Following a similar approach as described above for the 24-h and 48-h sampling periods, the effect of the number of 7-d periods was assessed. As a result, 138 (dataset 1) and 142 (dataset 2) 7-d moving averages were calculated, representing all possible 7 consecutive measuring day periods starting between 10 and 11 a.m. and not starting in weekends. The maximum of non-overlapping periods in dataset 1 and 2 were 27 and 28, respectively. Thus, a monitoring strategy could exist of n random non-overlapping and not necessarily consecutive weeks, with a maximum of n = 27 (dataset 1) or n = 28 (dataset 2). As for the 24-h and 48-h sampling strategy, all of the hourly ERs over the n 7-d periods were averaged for each n.