Examining different methods to assess feed intake, milk yield, and enteric methane emissions at smallholder dairy farms: A case study in West Java, Indonesia

• Two models estimating feed intake were combined with three CH 4 emission equations • CH 4 emission intensity estimates from models and feed intake measurements differed • CH 4 emission intensity estimates were similar across the three equations • Farm rankings for CH 4 emission intensity were consistent across method combinations • Model input quality and calibration need improvement for smallholder dairy farms Effective and accurate estimation of enteric methane (CH 4 ) emissions of dairy cows in data-scarce environments, such as smallholder dairy farms, is hindered because performance and applicability of different approaches to estimate these emissions have hardly been compared. The objective of this study was to examine whether different combinations of methods to assess feed intake, milk yield, and enteric CH 4 emissions affect results and conclusions when applied to smallholder dairy farms in Indonesia. Feed intake, milk yield, and body weight were measured bimonthly for 46 cows in Lembang, Indonesia, from December 2017 to October 2018. The NRC model and the LiGAPS-Dairy model were selected to estimate feed intake. Measured feed intake and estimated feed intake from the two models were fed to three empirical equations to estimate enteric CH 4 emissions from dairy cattle in the tropics. Farms were ranked based on their enteric CH 4 emission intensity (i.e., kg CH 4 per kg fat- and protein-corrected milk). Results showed the choice of methods to assess feed intake and milk yield is an important factor for the level of enteric CH 4 emissions intensity from smallholder dairy farms in Indonesia. Farm rankings based on their enteric CH 4 emission intensity using the six pairwise combinations were consistent. Modelling feed intake and milk yield could be more affordable in data-scarce environments than conducting measurements, but improvements in the accuracy of input data (feed quality) and model calibration with small datasets are still needed for accurate estimation of feed intake and the associated enteric CH 4 emissions.