Measurements and Modelling of Full-Scale Concentration Fluctuations : Open-field experiments using krypton-85 and tetrahydrothiophen as tracers

Abstract

The simulation of odour frequencies near animal housing facilities or industrial plants requires a model of the frequency distribution of con- centration. Generally, for this purpose two-parameter distribution func- tions are used, which, depending upon the mean value and the variance of odour concentration, provide information about the probability of oc- currence. In order to examine how well the simulations coincide with reality, the calculation results are compared with measurement data. In model construction this examination step is also termed validation. Open-field experiments with tracers serve to examine the dispersion and fluctuation behaviour of airborne material in the atmospheric wind field. If in particular one intends to analyze the concentration fluctuations of odours, measuring techniques with high temporal resolution must be employed because odours can be perceived by the human nose within one second. In this examination, the artificially produced gas krypton-85 ser- ves as a tracer, whose concentration was measured by a group of detec- tors each at different distances from the source in one-second intervals. Based on the measurement data, mean values, variance, and higher sta- tistical moments of the tracer concentration are calculated and compared with the corresponding results of model simulations. Observed and simu- lated exceedance probabilities are compared, and the statistical charac- teristics of the models used are discussed. Do the results of the tracer measurement can be directly applied to odour frequencies? In order to answer this question, in some field experiments, the odourant tetrahydrothiophen, whose odour effects were registered by test persons at selected detector positions, was released at the same time as the tracer krypton-85. In contrast to the precise, unambiguous tracer measurement, the perception and assessment of odours is dependent u- pon sensation so that the application of the exceedance probability de- termined through measurement to the frequency of sensory perception needs to be statistically validated.