calculated with the Robins (1979) approach, with atmosphericstability deduced from ultrasonic anemometer measurements, andwith the use of the factor 4 of the German TA Luft (2002).The separation distances are shown as isolines in the followingfigures, encompassing the area of exceedence of the given thresholds.The larger the area, the more unfavourable is the odour impactcriterion. In each of the figures showing the separation distances,AODM results (left) are compared to LASAT results (right) for thesame scenario. A statistical comparison in the form of scatter plotsand a regression analysis both for 3 and 8% exceedence probabilityis added. Separation distances smaller or equal to 50 m are set to50mas a lower limit (see Section 3.1); these data are excluded fromthe statistical analysis (Tables 5e7) and do not show up in the scatter plots (Figs. 6, 8 and 10). The amount of excluded data can bededuced indirectly from the