This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy Atmospheric Environment 42 (2008) 7050–7058 Contents lists available at ScienceDirect Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv A comparative and critical evaluation of odour assessment methods on a landfill site Laura Capelli a, *, Selena Sironi a, Renato Del Rosso a, Paolo Céntola a, Massimiliano Il Grande b a Politecnico di Milano, Olfactometric Laboratory, Department of Chemistry, Materials and Chemical Engineering ‘‘Giulio Natta’’, P.za Leonardo Da Vinci, 32, 20133 Milano, Italy b Progress S.r.l., Via N.A. Porpora, 147, 20131 Milano, Italy a r t i c l e i n f o a b s t r a c t Article history: Received 1 October 2007 Received in revised form 11 April 2008 Accepted 12 June 2008 This paper discusses the application of three different odour characterization techniques, i.e. chemical analyses, dynamic olfactometry and electronic noses, for the assessment of odour emissions from a landfill site. The results of the chemical analyses, which are useful to determine the chemical composition of odours, show no correlation with the odour concentration values measured by dynamic olfactometry. Olfactometric analyses enabled to measure odour concentration and thereby to quantify the sensory impact of odours. Finally, the continuous ambient air monitoring by electronic noses allowed to quantify the time percentage in which odours are detected at the landfill boundaries and at a receptor, which always turned out to be lower than 15%. This study represents a critical review of employing three different odour characterization methods for differing reasons on the same site, showing that, whilst the results don’t necessarily correlate, they do have an intrinsic value, and therefore demonstrating the complexity of environmental odour measurement. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Odour emissions Electronic nose Dynamic olfactometry Odour concentration Gas chromatography 1. Introduction Odour emissions are one of the major environmental problems that several industrial categories have to face. Complaints from the population living near plants characterized by unpleasant odour emissions (e.g. waste treatment and disposal plants, wastewater treatment plants, chemical industries, food industries, livestock activities, rendering plants, tanneries, etc.) are becoming more and more frequent, and they often are the origin of legal suits (in the case of existing plants), or they may represent the limiting factor for the construction of new plants (Schlegelmilch et al., 2005). * Corresponding author. Tel.: þ39 02 23993206; fax: þ39 02 23993291. E-mail addresses: laura.capelli@polimi.it (L. Capelli), selena.sironi@ polimi.it (S. Sironi), renato.delrosso@polimi.it (R. Del Rosso), paolo.centola@polimi.it (P. Céntola), m.ilgrande@olfattometria.com (M. Il Grande). 1352-2310/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2008.06.009 Given the high number of industrial typologies facing environmental odour problems, their relative abundance on the territory and the fact that the population is becoming more sensitive towards topics concerning air quality, regulatory and environmental protection bodies are in need of specific tools for odour nuisance assessment. The techniques available for odour nuisance characterization and quantification are substantially of three different kinds (Gostelow et al., 2001): - Analytical: chemical analyses; - Sensorial: dynamic olfactometry; and - Senso-instrumental: electronic nose. Analytical techniques allow to determine the qualitative and quantitative composition of a gas mixture using suitable separation and identification techniques, e.g. gaschromatography coupled with mass-spectrometry (GC– MS) (Davoli et al., 2003). This technique has the advantage Author's personal copy L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 of being consolidated and therefore considered objective, repeatable and accurate. The disadvantage relies in the difficulty of relating the chemical composition of an odorous mixture to its olfactory properties (Stuetz et al., 1999). This is mainly due to the effects of synergy or masking that can occur within the different compounds (odour properties are not necessarily additive). Another problem is given by the difficulty, when dealing with complex odours, of identifying a limited number of compounds, representative of the perceived odour. Moreover, in general it is also difficult to analytically detect the presence of some compounds at extremely low concentrations, which may nonetheless originate odour due to their low odour thresholds. As a result, good correlations between olfactory properties and analytical concentrations can be found for mixtures composed by few and well defined different compounds. Sensorial techniques, such as dynamic olfactometry (EN 13725, 2003), use the human nose as a sensor, and therefore enable to characterize odours by referring directly on their effects on a panel of qualified examiners. In addition to olfactory properties there are several factors that may influence odour perception. The most important one is the variability of human olfaction between different subjects. This problem is minimized by using a panel composed by several examiners, selected with precise criteria in order to have people with a standardized olfaction, and by averaging the single examiners’ responses. Senso-instrumental techniques use artificial noses, which perform instrumentally the functions of human olfaction. Electronic noses are complex systems with a human nose like structure (Pearce, 1997). An array of partially selective sensors gives a characteristic pattern of each analyzed odorous mixture, which is subsequently classified based on a reference database acquired by the instrument in a previous training phase (Persaud and Dodd, 1982). Performances of these instruments depend critically on a set of operational choices, i.e. on the sensor realization technique and working conditions, on the measurement settings, on the data processing methods and on the classification algorithms (Sironi et al., 2007a). This technology is widely used for odour recognition in the food industry (Schaller et al., 1998; Ampuero and Bosset, 2003), but its applications in the environmental sector are very recent and still under study (Micone and Guy, 2007). Among existing types of industrial installations that can cause odour nuisance, landfills represent one of the most common sources of odour emissions and complaint (Sironi et al., 2005). Odours from landfill sites originate principally from the atmospheric release of compounds that are formed during the biological and chemical processes of waste decomposition (El Fadel et al., 1997). The aim of this paper is to discuss the application of the above mentioned odour characterization and quantification techniques for the assessment of the odour emissions of a municipal solid waste (MSW) landfill site of average dimensions, having a surface of about 300,000 m2, located in the North of Italy. The monitoring campaigns were repeated in four different periods of the year (winter, spring, summer and fall), in order to improve the evaluation of the landfill odour 7051 impact by taking account of different meteorological conditions. For a more complete evaluation of the results, the meteorological data (especially wind speed and wind direction) relevant to the four monitoring campaigns, registered by a meteorological station installed inside the landfill at issue, were analyzed. This work represents a critical review of how three different odour assessment methods can be employed for evaluating the odour emissions from a landfill site. Through the critical analysis of the results it is possible on one hand to analyze the specificities of the three adopted odour characterization techniques and on the other hand to discuss the correlations between these techniques, thus demonstrating the complexity of odour measurement in the environment. 2. Experimental 2.1. Sampling Independently from the adopted analysis technique, the first problematic aspect of an odour monitoring campaign is sampling. Representative gas samples should be collected with the aim of characterizing the odour emissions associated with all the principal odour sources of the landfill at issue (Bockreis and Steinberg, 2005). For this purpose, specific strategies and suitable sampling equipment are needed (EN 13725, 2003). Sampling on point sources (i.e. conveyed emissions, e.g. through a stack) was carried out by sucking part of the odorous airflow into an 8-L sampling bag in NalophanÔ equipped with a TeflonÔ inlet tube by means of a depression pump. Sampling on passive area sources (i.e. liquid or solid surfaces without an outward flow) was performed using a wind tunnel system (Jiang and Kaye, 2001; Frechen et al., 2004), which consists of a PET hood that is positioned over the emitting surface. A neutral air stream is introduced at known airflow rate from an air bottle into the hood, simulating the wind action on the liquid or solid surface. Air samples are then collected in the outlet duct by means of a pump, which creates vacuum in a case in which the sampling bag is inserted, thereby sucking air into the bag. The wind tunnel used during the experimentation has a circular section inlet and outlet duct, of 0.08 m diameter. The central body of the hood used was a 0.25 m wide, 0.08 m high and 0.5 m deep rectangular section chamber. Inside the inlet duct there is a perforated stainless steel grid and inside the divergent that connects this duct to the central body of the hood there are three flow deflection vanes. Both these devices have the function of making the airflow as homogeneous as possible (Sironi et al., 2006). Gas samples were collected in different positions: - inside the landfill, in correspondence of the principal odour sources: landfill gas extraction wells (on exhausted and active parcels), exhausted landfill surface, waste grinder, oxygenated leachate collection tanks and cogeneration fumes abatement plant; - at the landfill boundaries: cogeneration zone (boundary NW), offices (boundary SW) and landfill entrance (boundary SSE); and Author's personal copy 7052 L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 - outside the landfill, at two receptors where the presence of malodours from the landfill were often lamented: receptor 1, a sports ground located at 2 km SSW from the landfill, and receptor 2, a farm at 1.2 km SW from the landfill. 2.2. Chemical analyses The chemical analyses were carried out with the aim of identifying and quantifying the atmospheric pollutants in some critical points inside and outside the landfill, focusing on the pollutants with a low odour detection threshold, which may originate odour nuisance. The ambient air samples were analyzed by gaschromatography and mass-spectrometry (GC–MS) preceded by cryogenic pre-concentration with liquid nitrogen, according to the methodology EPA TO15, modified in order to determine the concentration of 140 compounds belonging to different pollutant families. The instrument used for the analyses is a GC–MS system model 6890N-5973N (Agilent Technologies) equipped with a cryogenic pre-concentration system TDS2 (Gerstel) and workstation with processing software and spectral database Wiley with 275000 spectra. 2.3. Olfactometric analyses Dynamic olfactometry is a sensorial technique that allows to determine the odour concentration (cod) of an odorous air sample relating to the sensation caused by the sample directly on a panel of opportunely selected people. cod is expressed in European odour units per cubic meter (ouE m3), and it represents the number of dilutions with neutral air that are necessary to bring the odorous sample to its odour detection threshold concentration. The analysis is carried out by presenting the sample to the panel at increasing concentrations by means of a particular dilution device called olfactometer, until the panel members start perceiving an odour that is different from the neutral reference air. The cod is then calculated as the geometric mean of the odour threshold values of each panellist. As defined by the EN 13725, 2003, the individual threshold estimate is defined by the two presentations in one dilution series, sorted on growing odour concentration, where a certain change in response from ‘‘false’’ to a consistently ‘‘true’’ response occurs. The individual threshold estimate is calculated as the geometric mean of the dilution factors of the two defined presentations. An olfactometer ECOMA Mannebeck model TO7, based on the ‘‘yes/no’’ method, was used as a dilution device. This instrument with aluminium casing has 4 panellist places in separate open boxes. Each box is equipped with a stainless steel sniffing port and a push-button for ‘‘yes’’ (odour threshold). The measuring range of the TO7 olfactometer starts from a maximum dilution factor of 1:64,000, with a dilution step factor 2. All the measurements were conducted within 30 h after sampling, relying on a panel composed of 8 panellists (4 þ 4). 2.4. Electronic nose The instruments used for this work (Falasconi et al., 2005) include a pneumatic assembly for dynamic sampling (pump, electro-valve, and electronic flow meter), a thermally controlled sensor chamber with 35 cm3 of internal volume and an electronic board for controlling the sensor operational conditions. Each instrument has been equipped with an array of six thin film MOS (Metal Oxide Semiconductor) sensors, which make the system sensitive to a large spectrum of volatile compounds, and a humidity sensor. In order to use electronic noses for monitoring the ambient air and detecting the odours from the landfill at issue, it is necessary to train the instruments to recognize odours qualitatively, by attributing them to a specific olfactory class, and quantitatively, by estimating the cod of the analyzed air. During the training phase, which is particularly important and delicate, it is necessary to create a complete database that the instrument uses as a reference for the pattern recognition. The training consists in the analysis of different gas samples of known olfactory quality diluted at different cod values (Sironi et al., 2007b). During each monitoring, three electronic noses were used contemporaneously, in order to analyze the ambient air in three different positions. Two instruments were installed at the plant boundaries (respectively at the offices and at the landfill entrance), while the third instrument was installed outside the plant, at receptor 1, located at about 2 km SSW from the landfill. Each monitoring had a duration of 10 days, during which the instruments analyzed the air repeatedly every 15 min. For each of these time intervals, the electronic noses pumped in the external air for three consecutive minutes, and a recovery time of 12 min was left between each measurement. 3. Results 3.1. Determination of chemical composition by GC–MS The analyses by GC–MS allowed to determine the concentration trend of the chemical compounds for each sampling point being considered. As an example, Fig. 1 illustrates the concentration values of the pollutant families determined in the ambient air samples collected near the waste grinder. The results show that the ratio between the different pollutant families remains constant in all monitored points. The relative abundance of the pollutant families has the following trend: Hydrocarbons > oxygenated comp: > halogenated comp: > nitrogenous comp: > sulphured comp: Once the chemical composition of an odorous gas mixture is known, it is possible to estimate its cod based on the analytical concentration of the single odorous compounds and their odour threshold concentration values, according to the following equation (Pagé et al., 2007): cod;OT ¼ N X ci OT i i¼1 Author's personal copy 7053 L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 10,000 1,000 Hydrocarbons Oxygenated comp. Halogenated comp. Nitrogenous comp. Sulphured comp. µg/m3 100 10 1 0.1 I Camp. Feb. 2006 II Camp. May 2006 III Camp. Aug. 2006 IV Camp. Nov. 2006 Fig. 1. Waste grinder: analytical concentration of pollutant families. Where ci is the analytical concentration of the odorous compound i (in mg m3), OTi is the odour threshold concentration of compound i (in mg ou1 E ), N is the number of compounds in the mixture and cod,OT is the cod of the mixture (in ouE m3) calculated from the chemical concentration and the odour threshold of its components, which will be defined as ‘‘theoretical odour concentration’’. As an example, Fig. 2 reports the theoretical odour concentration values associated with the examined pollutant compounds relevant to the samples collected at the waste grinder. The results show that, even though the analytical concentration of the oxygenated compounds is lower than the analytical concentration of the hydrocarbons, the theoretical cod values of both pollutant families are comparable, due to the fact that the odour threshold concentration values relevant to the oxygenated compounds are generally lower than the odour threshold concentration values of the hydrocarbons. These two pollutant families are the ones that mostly contribute to the overall odour mixture. 3.2. Determination of odour concentration by dynamic olfactometry Figs. 3 and 4 show the trends of the cod values of the samples collected on the odour sources inside the landfill and in the ambient air at the boundaries and outside the landfill, respectively. Based on the results of the olfactometric analyses it is possible to make some consideration about the odour emissions associated with the olfactory classes being considered. cod,OT = conc/OT (ouE/m3) 1,000 100 Hydrocarbons Oxygenated comp. Halogenated comp. Nitrogenous comp. Sulphured comp. 10 1 0.1 I Camp. Feb. 2006 II Camp. May 2006 III Camp. Aug 2006 IV Camp. Nov. 2006 Fig. 2. Waste grinder: theoretical odour concentration of pollutant families. Author's personal copy 7054 L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 10,000,000 1,000,000 ouE/m3 100,000 10,000 1,000 100 10 I Camp. Feb. 2006 II Camp. May 2006 III Camp. Aug. 2006 Waste grinder Cog. fumes abatement plant Leachate collection tanks IV Camp. Nov. 2006 Surface of exhausted parcels LFG extraction wells Fig. 3. Geometrical mean of cod values measured in correspondence of the landfill odour sources. The cod values of the landfill gas (LFG) in the different monitoring campaigns are rather stable. A decrease of the cod level is observed in correspondence of the II and part of the III campaign, may be due to the high atmospheric pressure of the period. High atmospheric pressure may cause the infiltration of air inside the body of the landfill, which causes a dilution of the LFG and therefore of its cod. A significant increase of the cod values measured on the surface of the exhausted parcel was observed in correspondence of the III and IV monitoring campaign. This may be due to the fact that in these cases the samples were collected at the base of a LFG extraction well, where probably LFG leaks occurred causing a rise of the cod. On the leachate collection tank, particularly high cod values were measured in correspondence of prolonged periods of drought (II, III an IV campaign), which caused a concentration of the leachate. The highest cod values at the waste grinder were measured in correspondence of the III campaign. This fact can be explained considering that during the summer the waste is conferred to the landfill in an advanced fermentation status and therefore it is characterized by more intense odour emissions. This advanced fermentation level is due to the high temperatures, which accelerate the organic matter degradation processes, and to the higher content of organic matter, such as fruit and vegetable rests, whose consumption is increased during the summer. The cod values of the ambient air at the landfill boundaries and at the receptors were rather low and therefore not particularly relevant, except for a sample collected near the 1,000 ouE/m3 380 100 76 53 42 54 40 28 19 10 I Camp. Feb. 2006 II Camp. May 2006 III Camp. Aug. 2006 IV Camp. Nov. 2006 Fig. 4. cod values of ambient air at landfill boundaries ( ) and at receptors ( ). Author's personal copy L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 landfill entrance during the IV monitoring campaign, which had a cod of 380 ouE m3. 3.3. Continuous ambient air monitoring by electronic nose The results of the ambient air monitoring by electronic nose are represented by large tables that report the olfactory class and the cod value attributed by each instrument to the analyzed air for each measurement carried out during the monitoring period. Table 1 reports the relative recognition frequencies of the considered olfactory classes in the three monitoring positions. As an example, Fig. 5 is a graphical representation of the relative recognition frequencies of the different olfactory classes in the four monitoring campaigns at receptor 1, i.e. the sports ground located at 2 km SSW from the landfill at issue. The graph also reports the sum of the relative odour detection frequencies, i.e. the time percentage in which the electronic nose detected the presence of odours from the landfill in the analyzed ambient air. Based on the results of the ambient air monitoring by electronic noses, it is possible to make some considerations about the landfill odour impact on the three points where the instruments were installed. The high relative recognition frequency of odours attributed to the olfactory class ‘‘LFG’’, allows to affirm that the LFG emitted through the landfill surface or through the not perfectly airtight extraction wells represents the principal odour source of the monitored landfill. During Wednesday, 22nd November 2006, in correspondence of the IV monitoring campaign, the three electronic noses attributed most of the measures to the olfactory class ‘‘LFG’’. This may be due to the fact that during the night between Tuesday, 21st November 2006, and Wednesday, 22nd November 2006, a current drop Table 1 Relative recognition frequencies in the three monitoring positions Analyzed air quality I Camp. II Camp. III Camp. Feb.2006 May 2006 Aug.2006 Relative recognition frequency of the considered olfactory classes the offices (boundary SW) Neutral air 73.0% 95.4% 90.1% LFG 24.1% 4.1% 0.3% Fresh waste 1.4% 0.5% 7.1% Cog. fumes abat. pant – 0.0% 0.0% Leachate 1.5% 0.0% 2.5% IV Camp. Nov.2006 at 77.0% 22.7% 0.0% 0.3% 0.0% Relative recognition frequency of the considered olfactory classes at the landfill entrance (boundary SSE) Neutral air 79.2% 77.5% 44.9% 51.0% LFG 16.4% 20.0% 45.4% 44.1% Fresh waste 3.9% 1.9% 9.7% 3.1% Cog. fumes abat. pant – 0.4% 0.0% 1.6% Leachate 0.5% 0.2% 0.0% 0.2% Relative recognition frequency of the considered olfactory classes at receptor I Neutral air 90.7% 85.7% 91.6% 97.4% LFG 9.0% 10.6% 1.6% 1.4% Fresh waste 0.3% 3.2% 0.8% 0.0% Cog. fumes abat. pant – 0.1% 5.3% 1.1% Leachate 0.0% 0.3% 0.7% 0.0% 7055 caused the interruption of the functioning of the LFG extraction system for several hours. In the four monitoring campaigns the electronic nose installed at receptor 1 registered several episodes during which the odours from the landfill were detected. Nonetheless, the percentage of measures attributed to an olfactory class different from ‘‘neutral air’’ always turned out to be lower than 15%, which is the limit fixed by the German guideline ‘‘GIRL Geruchsimmission-Richtlinie’’ about odour immissions (LAI, 1998) for industrial or agricultural zones. 4. Discussion In this paragraph, the three different monitoring techniques adopted for odour impact assessment are compared. For a more complete evaluation, the meteorological data have also been considered and discussed together with the monitoring results. 4.1. Olfactometric analyses vs. chemical analyses In order to evaluate the results obtained with these two odour characterization techniques, the odour concentration values measured by dynamic olfactometry were compared with the theoretical cod values (see Section 3.1), which are based on the chemical composition of the odorous mixture (Fig. 6). It is possible to observe that there is no precise correlation between the results obtained with these two different odour characterization techniques. The main problem consists in the high degree of inaccuracy that characterizes the evaluation of the theoretical odour concentration of a gas sample based on the analytical concentration and the odour threshold concentration of its components. This may be due to the fact that this evaluation, based on a simple summation of the contribution of each compound to the overall odour of the mixture, doesn’t take account of the synergy and masking effects that occur between the single compounds in a complex mixture, which make their olfactory properties not additive. Moreover, it is important to highlight that the odour threshold concentration values that can be found in literature have unknown accuracy, as the values relevant to a single odorous compound often differ by several orders of magnitude. 4.2. Electronic nose vs. olfactometric analyses The comparison of the results of the olfactometric analyses and the results of the continuous ambient air monitoring campaigns by electronic noses enable the formulation of some considerations relevant to the different monitoring positions. In general, the olfactory class that was recognized more frequently by all the instruments is the olfactory class ‘‘LFG’’. This result is in agreement with the fact that the cod values measured in the LFG samples collected directly at the LFG extraction wells are by several orders of magnitude higher than the cod values relevant to the other monitored odour sources. This observation allows to identify the LFG Author's personal copy 7056 L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 16.0% 15.0% 14.3% 14.0% 13.0% 12.0% 11.0% 10.0% 9.3% 9.0% Landfill LFG Fresh waste Cog. fumes abat. plant Leachate 8.4% 8.0% 7.0% 6.0% 5.0% 4.0% 2.6% 3.0% 2.0% 1.0% 0.0% I Camp. Feb. 2006 II Camp. May 2006 III Camp. Aug. 2006 IV Camp. Nov. 2006 Fig. 5. Relative recognition frequencies of the different olfactory classes at receptor 1. emitted through the landfill surface or through the not perfectly airtight LFG extraction wells as the most critical odour source of the monitored landfill. An exception is represented by the results of the ambient air monitoring by electronic nose conducted at the offices in the month of August, during which the electronic nose attributed the majority of the odour detections to the olfactory class ‘‘fresh waste’’. The offices are located in proximity of the waste grinder, where the fresh waste is conferred, and the high relative detection frequency may be justified observing that in correspondence of the III monitoring campaign the ambient air samples collected near the waste grinder presented particularly high cod values (320 ouE m3 and 710 ouE m3, respectively). Another exception is represented by the monitoring at receptor 1 in the month of August, during which the majority of the odour detections where attributed to the olfactory class ‘‘cogeneration fumes abatement plant’’. These results are in agreement with the particularly high cod values that were measured in the samples collected at the stack through which the gases coming out from the cogeneration fumes abatement plant are emitted into the atmosphere in correspondence of the III monitoring campaign (4000 ouE m3 and 16,000 ouE m3, respectively). The monitoring campaign during which the odours from the landfill were detected at receptor 1 for the highest time percentage is the second one, conducted in the month of May. This is not justified by the determination of particularly high cod values in the samples collected in correspondence of the landfill principal odour sources. For this reason, it is reasonable to assume that the more frequent detection of odours from the landfill at receptor 1 500 450 400 ouE/m3 350 300 cod-OT cod-Olf 250 200 150 100 50 Waste grinder Landfill entrance Receptor 1 Receptor 2 Waste grinder Landfill entrance Receptor 1 Receptor 2 Waste grinder Landfill entrance Receptor 1 Receptor 2 Waste grinder Landfill entrance Receptor 1 Receptor 2 0 I Camp. Feb. 2006 II Camp. May 2006 III Camp. Aug. 2006 IV Camp. Nov. 2006 Fig. 6. Comparison between cod values measured by dynamic olfactometry and theoretical odour concentration values. Author's personal copy L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 is not due to an increase of the landfill odour emissions, but rather to meteorological factors. The landfill entrance is the monitoring position in which, in all four monitoring campaigns, the highest number of odour detections was registered, i.e. the highest number of measures attributed to one of the olfactory classes relevant to the landfill odour emissions. These results clearly indicate that the landfill entrance represents a critical zone as far as the presence of odours is concerned. 4.3. Electronic nose vs. meteorological data The in depth study of meteorological data can be in general very useful in order to evaluate the dispersion of odours emitted from the monitored landfill on the surrounding territory. In this case, a rough analysis of the meteorological data relevant to the four monitoring campaigns (especially wind speed and wind direction) was performed (Fig. 7), just with the aim of comparing them with the results of the ambient air monitoring by electronic noses, in order to look for possible connections between meteorological conditions and odour detections outside the landfill. It can be noticed that the highest odour detection percentage at receptor 1 was registered during the II monitoring campaign, which was characterized by unstable meteorological conditions, i.e. weak wind with unstable direction. On the contrary, during the remaining three monitoring campaigns, characterized by the clear presence of a predominant wind direction, the odour detection frequency at the same receptor was significantly lower. These results show that, in this case, the presence of unstable meteorological conditions is particularly unfavourable to the dispersion of odours into the atmosphere. 5. Conclusions and future work This study represents a critical review of employing three different odour evaluation techniques for the assessment of odour emissions from a landfill site. The discussion of the results achieved by the parallel application of GC–MS, dynamic olfactometry and electronic nose technology effectively show how these three odour characterization methods can be effectively used, for differing reasons, on the same site. Although there is no evidence of correlation between the chemical composition of the analyzed samples and their cod measured by dynamic olfactometry, chemical analyses can be useful to analyze odour composition, in order to design intervention and treatment strategies. The results of the olfactometric analyses enabled a quantification of the landfill odour emissions, giving indicative values of sensory impacts that are liable to affect off-site. Finally, electronic nose technology could be effectively used as a management tool in order to monitor site changes or operational failures. In this specific case, the results of the ambient air monitoring by electronic noses allowed to quantify the time percentage in which the presence of odours is perceived at the landfill boundaries and at a receptor located at 2 km from the landfill, which always turned out to be lower than the limit of 15% fixed by the German guideline ‘‘GIRL Geruchsimmission-Richtlinie’’ about odour immissions (LAI, 1998) for industrial or agricultural zones in all the four monitoring campaigns. Moreover, the analysis of the meteorological data relevant to the four monitoring campaigns, conducted in four different seasons (winter, spring, summer and fall) allowed to highlight that the presence of unstable meteorological conditions (i.e. weak wind with unstable direction) is N NNW 35.0% NNE I Campaign - February 2006 II Campaign - May 2006 III Campaign - August 2006 IV Campaign - November 2006 30.0% NW NE 25.0% 20.0% 15.0% WNW ENE 10.0% 5.0% 0.0% W E WSW ESE SE SW SSW 7057 Wind calms (v<0.1 m/s): I Campaign: 2.2% II Campaign: 4.1% III Campaign: 2.5% IV Campaign: 9.4% SSE S Fig. 7. Wind roses relevant to the four monitoring campaigns. Author's personal copy 7058 L. Capelli et al. / Atmospheric Environment 42 (2008) 7050–7058 particularly unfavourable to the dispersion of odours into the atmosphere. As a conclusion, this study shows that whilst the results of the three different odour characterization techniques don’t necessarily correlate, they do have an intrinsic value, thus demonstrating the complexity of odour measurement in the environment. References Ampuero, S., Bosset, J.O., 2003. The electronic nose applied to dairy products: a review. Sensors and Actuators B, Chemical 94, 1–12. Bockreis, A., Steinberg, I., 2005. Measurement of odour with focus on sampling techniques. Waste Management 25, 859–863. Davoli, E., Gangai, M.L., Morselli, L., Tonelli, D., 2003. Characterisation of odorants emissions from landfills by SPME and GC/MS. Chemosphere 51, 357–368. El Fadel, M., Findikakis, A.N., Leckie, J.O., 1997. Environmental impacts of solid waste landfilling. Journal of Environmental Management 50, 1–25. EN 13725, 2003. Air Quality – Determination of Odour Concentration by Dynamic Olfactometry. Comité Européen de Normalisation, Brussels, pp. 1–70. Falasconi, M., Pardo, M., Sberveglieri, G., Riccò, I., Bresciani, A., 2005. The novel EOS835 electronic nose and data analysis for evaluating coffee ripening. Sensors and Actuators B, Chemical 110, 73–80. Frechen, F.-B., Frey, M., Wett, M., Löser, C., 2004. Aerodynamic performance of a low-speed wind tunnel. Water Science and Technology 50 (4), 57–64. Gostelow, P., Parsons, S.A., Stuetz, R.M., 2001. Odour measurements for sewage treatment works. Water Research 35, 579–597. Jiang, J.K., Kaye, R., 2001. Sampling techniques for odour measurement. In: Stuetz, Richard, Frechen, Franz-Bernd (Eds.), Odours in Wastewater Treatment: Measurement, Modelling and Control. IWA Publishing, London, UK, pp. 98–117. LAI (Länderausschuss für Immissionsschutz), 1998. GeruchsimmissionsRichtlinie (GIRL – Odour Emissions Guidelines), Ministry of Environment, Environmental Planning, and Agriculture of the State Nordrhein-Westfalen, Germany. Micone, P.G., Guy, C., 2007. Odour quantification by a sensor array: an application to landfill gas odours from two different municipal waste treatment works. Sensors and Actuators B, Chemical 120, 628–637. Pagé, T., Narjoux, A., Guy, C., Caron, R.F., Fécil, B., 2007. Impact evaluation of landfills on air quality: odours and toxic compounds. In: Cossu, R., Diaz, L.F., Stegmann, R. (Eds.), Proceedings Sardinia 2007. CISA Pub., Cagliari, Italy, pp. 215–216. Pearce, T.C., 1997. Computational parallels between the biological olfactory pathway and its analogue ‘The Electronic Nose’: part II. Sensor-based machine olfaction. Biosystems 41, 69–90. Persaud, K., Dodd, G.H., 1982. Analysis of discrimination mechanisms in the mammalian olfactory system using a model nose. Nature 299, 352–355. Schaller, E., Bosset, J.O., Escher, F., 1998. Electronic Noses’ and their application to food. Lebensmittel-Wissenschaft und -Technologie 31, 305–316. Schlegelmilch, M., Streese, J., Stegmann, R., 2005. Odour management and treatment technologies: an overview. Waste Management 25, 928–939. Sironi, S., Capelli, L., Céntola, P., Del Rosso, R., Il Grande, M., 2005. Odour emission factors for the assessment and prediction of Italian MSW landfills odour impact. Atmospheric Environment 39, 5387–5394. Sironi, S., Capelli, L., Céntola, P., Del Rosso, R., Il Grande, M., 2006. Odour emission factors for the prediction of odour emissions from plants for the mechanical and biological treatment of MSW. Atmospheric Environment 40, 7632–7643. Sironi, S., Capelli, L., Céntola, P., Del Rosso, R., Il Grande, M., 2007a. Continuous monitoring of odours from a composting plant using electronic noses. Waste Management 27, 389–397. Sironi, S., Capelli, L., Céntola, P., Del Rosso, R., 2007b. Development of a system for the continuous monitoring of odours from a composting plant: focus on training, data processing and results validation methods. Sensors and Actuators B, Chemical 124, 336–346. Stuetz, R.M., Fenner, R.A., Engin, G., 1999. Assessment of odours from sewage treatment works by an electronic nose, H2S analysis, and olfactometry. Water Research 33, 453–461.