Since there are 8 data points available per day and since a sampling strategy lasts for about 30 weeks, the number of possible outcomes is very high (830). Therefore, the procedure was repeated 1000 times for each week day in order to get a representative sample of the full number of possible estimates. Doing so, 1000 estimated NH3 EFs were obtained for each weekday. This procedure was also applied to the scenario where a weekly sample was taken on a random working day instead of on a fixed day, again repeated 1000 times. For each of the fixed and random weekday strategies, the 1000 values were possible outcomes for the respective sampling strategy. All possible estimates were collected in histograms with a bin width of 0.25 kg NH3 year−1 (animal place)−1.