In a long-term weekly grab sampling strategy, a single grab sample was taken e.g. every Monday, for a certain number of weeks. In that case, the average of these samples is an estimate of the emission. This was mimicked as follows. For each Monday in the original dataset, a data point between 9 a.m. and 5 p.m. was randomly picked. This was done for each working day. Since not all days of the week are equally represented in the original dataset, not all simulated sampling strategies had the same length: the lengths varied between 31 and 34 weeks. The average of these 31 to 34 (depending on the day of the week) randomly picked data points is a possible emission estimate for a long-term weekly sampling strategy.