Comparison of different approaches for odour impact assessment: Dispersion modelling (CALPUFF) vs field inspection (CEN/TC 264)

Provided for non-commercial research and educational use only. Not for reproduction or distribution or commercial use. This article was originally published by IWA Publishing. IWA Publishing recognizes the retention of the right by the author(s) to photocopy or make single electronic copies of the paper for their own personal use, including for their own classroom use, or the personal use of colleagues, provided the copies are not offered for sale and are not distributed in a systematic way outside of their employing institution. Please note that you are not permitted to post the IWA Publishing PDF version of your paper on your own website or your institution’s website or repository. Please direct any queries regarding use or permissions to wst@iwap.co.uk © IWA Publishing 2013 Water Science & Technology 1731 | 68.8 | 2013 Comparison of different approaches for odour impact assessment: dispersion modelling (CALPUFF) vs field inspection (CEN/TC 264) Licinia Dentoni, Laura Capelli, Selena Sironi, Jean-Michel Guillot and Andrea N. Rossi ABSTRACT Odour impact assessment has become an important environmental issue. Different approaches can be used in order to evaluate the odour impact on receptors, and therefore to regulate it. Among the different possible regulation approaches, the use of dispersion modelling is suggested or required by several national or regional legislations. The wide diffusion of this approach is probably due to the fact that odour dispersion modelling is relatively cheap and results are easily understandable. Another kind of approach attempts to evaluate the odour impact directly in the field relying on a panel of trained human assessors (field inspection). The growing importance of this odour impact assessment method is proved by the current draft of a European Standard (CEN/TC 264), which defines two different methodologies of field inspection: grid measurement and plume measurement. In this study two different approaches were compared, i.e. odour dispersion modelling and field inspection by plume measurement (with specific adaptation for the studied site), the latter consisting in using a panel of examiners for determining the absence or presence of odour downwind relative to Licinia Dentoni Laura Capelli (corresponding author) Selena Sironi Politecnico di Milano, Department of Chemistry, Materials and Chemical Engineering ‘Giulio Natta’, Piazza Leonardo da Vinci 32, 20133 Milano, Italy E-mail: laura.capelli@polimi.it Jean-Michel Guillot Ecole des Mines d’Alès, Industrial Environment, 6 av. De Clavières, 30319 Alès cedex, France Andrea N. Rossi Progress S.r.l., via N.A. Porpora 147, 20131 Milano, Italy the source, in order to evaluate the plume extent. The comparison was based on application of both methods to the assessment of the odour impact of a plant for the composting of sludge from an Italian food industry. The results show that the odour impacts assessed by the two strategies turned out to be quite comparable, thus indicating that, if opportunely applied, both approaches may be effective and complementary for odour impact assessment purposes. Key words | atmospheric dispersion simulation, field measurements, human panellists, odour dispersion, olfactometry INTRODUCTION For several decades it has been known that the odours resulting directly or indirectly from human activities may cause adverse effects on citizens (Sucker et al. ; Aatamila et al. ), and have recently been considered as atmospheric contaminants. It is important to highlight that odours are, among atmospheric pollutants, the major cause of the population’s complaints to local authorities (Henshaw et al. ). For this reason, odours are nowadays subject to control and regulation in many countries (Nicell ). The need to regulate odour impacts entails the need for specific methods of odour measurement. Dynamic olfactometry (CEN ) is now a widespread and common technique for the doi: 10.2166/wst.2013.387 quantification of odour emissions (Muñoz et al. ). However, besides source characterization, it is important to evaluate the effective impact of odour on citizens. Different approaches can be used in order to evaluate the odour impact on receptors, and therefore to regulate it. Odour dispersion modelling is commonly applied to simulate how odour disperses into the atmosphere, and therefore to calculate ground odour concentration values in the simulation space-time domain (Capelli et al. ). This approach may be useful for the definition of specific odour regulation. In general, this kind of regulation defines a maximum impact criterion expressed as an odour concentration limit that has not to be exceeded for a certain period L. Dentoni et al. 1732 | Comparison of different approaches for odour impact assessment at receptors ( JORF ; Regione Lombardia ). The wide diffusion of this approach is due on one hand to the fact that odour dispersion modelling is relatively cheap and results are easily understandable. On the other hand, another compelling reason for the use of odour dispersion modelling is the ability to apply it in advance of construction of a new plant, thus being ‘predictive’, and not solely ‘descriptive’, as field measurement are. Nonetheless, besides dispersion modelling, direct odour measurement in the field (i.e. ‘field inspection’) is still an important approach for odour impact assessment. This ‘descriptive’ way of assessing odours, which relies on a panel of trained human assessors (Nicolas et al. ), entails the advantage of allowing the direct determination of ambient air concentration close to the odour detection threshold. The growing importance of this odour impact assessment method is proved by the current draft of a European Standard (CEN/TC 264). More specifically, this standard will include two methods of field inspection: grid and plume methods (Guillot et al. ). The grid method is a long period (1 year) statistical survey method to obtain a representative map of a recognizable odour exposure over a selected area, whereas the plume method is a short period experiment (several times of approximately half a day under meteorological conditions) to determine the extent of recognizable odour from a specific source. Both methods are based on odour detection and recognition by human panellists. In this study the two above described approaches were compared, i.e. odour dispersion modelling (CALPUFF), and field inspection, the latter by adaptation of the plume measurement method. Indeed, even though plume measurement is well described in the standard (Guillot et al. ), Figure 1 | Considered emission sources. Water Science & Technology | 68.8 | 2013 specific adaptations had to be proposed to make it applicable to the studied site, thus allowing the panel of examiners to determine the absence or presence of odour downwind relative to the source, in order to evaluate the plume extent. The comparison was based on application of both approaches (calculation by the model and field measurement) to the assessment of the odour impact of a plant for the composting of sludge from an Italian food industry. METHODS Site description The studied plant performs the composting of agro-alimentary sludge produced by a food industry for the production of cheese located in Northern Italy. The studied emission sources (Figure 1) were the composting sludge heaps and the final product storage. The composting sludge heaps have an emitting surface of about 1,600 m2, whereas the final product storage surface is about 320 m2. Frequent citizens’ complaints about odour nuisances coming from this site in the surrounding area have been reported to local authorities. Sampling and analysis Two odour sampling and measurement trials were conducted in order to characterize the emissions. Specifically, the composting sludge heaps emissions were evaluated in different periods of the composting process. Obtained odour concentration values were than averaged to obtain a 1733 L. Dentoni et al. | Comparison of different approaches for odour impact assessment representative value to be used for the odour dispersion modelling. Sampling on odour sources was performed using a wind tunnel system, which consists of a hood that simulates the wind action on the surface to be monitored. In this case, a specific wind tunnel made in polyethyleneterephthalate (PET) (Capelli et al. ) was used by positioning it over the emitting surface. A neutral air stream is introduced at known airflow rate from an air bottle into the hood. Air samples are then collected in the outlet duct using a depression pump. All collected samples were analysed by dynamic olfactometry, which is a sensorial technique that allows determination of 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. Odour concentration cod is expressed in European odour units per cubic metre (ouE m 3), and it represents the number of dilutions with neutral air that are necessary to bring the odorous sample to its odour detection threshold concentration (CEN ). An olfactometer model TO8 produced by ECOMA GmbH, based on the ‘yes/no’ method, was used as a dilution device. All the measurements were conducted within 30 h after sampling, relying on a panel composed of four panellists. Dispersion modelling The dispersion of emissions is determined by an atmospheric dispersion model, which calculates the pollutant concentration in ambient air at the ground level, by processing emission data, meteorological data and terrain profile. The model used is the CALPUFF model (Wang et al. ) and the data pre-processing and post-processing are realized by means of specific software developed at the Department of Chemistry, Materials and Chemical Engineering ‘Giulio Natta’ of the Politecnico di Milano. The dimensions of the spatial grid considered as the simulation domain are 4,000 m × 4,000 m, with a receptor every 100 m. The meteorological data used for the dispersion modelling are meteorological parameters of 1 complete year (2011) listed in Table 1, whereas the output parameters of the pre-processor used for the calculation of the micrometeorological variables are listed in Table 2. As emission data, the results of the olfactometric analyses conducted on the plant can be used. The data needed as input for the model are not the odour concentration values, but the Odour Emission Rate (OER) values, expressed in ouE s 1, Table 1 Water Science & Technology | | 68.8 | 2013 Meteorological parameters used for the dispersion modelling Meteorological parameter Type of data Unit of measurement Temperature Hourly average W Wind speed Hourly average ms C 1 Wind direction Predominant (1 h) Sexagesimal degree Net solar radiation Hourly average Wm Relative humidity Hourly average % Rainfall Total (1 h) mm Table 2 | 2 Parameters calculated by the micrometeorological model Micrometeorological parameter Symbol Type of data Calculation method Surface heat flux Qh Hourly average Thomson () Friction velocity u* Hourly average Monin–Obukhov length LMO Hourly average Convective velocity scale w* Hourly average Mixing height MH Hourly average Scire et al. () associated with each considered odour source. The evaluation of the OER relevant to considered emission sources requires the calculation of the Specific Odour Emission Rate (SOER), which is expressed in ouE s 1 m 2. Once the odour concentration value of a sample collected at the outlet of the wind tunnel is determined, it is possible to obtain the SOER by multiplying the odour concentration (ouE m 3) by the air flow rate over the surface (m3 s 1) and dividing by the base area of the central body of the hood (m2). The OER is finally obtained as the product of the SOER value and the emitting surface of the considered source (m2) (Sironi et al. ). Table 3 reports the characteristics of the considered odour sources. Based on the previous considerations, the OER is a function of the air flow rate over the surface, i.e. of the wind speed. Once the OER relevant to the sampling conditions is evaluated, the OER for any other air velocity (i.e. wind Table 3 | Odour sources considered W Section cod av OER at 20 C Emission ( C) (m2) (ouE/m3) (ouE/s) Composting 35 1,595 5,900 15,057 Storage 20 321 1,600 822 Temp. W L. Dentoni et al. 1734 | Comparison of different approaches for odour impact assessment speed) can be calculated for each hour of the simulation domain based on the current wind speed, according to the following expression (Sohn et al. ; Capelli et al. ): OERv2 ¼ OERv1  1=2 v2 v1 The OER values reported in Table 3 are referred to an air velocity of 0.01 m s 1. Besides the simulation of odour dispersion with the data of 1 complete year (2011) required for odour impact assessment purposes, the dispersion model was also applied in order to simulate the odour dispersion relevant to the hours during which the field inspection took place. Field inspection In order to determine the odour in the field by using human assessors, two methods can be used: the grid and the plume method. In both cases, human panel members characterize an area by the presence or absence of an odour. The first method (grid method) uses direct assessment of the ambient air by panellists to characterize odour exposure in a defined assessment area. The second method (plume method) uses panellists to determine the extent of the downwind odour plume under defined meteorological conditions, because the extent from a source is variable as a function of dispersion conditions. With the plume method the presence or absence of recognizable odours in and around the plume originating from a specific odour emission source, under a specified emission situation and meteorological conditions is determined. Figure 2 | Dynamic (left) and static (right) plume measurement. Water Science & Technology | 68.8 | 2013 The extent of the plume is assessed as the transition of absence to presence of recognizable odour. In this study we decided to use the plume method, by introducing some specific adaptations to make it suitable for the studied site. The plume measurement by field inspection can be performed in two different ways. The panel may zigzag and traverse the plume (dynamic method) or be located at specific points on perpendicular axes (static method) (Figure 2). The first intention of the study was to use the dynamic plume measurement method. Nonetheless, the dynamic method had to be re-adapted considering the studied site. The panel couldn’t really zig-zag to approach the plant due to the presence of 2 m height corn plantations they could hardly have walked through. For this reason, odour measurements could only take place on pedestrian paths. So, starting far away from the source and walking in the direction of the plant on such paths, panellists were asked to stop when odour was detected and to stay at this point to estimate exposure. As provided by the draft European Standard about field inspection, exposure estimation at each point lasted 10 min, during which, the panellists had to register on a specific form whether they perceived odour or not every 10 s, thus giving a total of 60 observations for each panellist at each point. A preliminary study was performed in order to set up the plume measurement. This study involved the simulation of the odour dispersion at different times using the historical meteorological data relevant to the same month in which the field inspection should be run. This allowed planning of the field inspection by evaluating the expected extension and direction of the odour plume as a function of the time of the L. Dentoni et al. 1735 | Comparison of different approaches for odour impact assessment day. In addition, the area to be studied was mapped in a detailed way, in order to identify the paths around the source that could be used by the panellists for the field inspection. The field inspection planning also involved the training of the panel to recognize the characteristic odour of the considered source. The field inspection took place in two different evenings, with different meteorological conditions. For each measurement five panellists divided into two groups were involved. More in detail, each group (A and B) was asked to go along the different paths identified during the preliminary study towards the odour emission and to indicate the point at which they started to perceive the characteristic odour from the plant on each path. The paths identified to be covered by the two groups of panellists (A and B) during the two field measurements are shown in Figure 3. So, each path was covered twice (once by group A and once by group B) the same day. The groups were present on the same path approximately with 80 min of difference in time. Globally, it took about 2.5 hours for the two groups to complete the field inspection. Water Science & Technology | 68.8 | 2013 map of the 98 percentile of the peak odour concentration values (Figure 4). As shown in Figure 4, the odour impact of the considered plant is considerable and, due to the absence of a prevalent direction of the wind; the entire considered domain could be affected by the presence of the odour coming from the composting plant. W Field measurements The simulation of the odour dispersion at different times using the historical meteorological data relevant to the same month in which the field inspection should be run showed that the most suitable timings to perform the field inspection (i.e. timings with major odour plume extent) are the early morning and the late afternoon. Moreover, as there is no prevalent wind direction expected, the field inspection has to be performed all around the area surrounding the plant, without any preferential direction. For this reason, the pedestrian paths for the field inspection were identified all around the plant, as shown in Figure 3. The results of the two field inspection campaigns (I and II) are shown on the left side of Figures 5 and 6, RESULTS AND DISCUSSION Modelled odour impact The plant overall odour impact obtained using the meteorological data of 1 complete year (2011) is represented by the Figure 4 Figure 3 | Paths identified for field inspection. | Map of the 98th percentile of the peak odour concentration values on a yearly basis (2011). L. Dentoni et al. 1736 | Comparison of different approaches for odour impact assessment Water Science & Technology Figure 5 | Odour impact relevant to the I campaign: assessed by field inspection (left) vs. simulated by odour dispersion modelling (right). Figure 6 | Odour impact relevant to the II campaign: assessed by field inspection (left) vs. simulated by odour dispersion modelling (right). respectively. More in detail, the maps report, for both campaigns, and for each panel group (A and B, in different colours on the maps) the points on the different paths where the odour started to be perceived. | 68.8 | 2013 Comparison between modelled odour impact and odour impact obtained by field inspection The right side of Figures 5 and 6 show the odour impact simulated by running the odour dispersion model at the 1737 L. Dentoni et al. | Comparison of different approaches for odour impact assessment same time as the I and the II field inspection campaigns took place, i.e. with the meteorological conditions that were present at the time of the field measurement, in order to make it possible to compare the two assessment approaches. In order to compare the dispersion modelling results with the outcomes of the field inspection, the lines connecting the points where the odour started to be perceived during the field measurements should be compared with the modelled iso-concentration lines. For this purpose, it is important to consider that the points resulting from the field inspection are in fact the points at which the panellists did not only perceive the odour, but where they were able to recognize the odour as coming from the studied site, which they were previously trained to identify. For this reason, the lines resulting from the field inspection (Figures 5 and 6, left) should not be compared with the iso-concentration line of 1 ouE m 3, which corresponds by definition to the odour detection threshold concentration, but rather with the isoconcentration line corresponding to the odour recognition threshold concentration, which is assumed to be about 3 ouE m 3. Thus, the odour impact assessed with the two different approaches seems to be quite comparable. Nonetheless, a slight overestimation of the odour impact may result from dispersion modelling, due to the experimental evidence of overestimation of odour impacts obtained when CALPUFF is applied to area sources. Moreover, based on the experience of our laboratory, in some cases, especially when dealing with odour emissions having an organic matrix located in an agricultural context, high odour concentrations are measured at the source, but odour becomes less perceivable when moving away from the source, probably because it tends to be confused with the background odour. In addition, the different shapes of the odour impacts may be due to an essential difference between the two compared approaches: while the odour dispersion modelling uses the hourly prevalent wind direction, the field measurements are affected by instantaneous wind direction changes. Better correspondence between the two assessment approaches may be possibly obtained by choosing a shorter averaging time for dispersion modelling (<1 h). CONCLUSIONS The main purpose of this work was the comparison of two different approaches that can be used for odour impact assessment. Water Science & Technology | 68.8 | 2013 The odour impacts resulting from application of the two strategies turned out to be quite comparable, thus indicating that both approaches may be effective and complementary for odour impact assessment purposes. The proposed adaptation of the dynamic plume method, where zig-zags of panellists to cross an odour plume are physically impossible, proved to be effective in order to obtain significant data on odour exposure for comparison with dispersion modelling. Better correspondence between the two assessment approaches (i.e. modelling vs. field inspection) could probably be obtained by choosing a shorter averaging time for dispersion modelling (<1 h), considering that field measurements are affected by instantaneous changes in the meteorological conditions. Of course, further studies will be necessary in order to consolidate the experience with field inspections and to verify these preliminary results. Nonetheless, it is important to highlight that odour dispersion modelling entails the possibility to be applied in advance of the construction of a new plant, thus allowing us to conduct ‘what if’ sensitivity analysis to define the appropriate controls systems, as well as the most beneficial configuration of a planned facility from the perspective of minimizing odour emissions, whereas field odour assessment methods alone can never fulfil this planning and design function. ACKNOWLEDGEMENTS The authors thank the panellists for the work and the Galilée (French–Italian) programme which allowed co-operation between the research groups. REFERENCES Aatamila, M., Verkasalo, P. K., Korhonen, M. J., Suominen, A. L., Hirvonen, M. R., Viluksela, N. K. & Nevalainen, A.  Odour annoyance and physical symptoms among residents living near waste treatment centres. Environmental Research 111, 164–170. Capelli, L., Sironi, S., Del Rosso, R. & Céntola, P.  Design and validation of a wind tunnel system for odour sampling on liquid area sources. Water Science and Technology 59 (8), 1611–1620. Capelli, L., Sironi, S., Del Rosso, R., Céntola, P., Rossi, A. & Austeri, C.  Olfactometric approach for the evaluation of citizens’ exposure to industrial emissions in the city of Terni, Italy. Science of the Total Environment 409, 595–603. 1738 L. Dentoni et al. | Comparison of different approaches for odour impact assessment CEN  EN 13725:2003. Air Quality – Determination of Odour Concentration by Dynamic Olfactometry. Brussels. Guillot, J. M., Bilsen, I., Both, R., Hangartner, M., Kost, W. J., Kunz, W., Nicolas, J., Oxbol, A., Secanella, J., Van Belois, H., Van Elst, T., Van Harreveld, T. & Milan, B.  The future European standard to determine odour in ambient air by using field inspection. Water Science and Technology 66, 1691–1698. Henshaw, P., Nicell, J. & Sikdar, A.  Parameters for the assessment of odour impacts on communities. Atmospheric Environment 40, 1016–1029. Journal Officiel de la République Française (JORF)  Arrêté du 22 avril 2008 fixant les règles techniques auxquelles doivent satisfaire les installations de compostage ou de stabilisation biologique aérobie soumises à autorisation en application du titre Ier du livre V du code de l’environnement (Order of 22 April 2008 laying down the technical rules that must be met by composting facilities or aerobic biological stabilization subject to authorization under Title I of Book V of the Environmental Code). JORF n 0114 du 17 mai 2008. Muñoz, R., Sivret, E. C., Parcsi, G., Lebrero, R., Wang, X., Suffet, I. H. & Stuetz, R. M.  Monitoring techniques for odour abatement assessment. Water Research 44, 5129–5149. Nicell, J. A.  Assessment and regulation of odour impacts. Atmospheric Environment 43, 196–206. Nicolas, J., Craffe, F. & Romain, A. C.  Estimation of odor emission rate from landfill areas using the sniffing team method. Waste Management 26, 1259–1269. W Water Science & Technology | 68.8 | 2013 Regione Lombardia  D.G.R. 15 febbraio 2012 – n. IX/3018. Determinazioni generali in merito alla caratterizzazione delle emissioni gassose in atmosfera derivanti da attività a forte impatto odorigeno (General determinations concerning the characterization of gas emissions into the atmosphere from high-impact odorous activities). Bollettino Ufficiale 20 febbraio 2012, pp. 20–49. Scire, J. S., Robe, F. R., Fernau, M. E. & Yamartino, R. J.  A User’s Guide for the CALMET Meteorological Model. Version 5, Earth Tech Inc., Concord, MA. Sironi, S., Capelli, L., Céntola, P., Del Rosso, R. & Il Grande, M.  Odour emission factors for the prediction of odour emissions from plants for the mechanical and biological treatment of MSW. Atmospheric Environment 40, 7632–7643. Sohn, J. H., Smith, R., Yoong, E., Leis, J. & Galvin, G.  Quantification of odours from piggery effluent ponds using an electronic nose and an artificial neural network. Biosystems Engineering 86, 399–410. Sucker, K., Both, R. & Winneke, G.  Review of adverse effects of odours in field studies. Water Science and Technology 59, 1281–1289. Thomson, D. J.  The Met Input Module ADMS 3 Documentation Paper Number P/05/01/00. The Met Office, Exeter, Devon, UK. Wang, L., Parker, D. B., Parnell, C. B., Lacey, R. E. & Shaw, B. W.  Comparison of CALPUFF and ISCST3 models for predicting downwind odor and source emission rates. Atmospheric Environment 40, 4663–4669. First received 20 April 2013; accepted in revised form 20 May 2013