September 9, 2008

Dispersion Modeling to Compare Alternative Technologies for Odor Remediation at Swine Facilities

By Schiffman, Susan S Graham, Brevick G; Williams, C Mike

ABSTRACT The effectiveness of 18 alternative technologies for reducing odor dispersion at and beyond the boundary of swine facilities was assessed in conjunction with an initiative sponsored through agreements between the Attorney General of North Carolina and Smithfield Foods, Premium Standard Farms, and Frontline Farmers. The trajectory and spatial distribution of odor emitted at each facility were modeled at 200 and 400 m downwind from each site under two meteorological conditions (daytime and nighttime) using a Eulerian-Lagrangian model. To predict the dispersion of odor downwind, the geographical area containing the odorant sources at each facility was partitioned into 10-m^sup 2^ grids on the basis of satellite photographs and architectural drawings. Relative odorant concentrations were assigned to each grid point on the basis of intensity measurements made by the trained odor panel at each facility using a 9-point rating scale. The results of the modeling indicated that odor did not extend significantly beyond 400 m downwind of any of the test sites during the daytime when the layer of air above the earth's surface is usually turbulent. However, modeling indicated that odor from all full-scale farms extended beyond 400 m onto neighboring property in the evenings when deep surface cooling through long-wave radiation to space produces a stable (nocturnal) boundary layer. The results also indicated that swine housing, independent of waste management type, plays a significant role in odor downwind, as do odor sources of moderate to moderately high intensity that emanate from a large surface area such as a lagoon. Human odor assessments were utilized for modeling rather than instrument measurements of volatile organic compounds (VOCs), hydrogen sulfide, ammonia, or particulates less than 10 [mu]m in diameter (PM^sub 10^) because these physical measurements obtained simultaneously with human panel ratings were not found to accurately predict human odor intensity in the field.


Over the last several decades, swine production has moved from pasture to confinement feeding with large numbers of hogs now reared indoors in confinement buildings.1 Odors emanating from these industrial-scale facilities that concentrate large numbers of animals in a small area of land present a major challenge to the swine industry because nearby community dwellers frequently complain about the malodorous emissions emanating from these confinement operations.1-6 Psychological studies of neighbors of confinement swine facilities indicate that they experience negative emotional responses when exposed to the malodors, including tension, depression, anger, fatigue, and confusion.7 Adverse physiological health parameters have also been associated with exposure to odorous emissions at levels that can occur in neighboring communities downwind of swine operations. These adverse health parameters include increased occurrences of eye, nose, and throat irritation, headaches, nausea, diarrhea, and symptoms of respiratory inflammation as well as suppressed mucosal immunity.3-5,8,9 Increased prevalence of asthma symptoms in children has also been associated with proximity to swine operations. 10 Workers at confined swine operations have long been known to suffer negative health effects including respiratory symptoms, reductions in pulmonary function, and increased bronchial responsiveness from exposure to the ambient air at these facilities.11,12 In addition to the psychological and physical health impacts from odorous emissions, potential or ongoing odor from large-scale swine production facilities raises concerns about environmental pollution and lowering of property values in the surrounding communities.1

The odors associated with traditional confinement swine production arise from anaerobic decomposition of waste from swine housing, lagoons that receive waste flushed from the houses, and land application of both solid and liquid waste. Lagoons are typically large, uncovered earthen basins with extensive surface areas that rely on anaerobic bacteria to decompose the organic waste.13 Volatile organic compounds (VOCs), organic dusts, and gases such as hydrogen sulfide and ammonia, are the main sources of odorous emissions from swine operations. 3,14-16 Hundreds of diverse chemical compounds have been quantified in the odorous air emitted from swine production facilities including acids, alcohols, al- dehydes, amides, amines, aromatics, esters, ethers, fixed gases, halogenated hydrocarbons, hydrocarbons, ketones, nitriles, other nitrogen-containing compounds, phenols, sulfur-containing compounds, and steroids.14,16-20 Although the vast majority of individual compounds in swine emissions are present at concentrations below published odor and irritation thresholds, the cumulative effect of hundreds of compounds can produce intense odors and even sensory irritation downwind of swine operations.16

Concern about malodorous aerial emissions and other environmental impacts associated with concentrated swine production has been especially poignant in North Carolina, the state that has experienced the most dramatic expansion of swine production over the last 20 yr. Between 1991 and 1997, the swine inventory in North Carolina increased from 2.7 million head to approximately 10 million head (over 300%), making North Carolina the second largest hog producer (behind Iowa) in the United States.21 However, since 1997 the number of swine statewide has remained relatively stable due, in part, to a state-mandated moratorium (North Carolina House Bill 515) on development of new facilities that utilize traditional lagoon waste management treatment processes. Because sustainability of confined swine production will necessitate the development of new technologies to mitigate odorous aerial emissions and other environmental issues, a program was developed to evaluate "environmentally superior technologies" (ESTs) under an initiative sponsored through agreements between the Attorney General of North Carolina and Smithfield Foods/Premium Standard Farms.21,22 Technology candidates for ESTs for this initiative were selected for evaluation on the basis of competitive review (by scientists and stakehold-ers) of their potential for reducing odor and ammonia emissions, disease-transmitting vectors, and airborne pathogens as well as discharge of other contaminants into surface and groundwater.22 An emissions analysis component of this program called Project OPEN (Odors, Pathogens, and Emissions of Nitrogen) was established in December 2000 to compare 18 ESTs selected by the competitive review just described as alternative waste technologies to traditional lagoon waste management. 17,22 For a potential EST to be considered environmentally superior from the standpoint of odor, according to the agreement with the Attorney General of North Carolina,22 it must substantially eliminate the emission of odor that is detectable beyond the boundaries of the parcel or tract of land on which the swine farm is located.

The purpose of the OPEN odor project was to evaluate the 18 different alternative technologies selected by competitive review for evaluation to compare their effectiveness in reducing odor dispersion at and beyond the boundary of the swine facility. Two farms using traditional lagoon technology were included as controls. Factors in addition to odor reduction efficiency, such as the efficacy of reduction of emissions of ammonia and pathogens and the economic feasibility of the technologies, were addressed in other OPEN projects.


Odorous emissions were assessed using two types of measurement techniques: human odor assessments (olfac-tometry) and instrument sampling methods (see Table 1 for overview). Both measurement techniques were utilized in the field as well as in the laboratory. Sampling performed in the field involved the assessment of ambient air at odor sources and at varying distances downwind. Physical samples collected in the field and tested in the laboratory included liquids and solids from the waste stream, fabric swatches used for odor and dust adsorption, glass fiber filters, and Tedlar air bags. A description of the assessment methods is given below.

Odors were characterized in the field at each swine facility, at odor sources and at varying distances downwind by trained human assessors using the following psychophysical techniques performed in this order: (1) odor threshold measurements using two different field olfactometers; (2) irritation assessments of ambient air using a lateralization device16 (see below for device de-scription); (3) sensory assessments of ambient air on odor intensity, irritation intensity, and pleasantness scales; and (4) odor description. All subjects on the odor panel were pretested and had normal butanol odor thresholds using methodology described by Stevens and Dadarwala.23 For the field measurements, the vast majority of sampling was performed by four trained odor panelists. In the field, odor panelists wore a 3M half facepiece mask (3M model 6000 series) fitted with the 3M multigas/vapor cartridge with P100 filter before and between odor assessments. This mask prevents exposure to vapors as well as dusts and eliminates olfactory adaptation. Each of the 18 different technologies was evaluated on 4 different days, two during a warm season and two during a cool season to control for possible seasonal variations. The number and locations of ambient air field assessments performed at odor sources were selected for each site, depending upon technologies and waste management systems used, site conditions, geographical orientation of odor sources, and meteorological conditions. Field evaluations were undertaken at all significant odor sources at each test site including technology components, swine houses, waste lagoons, and other waste storage and treatment facilities. Once a site was evaluated the first time all subsequent testing at that site included the same sample locations and types that occurred on the first visit. Each field location was evaluated during a 15- to 40-min time period. Odor sampling was interspersed with intertrial resting periods of 20 min or more with the facemask in place so that subjects could breathe clean (filtered) air between assessments. Wind direction dictated the actual location of downwind sampling; for example, if a certain odor point source was evaluated on the first site visit at a location 30- ft downwind, all subsequent analogous samples took place 30-ft downwind of that point source at a location determined by wind direction.

At each testing location in the field, wind speed, relative humidity, and ambient air temperature were monitored with a Kestrel 3000 pocket mini weather station. In addition, wind direction was recorded. Distances were measured in the field using a tape measure and a Nikon Laser 800 Rangefinder. All field air samples and meteorological data were evaluated at an elevation of approximately 5 ft (at nose level).

Olfactory Thresholds in the Field

Odor threshold measurements were obtained to determine how many times the odorous source needed to be diluted to reach threshold (D/ T or dilutions to threshold). Odor thresholds in the field were measured using two different portable olfactometers, the Scentometer (Barnebey and Sutcliffe) and Nasal Ranger olfactometer (St. Croix Sensory). The Scentometer is a hand-held plastic box with two glass nasal ports on one end and six inlets for the odorous air (inlet sizes of 1/32 in. (0.08 cm), 1/16 in. (0.16 cm), 1/8 in. (0.32 cm), 3/16 in. (0.48 cm), 1/4 in. (0.64 cm), and 1/2 in. (1.27 cm)) at the other end which are connected to a mixing chamber. These inlets correspond to 350, 170, 31, 15, 7, and 2 D/T, respectively. D/T values are measured with all but one of the six odorous air inlets closed, and the size of the open inlet determines the dilution of the odorous air. For example, if a threshold was found with the 1/ 16" inlet open, the concentration of the odor was considered to be 170 times above threshold.

In addition to the Scentometer, the Nasal Ranger olfactometer was used by the human odor panel to assess odor thresholds in the field. This olfactometer is a handheld device that has a dial that can be turned to deliver blank air (air that is only drawn through a filter) or to deliver air at the D/T levels of 60, 30, 15, 7, 4, and 2. An additional dial with the D/T levels of 500, 400, 300, 200, 100, and 60 was also used. An light-emitting diode display indicates when the subject is nasally breathing the air too quickly or too slowly based on static pressure, thus assuring a flow rate of 14-18 L/min. The presentation device is the AC'SCENT SWIFT nasal mask.

Lateralization Method for Assessing Nasal Irritation in the Field

Irritation, unlike olfaction, can be localized to one nostril or the other.3,16 That is, when an odorous stimulus that is not an irritant (i.e., only stimulates the olfactory nerve) is presented to one nostril and a nonodorous stimulus is presented to the other nostril, the subject cannot determine which nostril received the odor. However, if the odorous stimulus is also an irritant (i.e., stimulates the trigeminal nerve causing nasal pungency), the subject can identify the nostril that received the stimulus. Subjects used an irritation device16 that was designed in our laboratory to assess the presence of irritant properties in ambient air evaluated in the field. Subjects sniffed odorous ambient air into one nostril and charcoal-filtered air into the other nostril at different locations at the swine facilities. These locations were the same as those evaluated using the threshold assessments and the other field odor evaluations. The nostril that received the ambient air was randomized over trials. Subjects recorded which nostril received the stronger odor.

Odor Intensity, Nasal Irritation, Odor Pleasantness, and Odor Description in the Field

Odor panelists evaluated ambient air in the field for odor intensity, nasal irritation, odor pleasantness, and odor character. All samples were rated on three 9-point line scales numbered from 0 to 8.16,21 These included odor intensity, irritation intensity, and hedonic ratings. For odor and irritation intensity, the scale was labeled as follows: 0 = none at all, 1 = very weak, 2 = weak, 3 = moderately weak, 4 = moderate, 5 = moderately strong, 6 = strong, 7 = very strong, and 8 = maximal. The scale used for pleasantness/ unpleasantness was: 0 = extremely pleasant, 1 = very pleasant, 2 = moderately pleasant, 3 = slightly pleasant, 4 = neither pleasant nor unpleasant, 5 = slightly unpleasant, 6 = moderately unpleasant, 7 = very unpleasant, and 8 = extremely unpleasant. Subjects also provided a description of each sample for which they perceived an odor. Odor panelists were trained to detect and evaluate 146 different odor qualities using an odor-descriptive adjective scale.24 If a subject wanted to include an adjective descriptor that was not present among the 146 odor terms on the form, he or she provided other words and/or phrases to describe that odor.

Laboratory Odor Measurements

All physical samples were evaluated for odor intensity, irritation intensity, hedonics, and odor character by eight odor panelists in a laboratory that contained an exhaust hood designed to prevent accumulation of odors. The samples collected in the field and evaluated in the laboratory included the following: (1) Tedlar air bag samples, which were used for odor threshold determinations with an AC'SCENT olfactometer (St. Croix Sensory) as well as full strength evaluations taken directly from the sampling bag; (2) cotton swatch samples that were placed at odor sources at a swine facility for 1 hr to collect odorants and dusts and sniffed from glass bottles with airtight lids; (3) glass fiber samples from the Andersen Non-Viable Eight Stage Dust Impactor, which collected odors and dusts for a 24-hr period and were sniffed from the same type of glass bottles as those used for swatches; and (4) liquid and solid waste stream samples presented in the glass bottles. Sample collection locations were chosen at each site on the basis of analogous samples at other sites.

Olfactory Thresholds in the Laboratory. Air samples were collected in Tedlar bags (from SKC, Inc.) at and downwind from lagoons, animal houses, and other odorant sources at swine facilities. Odor thresholds for these Tedlar bag samples were determined using the AC'SCENT olfactom-eter (St. Croix Sensory), which meets CEN TC264 olfac-tometry standard requirements and is calibrated to 20 L/min. The presentation device was the AC'SCENT SWIFT nasal mask, and the presentation face velocity was 0.25 m/sec at 20 L/min. The sample delivery (sniffing) time was set to 3 sec.

Thresholds were determined using triangular forced-choice methodology. At each level in triangular forced-choice mode, the assessor was presented with one presentation of the dilute odor and two blank/control (filtered room air) presentations. The subject's task was to select the presentation that contained the strongest odor by declaring to the test administrator whether the selection was a "guess,""detection," or "recognition," as defined by ASTM E679-91.25 This method of declaration is called "forced choice." The test began with the presentation of an odor concentration below the odor threshold (sub-threshold). Each succeeding level had an odor concentration 2 times higher than the previous one. The assessor evaluated each successive dilution level until he/she detected (and ultimately recognized) the odor. For example, the assessor began at level 4 and sniffed all three presentations at least once each. The assessor then was asked to make a selection as to the strongest odor by guessing if necessary. The assessor then continued to level 5, sniffed the three presentations at least once, and made a selection. The assessor continued to levels 6, 7, and so on, until making at least three consecutive correct choices. The lowest concentration of a series of three or more correctly chosen levels was considered to be the subject's detection threshold, and the recognition threshold was the level, at or above the detection threshold, at which the subject correctly described the odor by giving a description such as animal, animal waste, or related swine operation odor terms. This statistical approach of increasing levels of sample presentation is called an "ascending concentration series." In cases in which the subject indicated the strongest odor and recognition, he/she provided a description of the odor. After threshold measurements were completed, air samples in the Tedlar bags were also evaluated by the subjects at full strength in the laboratory. They were delivered to the nostrils of the subjects at 4 L/min and evaluated using the measures described above for odor intensity, nasal irritation, odor pleasantness, and odor character.

Cotton Swatch Evaluations in the Laboratory. Cotton fabric swatches were used to adsorb odorants and particulates. After a 1- hr exposure to the odorants, the cotton swatches were placed in amber glass bottles. The 30-mL capacity glass bottles had airtight sealed closures, and subjects smelled the headspace above these samples at the level of the nostrils. In the laboratory, odor panelists evaluated the swatch samples for odor and irritation intensity, hedonic properties, and odor quality. Subjects waited at least 3 min between samples. Andersen Dust Impactor Evaluations in the Laboratory. Re-spirable dust (>/=10 [mu]m in diameter) was collected at exhaust fans of swine houses (or at a downwind house edge for naturally ventilated houses) over a 24-hr period using an Andersen Non-Viable Eight-Stage Impactor (Graseby Andersen). Nine different filters were used to adsorb particulates of various size ranges (0-0.43, 0.43-0.65, 0.65- 1.1, 1.1-2.1, 2.1-3.3, 3.3-4.7, 4.7- 5.8, 5.8 -9, and 9-10 [mu]m) for odor evaluations and for measurements of the weight of each particulate size range. Assessment of the glass fiber filters by the odor panel involved the same sample bottles and methodology as that for the cotton swatch samples (see above).

Liquid and Solid Waste Stream Evaluations in the Laboratory. Liquid and solid waste stream samples were collected in the field and evaluated by the odor panel in the laboratory. Samples were collected at each possible stage of the waste stream process. Liquid and solid samples were assessed using the same glass bottles as were used for the swatch and Andersen samples. Sample bottles contained 15 mL of each liquid sample or 5 g of each solid sample, and the same methods as described above for the swatch and Andersen filter samples were used except that the subjects waited 3-5 min between samples, depending on the strength of the samples.

Solid and liquid waste samples that had hydrogen sulfide levels higher than 15 parts per million (ppm) in the sample headspace (as measured by the Jerome meter with an ammonia filter, see Hydrogen Sulfide Measurements below) were not evaluated by the odor panel. Rather, these samples were mixed with deionized water (except for solid samples) to make a 1/10 dilution that was evaluated by the panel; none of these 1/10-diluted samples exceeded 15 ppm hydrogen sulfide in the headspace. In addition, liquid and semi-solid samples were mixed with deionized water to make 1/1000 dilutions for evaluation by the odor panel. In a few cases, 1/100 dilutions were tested.

Instrument Measurements

Most of the field locations evaluated by the subjects were simultaneously assessed for total VOCs, hydrogen sulfide, particulates, and ammonia. For certain physical samples collected in the field and tested in the laboratory, some of these instrument measurements were done as well (see Table 1). All instrument measurements were performed at an elevation of approximately 5 ft (at nose level). Each machine was calibrated regularly on a schedule suggested by the equipment manufacturer. Manufacturers' instructions for routine maintenance of equipment were followed.

VOC Measurements. The VOCs in ambient air in aggregate were measured in real time using a photoionization detector (PID), (ppbRAE, model PGM-7240, RAE Systems). The ppbRAE is a broad- spectrum monitor that can detect VOC concentrations as low as a few parts per billion. Two different discharge ultraviolet (UV) lamps (10.6 and 11.7 eV) were used as high-energy photon sources to ionize the odorants so that they could be measured with the detector. The 10.6-eV lamp can be used to detect a vast number of odorants because most of their ionization potentials are less than 10.6 eV; however, the 11.7-eV lamp was necessary to improve detection of acetic acid (a major contributor to swine odor) as well as to detect compounds with ionization potentials greater than 10.6 eV and lower than 11.7 eV. The ppbRAE calculates an average and a peak rating for each measurement period. The measurement period was the same period as when the field odor panel measurements were taken; the ppbRAE took VOC measurements every second during each measurement period. Before each day of testing with the ppbRAE, the machine was fully charged starting the night before. A water trap prefilter was inserted in the end of the inlet tube to prevent any water or dust from getting into the ppbRAE.

Hydrogen Sulfide Measurements. The Jerome hydrogen sulfide analyzer model 631-X (Arizona Instrument Corp.) was used to measure the airborne concentrations of hydrogen sulfide present in the air. The Jerome analyzer uses a gold film sensor selective for hydrogen sulfide without interference from sulfur dioxide, carbon dioxide, carbon monoxide, and water vapors. Hydrogen sulfide measurements in the field were taken during the same time period as human odor panel measurements. Sampling was done each minute during this time period. An average parts per billion level of hydrogen sulfide, plus the maximum and minimum readings, were recorded. Because ammonia can interfere with the Jerome meter's ability to accurately measure hydrogen sulfide levels, an ammonia filter containing silicon orthophosphate, silicon pyrophosphate, and quartz dust was attached to the sampling inlet during measurements. The accuracy of the Jerome instrument was assured through factory calibration, routine regeneration (heating the gold element of the analyzer to a sufficient temperature to ensure all hydrogen sulfide is eliminated from the sensor), functionality tests with a functional test module (FTM) (Arizona Instrument Corp.), and performance checks with calibration gases diluted with the Model 1010 precision gas diluter (Custom Sensor Solutions).

EPAM 5000 HazDust Particulates Measurements. Particulate concentrations of airborne dust particles 10 [mu]m and smaller (PM^sub 10^) were measured in real time by the HAZ-DUST EPAM-5000 environmental particulate air monitor (Environmental Devices Corp.) that utilizes aerodynamic particle sizing and an in-line filter cassette for gravimetric sampling. The PM^sub 10^ measurements were obtained during the same period of time as when the field odor panel measurements were taken. Sampling was continuous, and the EPAM-5000 took a measurement every second during this time period. The maximum, minimum, and time-weighted average (TWA) were calculated for each measurement period. The accuracy of the HAZ-DUST EPAM 5000 was assured by factory calibration using the National Institute for Occupational Safety and Health gravimetric method, proper charging of the machine, and careful cleansing of the optics in the light scattering chamber with pressurized air or a cotton swab and ethyl alcohol.

Ammonia Measurements. Ammonia measurements were accomplished using the Drager Accuro Gas Detection Pump with a Drager Tube (Drager Safety AG & Co. KGaA) specific for ammonia. The measuring range of one type of tube is 0.25-3 ppm, and the measuring range of the other type of tube is 2-30 ppm. Both tubes use a pH indicator to react with ammonia. The indicator changes from yellow to blue as it is exposed to ammonia. The length of blue discoloration is read on the exposed tube and on the side of the tube appears a series of marks that indicate concentration of ammonia in the tested ambient air. The Drager Accuro gas detection pump was used to draw air through the sampling tube. Each stroke of the pump draws 100 +- 5 mL of air. The tube that measures ammonia levels between 0.25 and 3 ppm requires 10 strokes of the pump and the tube that measures levels between 2 and 30 ppm requires 5 strokes of the pump.

Technologies Assessed

Eighteen different technology sites were evaluated during this project. The technologies included the following: (1) standard lagoon technology with naturally ventilated houses (Stokes farm, control farm 1); (2) standard lagoon technology with houses that utilized fan ventilation (Moore farm, control farm 2); (3) in- ground ambient temperature anaerobic digester with energy recovery and greenhouse vegetable production system (Barham farm); (4) manure solids conversion to insect biomass for value-added processing into animal feed protein meal and oil system (black soldier fly larvae demonstration project at North Carolina State University [NCSU]); (5) aerobic blanket system and aerobic digester (Carrolls farm); (6) solids separation/combustion for energy and ash recovery centralized system (Corbett farms 1 and 4); (7) upflow biofil-tration system (EKOKAN); (8) belt manure removal and gasification system to thermally convert dry manure to a combustible gas stream for liquid fuel recovery (demonstration project at Grinnells Laboratory, NCSU); (9) permeable cover anaerobic digester and aerobic digester (Har- rells farm); (10) solids separation/constructed wetlands system (Howard farm), (11) sequencing batch reactor (SBR) system (Hunt farm); (12) gasifier to treat belt-separated solids (Koger gasifier demonstration project at NCSU); (13) belt system for manure removal (demonstration project at Lake Wheeler Road, NCSU); (14) high solids high temperature anaerobic digester system (OR-BIT); (15) solids separation/reciprocating wetland technology system (ReCip); (16) solids separation/nitrification denitrification/soluble phosphorus removal/solids processing system (Super Soil Systems USA); (17) composting solids (Super Soils Composting Site); and (18) mesophilic digester, permeable lagoon cover, aerobic blanket and microturbine generator (Vestal Farms ISSUES Project-Innovative Sustainable Systems Utilizing Economical Solutions).

Computer Model Used to Predict and Compare Odor Dispersions

The model used to predict the trajectory of odorous emissions from the odor sources at each technology test site has been utilized to predict the long-distance dispersal of seeds by wind.26-29 This model was found to fit the experimental data downwind more effectively than other atmospheric dispersion models developed by the U.S. Environmental Protection Agency (EPA). The model utilized here, which is based on stochastic differential equations for turbulent diffusion that utilize a Eulerian-Lagrangian approach,27,28 has been adapted to predict dispersion of odor.29 The model utilizes the spatial distribution of odor intensities at multiple emission sources (in steady-state conditions) to predict the spatial distribution of odor downwind under a variety of meteorological conditions (e.g., during the daytime when the boundary layer is usually turbulent because of ground-level heating from solar short-wave radiation, and during the evening when deep surface cooling through long-wave radiation to space creates a stable boundary layer). To predict the trajectory and spatial distribution of odor downwind from each facility, the geographical area containing the odorant sources at each facility was partitioned into 10-m^sup 2^ grids on the basis of satellite photographs and architectural drawings.29 The relative odorant concentrations assigned to each grid point were determined from on-site intensity measurements made by the trained odor panel using the nine-point rating scale; that is, the input to the model was the average intensity rating on the nine-point rating scale of the human odor panel at odor sources. The intensity ratings rather than the threshold values were used for modeling because some of the threshold values could not be determined because of the detection limits of the Scentometer and Nasal Ranger instruments. Land application such as spraying at control (and alternative technology) farms was not included in the modeling because of the intermittent nature of the process. For each test site, the mean of the non-zero elements of the grid points at 200 and 400 m from the source of odor closest to the downwind property line was determined for four directions (north, east, south, and west). The reason for calculating mean odor intensity from only the positive (e.g., non- zero) odor intensities at each grid point at the 200- and 400-m distances was that any odorous plumes that reach these distances are significant. That is, any odorous plumes that reach these distances should not be negated or made less significant by the presence of nonodorous grid points; in reality these odorous peaks are present and perceptible. Calculations at 200- and 400-m distances were computed for all four directions (north, south, east, west). For north (i.e., with a wind blowing to the north), the mean odor intensity was calculated at 200 m north of the northernmost odor source. The mean over each of the four directions was then calculated to remove directional dependence in rankings.



The technology sites were compared with respect to mean odor intensity at 200 and 400 m from the edge of the last odorous source in the direction of the wind for both day and night averaged over the four directions. The complete farms were also modeled with and without the swine housing and ranked using the same model to determine the relative contribution of houses and waste technologies to the odor downwind. The results of the dispersion modeling are shown in Table 2. This table gives predicted odor intensities for all test sites, for night and day, at 200 and 400 m downwind averaged over north, south, east, and west. Sample dispersions for the Stokes farm, which utilizes standard lagoon technology with naturally ventilated houses, and the Super Soils farm, which utilizes a solids separation/nitrification-denitrification/soluble phosphorus removal/solids processing system, are shown as well in Figures 1 and 2.

Overall, Table 2 shows that the predicted odor dispersion is far less during daytime than nighttime at all farms, which is consistent with field reports from individuals living nearby. All test sites including technologies with and without houses had an average odor intensity below 1 (very weak) at 400 m downwind during the day. Ten sources had an average odor intensity below 1 (very weak) at 200 m during the day: Corbett 4 no houses, ReCip, Super Soils Technology, Super Soils Composting, Lake Wheeler Road Belt System, Grinnells, ReCip no houses, ORBIT, Koger gasifier, and Black Soldier Fly. The remainder of the sites had average odor intensities during the day at 200 m that were below 2 (weak). Odors were 2.9 times greater on average at nighttime compared with daytime at 200 m and 6.7 times greater at 400 m at nighttime. Only the Koger gasifier had an odor intensity below 1 at 200 m at night. The remainder of the sites had nighttime odor intensities at 200 m that reached a maximum of 4.41 (moderate to moderate strong). Only point sources (Koger gasifier, Black Soldier Fly, ORBIT, Grinnells) and ReCip without the houses had odor intensities below 1 at 400 m at night. The remainder of the sites had odor intensities at night at 400 m that reached a maximum of 3.64 (moderate).

It should be noted that in some cases in Table 2, the odor intensities predicted by the model are higher for certain farms without houses than with houses. This is because the location of the 200- and 400-m sites for downwind predicted odor intensities as calculated were based on the location of the odor source closest to the direction of the dispersion (north, east, south, or west), and sometimes when the houses are removed, a significant odor source (often with a large surface area) then becomes closer to the 200- and 400-m downwind location.

Comparisons among Human and Instrument Field Measures

Table 3 shows the mean values for human and instrument ratings in the field at house fans, house edges, lagoons and tanks, and downwind from sources. Both mean intensity judgments and mean odor threshold measurements indicate that odor was strongest at house fans followed by house edges, lagoons and tanks, and downwind measurements. VOCs, hydrogen sulfide, and ammonia were also greatest at house fans. The highest mean odor intensity for the panel recorded during the project (7.67) was found at house fans, followed by a rating of 7.5 at a solids separator. Odor thresholds seldom exceeded 350 D/T, which is the upper D/T limit of the Scentometer. However, D/T values in excess of 350 were found sporadically using the Nasal Ranger olfactome-ter with the odor thresholds; for example, at house fans, solids separator, feces cart, and a house edge reaching 500 D/T. The highest level of VOCs measured by the ppbRAE were at the house fans (191,000 parts per billion [ppb]), followed by the Black Soldier Fly larva basin (143,000 ppb), a manure reception pit (46,600 ppb), and a lagoon (24,700 ppb). The highest hydrogen sulfide measurements obtained with the Jerome meter (without an ammonia filter) were at a solids separator (2700 ppb), followed by house fans (460 ppb), a lagoon (390 ppb), and house edges (250 ppb). The highest hydrogen sulfide measurement obtained with the Jerome meter (with an ammonia filter) was 2200 ppb at a solids separator, which (along with the means in Table 3) suggests that ambient ammonia at levels of 0.5 ppm and above can elevate Jerome readings. The highest particulate values for PM^sub 10^ obtained with the EPAM were at a house edge (3.16 mg/m3) and at an equalization tank (2.83 mg/m 3).

Table 4 shows intercorrelations among the human odor measurements and the instrument measurements taken in the field, including measurements in the immediacy of odor sources (e.g., house fans) as well as at varying distances downwind. There were six noteworthy correlations (see footnote a in Table 4); the odor intensity ratings were significantly correlated with Scentometer odor threshold ratings (R2 = 0.60). When odor intensity ratings are plotted against threshold ratings, the equation of best fit is: T = 3.2109e0.7044I, where T is the odor threshold and I is the odor intensity. Thus an intensity rating of 1 corresponds to 6.49 D/T (or 6.49 times above threshold), 2 corresponds to 13.14 D/T, 3 corresponds to 26.57 D/T, 4 corresponds to 53.74 D/T, 5 corresponds to 108.70 D/T, and 6 corresponds to 219.85 D/T. Odor intensity and odor detection thresholds were also moderately correlated with hydrogen sulfide levels with (r = 0.30 and 0.32, respectively) and without (r = 0.34 and 0.36, respectively) the ammonia filter. Hydrogen sulfide Table 4. Correlation matrix for human and instrument field data. levels were very highly correlated (0.98) with and without the NH3 filter. When correlations were performed between human odor assessments and instrument ratings directly at the swine houses (without inclusion of downwind or lagoon assessments), the values increased to r = 0.56 (for odor thresholds and hydrogen sulfide levels at houses) and r = 0.32 (for odor thresholds and NH3 levels at houses). That is, the relationship of odor to hydrogen sulfide and ammonia is strongest at the houses, is much lower at lagoons, and dissipates downwind.

Two types of dust measurements obtained at the houses were statistically equivalent; that is, PM^sub 10^ using the EPAM and PM^sub 10^ using the Andersen impactor. The mean PM^sub 10^ concentration of particles measured in real time by the HAZ-DUST EPAM-5000 at the house fans or house edge (for naturally ventilated houses) for all facilities was 0.195 +- 0.165 mg/m3 . The mean PM^sub 10^ concentration at the houses for all facilities obtained by summing the weight of the particle sizes across the Andersen filters was 0.237 +- 0.306 mg/m3.

Additional comparisons indicated that field odor intensity ratings by the trained odor panel were significantly greater in magnitude (mean = 5.83) than mean odor intensity ratings of air bag samples (from the same locations) evaluated by the trained panel in the laboratory (mean = 3.94). Furthermore, odor field intensity ratings were greater in magnitude (mean = 5.72) than intensity ratings of swatch samples (from the same locations) evaluated in the laboratory (mean = 2.90). The correlation between field odor intensity ratings and intensity ratings from air bag samples was 0.45. The correlation between field odor intensity and swatch samples was 0.44. Two other correlations were insignificant: (1) irritation intensity ratings in the field versus irritation measured with lateralization device (r = 0.001), and (2) odor intensity ratings and particle size up to 10 [mu]m from the Andersen dust impactor (r = 0.002). Respirable particulate (PM^sub 10^) levels, as measured by both the Andersen dust impactor and the EPAM HazDust, were not at all correlated with human odor data (r = 0.002). Furthermore, ambient temperature was unrelated to odor intensity (r = 0.000). DISCUSSION

Several major findings emerged from this study. First, an overview of the dispersion results from the Eulerian-La-grangian model indicates that odors occurred offsite at all full-scale farms at night but not during the daytime independent of the waste technology used. Specifically, although odor was not predicted to extend beyond 200 m (with a few exceptions) or 400 m downwind during the daytime at any of the test sites, the modeling indicated that odor did extend onto neighboring property in the evenings (with the extent dependent upon specific property line orientations for each site). The elevated odor during the evening results from diurnal changes in meteorological conditions; that is, the layer of air above the earth's surface is usually turbulent during the daytime, but creates a stable (nocturnal) boundary layer during the evening when deep surface cooling occurs through long-wave radiation to space. For a technology to be considered environmentally superior from the standpoint of odor, according to the agreement with the Attorney General of North Carolina,22 it was mandated that the technology "substantially eliminates the emission of odor that is detectable beyond the boundaries of the parcel or tract of land on which the swine farm is located." If an operational definition of "substantial elimination" of odor emissions at the property boundary is defined as odorous emissions that are equal to or less than the weak rating of 2 on the 0-8 odor intensity scale, then all facilities met the odor performance criteria during the daytime at 200 and 400 m. However, at night only the demonstration projects (Koger gasifier, Black Soldier Fly, ORBIT, Lake Wheeler Road Belt System, Grinnells) and ReCip technology modeled without the swine houses met the performance criteria at 200 m along with Super Soils Composting and Super Soils Technology at 400 m. These findings indicate that, even at 400 m downwind, odors emanating from swine facilities during nighttime can be significant. This is in part because the swine housing accounts for a substantial amount of the odor downwind.

Second, the trajectory and spatial distribution of the odor dispersions were found to depend on multiple factors in addition to the type of technology and the meteorological conditions (daytime vs. nighttime) including: (1) the amount of surface area of odor sources at each facility, and (2) the geometrical or spatial distribution of the relative concentrations of odor intensity. When odor sources were aligned extensively in the direction of the wind, the odor carried further but affected a narrower cross-section of neighbors than when the sources were spatially spread out.

Third, the data show that there are no surrogates or proxies for field odor measurements; that is, no instrument or collection device accurately predicts odor intensity in the field downwind. Although correlations between human odor data and hydrogen sulfide and ammonia are greatest at houses, this relationship disappears downwind. The correlations between odor and hydrogen sulfide found here are lower than those reported by other researchers at houses and lagoons.30,31 The finding in this study that correlations between odor and hydrogen sulfide are greater at the houses than at other odor sources or downwind indicates that hydrogen sulfide levels cannot be used as a proxy for odor levels downwind. Overall, instrument measurements in this study were not predictive of odor downwind, which is consistent with a previous study using gas chromatography, coupled with mass spectrometry, in which it was difficult to correlate individual compounds with swine odor.16 Importantly, the true odor intensity can only be measured at present by human observers in the field. Use of cotton swatches and Tedlar air bags to collect odorants for evaluation in the laboratory always yields significantly lower odor ratings than field odor data from the same locations and odor sources. Thus cotton swatches and Tedlar air bags can only be used to give relative differences between technologies but not absolute intensity values. Overall, this study indicates that there is a need to develop a reliable, portable, and sensitive instrument (e.g., an "electronic nose"32) for field measurement of odorous emissions in real time to assist in the development and implementation of cost-effective odor-abatement techniques that will enable operators of livestock operations to meet performance standards.


This paper describes a rational method to compare alternative technologies for reducing odor emissions using a model for odor dispersion. The main conclusion from these data is that odor ratings obtained by human panelists in the field are the best input for modeling odor downwind of swine facilities. Instrumental measurements including concentrations of VOCs, hydrogen sulfide, ammonia, and PM^sub 10^ were not found to accurately predict human odor intensity at most odor sources or downwind. Although strong odor ratings were always associated with elevated concentrations of hydrogen sulfide, strong odor ratings also occurred when hydrogen sulfide ratings were low. Thus, hydrogen sulfide concentrations are not a reliable surrogate for odor intensity downwind of swine operations.


1. Reisner, A.E. Newspaper Coverage of Controversies about Large- Scale Swine Facilities in Rural Communities in Illinois; J. Animal Sci. 2005 , 83, 2705-2712.

2. Schiffman, S.S. Livestock Odors: Implications for Human Health and Well-Being; J. Animal Sci. 1998, 76, 1343-1355.

3. Schiffman, S.S.; Walker, J.M.; Dalton, P.; Lorig, T.S.; Raymer, J.H.; Shusterman, D.; Williams, C.M. Potential Health Effects of Odor from Animal Operations, Wastewater Treatment, and Recycling of Byprod-ucts; J. Agromed. 2000, 7, 7-81.

4. Wing, S.; Wolf, S. Intensive Livestock Operations, Health, and Quality of Life among Eastern North Carolina Residents; Environ. Health Per-spect. 2000, 108, 233-238.

5. Thu, K.; Donham, K.; Ziegenhorn, R.; Reynolds, S.; Thorne, P.S.; Sub-ramanian, P.; Whitten, P.; Stookesberry, J. A Control Study of the Physical and Mental Health of Residents Living near a Large-Scale Swine Operation; J. Ag. Safety Health 1997, 3, 13-26.

6. Iowa Concentrated Animal Feeding Operations Air Quality Study; Prepared for Governor Tom Vilsack, Des Moines, IA, by Iowa State University and the University of Iowa Study Group; University of Iowa: Iowa City, IA, 2002.

7. Schiffman, S.S.; Sattely-Miller, E.A.; Suggs, M.S.; Graham, B.G. The Effect of Environmental Odors Emanating from Commercial Swine Operations on the Mood of Nearby Residents; Brain Res. Bull. 1995, 37, 369-375.

8. Avery, R.C.; Wing, S.; Marshall, S.W.; Schiffman, S.S. Odor from Industrial Hog Farming Operations and Mucosal Immune Function in Neighbors; Arch. Environ. Health 2004, 59, 101-108.

9. Schiffman, S.S.; Studwell, C.E.; Landerman, L.R.; Berman, K.; Sundy, J.S. Symptomatic Effects of Exposure to Diluted Air Sampled from a Swine Confinement Atmosphere on Healthy Human Subjects; Environ. Health Perspect. 2005, 113, 567-576.

10. Merchant, J.A.; Naleway, A.L.; Svendsen, E.R.; Kelly, K.M.; Burmeister, L.F.; Stromquist, A.M.; Taylor, C.D.; Thorne, P.S.; Reynolds, S.J.; Sand-erson, W.T.; Chrischilles, E.A. Asthma and Farm Exposures in a Cohort of Rural Iowa Children; Environ. Health Perspect. 2005, 113, 350-356.

11. Dosman, J.A.; Lawson, J.A.; Kirychuk, S.P.; Cormier, Y.; Biem, J.; Koehncke, N. Occupational Asthma in Newly Employed Workers in Intensive Swine Confinement Facilities; Eur. Respir. J. 2004, 24, 698 702.

12. Von Essen, S.; Donham, K. Illness and Injury in Animal Confinement Workers; Occup. Med. 1999, 14, 337-350.

13. Design of Anaerobic Lagoons for Animal Waste Management; ASAE Standard EP403.3; American Society of Agricultural and Biological Engi-neers: St. Joseph, MI, 2004.

14. Zahn, J.A.; Hatfield, J.L.; Do, Y.S.; DiSpirito, A.A.; Laird, D.A.; Pfeiffer, R.L. Characterization of Volatile Organic Emissions and Wastes from a Swine Production Facility; J. Environ. Qual. 1997, 26, 1687-1696.

15. Cai, L.S.; Koziel, J.A.; Lo, Y.C.; Hoff, S.J. Characterization of Volatile Organic Compounds and Odorants Associated with Swine Barn Particulate Matter Using Solid-Phase Microextraction and Gas Chromatography Mass Spectrometry- Olfactometry; J. Chromatogr. A 2006 , 1102, 60-72.

16. Schiffman, S.S.; Bennett, J.L.; Raymer, J.H. Quantification of Odors and Odorants from Swine Operations in North Carolina; Ag. Forest Meteorol. 2001, 108, 213-240.

17. Blunden, J.; Aneja, V.P.; Lonneman, W.A. Characterization of Non Methane Volatile Organic Compounds at Swine Facilities in Eastern North Carolina; Atmos. Environ. 2005, 39, 6707-6718.

18. O'Neill, D.H.; Phillips, V.R. A Review of the Control of Odour Nuisance from Livestock Buildings. Part 3: Properties of the Odorous Substances which have been Identified in Livestock Wastes or in the Air around Them; J. Ag. Engineer. Res. 1992, 53, 23-50.

19. Wright, D.W.; Eaton, D.K.; Nielsen, L.T.; Kuhrt, F.W.; Koziel, J.A.; Spinhirne, J.P.; Parker, D.B. Multidimensional Gas Chromatography Olfactometry for the Identification and Prioritization of Malodors from Confined Animal Feeding Operations; J. Ag. Food Chem. 2005 , 53, 8663-8672. 20. Hammond, E.G.; Heppner, C.; Smith, R. Odors of Swine Waste La-goons; Ag. Ecosyst. Environ. 1989, 25, 103-110.

21. Schiffman, S.S.; Williams, C.M. Science of Odor as a Potential Health Issue; J. Environ. Qual. 2005, 34, 129-138.

22. Attorney General of North Carolina Agreement; 2000; available at http:// agreement.pdf (accessed May 11, 2006).

23. Stevens, J.C.; Dadarwala, A.D. Variability of Olfactory Threshold and its Role in Assessment of Aging; Percept. Psychophys. 1993, 54, 296-302.

24. Atlas of Odor Character Profiles; DS 61; American Society for Testing and Materials: Philadelphia, PA, 1992.

25. Standard Practice for Determination of Odor and Taste Thresholds by a Forced-Choice Ascending Concentration Series Method of Limits; E679-91; American Society for Testing and Materials: Philadelphia, PA, 1991.

26. Nathan, R.; Katul, G.G.; Horn, H.S.; Thomas, S.M.; Oren, R.; Avissar, R.; Pacala, S.W.; Levin, S.A. Mechanisms of Long-Distance Dispersal of Seeds by Wind; Nature 2002, 418, 409-413.

27. Hsieh, C.I.; Katul, G.G.; Schieldge, J.; Sigmon, J.T.; Knoerr, K.K. The Lagrangian Stochastic Model for Fetch and Latent Heat Flux Estimation above Uniform and Nonuniform Terrain; Water Resources Res. 1997, 33, 427-438.

28. Katul, G.G.; Albertson, J.D. An Investigation of Higher- Order Closure Models for a Forested Canopy; Boundary Layer Meteorol. 1998, 89 , 47-74.

29. Schiffman, S.S.; McLaughlin, B.; Katul, G.G.; Nagle, H.T. Eulerian- Lagrangian Model for Predicting Odor Dispersion Using Instrumental and Human Measurements; Sensors Actuators B 2005, 106, 122-127.

30. Gay, S.W.; Schmidt, D.R.; Clanton, C.J.; Janni, K.A.; Jacobson, L.D.; Weisberg, S. Odor, Total Reduced Sulfur, and Ammonia Emissions from Animal Housing Facilities and Manure Storage Units in Minnesota Source; Appl. Engineer. Ag. 2003, 19, 347-360.

31. Lim, T.T.; Heber, A.J.; Ni, J.Q.; Sutton, A.L.; Shao, P. Odor and Gas Release from Anaerobic Treatment Lagoons for Swine Manure; J. Environ. Qual. 2003, 32, 406-416.

32. Handbook of Machine Olfaction: Electronic Nose Technology; Pearce, T.C.; Schiffman, S.S.; Nagle, H.T.; Gardner, J.W., Eds.; Wiley-VCH: Wein-heim, Germany, 2003.

Susan S. Schiffman and Brevick G. Graham

Department of Psychiatry, Duke University Medical Center, Durham, NC

C. Mike Williams

Department of Poultry Science, North Carolina State University, Raleigh, NC

About the Authors

Dr. Susan Schiffman has spent her entire career with Duke University, ultimately as a professor of Medical Psychology in the medical school as well as a professor in the Department of Psychological and Brain Sciences. The focus of her olfactory research over the last 13 yr has been assessment of malodorous aerial emissions, predominantly those from swine operations. Brevick Graham has served as a research technician in Dr. Schiffman's laboratory. Dr. C. Mike Williams is a professor of Poultry Science at North Carolina State University. Please address correspondence to: Susan Schiffman, Department of Psychiatry, Duke University Medical Center, 54212 Woodhall Building, Durham, NC 27710 3259; phone: 1-919-660-5657; e-mail: [email protected]

Copyright Air and Waste Management Association Sep 2008

(c) 2008 Journal of the Air & Waste Management Association. Provided by ProQuest LLC. All rights Reserved.