indices as routinely as weather forecasts, and construction companies face real financial penalties for exceeding dust limits.

Dense sensor networks: closing the data gap

The solution is straightforward: deploy more monitoring points. Not to replace reference stations, but to complement them with high-density networks of compact, solar-powered, cellular-connected sensors that deliver continuous data from dozens or hundreds of locations simultaneously.

This approach has matured considerably since the early experiments with low-cost sensors a decade ago. The key problems that plagued early devices — poor sensitivity, drift, cross-interference between gases, and unreliable calibration — have been addressed through better electrochemical cell design, individual sensor calibration, and intelligent firmware that compensates for temperature and humidity effects.

Sensorbee’s Air Pro 2 Multi-Sensor Monitoring Station is designed for exactly this use case. Weighing just 1.9 kg and housed in a weather-resistant polyamide enclosure, the Pro2 runs on solar power or external DC supply, communicates via LTE-M or NB-IoT, and supports up to six modular gas sensor slots alongside built-in particulate matter monitoring (PM1, PM2.5, PM10), temperature, humidity, and barometric pressure. An expansion port with Modbus RS-485 allows integration of external sensors such as vibration monitors, wind sensors, and sound level meters — making it a genuine all-in-one environmental monitoring station.

Critically, the Sensorbee Pro2 (SB8202) is MCERTS certified, meaning its particulate matter measurements meet the standards required by the UK Environment Agency for regulatory monitoring. MCERTS certification is not a marketing badge — it is a rigorous performance standard that confirms the sensor delivers accurate, repeatable results under real-world conditions. For construction companies operating in the UK, this means Sensorbee data can be used directly for compliance reporting without the expense of separate reference-grade equipment.

Construction site monitoring: a practical necessity

Construction and demolition activities are among the largest sources of particulate matter in urban areas. Excavation, concrete cutting, material handling, and vehicle movement on unpaved surfaces all generate dust. Diesel engines powering heavy machinery add NO₂, CO, and further PM2.5 to the mix. Without monitoring, site managers are operating blind — they have no way of knowing whether their dust suppression measures are working or whether they are about to breach a permit condition.

Regulations are tightening across Europe and North America. Cities increasingly require construction sites to deploy continuous monitoring with automated alerts when dust levels approach threshold values. The expectation is shifting from periodic spot-checks to real-time oversight with a documented data trail.

Sensorbee addresses this with a turnkey solution. A typical construction site deployment consists of one or more Pro2 units positioned at the site boundary, each equipped with the SB4102 Particle Matter Module (individually 3-point calibrated with its own certificate), the SB4652 Sound Level Meter for noise compliance, and the SB3641 Vibration Sensor for monitoring ground vibrations from piling or demolition. Wind speed and direction data from the SB3611 Wind Sensor helps correlate dust readings with meteorological conditions — essential for understanding whether elevated readings are caused by on-site activity or background sources.

This all-in-one approach eliminates the need to procure, integrate, and maintain separate instruments from different vendors. One device, one data stream, one cloud platform. Setup takes minutes: mount the unit on a pole or wall, connect the solar panel, insert a SIM card, and scan the QR code to register it on the Sensorbee Cloud. Data flows immediately.

From data to decisions: what AI actually does

There is a lot of noise around AI in environmental monitoring — much of it vague. Here is what it concretely delivers in 2026:

• Sensor health analysis. AI algorithms continuously evaluate sensor readings against expected patterns. A gradual baseline drift, an unusual spike, or a flat-line reading that suggests a blocked inlet — these are detected automatically and flagged before they corrupt your dataset. This is especially valuable in networks of dozens or hundreds of sensors where manual inspection of every data stream is impractical.

• Source identification. When multiple sensors report elevated readings simultaneously, AI can cross-reference with wind direction and speed to triangulate the pollution source. When only one sensor spikes, the cause is most likely local. When a cluster of sensors show elevated values downwind of a particular area, you are looking at a point source. This transforms monitoring from passive recording into active investigation.

• Predictive alerts. By learning from historical patterns — time of day, day of week, weather conditions, construction schedule — AI models can predict when pollution levels are likely to approach threshold values and issue warnings before exceedances occur. A construction site manager who receives an alert 30 minutes before a predicted dust exceedance has time to activate water suppression or pause a dusty operation. An alert that arrives after the exceedance is just documentation of a compliance failure.

• Automated reporting. Regulatory compliance generates paperwork. AI-powered platforms can automatically compile monitoring data into the format required by local authorities, flag any periods of exceedance, and generate trend analysis that demonstrates whether mitigation measures are working over time. This saves consultants and site managers hours of manual data processing.

• Sensorbee is integrating AI capabilities into the Sensorbee Cloud platform to deliver these insights directly to users — not as abstract dashboards, but as actionable notifications and reports that lead to real decisions on site.

What well-designed networks reveal

A dense network of sensors provides insights that no single monitoring station can match, regardless of its accuracy. Spatially distributed measurements reveal:

• Pollution hotspots — specific intersections, building facades, or site boundaries where concentrations are consistently elevated.

• Temporal patterns — morning rush-hour NO₂ peaks, afternoon ozone buildup, overnight industrial emissions.

• Far-field vs. local sources — distinguishing between pollution generated on your site and regional background levels that you cannot control.

• Mitigation effectiveness — quantifiable evidence that a new traffic management measure, a dust suppression system, or a low-emission zone is actually reducing exposure.

• Network health — if one sensor reads differently from its neighbours, it may indicate a sensor fault rather than a real pollution event, improving data reliability across the entire network.

This kind of hyperlocal data is exactly what city planners need to make infrastructure decisions that deliver measurable improvements rather than guessing where to allocate budgets.

The Sensorbee ecosystem

The Sensorbee product line is designed around a simple principle: one base station, modular sensors, one cloud platform. The Pro2 base unit handles power management, data logging (with built-in buffer for connectivity outages), and cellular communication. Sensor modules plug in without tools and are individually calibrated.

Available sensor modules:

Model	Parameter	Range	Key Spec
SB4202	Nitrogen Dioxide (NO₂)	0–10,000 ppb	±7 ppb accuracy
SB4242	Nitric Oxide (NO)	0–5,000 ppb	±7 ppb accuracy
SB4262	Carbon Monoxide (CO)	0–7,000 ppb	±80 ppb accuracy
SB4272	Ozone (O₃ + NO₂)	0–10,000 ppb	±8 ppb accuracy
SB4252	Sulphur Dioxide (SO₂)	0–10,000 ppb	±15 ppb accuracy
SB4282	Hydrogen Sulphide (H₂S)	0–2,000 ppb	±10 ppb accuracy
SB4292	VOC	0–2,000 ppb	±130 ppb accuracy
SB4232	Ammonia (NH₃)	0–100 ppm	±5 ppm accuracy
SB4212	Carbon Dioxide (CO₂)	400–5,000 ppm	Photoacoustic, 3yr life
SB4102	Particulate Matter	PM1, PM2.5, PM10	±5% precision, MCERTS
SB4652	Sound Level Meter	40–100 dBA	±1 dB(A) accuracy
SB3641	Vibration Sensor	±200 mm/s	BS 7385 / BS 6472
SB3611	Wind Speed & Direction	0.5–45 m/s	±1° direction accuracy
SB3602	Wind & Rain Combo	Wind + 0.2mm rain	Ultrasonic + IR optical

All gas sensors use electrochemical cells with ppb-level resolution and are tested according to CEN/TS 17660-1:2022 methodology. The CO₂ sensor uses photoacoustic technology with a 3-year operational life. Each sensor module ships with individual calibration documentation.

Beyond sensors, the ecosystem includes the SB6103 Solar Panel (14W, with adjustable mounting bracket), EU and UK power adapters for mains-powered installations, extension cables, and Y-splitters for connecting multiple external sensors. For sites requiring Modbus integration with third-party systems, the Pro2’s M12 8-pin connector supports RS-485 at both 5V and 12V.

For cost-sensitive mass deployments, the Sensorbee Air Lite (SB3352) provides particulate matter, temperature, humidity, and noise monitoring in a compact 300g polycarbonate housing with RS-485 Modbus output — ideal for integration into third-party systems or large-scale urban networks.

Sensorbee Cloud: data you can act on

The Sensorbee Cloud platform manages devices, stores data, and delivers insights through a web interface, REST API, and push notifications. Devices communicate using LwM2M (version 1.1 or later), an industry-standard protocol for IoT device management that ensures secure, efficient over-the-air configuration and firmware updates.

Project management features let you group sensors by site, set threshold-based alerts via email and SMS, download data as CSV for further analysis, and integrate with third-party platforms via PUSH or REST API. For organisations that require data sovereignty, Sensorbee supports deployment on customer-specific cloud infrastructure.

What comes next

The trajectory is clear. Regulations will continue to tighten. The EU’s updated Ambient Air Quality Directive sets stricter limits on PM2.5 and NO₂, with member states required to bring levels down to WHO-aligned targets. Construction site monitoring requirements are expanding in scope, with more jurisdictions mandating continuous noise, dust, and vibration monitoring with real-time reporting.

At the same time, the cost and complexity of monitoring is falling. Solar-powered cellular sensors eliminate the need for mains power and wired networks. Cloud platforms replace manual data collection. AI replaces the environmental consultant who used to spend days reviewing spreadsheets.

The result is that continuous environmental monitoring is no longer reserved for large corporations with dedicated environmental teams. A mid-size construction company can deploy an MCERTS-certified monitoring system on a project site in under an hour and have real-time dust, noise, and vibration data accessible from any web browser.

That is not a vision of the future. That is what Sensorbee delivers today.

Sensorbee AB

Jägarvallsvägen 8B, 584 22 Linköping, Sweden

Phone: +46 13 390 95 37

Email: hello@sensorbee.com

Web: www.sensorbee.com