The measuredconcentrations have been converted into non-dimensional values of K = (Cmeas-ured/Csource) · uref · H2 / Q for comparison with similar results from numericalsimulations and field measurements. Cmeasured is the measured tracer concentra-tion at a certain measurement point, Csource stands for the tracer concentration atthe source, uref defines a reference wind speed, H is a reference height (buildingheight) and Q indicates the emission source strength.  Defining a maximum al-lowable concentration of a given air pollutant and a given source strength, safedistances can be derived directly from contour plots of measured K-values.  Theresults of high-resolution concentration measurements indicated that close to thewalls of the source building high concentration gradients can be found and asignificant transport of pollutants along the surface of the building occurs.A comparison with similar results from numerical simulations (TÜV Nord,1996) shows that safe distances calculated from numerical simulations are con-stantly smaller than the values derived from our wind tunnel experiments.  Ob-viously, the limited spatial resolution of discrete numerical models affects thepredicted pollutant concentration especially in regions with strong concentrationgradients found close to building surface.  Considering, that apartment buildingsand office complexes often have windows near to ventilation outlets of parkinggarages it becomes evident that numerical model results are likely to underesti-mate the air quality problems and health risks around ventilation outlets. 
Ana-lysing the time series of measured concentrations it was found that the dispersionin structured arrays of buildings is characterized by low frequency, aperiodicchanges of the concentration distribution.  As a result, the instantaneous disper-sion patterns substantially differ from what a time averaged concentration plotshows.  Especially in the street canyons parallel to the wind (setup B) areas withrather low mean immission concentrations were characterized by very highshort-term concentrations.4.  ConclusionsBased on results of two simplified test configurations the capability of windtunnel modelling of emissions from underground parking garages has been dem-onstrated.  It could be shown that physical modelling leads to reliable resultswith high resolution in space and time even in areas close to the emissionsources.  Effects like pollutant transport within the internal boundary layer onbuilding surfaces could be resolved and quantified.  The instantaneous characterof flow and dispersion patterns was investigated by analysing high frequencytime series of point-wise flow and dispersion measurements.  A comparison withsimilar results from numerical simulation gives evidence that conventional nu-merical modelling is substantially underestimating concentration gradients closeto the modelled emission sources.  To avoid risks for human health, emissions from naturally ventilated underground parking garages should be investigated inmore detail by means of systematic wind tunnel experiments and backing fieldmeasurements. In addition, the results of systematic wind tunnel measurementscould be used for the development, improvement and validation of numericaldispersion models applied to urban-type dispersion problems.AcknowledgmentThe authors wish to express their appreciation for support from the German Fed-eral Environmental Agency (grant 296 43 831).ReferencesB.A.U.CH.: 1993, Sachbericht zum Projekt “Belastung der Innenraumluft durch Emissionen ausTiefgaragen, Beratung und Analyse - Verein für Umweltchemie e.V., Wilsnacker Str.  15, D-10559 Berlin, p. 47Environ Corporation: 1984, Sensitivity Analysis of the California Department of Health ServicesRisk Assessment of Benzene, Prepared for the American Petroleum Institute, WashingtonARC MONOGRAPHS: 1982, Benzene and Annex. In: Some industrial chemicals and dyestuffs.Lyon, International Agency for Research on Cancer, IARC Monographs on the Evaluation ofthe Carcinogenic Risk of Chemicals to Humans, Vol. 29Kaimal, J.C., Wyngaard, J.C., Izumi, Y., Cote, O.R.: 1971, Spectral characteristic of surface layerturbulence, Quart. J. Roy. Meteorol. Soc., Vol. 98, No. 417, pp. 563 – 589Simiu, E., Scanlan, R. H.: 1986, Wind Effects on Structures – Part A: The Atmosphere, John Wiley& Sons, Inc.TÜV Nord e.V.: 1996, Studie - Anforderungen an die Lüftung von Tiefgaragen unter demGesichtspunkt von Schadstoffimmissionen,  Technischer Überwachungs-Verein Nord e.V.,Große Bahnstraße 31, D-22525 Hamburg, p.70VDI 3783: 1999, Physical modelling of flow and dispersion processes in the atmospheric boundarylayer, VDI, Environmental Meteorology, VDI/DIN Manual of the Commission on Air PollutionPrevention, VDI 3783 - 12 (Draft), p.22 地下车库的污染物扩散英文文献和中文翻译(3):http://www.youerw.com/fanyi/lunwen_69929.html