Objectives Particulate concentrations within poultry barns pose a potential occupational and environmental health concern. We aimed to identify determinants of PM₁₀ and PM_2.5 levels in poultry barns and evaluate the effectiveness of electrostatic precipitators (ESP) to reduce environmental air pollution.

Methods PM₁₀ and PM_2.5 Harvard impactor samples were collected in “side-by-side” barns (one with an ESP, one without) at five poultry farms in British Columbia, Canada from 2008 to 2012. Measured particulate levels were analysed using multimodel inference and linear mixed-effects models after log-transformation. Random effects were added to account for clustering within farms, barns, and rearing cycles.

Results A total of 234 PM_2.5 and 230 PM₁₀ valid samples were modelled. Geometric means of PM_2.5 and PM₁₀ were significantly lower in barns with ESP control in place (151 and 427 µg/m³, respectively) compared to barns with no control (334 and 969 µg/m³), resulting in unadjusted % reductions of 47% and 50% respectively.

In statistical models, the fixed-effects explained 57% and 72% of the total variance in the PM_2.5 and PM₁₀ measurements, respectively. The residual (i.e. within rearing cycle) and between rearing cycle variance were the most affected by adding the fixed effects structure. Strongest determinants in the models for both dust types were ESP use (i.e. approximately halving particulate levels for both PM_2.5 and PM₁₀), bird age (i.e. 10–30 fold increase in particulate levels depending on bird and particulate type), and type of bird (i.e. approximately a 2.5 fold increase for PM_2.5, and four fold increase in PM₁₀ for turkey compared to chicken). Interactions were suggested in PM₁₀ models between type of bird, bird age, and ESP use.

Conclusions The use of ESP resulted in significant reductions in in-barn particulate even after controlling for other determinants such as bird age, type of bird, ventilation, and date.