SAMPLING DURATION AND FREQUENCY FOR DETERMINING EMISSION RATES FROM NATURALLY VENTILATED DAIRY BARNS

摘要

The ideal practice for determining gas emission rates from concentrated animal feeding operations (CAFOs) is continuous measurement throughout the year to capture diurnal and seasonal variations. However, data that meet this criterion are scarce because this approach is costly and technically challenging. A practical approach with reduced sampling time and frequency without compromising integrity is thus necessary. In this study, we examined five reduced sampling protocols for determining emission rates from naturally ventilated dairy barns: (1) six sampling events, during even months, each event running continuously for 24 h (144 hourly data points); (2) six sampling events, during odd months, each event running continuously for 24 h (144 hourly data points); (3) six sampling events, during even months, each event running continuously for 7 d (1,008 hourly data points); (4) six sampling events, during odd months, each event running continuously for 7 d (1,008 hourly data points); and (5) 12 sampling events, one event every month, each event running continuously for 24 h (288 hourly data points). These five reduced sampling protocols were evaluated against baseline emission rates obtained from a protocol consisting of 12 sampling events, one event every month, each event running continuously for 7 d (2,016 hourly data points). The average baseline emission rates for carbon dioxide (CO2), ammonia (NH3), and hydrogen sulfide (H2S) ranged from 7.0 to 10.8 g d(-1) AU(-1), from 16.9 to 24.3 g d(-1) AU(-1), and from 0.40 to 0.63 g d(-1) AU(-1), respectively. Emission rates for CO2 and NH3 obtained from all five reduced sampling protocols had relative biases of less than 20% from the respective baseline emission rates. This implied that even the most reduced sampling protocol (six sampling events per year, each event running continuously for 24 h) would be adequate for determining CO2 and NH3 emission rates. However, for H2S, relative biases of the reduced sampling protocols ranged from 2% to 45%, with a 50% chance of emission rates falling outside +/- 20% of the baseline emission rates. These protocols, however, are only applicable for quantifying emission rates from barns with steady-state live animal mass.