A naive read of the dashboard treats each parameter as independent. The chemistry and the sensors disagree. Several well-documented interactions affect both the readings and the physiology. The interpretation layer encodes them so it does not confidently report a false event.
Sensor-side interactions: humidity above 80% nudges the VOC index up by 10 to 30 even with no actual VOC change; very humid air also scatters infrared light, inflating optical PM readings until the droplets evaporate. NOx index responds to ozone in addition to nitrogen oxides; outdoor-ozone infiltration on a sunny afternoon can produce a small NOx rise with no combustion source. Sensor cross-sensitivities covers the detail.
Indoor chemistry interactions are the more interesting case. Weschler reviewed the field: ozone reacts with terpenes (limonene from citrus cleaners, alpha-pinene from pine products) to produce formaldehyde and small ultrafine particles. The reaction is fast; opening a window on a high-outdoor-ozone day right after you cleaned with a citrus product can produce a measurable formaldehyde signature indoors that neither the ozone nor the cleaner produced alone.
Physiological interactions follow similar logic. Cold-dry winter air potentiates particle and pollen irritation (dry mucous membranes have less buffering capacity). High humidity at high temperature impairs heat-stress recovery (see apparent temperature). Combined cognitive impact studies (CO2 + temperature + noise) show additive rather than independent effects. The AI uses these to weight its notifications: a moderate VOC spike during a cold-dry winter morning gets a slightly stronger flag than the same spike on a temperate day, because the symptom risk is higher in the first context.
References
- Sensirion - SEN66 datasheet and VOC index info sensirion.com
- Weschler - Indoor reactive chemistry (Environmental Health Perspectives) doi.org
- EPA - Volatile organic compounds and indoor air www.epa.gov
- EPA - Ground-level ozone basics www.epa.gov
