Why Low-Cost Air Quality Sensors Drift, Skew and Sometimes Lie — and What to Do About It

Low-cost air quality sensors are transforming how environmental scientists and community networks observe the atmosphere, but their readings are, by design, of lower quality than those from reference instruments. A peer-reviewed comparison of correction techniques published in npj Climate and Atmospheric Science identifies three recurring culprits: interference from environmental factors such as humidity, interference from other atmospheric species, and measurement drift. The good news is that each has a recognizable signature, and each can be managed through co-location testing and principled data correction.

The Promise and Pitfalls of Low-Cost Air Quality Sensors

The appeal of these devices is straightforward: dense networks of inexpensive sensors can fill spatial gaps that a handful of expensive monitoring stations leave wide open. Yet the authors of the npj Climate and Atmospheric Science review, led by Idris Hayward and colleagues, caution that low-cost sensor measurements are "usually of lower quality than those made by reference instruments." Because the technology remains under development, the EPA notes that little information exists on the quality of data these sensors produce — which is precisely why its researchers systematically evaluate how well they measure air pollutants and how easy they are to use.

Environmental Interference: A Major Source of Inaccuracy

The first culprit is the environment itself. Humidity — flagged explicitly in the review as an example — can significantly alter sensor readings, and it is far from alone. For electrochemical (EC) sensors, which low-cost devices commonly use to measure trace gas concentrations, other atmospheric species can interfere with the target analyte, producing readings that look plausible but are chemically confounded.

The interference problem compounds in the field. The review emphasizes that sensors deployed in an environment different from the one in which they were calibrated perform worse, because laboratory calibrations rarely span the full range of pollutant characteristics and environmental conditions a device will actually encounter.

Understanding Calibration Drift and Deployment Mismatch

The second culprit is time. The review lists measurement drift among the significant interferences affecting low-cost sensors, meaning a device that read accurately at installation may report progressively less trustworthy values later in its life. That is why the authors conclude these sensors "require regular corrections to improve their lifetime and the quality of their measurements, to enhance confidence in their reported data."

For optical particle counters (OPCs), the drift problem has a distinctive chemical twist. Changes in the composition of ambient particulate matter can bias OPC measurements because different particles have different refractive indices — a property the optical detection principle depends on. A sensor calibrated against one aerosol mixture may misjudge another, even if the instrument itself is mechanically sound.

The Role of Co-location Studies in Sensor Validation

So how does anyone know what a low-cost sensor is actually reporting? The EPA's answer is the co-location study. Researchers place the sensors near a regulation-grade monitor and collect data on air quality with both technologies simultaneously. Because both instruments observe the same air under the same weather conditions, researchers can compare how accurate and reliable the low-cost devices are relative to traditional methods.

Co-location is not just an evaluation exercise; it is the foundation for calibration. The parallel datasets reveal the systematic biases a particular device exhibits in a particular setting, providing the empirical basis for correction factors. One caution from the review: co-location study results themselves vary, influenced by the specific devices used, intra-sensor variability, secondary environmental conditions of each study, validation set sizes, and the black-box calibrations manufacturers apply internally. Two co-location studies of the same sensor model can therefore disagree, and the review's authors call for further work to understand those variances so results can be compared more meaningfully.

Principles of Data Correction for Reliable Measurements

Correction, in principle, means using the relationship observed during co-location to transform raw sensor output into calibrated estimates. If humidity inflates a device's response, the co-located record reveals that dependence; if an OPC skews under a changed aerosol composition, the comparison quantifies the bias. The review's central finding is that the right correction technique depends on both the device type and the deployment environment — there is no universal fix.

Crucially, correction is not a one-time event. Given measurement drift and the environmental sensitivities described above, regular re-evaluation is what sustains confidence in reported data over a sensor's lifetime.

Decision Framework: Diagnosing Low-Cost Air Sensor Inaccuracies

When your sensor's data looks questionable, match your sensor type and deployment context against this framework to isolate the most likely cause and the appropriate response.

Diagnostic mapping of common low-cost air sensor error sources to their causes and mitigation principles.
Source of Error	Affected Sensor Type(s)	Example Environmental Factor	Mitigation Principle
Environmental interference	Low-cost sensors generally; EC trace-gas sensors in particular	Humidity altering sensor response	Quantify the dependence through co-location and correct readings against measured environmental conditions
Cross-gas interference	Electrochemical sensors measuring trace gases	Other atmospheric species present in ambient air	Characterize the interfering response during co-location and subtract or model its contribution
Particulate composition variance	Optical particle counters	Changes in aerosol chemical composition, shifting refractive index	Calibrate against the ambient PM actually present; re-check calibration when the aerosol mixture changes
Measurement drift	Low-cost sensors generally	Deployment in an environment different from the calibration environment, plus time in service	Apply regular corrections throughout the sensor's lifetime; repeat co-location periodically
Manufacturer black-box calibration	Pre-calibrated commercial devices	Differing validation conditions across studies	Treat reported values as provisional; verify locally rather than assuming factory settings transfer to your site

Use the table prospectively as well as diagnostically: a humid coastal site with an EC sensor raises the environmental-interference and cross-gas flags before deployment, while an OPC downwind of changing emission sources flags the composition-variance row.

Making Informed Decisions with Low-Cost Sensor Data

Environmental scientists and engineers should implement co-location studies and apply appropriate data correction principles to ensure the reliability of low-cost air quality sensor measurements. Treat raw output as a hypothesis, not a measurement: the EPA's own evaluation practice — comparing sensors against regulation-grade monitors under identical conditions — is the model to follow, at whatever scale your project can support.

The measurable indicators of success are concrete: improved correlation coefficients between low-cost sensor data and reference monitor data, and reduced unexplained variance in long-term deployments. If your co-location checks show those two numbers moving in the right direction after correction, your sensor network is earning the confidence its data will be trusted with. If they do not, the framework above tells you which row to investigate next.

Why Low-Cost Air Quality Sensors Drift, Skew and Sometimes Lie — and What to Do About It

The Promise and Pitfalls of Low-Cost Air Quality Sensors

Environmental Interference: A Major Source of Inaccuracy

Understanding Calibration Drift and Deployment Mismatch

The Role of Co-location Studies in Sensor Validation

Principles of Data Correction for Reliable Measurements

Decision Framework: Diagnosing Low-Cost Air Sensor Inaccuracies

Making Informed Decisions with Low-Cost Sensor Data

Sources

Tags

Related articles

Matching the Quantum Sensor to the Measurement: Magnetometers, Gravimeters, and Atomic Clocks

How Quantum Error Correction Works: Redundancy, Syndromes, and the 3-Qubit Bit-Flip Code

Quantum Entanglement Explained: How It Powers Quantum Key Distribution—and Why the NSA Still Urges Caution

Four Kinds of Wearables, Four Different Jobs: How to Tell Them Apart