English Abstract In heavy loaded traffic situations, large negative NO2
measurement data are regularly encountered. This is directly due to the
technical construction of the employed standard Philips PW9762
nitrogen-oxide monitor. Upon presentation of the cumulative frequency
distribution, the higher percentile values will be overestimated
consequently ; particularly 10 and 15% for the 95 and 98-percentile one
respectively. The use of 3-hourly floating averages, instead of the
customary hourly ones enables a realistic estimate of these higher NO2-
percentiles. Appropriate adaption of the shape of the sample delay volume
enables the proper establishment of NO2, even in the afore-mentioned traffic
situations.
