142
© IWA Publishing 2014 Water Science & Technology
|
69.1
|
2014
Comparison of two standard odor intensity evaluation methods for odor problems in air or water Jane Curren, Cherie L. (Cher) Snyder, Samantha Abraham and I. H. (Mel) Suffet
ABSTRACT Government agencies responsible for ensuring healthful water and/or air quality are often faced with resolving public complaints of nuisance odors. Understanding variations in odor intensity may ultimately lead to the establishment and application by such agencies of quantitative limits for effective odorant control. An odor panel was trained in suprathreshold odor intensity evaluation using both the ASTM Method E544 (Butanol Method) and the APHA Method 2170 (Flavor Profile Analysis (FPA) Method). A linear mixed model was fitted to the panel data, taking into account the fixed effects of concentration levels and the random effects of panelists and sessions. The FPA method proved easier to administer and revealed less inter-session variance than the ASTM Method, suggesting its greater utility in applications involving odor panels. For both methods, there was a high standard deviation, relative to the mean. This finding indicates that the intensity scales may be useful
Jane Curren (corresponding author) Samantha Abraham I. H. (Mel) Suffet Environmental Science and Engineering Program and Department of Environmental Health Sciences, School of Public Health, University of California, Los Angeles, CA 90095, USA E-mail: [email protected] Cherie L. (Cher) Snyder South Coast Air Quality Management District, 21865 Copley Dr, Diamond Bar, CA 91765, USA
for understanding relative odor intensities, but should not be used as a precise measure, or as a basis for establishing regulatory limits. Key words
| butanol, FPA Method, intensity method, nuisance, odor intensity method, odor intensity scales
INTRODUCTION The ability to detect and perceive odorous compounds appears to vary not only with the sensory acuity and experience of the observer, but also with the intensity, character, and hedonic value of the odorant itself, as well as the context in which the odorant is presented. For example:
• • • •
Those who are anosmic to one or more odor types may be up to a thousand times less sensitive to a given odor than others with average odor detection capabilities (Amoore ). Saturation of the odor receptors due to prolonged exposure to an odorant can result in odor fatigue such that the odor becomes less noticeable or may cease to be noticed. Formal training or experience in something such as wine tasting can result in improved chemosensory ability (Lawless ). Environmental cues may affect perceived odor intensity. The reported intensity of odors caused by cigarette
doi: 10.2166/wst.2013.567
smoke increased when the smoker was in the sight of the observer (Moschandreas & Relwani ). Both et al. () determined that once an odor becomes recognizable, it has the potential to cause a nuisance depending on hedonic tone and odor character. Government agencies responsible for ensuring healthful water and/or air quality are often faced with resolving public complaints of nuisance odors. Understanding the variation in odor intensity may help inform the development of reliable methods to rate the perceived strength of an odorant and to establish and apply quantitative limits for odorant control. The standard method for evaluating odor intensity in air is ‘Standard Practices for Referencing Suprathreshold Odor Intensity,’ also known as the Butanol Method (ASTM Standard Test Method , published ). This method measures the ‘total odor’ intensity of all combined odor characteristics by comparing a sample odor intensity to the intensity of standard solutions of n-butanol in either air or water.
143
J. Curren et al.
|
Comparison of two standard odor intensity evaluation methods
A parallel method for evaluating odor intensity in water is the ‘Standard Methods for the Evaluation of Water and Wastewater Method 2170,’ also known as the Flavor Profile Analysis (FPA) Method (American Public Health Association et al., Method 2170, revised ), used as a troubleshooting tool since the 1980s in the drinking water industry. Odor problems are identified by odor character using the ‘drinking water wheel’ and the intensity of each odor character is established using a seven-point intensity scale to measure the strength of individual flavor and odor characters (Suffet et al. ). Four of the intensity points are anchored to flavor standards, typically solutions of sugar (sucrose) in water. Other scales using verbal descriptors (e.g., weak, mild, strong) and/or numeric rankings of odor intensity typically require less training than the use of methods involving sensory intensity anchors, and do not lead to olfactory fatigue. The lack of equal intervals between adjacent points on these ordinal scales, however, limits their suitability for statistical analysis (Lawless & Heymann ), making the use of the Butanol and FPA methods preferable from the standpoint of data reduction. The FPA and Butanol methods each require the use of an odor panel of at least four or eight people, respectively, to reduce response variability (APHA et al. ). While methods for panel selection and training in the use of each method are available in the literature (Meng & Suffet ; ASTM ), they are not outlined by either method. The objective of this study is to evaluate and compare the Butanol and FPA methods with respect to the precision of the data provided by their use.
METHOD Odor intensity evaluation Two odor intensity scaling methods were evaluated in this study:
•
ASTM Method E544 ‘Standard Practices for Referencing Suprathreshold Odor Intensity (ASTM )’. ASTM Method E544, known informally as the ‘Butanol method’, measures perceived odor intensity when participants compare samples to standard solutions of n-butanol (99.9% pure) in either air or water. An 11-point scale corresponded to concentrations of 0, 2.5, 5, 10, 20, 40, 80, 160, 320, 640, and 1,280 mg/L of n-butanol in odor-free water prepared freshly from n-butanol before each panel session.
Water Science & Technology
|
69.1
|
2014
Panelists rated odor intensity by selecting a scaled value (or a point between two such values) corresponding to the equivalent concentration of n-butanol. The rated odor intensity was directly proportional to the concentration of the combined odor-causing compounds present in the sample.
•
American Public Health Association Method 2170 ‘Flavor Profile Analysis’, in ‘Standard Methods for the Evaluation of Water and Wastewater (APHA et al. )’. The FPA Method uses a seven-point intensity scale to measure the intensity of individual flavor and odor characters in drinking water (APHA et al. ), as shown in Table 1 below. Three of the intensity points were anchored to flavor standards; in this case, to concentrations of sugar (sucrose) in water. Panelists based their evaluations of sample odor intensity on the perceived intensity of the flavor of the sugar solutions.
Both the Butanol and FPA Odor Intensity Scaling Methods were used to evaluate two concentration series of butyric acid (500, 1,500, 5,000, and 10,000 μg/L) and dimethyl sulfide (5, 50, 250, and 750 μg/L). The characteristic odor of butyric acid derives from microbiological decay of fats found in spoiled milk (Friedrich & Acree ) and is perceived as a rancid, sour milk odor. Dimethyl sulfide (DMS) is a natural component of the odor of canned corn at dilute concentrations (Bills & Keenan ) and decaying vegetation at higher concentrations (APHA et al. ) and is commonly formed by plants, bacteria, and algae (Bentley & Chasteen ). These compounds were chosen because they have distinct odors and are represented on the drinking water, wastewater, compost and urban odor wheels (Suffet & Rosenfeld ), and because approximate odor intensities for these compounds were available in the literature (APHA et al. ). Samples and standards were presented at room temperature (21–25 C). W
Table 1
|
Relationships of scale values, verbal descriptors, and flavor standards for solutions of sugar in water
Scale valuea
Verbal descriptor
Flavor standard
1
Threshold
–
2
Very weak
–
4
Weak
5%
6
Moderate
–
8 10 12 a
10% Strong
– 15%
Zero is not included as a scale value but indicates the absence of detectable odor.
J. Curren et al.
144
|
Comparison of two standard odor intensity evaluation methods
Water Science & Technology
|
69.1
Panel recruitment and selection
to a log scale, using the following equation:
Of the 12 UCLA staff and students volunteering to participate in this study, eight individuals aged 18 to 42 years old were identified as primary odor panelists based on their more frequent attendance of study sessions; four served as alternates whenever a primary panelist was unavailable due to illness or scheduling conflicts. Panelists were not specifically screened for their ability to smell n-butanol, dimethyl sulfide, or butyric acid. All participants were screened for anosmia using the University of Pennsylvania Smell Identification Test (UPSIT; Doty et al. ); none were disqualified as anosmic.
butanol rescaled ¼ 1 þ lnðy1 =2:5Þ=lnð2Þ if y1 > ¼ 2:5
Panel sessions Panelists were trained in the use of the Butanol and FPA methods at the first two panel meetings. Data gathered during these meetings were excluded from analysis. Data collection sheets provided instructions for evaluating study samples. Prior to each session panelists were directed to avoid using scented personal care products; to refrain from smoking or eating at least 30 minutes before panel sessions; to rest between smelling samples; and to use odor-free water as a reference to refresh their noses. These directions were intended to minimize odor fatigue and prevent odor interference from uncontrolled sources. Eight panel sessions were held during which panelists evaluated a total of eight odorant samples each. In any given panel session, panelists were instructed to use only one of the two methods to evaluate four samples from the concentration series of butyric acid in water and four samples of dimethyl sulfide in water.
RESULTS AND DISCUSSION The two odor intensity methods were rescaled to allow for their direct comparison. The Butanol Method was adjusted Table 2
|
|
2014
(1)
where y1 is the odor intensity value ascribed to a sample by a panelist using the Butanol Method. After this rescaling the Butanol Method provided an 11point scale from 0 to 10 of n-butanol in odor-free water. The FPA Method was also adjusted so that the range of the scale would be equivalent to the butanol scale, using the following equation: FPA rescaled ¼ ðy2 =6Þ5
(2)
where y2 is the odor intensity value ascribed to a sample by a panelist using the FPA Method. After this rescaling the FPA Method provided an eightpoint scale from 0 to 10. The average intensity values and standard deviations for these methods at each concentration level of butyric acid and dimethyl sulfide, after rescaling, are shown in Table 2. The standard deviation is high, relative to the mean value at each concentration level. This indicates these data can be used to qualitatively compare odor intensity between odor events if the measurement is taken by the same person or panel, but cannot be used as a precise measure of intensity at the present time. The data were modeled using a linear mixed model (STATA; StataCorp ), taking into account the fixed effects of the concentration levels and the random effects of the panelists and sessions. Data from both methods were a significant fit to the models for both butyric acid and dimethyl sulfide, confirming their validity as tools to evaluate the intensity of these odors. Data were evaluated across sessions and among panelists to determine the possible effects of these variables on the odor intensity methods and compounds used. Findings from this evaluation appear in Table 3.
Average intensity value at each concentration level after adjustment of the intensity scales
Butyric acid Concentration (μg/L)
Dimethyl sulfide Butanol
FPA
500
1.7±1.4
1.4±1.9
1,500
2.9±1.7
5,000
5.1±1.5
10,000
6.1±1.6
Concentration (μg/L)
Butanol
FPA
5
3.1±2.2
1.1±1.3
2.3±1.9
50
5.1±2.2
3.4±1.9
5.4±1.9
250
6.6±2.0
6.1±2.5
7.3±2.4
750
7.7±1.7
7.9±2.3
J. Curren et al.
145
Table 3
|
|
Comparison of two standard odor intensity evaluation methods
Water Science & Technology
|
69.1
|
2014
Variance in linear mixed models by type of random effect
Butanol
FPA
Random effect
Odorant
Variance
SE
95% CI
Variance
SE
95% CI
Session
Dimethyl sulfide Butyric acid
1.58 1.07
0.58 0.45
0.77–3.25 0.47–2.48
0.39 0.15
0.32 0.33
0.08–1.93 0.002–11.22
Panelist
Dimethyl sulfide Butyric acid
1.62 0.27
1.05 0.44
0.45–5.83 0.01–6.51
1.78 1.29
0.95 0.71
0.63–5.09 0.44–3.81
SE ¼ standard error, 95% CI ¼ 95% confidence interval.
Session variance Eight panel sessions were held over a period of 8 weeks. During each session, panelists evaluated four samples each of butyric acid and dimethyl sulfide at different concentrations. The method used during each session was alternated weekly and samples were presented in a random order in order to reduce effects on the data due to participants’ increasing familiarity and experience with the use of either methodology, changes in attentiveness or willingness to follow instructions, the progression of concentrations of either compounds, or various other time- or sequence-related factors. Findings show less variance across sessions in the panels’ abilities to evaluate odor intensities:
• •
of butyric acid relative to dimethyl sulfide, regardless of the odor intensity method employed, and using the FPA Method relative to the Butanol Method, regardless of the odorant evaluated.
Panelist variance Although eight panelists participated in each session, up to four alternates substituted for regular panelists on occasions when one or more regular panelists were unavailable to participate. Findings show less variance across panelists in their ability to evaluate odor intensities:
• •
of butyric acid relative to dimethyl sulfide, regardless of the odor intensity method employed, and using the Butanol Method relative to the FPA Method, regardless of the odorant evaluated. These findings suggest that:
•
differences in odorants (in this case, possibly the greater volatility of dimethyl sulfide relative to butyric acid) can impact the relative effectiveness of a particular odor intensity scale,
• •
the FPA Method appears to show less variance across sessions, and the Butanol Method appears to show less variance across panelists.
Further studies evaluating a variety of odorants in addition to butyric acid and dimethyl sulfide are recommended to verify these findings. One notable advantage of the FPA Method is its ease of use. Because butanol standards must be used within two hours of formulation to ensure no diminution in odor strength, they are less amenable to use in the field. FPA odor intensity standards can be formulated a day in advance of their use if kept refrigerated during that time, and can also be prepared in the field with pre-measured amounts of water and sugar.
CONCLUSIONS Both the Butanol and FPA methods are valid for evaluating odor intensity using an odor panel. A linear mixed model was fitted to the data, taking into account both session and panelist effects. The FPA Method produced more consistent results across sessions than the Butanol Method, regardless of odorant, and was easier to administer and to use as a calibration standard on site. Although the future use of the FPA Method appears warranted, a caveat must be raised about the size of the data set obtained in this investigation. Additional investigations with a larger sample size that account for the level of panelist training and experience, the types and characters of odorants used, and other factors are recommended to verify the present findings. Additionally, the high standard deviation in the intensity scales relative to the mean indicate that they may be useful for understanding relative odor intensities, but should not be used as a precise measure. The intensity scale can be used to describe an odor problem but regulatory limits should not be based on these odor intensity values.
146
J. Curren et al.
|
Comparison of two standard odor intensity evaluation methods
In sum, the Butanol and FPA methods both appear to provide useful data about the perceived intensity of odors. Future studies are required to determine the applicability and efficacy of these tools in the realm of odor regulation.
ACKNOWLEDGEMENT Funded and completed with the help of the South Coast Air Quality Management District.
REFERENCES APHA Method 2170: Flavor Profile Analysis (FPA) Standard Methods for the Evaluation of Water and Wastewater. 22nd edn, American Public Health Association (APHA)/American Water Works Association (AWWA)/Water Environment Federation (WEF), Washington, DC, USA. Amoore, J. E. The chemistry and physiology of odor sensitivity. Journal of the American Water Works Association 78 (3), 70–76. ASTM The selection of judges for odor discrimination panels. Selected Technical Papers 440, 49–70. ASTM E544-10 Standard Practices for Referencing Suprathreshold Odor Intensity ASTM Standard. American Society for Testing and Materials (ASTM) West Conshohocken. Pennsylvania, USA. Bentley, R. & Chasteen, T. G. Environmental VOSCs – formation and degradation of dimethyl sulfide, methanethiol and related materials. Chemosphere 55 (3), 291–317.
Water Science & Technology
|
69.1
|
2014
Bills, D. D. & Keenan, T. W. Dimethyl sulfide and its precursor in sweet corn. Journal of Agricultural and Food Chemistry 16 (4), 643–645. Both, R., Sucker, K., Winneke, G. & Koch, E. Odour intensity and hedonic tone – important parameters to describe odour annoyance to residents? Water Science and Technology 55 (4), 83–92. Doty, R., Shaman, P. & Dann, M. University of Pennsylvania Smell Identification Test: a rapid quantitative olfactory function test for the clinic. The Laryngoscope 94 (2), 176–178. Friedrich, J. E. & Acree, T. E. Gas chromatography olfactometry (GC/O) of dairy products. International Dairy Journal 8 (3), 235–241. Lawless, H. T. Flavor description of white wine by expert and nonexpert wine consumers. Journal of Food Science 49, 120–123. Lawless, H. T. & Heymann, H. Sensory Evaluation of Food: Principles and Practices. Kluwer Academic/Plenum Publishers, New York, USA. Meng, A.-K. & Suffet, I. H. Assessing the quality of flavor profile analysis data. Journal of the American Water Works Association 84 (6), 89–96. Moschandreas, D. J. & Relwani, S. M. Perception of ETS odors: a visual and olfactory response. Atmospheric Environment, 26B, 263–269. StataCorp Stata Statistical Software: Release 11. College Station, TX. Suffet, I. H., Mallevialle, J. & Kawczynski, E. Advances in Taste and Odor Treatment and Control in Drinking Water. American Water Works Research Foundation, American Water Works Association, Denver, CO, USA. Suffet, I. H. & Rosenfeld, P. E. The anatomy of odour wheels for odours of drinking water, wastewater, compost and the urban environment. Water Science and Technology 55 (5), 335–344.
First received 29 April 2013; accepted in revised form 25 September 2013. Available online 24 October 2013
© IWA Publishing 2014 Water Science & Technology
|
69.1
|
2014
Comparison of two standard odor intensity evaluation methods for odor problems in air or water Jane Curren, Cherie L. (Cher) Snyder, Samantha Abraham and I. H. (Mel) Suffet
ABSTRACT Government agencies responsible for ensuring healthful water and/or air quality are often faced with resolving public complaints of nuisance odors. Understanding variations in odor intensity may ultimately lead to the establishment and application by such agencies of quantitative limits for effective odorant control. An odor panel was trained in suprathreshold odor intensity evaluation using both the ASTM Method E544 (Butanol Method) and the APHA Method 2170 (Flavor Profile Analysis (FPA) Method). A linear mixed model was fitted to the panel data, taking into account the fixed effects of concentration levels and the random effects of panelists and sessions. The FPA method proved easier to administer and revealed less inter-session variance than the ASTM Method, suggesting its greater utility in applications involving odor panels. For both methods, there was a high standard deviation, relative to the mean. This finding indicates that the intensity scales may be useful
Jane Curren (corresponding author) Samantha Abraham I. H. (Mel) Suffet Environmental Science and Engineering Program and Department of Environmental Health Sciences, School of Public Health, University of California, Los Angeles, CA 90095, USA E-mail: [email protected] Cherie L. (Cher) Snyder South Coast Air Quality Management District, 21865 Copley Dr, Diamond Bar, CA 91765, USA
for understanding relative odor intensities, but should not be used as a precise measure, or as a basis for establishing regulatory limits. Key words
| butanol, FPA Method, intensity method, nuisance, odor intensity method, odor intensity scales
INTRODUCTION The ability to detect and perceive odorous compounds appears to vary not only with the sensory acuity and experience of the observer, but also with the intensity, character, and hedonic value of the odorant itself, as well as the context in which the odorant is presented. For example:
• • • •
Those who are anosmic to one or more odor types may be up to a thousand times less sensitive to a given odor than others with average odor detection capabilities (Amoore ). Saturation of the odor receptors due to prolonged exposure to an odorant can result in odor fatigue such that the odor becomes less noticeable or may cease to be noticed. Formal training or experience in something such as wine tasting can result in improved chemosensory ability (Lawless ). Environmental cues may affect perceived odor intensity. The reported intensity of odors caused by cigarette
doi: 10.2166/wst.2013.567
smoke increased when the smoker was in the sight of the observer (Moschandreas & Relwani ). Both et al. () determined that once an odor becomes recognizable, it has the potential to cause a nuisance depending on hedonic tone and odor character. Government agencies responsible for ensuring healthful water and/or air quality are often faced with resolving public complaints of nuisance odors. Understanding the variation in odor intensity may help inform the development of reliable methods to rate the perceived strength of an odorant and to establish and apply quantitative limits for odorant control. The standard method for evaluating odor intensity in air is ‘Standard Practices for Referencing Suprathreshold Odor Intensity,’ also known as the Butanol Method (ASTM Standard Test Method , published ). This method measures the ‘total odor’ intensity of all combined odor characteristics by comparing a sample odor intensity to the intensity of standard solutions of n-butanol in either air or water.
143
J. Curren et al.
|
Comparison of two standard odor intensity evaluation methods
A parallel method for evaluating odor intensity in water is the ‘Standard Methods for the Evaluation of Water and Wastewater Method 2170,’ also known as the Flavor Profile Analysis (FPA) Method (American Public Health Association et al., Method 2170, revised ), used as a troubleshooting tool since the 1980s in the drinking water industry. Odor problems are identified by odor character using the ‘drinking water wheel’ and the intensity of each odor character is established using a seven-point intensity scale to measure the strength of individual flavor and odor characters (Suffet et al. ). Four of the intensity points are anchored to flavor standards, typically solutions of sugar (sucrose) in water. Other scales using verbal descriptors (e.g., weak, mild, strong) and/or numeric rankings of odor intensity typically require less training than the use of methods involving sensory intensity anchors, and do not lead to olfactory fatigue. The lack of equal intervals between adjacent points on these ordinal scales, however, limits their suitability for statistical analysis (Lawless & Heymann ), making the use of the Butanol and FPA methods preferable from the standpoint of data reduction. The FPA and Butanol methods each require the use of an odor panel of at least four or eight people, respectively, to reduce response variability (APHA et al. ). While methods for panel selection and training in the use of each method are available in the literature (Meng & Suffet ; ASTM ), they are not outlined by either method. The objective of this study is to evaluate and compare the Butanol and FPA methods with respect to the precision of the data provided by their use.
METHOD Odor intensity evaluation Two odor intensity scaling methods were evaluated in this study:
•
ASTM Method E544 ‘Standard Practices for Referencing Suprathreshold Odor Intensity (ASTM )’. ASTM Method E544, known informally as the ‘Butanol method’, measures perceived odor intensity when participants compare samples to standard solutions of n-butanol (99.9% pure) in either air or water. An 11-point scale corresponded to concentrations of 0, 2.5, 5, 10, 20, 40, 80, 160, 320, 640, and 1,280 mg/L of n-butanol in odor-free water prepared freshly from n-butanol before each panel session.
Water Science & Technology
|
69.1
|
2014
Panelists rated odor intensity by selecting a scaled value (or a point between two such values) corresponding to the equivalent concentration of n-butanol. The rated odor intensity was directly proportional to the concentration of the combined odor-causing compounds present in the sample.
•
American Public Health Association Method 2170 ‘Flavor Profile Analysis’, in ‘Standard Methods for the Evaluation of Water and Wastewater (APHA et al. )’. The FPA Method uses a seven-point intensity scale to measure the intensity of individual flavor and odor characters in drinking water (APHA et al. ), as shown in Table 1 below. Three of the intensity points were anchored to flavor standards; in this case, to concentrations of sugar (sucrose) in water. Panelists based their evaluations of sample odor intensity on the perceived intensity of the flavor of the sugar solutions.
Both the Butanol and FPA Odor Intensity Scaling Methods were used to evaluate two concentration series of butyric acid (500, 1,500, 5,000, and 10,000 μg/L) and dimethyl sulfide (5, 50, 250, and 750 μg/L). The characteristic odor of butyric acid derives from microbiological decay of fats found in spoiled milk (Friedrich & Acree ) and is perceived as a rancid, sour milk odor. Dimethyl sulfide (DMS) is a natural component of the odor of canned corn at dilute concentrations (Bills & Keenan ) and decaying vegetation at higher concentrations (APHA et al. ) and is commonly formed by plants, bacteria, and algae (Bentley & Chasteen ). These compounds were chosen because they have distinct odors and are represented on the drinking water, wastewater, compost and urban odor wheels (Suffet & Rosenfeld ), and because approximate odor intensities for these compounds were available in the literature (APHA et al. ). Samples and standards were presented at room temperature (21–25 C). W
Table 1
|
Relationships of scale values, verbal descriptors, and flavor standards for solutions of sugar in water
Scale valuea
Verbal descriptor
Flavor standard
1
Threshold
–
2
Very weak
–
4
Weak
5%
6
Moderate
–
8 10 12 a
10% Strong
– 15%
Zero is not included as a scale value but indicates the absence of detectable odor.
J. Curren et al.
144
|
Comparison of two standard odor intensity evaluation methods
Water Science & Technology
|
69.1
Panel recruitment and selection
to a log scale, using the following equation:
Of the 12 UCLA staff and students volunteering to participate in this study, eight individuals aged 18 to 42 years old were identified as primary odor panelists based on their more frequent attendance of study sessions; four served as alternates whenever a primary panelist was unavailable due to illness or scheduling conflicts. Panelists were not specifically screened for their ability to smell n-butanol, dimethyl sulfide, or butyric acid. All participants were screened for anosmia using the University of Pennsylvania Smell Identification Test (UPSIT; Doty et al. ); none were disqualified as anosmic.
butanol rescaled ¼ 1 þ lnðy1 =2:5Þ=lnð2Þ if y1 > ¼ 2:5
Panel sessions Panelists were trained in the use of the Butanol and FPA methods at the first two panel meetings. Data gathered during these meetings were excluded from analysis. Data collection sheets provided instructions for evaluating study samples. Prior to each session panelists were directed to avoid using scented personal care products; to refrain from smoking or eating at least 30 minutes before panel sessions; to rest between smelling samples; and to use odor-free water as a reference to refresh their noses. These directions were intended to minimize odor fatigue and prevent odor interference from uncontrolled sources. Eight panel sessions were held during which panelists evaluated a total of eight odorant samples each. In any given panel session, panelists were instructed to use only one of the two methods to evaluate four samples from the concentration series of butyric acid in water and four samples of dimethyl sulfide in water.
RESULTS AND DISCUSSION The two odor intensity methods were rescaled to allow for their direct comparison. The Butanol Method was adjusted Table 2
|
|
2014
(1)
where y1 is the odor intensity value ascribed to a sample by a panelist using the Butanol Method. After this rescaling the Butanol Method provided an 11point scale from 0 to 10 of n-butanol in odor-free water. The FPA Method was also adjusted so that the range of the scale would be equivalent to the butanol scale, using the following equation: FPA rescaled ¼ ðy2 =6Þ5
(2)
where y2 is the odor intensity value ascribed to a sample by a panelist using the FPA Method. After this rescaling the FPA Method provided an eightpoint scale from 0 to 10. The average intensity values and standard deviations for these methods at each concentration level of butyric acid and dimethyl sulfide, after rescaling, are shown in Table 2. The standard deviation is high, relative to the mean value at each concentration level. This indicates these data can be used to qualitatively compare odor intensity between odor events if the measurement is taken by the same person or panel, but cannot be used as a precise measure of intensity at the present time. The data were modeled using a linear mixed model (STATA; StataCorp ), taking into account the fixed effects of the concentration levels and the random effects of the panelists and sessions. Data from both methods were a significant fit to the models for both butyric acid and dimethyl sulfide, confirming their validity as tools to evaluate the intensity of these odors. Data were evaluated across sessions and among panelists to determine the possible effects of these variables on the odor intensity methods and compounds used. Findings from this evaluation appear in Table 3.
Average intensity value at each concentration level after adjustment of the intensity scales
Butyric acid Concentration (μg/L)
Dimethyl sulfide Butanol
FPA
500
1.7±1.4
1.4±1.9
1,500
2.9±1.7
5,000
5.1±1.5
10,000
6.1±1.6
Concentration (μg/L)
Butanol
FPA
5
3.1±2.2
1.1±1.3
2.3±1.9
50
5.1±2.2
3.4±1.9
5.4±1.9
250
6.6±2.0
6.1±2.5
7.3±2.4
750
7.7±1.7
7.9±2.3
J. Curren et al.
145
Table 3
|
|
Comparison of two standard odor intensity evaluation methods
Water Science & Technology
|
69.1
|
2014
Variance in linear mixed models by type of random effect
Butanol
FPA
Random effect
Odorant
Variance
SE
95% CI
Variance
SE
95% CI
Session
Dimethyl sulfide Butyric acid
1.58 1.07
0.58 0.45
0.77–3.25 0.47–2.48
0.39 0.15
0.32 0.33
0.08–1.93 0.002–11.22
Panelist
Dimethyl sulfide Butyric acid
1.62 0.27
1.05 0.44
0.45–5.83 0.01–6.51
1.78 1.29
0.95 0.71
0.63–5.09 0.44–3.81
SE ¼ standard error, 95% CI ¼ 95% confidence interval.
Session variance Eight panel sessions were held over a period of 8 weeks. During each session, panelists evaluated four samples each of butyric acid and dimethyl sulfide at different concentrations. The method used during each session was alternated weekly and samples were presented in a random order in order to reduce effects on the data due to participants’ increasing familiarity and experience with the use of either methodology, changes in attentiveness or willingness to follow instructions, the progression of concentrations of either compounds, or various other time- or sequence-related factors. Findings show less variance across sessions in the panels’ abilities to evaluate odor intensities:
• •
of butyric acid relative to dimethyl sulfide, regardless of the odor intensity method employed, and using the FPA Method relative to the Butanol Method, regardless of the odorant evaluated.
Panelist variance Although eight panelists participated in each session, up to four alternates substituted for regular panelists on occasions when one or more regular panelists were unavailable to participate. Findings show less variance across panelists in their ability to evaluate odor intensities:
• •
of butyric acid relative to dimethyl sulfide, regardless of the odor intensity method employed, and using the Butanol Method relative to the FPA Method, regardless of the odorant evaluated. These findings suggest that:
•
differences in odorants (in this case, possibly the greater volatility of dimethyl sulfide relative to butyric acid) can impact the relative effectiveness of a particular odor intensity scale,
• •
the FPA Method appears to show less variance across sessions, and the Butanol Method appears to show less variance across panelists.
Further studies evaluating a variety of odorants in addition to butyric acid and dimethyl sulfide are recommended to verify these findings. One notable advantage of the FPA Method is its ease of use. Because butanol standards must be used within two hours of formulation to ensure no diminution in odor strength, they are less amenable to use in the field. FPA odor intensity standards can be formulated a day in advance of their use if kept refrigerated during that time, and can also be prepared in the field with pre-measured amounts of water and sugar.
CONCLUSIONS Both the Butanol and FPA methods are valid for evaluating odor intensity using an odor panel. A linear mixed model was fitted to the data, taking into account both session and panelist effects. The FPA Method produced more consistent results across sessions than the Butanol Method, regardless of odorant, and was easier to administer and to use as a calibration standard on site. Although the future use of the FPA Method appears warranted, a caveat must be raised about the size of the data set obtained in this investigation. Additional investigations with a larger sample size that account for the level of panelist training and experience, the types and characters of odorants used, and other factors are recommended to verify the present findings. Additionally, the high standard deviation in the intensity scales relative to the mean indicate that they may be useful for understanding relative odor intensities, but should not be used as a precise measure. The intensity scale can be used to describe an odor problem but regulatory limits should not be based on these odor intensity values.
146
J. Curren et al.
|
Comparison of two standard odor intensity evaluation methods
In sum, the Butanol and FPA methods both appear to provide useful data about the perceived intensity of odors. Future studies are required to determine the applicability and efficacy of these tools in the realm of odor regulation.
ACKNOWLEDGEMENT Funded and completed with the help of the South Coast Air Quality Management District.
REFERENCES APHA Method 2170: Flavor Profile Analysis (FPA) Standard Methods for the Evaluation of Water and Wastewater. 22nd edn, American Public Health Association (APHA)/American Water Works Association (AWWA)/Water Environment Federation (WEF), Washington, DC, USA. Amoore, J. E. The chemistry and physiology of odor sensitivity. Journal of the American Water Works Association 78 (3), 70–76. ASTM The selection of judges for odor discrimination panels. Selected Technical Papers 440, 49–70. ASTM E544-10 Standard Practices for Referencing Suprathreshold Odor Intensity ASTM Standard. American Society for Testing and Materials (ASTM) West Conshohocken. Pennsylvania, USA. Bentley, R. & Chasteen, T. G. Environmental VOSCs – formation and degradation of dimethyl sulfide, methanethiol and related materials. Chemosphere 55 (3), 291–317.
Water Science & Technology
|
69.1
|
2014
Bills, D. D. & Keenan, T. W. Dimethyl sulfide and its precursor in sweet corn. Journal of Agricultural and Food Chemistry 16 (4), 643–645. Both, R., Sucker, K., Winneke, G. & Koch, E. Odour intensity and hedonic tone – important parameters to describe odour annoyance to residents? Water Science and Technology 55 (4), 83–92. Doty, R., Shaman, P. & Dann, M. University of Pennsylvania Smell Identification Test: a rapid quantitative olfactory function test for the clinic. The Laryngoscope 94 (2), 176–178. Friedrich, J. E. & Acree, T. E. Gas chromatography olfactometry (GC/O) of dairy products. International Dairy Journal 8 (3), 235–241. Lawless, H. T. Flavor description of white wine by expert and nonexpert wine consumers. Journal of Food Science 49, 120–123. Lawless, H. T. & Heymann, H. Sensory Evaluation of Food: Principles and Practices. Kluwer Academic/Plenum Publishers, New York, USA. Meng, A.-K. & Suffet, I. H. Assessing the quality of flavor profile analysis data. Journal of the American Water Works Association 84 (6), 89–96. Moschandreas, D. J. & Relwani, S. M. Perception of ETS odors: a visual and olfactory response. Atmospheric Environment, 26B, 263–269. StataCorp Stata Statistical Software: Release 11. College Station, TX. Suffet, I. H., Mallevialle, J. & Kawczynski, E. Advances in Taste and Odor Treatment and Control in Drinking Water. American Water Works Research Foundation, American Water Works Association, Denver, CO, USA. Suffet, I. H. & Rosenfeld, P. E. The anatomy of odour wheels for odours of drinking water, wastewater, compost and the urban environment. Water Science and Technology 55 (5), 335–344.
First received 29 April 2013; accepted in revised form 25 September 2013. Available online 24 October 2013
