📄 earth_science

Nephelometer-based air sensor networks systematically underestimate PM2.5 in dust-influenced West Africa

A multi-year study of air sensor networks in Ghana reveals that standard nephelometer-based sensors systematically underestimate PM2.5 and PM10 concentrations by significant margins during dusty Harmattan conditions, highlighting critical limitations in relying solely on optical sensors for air quality monitoring in dust-influenced regions like West Africa.

Original authors: Daniel M. Westervelt, Abhishek Anand, Selina Amoah, Benjamin Essien, Kingsley Mawuli Amegah, Clement Mensah Ackaah, Esi Nerquaye Tetteh, Nana Ama Browne Klutse

Published 2026-07-08

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Daniel M. Westervelt, Abhishek Anand, Selina Amoah, Benjamin Essien, Kingsley Mawuli Amegah, Clement Mensah Ackaah, Esi Nerquaye Tetteh, Nana Ama Browne Klutse

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to count the number of people in a crowded room, but you are using a camera that only sees tiny specks of dust and completely misses the big, heavy backpacks people are carrying. If you only count the specks, your total count will be way too low, especially when everyone is wearing heavy backpacks.

This is essentially what Daniel Westervelt and his team discovered about air quality sensors in West Africa. Here is the story of their findings, broken down simply:

The Problem: The "Blind" Sensors

For years, scientists and governments have been using cheap, low-cost air sensors to measure pollution (specifically PM2.5, which are tiny particles that can get deep into our lungs). Most of these popular sensors work like a nephelometer—think of it as a flashlight that shines through the air. It counts how much the light scatters to guess how much dirt is in the air.

The problem is that these sensors are great at seeing tiny, invisible specks (like smoke from a fire or car exhaust), but they are terrible at seeing larger, heavier particles (like sand and dust).

The Setting: West Africa's "Dusty Season"

West Africa, particularly Ghana, has a unique weather pattern called the Harmattan season (roughly November to March). During this time, massive clouds of dust blow in from the Sahara Desert.

The Analogy: Imagine a room where, for a few months, people are throwing handfuls of sand into the air. The air is thick with big, heavy grains.
The Reality: In this region, the air isn't just filled with tiny smoke; it's filled with these big, heavy dust grains.

The Experiment: A 2-Year Road Trip

The researchers set up 16 "smart" air monitors across four major cities in Ghana, from the dusty north (Tamale and Kumasi) to the coastal south (Accra and Sekondi-Takoradi). They ran this for about two years.

Crucially, each "smart" monitor had two different eyes inside it:

The "Flashlight" Eye (Plantower): The standard, cheap sensor that most people use.
The "Microscope" Eye (Optical Particle Counter): A more advanced sensor that can actually see and size the big dust grains.

They compared what the "Flashlight" saw against what the "Microscope" saw to see who was telling the truth.

The Big Discovery: The "Missing" Pollution

The results were a wake-up call:

In the North (Dusty Areas): When the Harmattan dust was blowing, the cheap "Flashlight" sensors were blind. They missed about 33% of the dangerous fine dust and a staggering 91% of the coarse dust.
- The Metaphor: It's like trying to weigh a truckload of sand using a scale that only registers the weight of the individual grains of sand, ignoring the truck itself. The scale says the load is light, but the truck is actually crushing the road.
In the South (Coastal Areas): The sensors worked much better here because the air was cleaner and had less of that heavy desert dust.
The Seasonal Shift: Pollution in Ghana is like a tide. It is much worse in the dry, dusty season (double or triple the pollution of the wet season). The sensors failed the most when they were needed the most.

Why This Matters

The paper points out that many previous studies and public air quality maps for West Africa have relied only on the "Flashlight" sensors.

The Result: These studies have been systematically underestimating how bad the air is. They are telling people the air is cleaner than it actually is.
The Conclusion: You cannot trust these cheap sensors to give an accurate picture of air quality in dusty environments. They need to be paired with better technology (like the "Microscope" eye) or carefully calibrated to account for the dust they can't see.

Summary

In short, the paper says: "Don't trust the cheap air sensors in West Africa during the dusty season. They are like a pair of glasses that only see the small stuff, making the big, dangerous dust clouds invisible. Because of this, we have been underestimating the health risks in this region."

Technical Summary: Nephelometer-based Air Sensor Networks Systematically Underestimate PM2.5 in Dust-Influenced West Africa

Problem Statement
Particulate matter (PM2.5) exposure is a leading cause of premature mortality globally, yet data availability in Africa remains limited, hindering effective policy and mitigation. While low-cost optical air sensors have proliferated to close this data gap, their reliability in specific environments is uncertain. A critical issue is that many widely used "low-cost" sensors, such as the Plantower PMS5003, operate as nephelometers. These devices measure light scattering from an ensemble of particles rather than sizing individual particles. In environments dominated by coarse, supermicron aerosols—such as the Saharan dust transported during the West African Harmattan season—nephelometer-based sensors may fail to accurately detect larger particles, leading to significant underestimation of PM2.5 and PM10 concentrations. Previous studies in the region have largely relied on these sensors without sufficient validation against reference-grade instruments during dusty conditions.

Methodology
To address these uncertainties, the authors deployed a multi-year monitoring network (January 2024 – May 2026) across four major metropolitan regions in Ghana: Accra-Tema, Sekondi-Takoradi, Kumasi, and Tamale. The network consists of 16 QuantAQ Modulair-PM (MOD-PM) units. Each MOD-PM unit integrates three distinct measurement capabilities:

Plantower PMS5003: A nephelometer-based sensor providing raw, uncalibrated PM estimates.
Alphasense OPC-N3: An optical particle counter (OPC) capable of sizing individual particles across 24 bins (0.35–40 µm).
Calibrated Product: A locally validated output combining data from both sensors with corrections for density, aspiration efficiency, and hygroscopicity (using $\kappa$ -Köhler theory based on local temperature and humidity).

The calibrated MOD-PM product was previously validated against Federal Equivalent Methods (FEM) in Accra, showing high correlation ( $r^2 = 0.86$ ) and low bias. For this study, the authors classified data points as "dust-influenced" using a PM1:PM10 ratio threshold of < 0.1, a metric aligned with literature on Saharan dust. They compared the raw Plantower-only estimates against the calibrated MOD-PM product to quantify bias using Mean Bias Error (MBE), Mean Normalized Bias (MNB), and Mean Absolute Error (MAE).

Key Results

Geographic and Seasonal Disparities: The study confirms a north-south gradient in air quality. Northern and central cities (Tamale, Kumasi) exhibit significantly higher PM2.5 concentrations and more frequent exceedances of Ghana EPA standards compared to coastal cities (Accra-Tema, Sekondi-Takoradi). A strong seasonality is observed, with dry season (Harmattan) PM2.5 levels being 2 to 3 times higher than wet season levels.
Systematic Underestimation by Nephelometers: The core finding is that Plantower-only sensors systematically underestimate PM concentrations during dusty conditions.
- PM2.5: During dusty conditions (PM1:PM10 < 0.1), the Plantower-only sensor underestimates PM2.5 by approximately 33% (Mean Normalized Bias of -32.8%). The regression slope for dusty conditions is 0.42, indicating that more than 30% of the actual mass is missed.
- PM10: The underestimation is far more severe for coarse particles. In dusty conditions, the Plantower sensor misses 91% of the PM10 mass (MNB = -91%).
- Non-Dusty Conditions: When coarse particles are not dominant (PM1:PM10 > 0.1), the Plantower sensor exhibits a slight positive bias (+27.9% for PM2.5), leading to partial error cancellation in the aggregate dataset, which masks the severity of the underestimation during dust events.
Exceedance Rates: While coastal areas exceed the Ghana EPA 24-hour standard (35 µg m⁻³) on 15% of days or less, central and northern cities exceed this on 30–49% of days. However, all sites exceed the stricter WHO Air Quality Guideline (15 µg m⁻³) on at least 50% of days.

Significance and Claims
The paper asserts that relying exclusively on nephelometer-based sensors (like the Plantower PMS5003) in dust-influenced environments leads to a "blind spot" in air quality monitoring. The authors claim that previous estimates of PM2.5 and PM10 in West Africa, and potentially other dusty regions, that rely solely on such sensors are likely underreporting exposure levels.

The study emphasizes that the bias introduced by these sensors in dusty conditions cannot be easily corrected using post-hoc statistical methods because the coarse particles are not being detected even indirectly. Consequently, the authors conclude that accurate air quality monitoring in these regions requires sensors capable of sizing individual particles (like optical particle counters) or rigorous, locally calibrated hybrid approaches. This work challenges the validity of existing crowdsourced data portals in the region and calls for careful performance evaluation before deploying sensor networks in environments where supermicron dust is prevalent.