Measurement optimization for determining concentrations and emission rates of particulate matter and ammonia from a large size dairy building

Total suspended particle (TSP), particulate matter less than 2.5 μm in aerodynamic diameter (PM_2.5) and ammonia (NH₃) in a large-size dairy building were continuous measured for one year with high sampling frequency (5 min-PM, 20 min-NH₃) at multi points (12-PM, 10-NH₃). A total of 64 and 16 measurement methods for each pollutant were obtained to calculate the annual mean concentrations (AMC) and emission rates (AME), respectively, which was combined by different levels of sampling duration (D1-D4: 30, 15, 7, 1 d month⁻¹), sampling frequency (F1–F4: 5/20, 60, 180, 360 min) and number of sampling points (N1–N4: 12/10, 5, 3, 1). Data measured by D1F1N1 was considered as the true value. Error analysis and hierarchical clustering were applied to select the optimal combinations. Results showed that the baseline AMC of TSP, PM_2.5 and NH₃ in the large dairy building were 86.4, 28.5, 875.0 μg m⁻³, respectively, and the corresponding AME were 140.6, 28.5, and 3461.1 mg h⁻¹·cow⁻¹. The effect of sampling frequency on AMC was less significant than the other two factors. The D1/D2 and N1/N2 levels were essential to ensure the monitoring accuracy of AMC of PM and NH₃, respectively, while D1-D3 levels were essential to quantify the AME. Measurement of AMC could be optimized from the maximum of D1F1N1 to D4, F4, N4 level, respectively, and corresponding measurement of AME could be optimized up to D2 and F4 level. Optimized measurement methods were varied among different monitoring purposes (AMC or AME) and different pollutant types.