In order to reduce the relative error to maximum ±15% for the single grab sampling strategy (Fig. 10), 84 (dataset 1) and 27 (dataset 2) random grab samples need to be taken. Taking 100 random grab samples reduced the relative error further down to −1% and 14% (dataset 1) and −8% and 7% (dataset 2). If more grab samples were taken, the precision only improved slightly. In order to reduce the relative error to maximum ±15% for the long-term weekly grab sampling strategy (Fig. 11), weekly grab samples had to be taken during 32 (dataset 1) and 28 (dataset 2) consecutive weeks.