The long-term weekly grab sampling strategy was simulated by taking weekly grab samples on a random working day. So, this means that a grab sample would be taken on n consecutive weeks on a random working day. This was simulated by combining the daytime values of the weekdays per week and taking n weekly grab samples from that combination during 1 to n weeks, starting with a random week. For each value of n, the values from the n weekly grab samples were averaged and then converted to kg year−1 (animal place)−1. This resulted in 35 (dataset 1) and 34 (dataset 2) estimated EFs.