EMISSION FACTOR DEVELOPMENT FOR PARTICULATE MATTER FROM A BROILER HOUSE

The Canadian Society for Bioengineering The Canadian society for engineering in agricultural, food, environmental, and biological systems. La Société Canadienne de Génie Agroalimentaire et de Bioingénierie La société canadienne de génie agroalimentaire, de la bioingénierie et de l environnement 1 Paper No. 0-1S EMISSION FACTOR DEVELOPMENT FOR PARTICULATE MATTER FROM A BROILER HOUSE Taylor Roumeliotis; Bill Van Heyst School of Engineering, University of Guelph, 0 Stone Rd. Guelph, ON. N1G W1 Written for presentation at the CSBE/SCGAB 00 Annual Conference Edmonton Alberta July 1-1, 00 1 1 1 1 1 1 1 1 0 1 Abstract The recent intensification of the livestock industry has raised concerns regarding the air pollution generated from the various animal housing operations. One pollutant of concern is particulate matter (PM), which is capable of lodging itself deep in the respiratory tract and causing serious detrimental respiratory effects to the workers and livestock. Little information is available in the literature that characterizes the emissions of PM from various types of livestock houses that also incorporate daily and seasonal variances typical of Canadian climates. A reproducible and reliable methodology for characterizing the emissions of PM has been developed based on various operational settings in a broiler house. Three DustTrak aerosol monitors measured PM concentrations for various particle size ranges and an Electronic Balancing Tool (EBT) with capture hood measured exhaust rates. Continuous PM monitoring from several bird production cycles has verified the reliability of the methodologies and established a procedure for characterizing the emissions from the poultry house. The measured results from the broiler house indicate that the estimates in the literature are not representative of the colder climate in south western Ontario. Findings also indicate that the bird s age and their activity levels strongly influence the emissions from the broiler house suggesting that the use of constant emission rates do not accurately represent the trend in PM emissions. Emission factors and emission inventories have been developed to describe the variations in daily and seasonal emissions. Papers presented before CSBE/SCGAB meetings are considered the property of the Society. In general, the Society reserves the right of first publication of such papers, in complete form; however, CSBE/SCGAB has no objections to publication, in condensed form, with credit to the Society and the author, in other publications prior to use in Society publications. Permission to publish a paper in full may be requested from the CSBE/SCGAB Secretary, PO Box 1, RPO McGillivray, Winnipeg MB RT S or contact bioeng@shaw.ca. The Society is not responsible for statements or opinions advanced in papers or discussions at its meetings.

1 1 1 1 1 1 1 1 0 1 INTRODUCTION The recent intensification of livestock housing operations in Canada has led to the initiation of concentrated animal feeding operations (CAFOs). CAFOs provide automated maintenance of the facility, which allows livestock operators to output considerably more product from one building. However, with this increase in production comes a decrease in indoor air quality. Of growing concern is the animal housing operations contribution to the local air quality in the form of particulate matter (PM). PM produced from livestock housing operations can contain viable particles (bioaerosols), such as pathogenic bacteria, viruses, and endotoxins, which are capable of lodging deep within the lungs and have the potential to cause serious detrimental respiratory effects (Pillai and Ricke 00). It has been established that bioaerosols can survive for a considerable length of time and distance, which raises concerns for the quantity of PM emissions from livestock facilities. It is believed that poultry operations produce the highest concentrations of bioaerosols of all livestock housing operations (Pallai and Ricke 00; CAC 000). The amount of PM emitted from poultry houses is dependent on several operational and maintenance conditions. In an attempt to characterize the emissions from several poultry operations, averaged yearly emission factors for some of the common size classifications have been developed based on limited experimental data from various geographic regions (Van Heyst 00; Lacey et al. 00; Takai et al. 1; Wathes et al. 1). However, variations in emissions exist from a range of seasonal operating conditions and animal activity levels which may bias the emission factors. Since the yearly emission factors were averaged from limited data, their use in accurately representing a livestock emission inventory outside of Canada is questionable. Also, the collected data from several geographic regions likely do not accurately represent the unique Canadian climate. Since little information is available for PM and no threshold levels have been established, experiments 1

are needed to assess the current emission levels in order to understand the future benefits of introducing best management practices and/or PM controls in the agricultural sector. Study Objectives One aspect of livestock emissions that is not well characterized is the lack of PM size classifications data. In order to advance the current knowledge regarding livestock emissions, an intensive study at a broiler house was initiated to develop consistent emission factors by continuously monitoring PM emissions for three size classifications. The objectives of this study were threefold: 1. Standardization of instrumentation and measurement techniques for the characterization of airborne particulate matter emissions from a livestock building. This objective 1 1 1 1 1 1 1 1 0 1 encompasses the measurements of both the airborne PM concentrations and the volumetric exhaust rates from the livestock building;. Quantification of emissions rates characterized by the changes in daily and seasonal activity in a livestock house; and. Development of PM emission factors from a livestock house which can be used to estimate an emission inventory for a portion of the Canadian agricultural sector. Emission factors will be developed for several common size classifications: specifically PM, PM., and PM 1. If rational and fair standards for environmental air quality are going to be introduced for the agricultural sector, there must first be improvements on data quantity and quality. The completion of these objectives should provide the first steps required to initiate the development of appropriate emission inventories for the poultry industry.

1 1 1 1 1 1 1 1 0 1 MATERIALS AND METHODS Since an overwhelming quantity of data is required to describe the emissions from the entire agricultural industry, it was decided to conduct the initial experiments at a broiler house with the anticipation that this methodology could be transferred to other livestock housing operations at a future date. The selected facility was a commercial broiler house constructed in 00, that is located off of county road at the Southwestern corner of Guelph, Ontario, Canada. The broiler house is a single storey, litter floor facility that is. m long by 1. m wide by. m tall (refer to Fig. 1 for the broiler house schematic). The house was carefully selected to be mechanically ventilated with no natural ventilation. Twenty-eight ducts with exhaust fans run along the length of the southern side of the building. There are fans with a diameter of 1. m, 1 fans with a diameter of 0. m, and fans with a diameter of 0. m. Of these fans, four of the smallest fans are variable speed fans, while the remainder are single speed fans with on/off controls. Automated adjustable louvers are located along the northern side of the house. Ventilation rates and three pipe heaters control the internal temperature of the house. Temperature is maintained at. o C and reduced by an average of 0. o C/day until the temperature reaches 1. o C. Feed and water supply to the birds is also fully automated. This commercial house can accommodate 1,000 to,000 birds during one production cycle. The average stocking density of birds in the house is. birds/m with an average residence time of approximately days (½ weeks). The broilers are free to roam throughout the entire house for the entire duration of the production cycle. During the ½ weeks, the birds will mature from 0 g to an average weight of. kg. After each production cycle, a week period is used to clean the house for biosecurity reasons and reapply a new litter floor. During this time, the louvers are fully opened to replenish the house with clean ambient air.

1 1 1 1 1 1 1 1 0 1 Instrumentation Indoor PM concentrations were measured with optical particle counting instruments, namely three DustTrak aerosol monitors. Three locations along the length of the house were used to verify the uniformity of PM concentrations throughout the house. It was established that PM concentrations are relatively uniform along the length of the house and hence, measurements were taken m from the control room to minimize broiler disturbances. Measurements were taken from an environmental enclosure with a protruding multipoint sampling line. An adjustable stand allowed the sample inlet to be reposition to different heights. At all three locations along the length of the building, the multipoint sample inlet was located. m from the exhaust side of the broiler house. It was assumed that the concentration measured at this distance would characterize the concentrations being exhausted from the fans. A PIXE cascade impactor (PCI) was also used to verify the size fraction of the PM. The PCI provided the precise particle size distribution in the broiler house. The impactor consisted of stages including an after-filter (AF). The PCI separates particles into the following size fractionations: greater than 1, 1-, -, -, -1, 1-0., 0.-0., 0.-0.1, and 0.1-0.0 µm (PIXE International Corporation 00). With these size fractionations, the relative mass fractions of each of the common size classifications could be used to evaluate the size fractionations from the particle counting instruments by manipulating the results from the PCI. The Alnor balometer kit, consisting of an electronic balancing tool (EBT) and capture hood, was used to measure the exhaust rates from the exhaust ducts. The instrument s capture hood was connected directly to an exhaust duct, downstream of the fans. A customized skirt isolated the flow sensors from influences on flow rate measurements from the surrounding air currents. Pressure differentials were measured across a 1 point velocity matrix to determine the average velocity over a given area. A complete range of flow rates from each fan type were fully

quantified during the bird production cycles. The broiler house s control room dictated which fans were used at any given time, and this information was used in conjunction with the balometer s measurements for each specific fan to estimate the overall ventilation rate for the house. 1 1 1 1 1 1 Estimates of PM Emissions Sampling was conducted throughout four complete bird production cycles from October th, 00 to June rd, 00. The four production cycles characterized the environmental conditions typically encountered in southern Ontario. The instruments were programmed to collect running averages of PM concentrations and flow rates every minutes. Hence, an estimate of the PM emissions could be calculated each minutes for the duration of each bird production cycle. The emissions were estimated with the equation: E = Q C (1) vent Where E is the emission rate of PM (g/hr), Q vent is the total house ventilation rate from all the exhaust ducts (m /hr), and C is the mean concentration of PM (g/m ). The emission factors developed from this study were expressed in units of kg of PM per 00 birds from a single production cycle. 1 1 RESULTS 0 1 Ventilation Rates The rate at which air was exhausted from the house varied with the age of the birds and ambient temperatures. During the initial days of a production cycle, the four small variable speed fans were the only fans used to ventilate the house. Their speed was managed to preserve the

1 1 1 1 1 1 1 internal temperature, which was highest at the beginning of a production cycle and consequently ventilation rates were lowest to prevent cool incoming air from reducing the internal temperature. As the birds aged, the internal temperature setting was gradually decreased and concurrently, the bird weight increased, resulting in the production of more body heat. Consequently, higher ventilation rates were needed to cool the internal temperature. The four variable speed fans were used until they reach their maximum ventilation rate. After this time, additional on/off fans were used to continue increasing the house ventilation rate. The computer control panel for the house dictated the number of fans used as well as the rate of the variable speed fans. Since the control panel records the percentage of the house s ventilation capacity, it was possible to estimate the house ventilation if the flow rate for each size of fan was fully quantified and the number of fans used was correlated to a particular ventilation percentage. With this data, an accurate real time house ventilation rate was predicted for use in emission estimates. Hence, the balometer was used to estimate the range of flow rates for the variable speed fans and a mean ventilation rate for the on/off fans. Two small variable speed fans were quantified; one fan was used continually through a production cycle while the other was used less frequently. The measurements for the entire range of flow rates from both of these small fans are shown in Fig.. The medium and large sized fans were quantified in a similar manner. The use of the larger fans varied greatly since their operation was dependent on the temperature difference between the 1 inside and outside the house. During the winter months, several medium and large sized fans 0 1 were not used at all. Moreover, even in the summer months, several fans were only used in the latter stages of a production cycle. For this reason, the flow rates from three medium sized fans and two large exhaust fans were measured to quantify any differences in flow rates due to fan wear. Figures and give the flow rate measurements as well as the calculated mean flow rate for the medium and large exhaust fans, respectively.

1 1 1 1 It was evident from the flow measurements from several fans that there is some extent of wear on the fan: flow rates were reduced up to 1% for frequently used fans compared to infrequently used fans. For this reason, the correlation between the percentage of house ventilation capacity and the actual flow rate from the broiler house is not linear. Since there is greater wear on frequently used fans, their flow rate will be overestimated by the control panel. Figure displays the correlation between the percentage of house ventilation capacity and the estimated flow rate. The correlation was determined to be: Q 0.1V +.V +.1 () vent = p p Where V p is the percentage of house ventilation (%). This correlation was used for estimating Q vent with Eq. 1 for real time emission estimates. The correlation between the percentage of house ventilation capacity and estimated flow rate has a coefficient of determination (R ) of 0.. Once the correlation was established, the total flow rate for three bird production cycles was quantified. Figure displays the complete real time ventilation rates for three bird production cycles. Since the ventilation rates are dependent on the difference in temperature inside and outside the house, the ventilation is highest in the summer and lowest in the winter. 1 1 1 1 0 1 Particle Concentrations Diurnal patterns were apparent during all seasons. There was a prompt reduction in airborne particulates during the night followed by a rapid increase early in the morning. Due to the promptness of the changing PM concentrations, it was established that this pattern was not a result of environmental deviations but rather a result of the change in bird activity from the lighting schedule. The house control panel switched the lights off for one hour per day for the first five days and roughly the last 1 days of a production cycle while six hours per day was

1 1 1 1 1 1 1 1 0 1 delegated for the days in between. Since the diurnal pattern was relatively brief during the first five days and last 1 days, it was confirmed that the diurnal variations in PM concentrations were connected to the lighting schedule. Figure displays two plots of daily concentrations that emphasize the changes in the diurnal patterns for one hour and six hours of dark. From these plots, it is clear that diurnal patterns are strongly correlated to the lighting schedule. Environmental parameters may cause the PM concentration in the broiler house to decrease at night slightly as evident from the 1 hour of dark plot but its effect is minimal in comparison to effects from lighting. This diurnal trend must be characterized in the emission factors if they are going to accurately describe the behaviour in emissions from the broiler house. In all four production cycles, airborne PM concentrations increased as the birds aged. The initial concentrations observed during all seasons were consistent at roughly 0., 0., and 0.0 mg/m for PM, PM., and PM 1, respectively. The concentrations of all particle size ranges increased steadily for the first 1 days of the production cycle. After this time, the respective PM concentrations increased up to their peak concentration. This concentration was essentially sustained for the remainder of the production cycle. However, there was a significant discrepancy between the winter and summer concentrations especially towards the end of the production cycle. This was likely a result of the increased ventilation rates in the summer that exhausted more PM from the broiler house. PM concentrations for winter and summer production cycles are displayed in Fig.. In order to verify the accuracy of the particle counter instruments, the PCI was used to confirm the precise particle size distribution. The results from the PCI were used in combination with the particle counter instrumentation to estimate the contribution of PM. and PM 1 from the PM emission estimates during the summer production cycle. Three PCI tests were performed during two bird production cycles. The results from each of the tests were consistent with the

measurements obtained from the DustTrak aerosol monitor. Figure displays the particle size distribution represented as a cumulative percent less than the particle size (dp) obtained from the three PCI results as well as the DustTrak measurements during the same sampling period. Since the DustTrak s do not measure the total suspended particulates (TSP), PM is normalized to the mean PCI results for µm particles using the equation in Fig.. From the results, it was established that the amount of PM. relative to PM was approximately.% on a mass basis. Similarly, the amount of PM 1 relative to PM was approximately.% on a mass basis. With this mass fractionation, it was possible to approximate all the common size classifications from a single size range. 1 1 1 1 1 1 1 1 0 1 PM Emission Rates and Emission Factors The real time concentrations from each bird production cycle were suitable measurements for the concentration values (C) in Eq. 1. These real time measurements were combined with the estimates of total house ventilation (Q vent ) to determine the real time estimates of PM emissions for the various common particle size classifications. The PM emission rates were estimated from three production cycles. The winter and spring emissions were estimated from measurements from the DustTrak aerosol monitor for all three size classifications. The only measured size classification for summer emissions was PM and as such, the mass fractionations established from the PCI were used to estimate the remaining size classification. Emission rates for all seasons follow similar patterns throughout the bird production cycle. During all bird production cycles, PM emissions increased as the bird s aged. This was result of an increase in both the PM concentrations and ventilation rates as the bird s aged. Furthermore, emission rates were far greater in the day compared to the night because of increased bird

1 1 1 1 1 1 1 1 0 1 activity. Since the PM emissions are interrelated to the indoor PM concentrations, they both follow a comparable diurnal pattern. Figure displays the PM emissions from three different seasons. From Fig., it was evident that PM emission rates were generally the same for all seasons throughout the entire bird production cycle. In some instances, summer emissions were occasionally higher at the end of the production cycle due to the extreme ventilation rates used during this period. Since there was a large diurnal variation in emissions, the developed emission factors must reflect this trend. The diurnal emission factors were generated from a six hour lighting schedule but are reported on an hourly basis which will allow them to alter the total particulate mass emitted per production cycle for any lighting schedule at any broiler house. The seasonal and diurnal emission factors for all common particulate size classifications are reported in Tables 1 and. The emission factors were developed from the number of birds and length of time for each specific production cycle as a total mass of particulates per production cycle per 00 birds: an average of, birds were raised for days in the winter,, birds were raised for days in the spring, and,1 birds were raised for days in the summer. The average emission factors were determined to be 0.1, 0., and 0.1 kg PM 00 birds -1 production cycle -1 for PM 1, PM., and PM, respectively. These emission factors provide a means to estimate an emission inventory of PM for an entire year of broiler production at a single commercial facility. Based on the length of the studied production cycles, this broiler house should be able to output batches of broilers per year. The average number of birds in a production cycle was, birds. From these averages, the emission inventory of PM for a single facility would be.,., and. kg PM /year for PM 1, PM., and PM, respectively.

1 1 1 1 1 CONCLUSIONS An emission inventory of PM for a single commercial broiler house was characterized. The methodology for estimating PM emissions involved obtaining measurements of continuous real time indoor PM concentrations from optical particle counting instruments located near the inlet of the broiler house s exhaust fan as well as real time ventilation estimates calculated from flow instrumentation and a capture hood. This methodology was applied to four bird production cycles, which provided estimates of emission rates that were representative of a southern Ontario climate during three seasons. It was established that seasonal emissions only slightly varied towards the end of a bird production cycle. Increasing emission rates with bird age were observed during all bird production cycles. From these emission rates, averaged PM emission factors were developed and reported as the mass of PM emitted per 00 birds per production cycle. The emission factors were found to be 0.1, 0., and 0.1 kg PM 00birds -1 production cycle -1 for PM 1, PM., and PM, respectively. Based on these emission factors and the average number of broilers produced per year, it was possible to generate a complete emission inventory of PM for a single commercial broiler house. The emission inventory was.,., and. kg PM /year for PM 1, PM., and PM, respectively. 1 1 1 0 1 REFERENCES Canadian Agricultural Census (CAC). 000. Area source agriculture: section O: agriculture (animals). Not published. Lacey, R.E., J.S. Redwine, and C. B. Jr. Parnell. 00. Emission factors for broiler production operations: a stochastic modeling approach. The Society for Engineering in Agricultural, Food, and Biological Systems. Paper #: 01. Presented at 00 ASAE Annual International Meeting.

1 1 1 1 1 1 1 1 0 1 Pillai, S.D. and S.C. Ricke. 00. Bioaerosols from municipal and animal wastes: background and contemporary issues. Canadian Journal of Microbiology : 1-. PIXE International Corporation. 00. Inertial Impactor Models I-L and I-1L. Tallahassee, FL. http://pixeintl.com Takai, H., S. Pedersen, J.O. Johnsen, J.H.M. Metz, P.W.G. Groot Koerkamp, G.H. Uenk, V.R. Phillips, M.R. Holden, R.W. Sneath, J.L. Short, R.P. White, J. Hartung, J. Seedorf, M. Schroder, K.H. Linkert, and C.M. Wathes. 1. Concentrations and emissions of airborne dust in livestock building in northern Europe. Journal of Agricultural Engineering Resources 0: -. Van Heyst, B.J. 00. Final report: Evaluation of emission factors for the improvement of the estimation methodology for particulate matter from agricultural poultry industry. University of Guelph. Contract: K1-0-0. Not published. Wathes, C.M., V.R. Phillips, M.R. Holden, R.W. Sneath, J.L. Short, R.P. White, J. Hartung, J. Seedorf, M. Schroder, K.H. Linkert, S. Pedersen, H. Takai, J.O. Johnsen, P.W.G. Groot Koerkamp, G.H. Uenk, J.H.M. Metz, T. Hinz, V. Caspary, and S. Linke. 1. Emissions of aerial pollutants in livestock building in northern Europe: overview of a multinational project. Journal of Agricultural Engineering Resources 0: -. ACKNOWLEDGEMENTS This work was assisted by funds provided from the Poultry Industry Council. The authors would also like to thank the support from the School of Engineering at the University of Guelph. 1

FIGURES Figure 1 A schematic of the commercial broiler house used in this study, showing the location of the exhaust ducts and control room Figure Complete range of flow rates for the small variable speed fans 1

Figure Measured flow rates from three medium sized exhaust fans Figure Measured flow rates from two large sized exhaust fans 1

Figure Correlation describing the complete range of house ventilation capacity as a percentage to the exhaust flow rate Figure Estimated flow rate for three bird production cycles during three different seasons 1

Figure A comparison of the effects of the two lighting schedules (1 hour and hours of dark) used at the broiler house on PM concentration Figure Complete measurements of winter and summer indoor PM concentrations 1

Figure Mass fractions of PM measured from the DustTrak aerosol monitor and the PCI Figure PM Emissions estimated for the duration of three entire bird production cycles during three seasons 1

TABLES Table 1 Seasonal emission factors reported as total mass of particulates emitted per production cycle per 00 birds Emission Factors Season PM 1 kg PM 00 birds -1 PM. kg PM 00 birds -1 PM kg PM 00 birds -1 production cycle -1 production cycle -1 production cycle -1 Winter 0. 0.1 0. Spring 0.0 0. 0.0 Summer 0.1 0.0 0. Table Hourly emission factors estimated during the hours of light and dark in the broiler house Averages from all seasons Hourly Emission Factors PM 1 PM. PM kg PM hr -1 00 kg PM hr -1 00 kg PM hr -1 00 birds -1 birds -1 birds -1 Lights off.x -.x -.x - Lights On 0.x - 0.1x - 0.x - 1