# Air Quality and Environmental Monitoring Networks

Urban air quality monitoring is an underserved application for community mesh networks. Low-cost sensor nodes can build hyperlocal air quality maps that government monitoring stations - typically spaced miles apart - cannot provide. Keep in mind throughout that low-cost air-quality sensors are **not regulatory- or reference-grade**: their readings are indicative and supplementary, not authoritative.

## Why Low-Cost Sensors Matter

Regulatory air quality monitors are sparse - often only a handful per metropolitan area - so a single station may represent tens of square miles (see EPA AQS network design criteria, 40 CFR Part 58 App. D). Regulatory reference monitors cost roughly $15K-$40K each, and a fully-equipped multi-pollutant station including siting and maintenance can run into the six figures; each represents a wide catchment area. Community mesh nodes with low-cost sensors (roughly $50-200 per node as a BOM estimate) can provide neighborhood-level data that reveals hotspots, traffic corridors, and industrial emission events invisible to the regional monitoring network. Community low-cost-sensor networks such as PurpleAir have demonstrated this hyperlocal value, but their readings are indicative rather than reference-grade.

## Sensor Options for Air Quality

<table id="bkmrk-parametersensorcosta"><thead><tr><th>Parameter</th><th>Sensor</th><th>Cost</th><th>Accuracy</th><th>Notes</th></tr></thead><tbody><tr><td>PM2.5 / PM10</td><td>Plantower PMS5003 or SDS011</td><td>$15-25</td><td>SDS011 ≈ max(±15%, ±10 µg/m³)</td><td>Most important for health; needs temperature correction. PMS5003 has no published absolute-accuracy spec, only consistency. Both are UART sensors - see the build note below on Meshtastic support.</td></tr><tr><td>CO₂</td><td>Sensirion SCD40 or SCD41</td><td>$35-50</td><td>±(50 ppm + 5% of reading)</td><td>True NDIR sensor; factory-calibrated, but relies on periodic fresh-air exposure (automatic self-calibration) to hold long-term accuracy.</td></tr><tr><td>VOCs</td><td>SGP30 or SGP41</td><td>$15-20</td><td>Semi-quantitative</td><td>Good for trend and event detection; not precise absolute levels</td></tr><tr><td>NO₂ / O₃</td><td>SPEC Sensors electrochemical</td><td>$50-100 each</td><td>±20 ppb (see SPEC Sensors datasheet)</td><td>Higher cost; good for near-road monitoring. Electrochemical sensors drift over time and need temperature/humidity compensation.</td></tr><tr><td>Temp + Humidity</td><td>BME280</td><td>$3-8</td><td>±0.5°C, ±3% RH</td><td>Essential for correcting PM sensor readings</td></tr><tr><td>Temp + Humidity</td><td>SHT31</td><td>$3-8</td><td>±0.3°C, ±2% RH</td><td>More accurate alternative for PM correction</td></tr></tbody></table>

## Building a PM2.5 Monitoring Node

The following is a **design sketch, not a turnkey recipe**. The PMS5003 and SDS011 are UART laser particulate sensors, and stock Meshtastic's Telemetry module does **not** support them. For particulate matter in stock Meshtastic, use the I2C **PMSA003I** (I2C address 0x12), which the Telemetry module's air-quality metrics path supports. A UART PMS5003 requires custom firmware or a separate MCU/co-processor that injects readings into the mesh.

```
# Hardware: RAK4631 + RAK1906 (BME680) for temp/humidity correction
# For stock Meshtastic PM support, use the I2C PMSA003I (addr 0x12),
# NOT the UART PMS5003/SDS011 (those are not natively supported).

# Path A (recommended, no custom firmware):
#   PMSA003I (I2C) on a Meshtastic node -> AirQualityMetrics telemetry

# Path B (UART PMS5003): requires a co-processor or custom firmware.
#   Meshtastic environment telemetry uses fixed protobuf fields, so a
#   raw PMS5003 needs either:
#   - an ESP32 (e.g. T-Beam) running custom Arduino firmware that reads
#     PMS5003 + BME280 over UART/I2C and formats the data, or a custom
#     portnum, or
#   - a Raspberry Pi co-processor reading PMS5003 via UART and injecting
#     messages into the mesh via the Meshtastic Python API (see the
#     Meshtastic Python API send-message docs; clarify whether you send
#     plain text or structured telemetry).
# MeshCore note: there is no deployable "SENSOR firmware" that broadcasts
# PM telemetry; treat the working paths above as the only buildable ones.
```

## Community Air Quality Network Design

For a neighborhood-scale (10-50 node) air quality network:

- **Spatial coverage:** 1 node per 0.5-1 km² in residential areas; denser near industrial sources and major roads
- **Transmission interval:** a recommended 5-15 minutes (PM sensors need averaging to reduce noise; see EPA sensor-performance guidance on averaging)
- **Data aggregation:** Central room server or MQTT gateway with InfluxDB backend
- **Public dashboard:** Grafana public dashboard showing real-time map (use Leaflet.js for geographic visualization)
- **Calibration:** Collocate 2-3 nodes with the nearest EPA monitoring station to develop calibration coefficients. Per EPA Air Sensor Toolbox collocation guidance, recommended collocation can run weeks to months and ideally spans multiple seasons.

## Community Value and Partnerships

- Share data with PurpleAir, AirNow, or OpenAQ platforms for broader visibility
- Partner with local universities for calibration studies and data analysis
- Provide data to neighborhood environmental justice organizations
- Submit findings to local air quality district as citizen science data

Important: low-cost PM and gas sensors are not regulatory-grade. Readings require collocation/calibration and should always be presented as indicative, supplementary data - not as authoritative air-quality measurements. Do not use uncalibrated node readings as the basis for individual health decisions; defer to official AirNow/EPA data for that.