The results of these three approaches are dramatically different because the fixed intercept forces all lines to the same point while the floating intercept allows for very different linear fits to the data. Further research is needed into the equation that best fits the relationship between concentration and intensity scales.The predicted ODTC50 using the 0-12 scale for several chemicals was within the range of the values from more traditional methods . Although the intensity scale is ordinal , it is handled as a metric scale , presumably because it simply works. As with analytical instruments, the zero intensity score is not included in the linear regressions. Note that the above equations use the base-10 logarithm rather than the natural logarithm, which is used in most science, perhaps because decibels use the base-10 logarithm. Another method of finding the intensity of an odorant is through repeated dilutions of the sample . These dilutions are presented to panelists from high concentration to low concentration using continuous airflow to anose port by a dynamic dilution instrument . Two odorless blanks and the sample are presented, and the panelist chooses which one is odorous . In Asia, the standard method is 3 bags instead of continuous flow. Points along the intensity curve can be observed, but it is typically the ODTC50 that is sought. Both methods,flood tray using an intensity scale or dilutions, are used to analyze mixtures as well. Mixtures are more complex and do not necessarily follow the above associations. Specifically, the number of dilutions required to reach the odor-detection threshold for a mixture does not properly reflect the actual sensory intensity , as demonstrated by a study of fecal odorants .
In other words, odor intensities increase and decrease with concentration at different rates for different odorants, not to mention the antagonistic and synergistic effects that also occur.One method used various vials of n-butanol as reference points to provide a sniffed intensity scale for panelists , while the other used three tasted solutions of sugar at different concentrations . The latter approach was applied to odor research based on the cross-modal premise that our senses are linked, so taste can be used to inform the sense of smell . Although the two anchors were not necessarily parallel at the high end of their scales , they compared favorably . Therefore, the result of this study cannot be used to say the two scales are similar beyond this one study’s approach. Further, the n-butanol method advises against translating the results into perceived odor intensities . A comparison with the other anchors would be informative .Note the similarities between odor-detection thresholds and method detection limits for analytical chemical analyses. Both are statistical measures that become less precise and highly variable when near these limits, with sensory thresholds showing greater variability, often of two orders of magnitude . Both the physical setting and panelist’s unique situation can lead to such large variability, even from one day to the next. Therefore, presenting an ODTC50 as a range rather than a single value is recommended . If data are only generated suprathreshold, then a Weber-Fechner plot can be used to crudely estimate an ODTC50. Quality controls should include method blanks and spiked samples, and analyzing samples in triplicate allows for the calculation of standard deviations .For risk assessment of conventional air pollutants, frequency and duration are temporal aspects used in the calculation of exposure and indicate the appropriate hazard benchmark .
For odor assessment, frequency and duration also are part of the sensory experience because they alter the perception of odor . An individual can become “used” to an odor and no longer able to detect it. Because odors are sensed within a few seconds or less, any averaging times applied to measurements may miss the peaks that trigger the complaints. Noting the frequency and duration of odor events can help inspectors and facility operators identify odor sources based on operation schedules and weather patterns.English speakers typically refer to a smell by its source . When forced to avoid using source terms, the descriptors “stinky,” “fragrant” and “musty” are used most often. Subsistence hunter-gatherers in the Malay Peninsula, by comparison, have a rich olfactory language, naming smells as easily as colors . Dutch participants in a study had the same facial expressions as the hunter-gatherers when smelling odorants; however, the words selected to describe the odors were vastly different, the Dutch words tending be vaguer . Overall, smell tends to be poorly coded in languages as compared to other senses, yet this seems to be based in cultural norms rather than any neurological reason .As a starting point, the previous reviews of odor-measurement methods conducted by doctoral students were consulted . These reviews built upon a substantial review , an effort sponsored by groups in Australia and Spain. To update the prior work, a literature search was conducted post-2010 to gather the most recent methods and critiques. The search was conducted online and at the UCLA and CARB physical libraries. When relevant articles or books were found, the “cited by” function was used to discover even more up-to-date information. Reviews of the latest approaches to odor exposure measurement and risk assessment were sought. Finally, relevant websites and posted materials that are not typically available in scientific journals were searched.
The starting point for risk assessment methodology was the foundational work by the National Research Council , now updated for the 21st century . The programs that grew out of the original work have issued their own guidance, which was also consulted. Such programs include pesticides , site remediation , exposure factors and California-specific work . International efforts, such as the review of the chemicals in commerce in the European Union , were also consulted. The 1,556-page tome edited by Dennis Paustenbach titled “Human and Ecological Risk Assessment: Theory and Practice” was an additional starting point. Risk assessment principles and terminology will be used to organize and structure the field of odor exposure assessment. The risk assessment framework will not be applied per se to environmental odor cases but, rather, offer well-established concepts, conventions and terminology that can be applied to odors. As one example, the challenge of evaluating real-world chemical mixtures rather than a single chemical at a time applies to both fields. At the core of the evaluation will be the scientific merits of the various approaches to ensure any recommendations are evidence-based. The strengths and shortcomings of popular sensory and instrument methods will be reviewed.Sensory methods use the human nose as the detector. Because odor complaints arise from this same detector, it is the “gold standard.” As discussed in Section 2.3, our sense of smell can detect odor notes,grow table although our vocabulary may struggle to supply the right words. The odor hedonic tone is more easily assigned. The odor intensity may be assigned categorical words or scores or placed on a scale. The scaling of intensity has led to the techniques discussed in this section. As with all else pertaining to odors, evaluating individual odorants is an oversimplification of their contribution to the total odor of a mixture.While chemical analyses can identify a subset of odorants and their concentrations in ambient air, chemical analyses often do not directly relate to human sensory experiences . Odor samples usually contain multiple odorants that can have synergistic and/or antagonistic effects on each other, altering the overall perception.
Odor is not simply additive, unlike concentration. Further, the human nose is usually more sensitive than analytical techniques. Therefore, the most accurate way to evaluate an odor it to judge its properties “as is.” This direct approach is called “odor profiling” and can be performed at the location where the odor is observed or by capturing a sample and transporting it to where a trained observer is located . A specific version of this approach, the Odor Profile Method , has been developed based largely on flavor profiling for the food and drinking water industries, specifically Method 2710 “Flavor Profile Analysis” by the APHA . OPM includes two parts: first, identifying one or more odor notes in the sample and, second, determining the odor intensity for each odor note. Duration of the odor at the site fenceline can also be included as a third factor . OPM can be part of a diagnostic investigation or an ongoing monitoring program. Rather than have panelists use their own natural, naïve language to describe the odor notes, a standardized vocabulary has been developed. Note, however, that only up to 2 odorants per mixture can be recognized by trained panelists . The standardized vocabulary has been tabulated for several industries and displayed graphically as wheels. Appendix A includes a collection of odor wheels, and Table 3.4 is a side-by-side presentation of their content. Over the years, such wheels have contributed a standardized way to classify, communicate, and identify odor notes, and sometimes the underlying odorants, in emissions . Odor wheels consist of three rings: an inner ring of general odor notes, a middle ring of more specific odor notes within each segment and an outer ring of potential odorants associated with each odor note. Their development has been described . A taste and odor wheel for drinking water has been a major contributor to that field .The OPM intensity scale, described and critiqued in Section 2.3 , is used for each odor note. This scale is anchored on three sugar solutions tasted by mouth with cross-modal sensory translation to smell. The 7-point OPM intensity scale is: threshold , slight , weak , medium , medium strong , strong , and very strong . Before an individual becomes a panelist, their sense of smell is verified. The OPM uses the University of Pennsylvania Smell Identification Test , which is the best-known test of smell that uses micro-encapsulated odorants . As such, it is highly portable and often used in field studies. It has been well-standardized against age, gender and correlates well with the results of quantitative odor-detection tests . A passing score is 70% of the 40 questions. Further guidance on the selection of panelists is provided by ASTM Method E1440 . A minimum of four panelists is required by OPM. They may not have a cold, mustache, wear perfume, eat food or drink during the session. They are trained using the applicable odor wheel as well as the intensity scale . The panel is held in an air conditioned room with no scents or in the field at the location of the odor. Panelists are trained to distinguish odor mixtures and the intensity of each odorant. Panelists are presented with mixtures of 2, 3, and 4 standard odorants and are asked to identify the odor characteristics and the intensity of each odorant. Panelists are then ready to analyze actual environmental odor samples.Before panelists are exposed to samples, a safety evaluation determines that the odorant concentrations are below safety thresholds . Samples are presented in bags and smelled by panelists, who then write down the odor notes and odor intensities. The leader then leads a discussion, and panelists may alter their decisions. Breaks are taken to avoid odor fatigue or health symptoms . The first challenge with OPM is that it usually requires confirmation by GC-sensory analysis to identify the odorant . The second challenge is that OPM usually operates outside the range of its calibration. The calibrated intensity scores are in the range 4 to 12, yet individual panelist scores – as well as the overall geometric mean – are usually in the 1 to 3 range. Such below-range extrapolations are not allowed in sound analytics. Data sets from analytical chemistry instruments have the same challenge and the temptation to extrapolate lower than the method detection limit. A concern is that there is no evidence that OPM is reproducible across panels . The group discussion may lead to the dominant person biasing others despite instructions to conduct independent evaluations . Substantial training is required.