Air emissions from sour-gas processing plants and dairy-cattle reproduction in Alberta, Canada

The dispersion of air pollutants from all 231 licensed sour-gas processing plants in Alberta, Canada, was modeled on a monthly basis over a 10-year period (1985-1994). Exposure estimates for sulfur dioxide (SO(2)-used as a surrogate for exposure to combusted emissions) then were assigned to 1382 provincial dairy farms using a geographical-information system. Individual average and peak exposure for periods prior to each of 15 months of age and conception (four exposure-averaging periods for each of two dispersion models) were estimated for 163,988 primiparous female dairy-cattle between 1986 and 1994. Monthly or annual average farm-site exposure estimates likewise were assigned to associated herd-level data sets for the biologically relevant period of interest for each of three additional reproductive outcomes: monthly herd-average calving interval, stillbirth risk, and twinning risk. In one of the main-effects models, the maximum (i.e., peak) monthly sour-gas exposure experienced by individual-animals from birth to conception was associated with an increased time to first-calving in the very-highest exposure category (hazard ratio=0.86, 95% CI=0.80, 0.92). This equates to a decreased hazard (lambda) of calving (in each month subsequent to 22 months of age) for the highest-exposure animals (lambda=0.170) versus the zero-exposure animals (lambda=0.198) in a model with referent values for agro-ecological region and season of birth. The dose-response was not consistent across the full range of exposure categories. There was significant (P=0.003) interaction of emissions with agro-ecological region. After accounting for the interaction, a more-consistent dose-response was evident for some (but not all) agro-ecological regions. This suggests that any effect of emissions on dairy-heifer reproduction is subject to modification by features of soil type, vegetative cover, and/or climate. The increase in monthly herd-average calving interval on farms exposed to the very-highest levels of emissions appeared quite small and of limited practical importance within the range of expected exposures. There was no association between exposure and the risk for twinning. Herds exposed to higher emissions exhibited a slight decrease in risk for stillbirth.