Methane emissions from a beef cattle feedyard: measurements and models

Methane (CH4) emissions from enteric fermentation by livestock account for about 2.1% of U.S. greenhouse gas (GHG) emissions, with beef and dairy cattle the most significant sources. Most current approaches to estimate the contribution of cattle to GHG emissions use emission factors based on production scenarios or calculate enteric emissions based on animal gross energy intake (GEI) and a methane conversion factor (Ym). Some models rely on statistical relationships between enteric methane emission and key dietary variables. More complex approaches incorporate ruminal processes to calculate enteric methane production at animal scale. Another class of models combines statistical, empirical and process based approaches with mass balance accounting to estimate methane emissions at farm scale. A better understanding of CH4 emissions from beef cattle feedyards can help build more accurate emission inventories, improve predictive models, and meet potential regulatory requirements.

Methane emissions during winter and summer at a commercial beef cattle feedyard on the southern High Plains in Texas were quantified during 32 days in winter and 44 days in summer using open path lasers, characterization of atmospheric turbulence and inverse dispersion analysis. The experimental uncertainty of this method, quantified using a Monte Carlo approach, was 17%. Mean monthly observed methane emissions were compared to methane emission estimates from diverse models that included statistical models, models based on ruminal processes, and a whole farm hybrid model. Models were on monthly time scale except for the whole farm model, which used daily time steps. Model inputs, including production and management practices, dry matter intake (DM) and dietary variables were derived from data collected at the commercial feedyard. Cattle diets differed between winter and summer. During winter, cattle diets averaged 70% steam flaked corn. During summer, the corn fraction was reduced and 21 to 27% wet distillers grains (a byproduct of ethanol production) was added to diets.

Methane per capita emission rates (PCER) ranged from 71 to 118 g animal -1 d-1 in winter and from 70 to 130 g animal -1 d-1 in summer. Mean monthly CH4 PCER was 84.1 ± 10.9, 85.2 ± 13.1, 85.9 ± 14.9 and 93.4 ± 15.4 g animal -1 d-1 in January, February, May and June, respectively. Methane loss ranged from 9.2 to 11.4 g CH4 kg-1 DMI, with lower values during winter. The GEI ranged from 135.2 to 164.5 MJ animal-1 d-1. Conversion of GEI to CH4 (Ym) averaged 2.8% in winter, 3.2% in summer and 3.0% overall. Observed mean monthly enteric CH4 energy loss ranged from 4.5 to 4.9 MJ animal-1 d-1, and averaged 4.6 ± 0.2 MJ animal-1 d-1.

Statistical models based on DMI, dietary metabolizable energy or fiber content of rations overestimated enteric CH4 emissions by 45 to 125% with root mean square errors (RMSE) that ranged from 2.0 to 5.8 MJ animal-1 d-1. Statistical models that incorporated the percentage of forage in diets, considered as the hay fraction of the feedyard diets, or also included DMI and ether extract (EE, diet fat content) agreed better with observed methane emissions, estimating observed emissions within -11 to 25% with RMSE that ranged from 0.2 to 0.6 MJ animal-1 d-1. A detailed model of ruminal processes developed for dairy cows yielded enteric CH4 estimates that averaged 21% greater than observations (RMSE = 1.1 MJ animal-1 d-1). Methane emission estimates from a farm scale model agreed within 10% of observed emissions, but a closer examination of the model showed that it was based on an Intergovernmental Panel on Climate Change (IPCC) Tier 2 method relevant to grazing cattle, not cattle in feedyards. Model estimates would have decreased to less than half of observed values if the conversion factor recommended for feedyard cattle (Ym = 3.0%) was used instead of that for grazing cattle (Ym = 6.5%).

Most effective statistical models were functions of forage, DMI and EE. The ruminal model, although it consistently overestimated enteric CH4 emissions, had monthly means that were within a standard deviation of observed emissions. It is encouraging that statistical models that estimated enteric CH4 emission as a function of forage content, DMI and EE performed reasonably well (within 25%), given that they were developed using data from grazing beef cattle, not cattle in feedyards on high concentrate diets. It suggests that enteric CH4 production can be estimated knowing the quantity of feed that an animal consumes, along with some simple measures of the composition of the DMI. Studies show that increasing the roughage, or forage, component of a diet increases CH4 production, and that increasing fat (measured as EE) decreases CH4 production. Including these two countervailing factors along with DMI could help improve statistical estimates of enteric CH4. Further refinement of ruminal process models is warranted, given the adequate performance of the ruminal model tested, which was largely based on data from dairy cows. The observed CH4 conversion factors (from 2.7 to 3.2%) confirm the Ym = 3.0 ± 1.0% currently recommended by IPCC for Tier 2 estimates of enteric CH4 from feedlot cattle. However, enteric CH4 emission calculated using this approach should be considered somewhat coarse because of the uncertainty in both measurements and Ym, and improved and refined models pursued.