Mitigating Gas Emissions from Liquid Manure Storage Systems: Management Practices, Measurements and Modeling
Livestock agriculture is a source of methane (CH4), nitrous oxide (N2O) and ammonia (NH3) emissions. A comprehensive approach aimed at mitigating these emissions is needed to reduce the environmental footprint of agriculture. This thesis examined three aspects of a comprehensive mitigation strategy which included: evaluating a management consideration, improving flux measurement techniques, and modeling emissions. The relationships between total solids (TS), gas emissions and surface crust dynamics were assessed. Diurnal and long–term CH4 and N2O flux variations were examined in the context of improving discrete sampling protocols. Finally, a mechanistic model that predicts CH4 emissions from manure slurries was evaluated. Over long–term storage, the crusts were not effective in mitigating total gas emissions because the slurries remained open to the atmosphere for a significant portion of the time. Total CH4 and NH3 emissions were related linearly to TS, while N2O exhibited a sigmoid response. The linear response to TS observed for CH4 and NH3 occurred despite varying crust conditions suggesting that the availability of substrates in slurries is the more important regulator of emissions over long–term storage. Diurnal CH4 and N2O flux variations were linked with the diurnal surface temperature (T0) cycle, with the strength of the relationship depending on surface crust conditions. An assessment of discrete sampling protocols revealed that sampling intervals should be ≤7 d. In terms of the timing, it is best to sample these gases when the T0 is closest to the daily mean, which would typically be before 0900 h or after 1700 h. A mechanistic model of CH4 emissions from slurries was evaluated. The model was most accurate during the first 47 d of storage, after which the accuracy decreased. However, total emissions estimated over 145 d were within –21.1 to 6.0% of measured emissions for slurries with TS 3.2%, 5.8% and 8.2%. Emissions were also modeled using the USEPA inventory method. There was relatively good agreement between the USEPA and mechanistic models, with relative percent differences ranging from 19.9 to 37.3%, which is encouraging from the standpoint of advancing greenhouse gas inventory methods.