The distribution of daily PM2.5 concentration (hourly) is bimodal in the city of Coyhaique (South of Chile). Coyhaique is the city more polluted by particulate matter accord the WHO. The major source of emission is the burning wood used for cooking and heating, principally at the night. The health risk estimation using longitudinal models could be biassed using the arithmetic mean each 24hrs as an exposure variable. Exist another function for PM2.5 different from the arithmetic mean?.
For ambient air quality data, log normal distributions are commonly used for some of the reasons you are concerned with. In many cases, in my opinion, the mode may be a more conservative measure risk and exposure analyses. Unfortunately I do not have references to peer reviewed work comparing the errors of one or the parameter used.
Thank you, Ra.
Maybe a time-discretionary function could be a possible "robust" solution to understand the (short-term) effect of the exposure to critical concentrations of PM2.5.
Best regards,
- Badges
- Science topic
- Similar topics
- Public Health
- Health