HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (11) Citing Articles via Google Scholar Google Scholar Articles by McGinn, S. M. Articles by Beauchemin, K. A. Search for Related Content PubMed PubMed Citation Articles by McGinn, S. M. Articles by Beauchemin, K. A. Agricola Articles by McGinn, S. M. Articles by Beauchemin, K. A. Related Collections Dispersion

TECHNICAL REPORTS

An Approach for Measuring Methane Emissions from Whole Farms

^a Agriculture and Agri-Food Canada, P.O. Box 3000, Lethbridge, AB, Canada T1J 4B1
^b Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, AB, Canada T6G 2E3
^c Agricultural Research Service, United States Department of Agriculture, Watkinsville, GA 30677

^* Corresponding author (mcginn{at}agr.gc.ca)

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Estimates of enteric methane (CH₄) emissions from ruminants are typically measured by confining animals in large chambers, using head hoods or masks, or by a ratiometric technique involving sampling respired air of the animal. These techniques are not appropriate to evaluate large-scale farm emissions and the variability between farms that may be partly attributed to different farm management. This study describes the application of an inverse-dispersion technique to calculate farm emissions in a controlled tracer-release experiment. Our study was conducted at a commercial dairy farm in southern Alberta, Canada (total of 321 cattle, including 152 lactating dairy cows). Sulfur hexafluoride (SF₆) and CH₄ were released from 10 outlet locations (barn and open pens) using mass-flow controllers. A Lagrangian stochastic (LS) dispersion model was then used to infer farm emissions from downwind gas concentrations. Concentrations of SF₆ and CH₄ were measured by gas chromatography analysis and open path lasers, respectively. Wind statistics were measured with a three-dimensional sonic anemometer. Comparing the inferred emissions with the known release rate showed we recovered 86% of the released CH₄ and 100% of the released SF₆. The location of the concentration observations downwind of the farm was critically important to the success of this technique.

Abbreviations: (C/F)_sim, theoretical ratio of concentration to emission • (F_area/F_point), ratio emissions based on area to that based on point configure of source • GC, gas chromatograph • h, obstacle height • LS, Lagrangian stochastic • x_s, source separation distance

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
THE GLOBAL RELEASE of methane (CH₄) from agricultural sources accounts for two-thirds of the anthropogenic CH₄ sources (Moss et al., 2000). These sources include rice growing, fermentation of feed by ruminants (enteric CH₄), biomass burning, and animal wastes. Globally, ruminant livestock are responsible for about 85 Tg of the 550 Tg CH₄ released annually. There is, however, considerable uncertainty associated with the estimates of CH₄ from livestock due to the scant availability of data to document the variability that exists at the farm scale (e.g., due to the significant impact of diet on enteric CH₄ production). Part of the problem is the difficulty of making emissions measurements from livestock facilities.

A few studies have measured CH₄ emissions using close approximations to real farms. Kinsman et al. (1995) calculated CH₄ emitted from a barn housing dairy cows, by monitoring inflow and exhaust air concentration and airflows. However, this technique, aside from being technologically challenging, cannot be applied to the majority of more modern "open" dairy barns or beef feedlots where the airflow is passive. A more appropriate approach for open facilities is to release a tracer gas at a known rate (Marik and Levin, 1996; Kaharabata et al., 2000). This assumes the relationship of the tracer flux to the tracer concentration in the downwind plume is similar for that of the target gas (where the concentration is also measured). Although this ratiometric technique is theoretically straightforward, it relies on matching the tracer and target gas sources that can be difficult to achieve (because the true emission configuration is unknown) and requires extensive setup time.

A more flexible technique for quantifying emissions is to model the dispersion of a target gas from the source (e.g., Kaharabata et al., 2000; Flesch et al., 2004), so that a downwind concentration of gas can establish the emission rate. This "inverse-dispersion" technique has the potential advantage of simplicity, as it requires only a single gas concentration measurement and basic wind information. However, most of the applications of this technique have been for emission sources over a uniform surface (e.g., Wilson et al., 1982). For more complex settings, like a farm with buildings and trees, further evaluation of the inverse-dispersion approach is needed.

The objective of our study was to examine the application of an inverse-dispersion technique that could be easily deployed to determine farm-scale CH₄ emissions (or any other trace gas). Such a technique would allow the quick evaluation of mitigation practices of entire farms and provide a means of understanding the variability that exists between different farm types.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Study Site
The study was conducted at a commercial dairy farm (49°48' N, 112°22' W) near Lethbridge, Alberta, where cows were located in an open pen (total area 9560 m²) and a large barn (stall area 7680 m²) (Fig. 1 ). The upper half of the barn walls were curtains that opened and closed depending on air temperature in the barn. These curtains were opened during the daytime allowing natural ventilation of the barn air. The peak of the barn roof (approximate 8-m height) was opened allowing natural air exchange as well. Manure within the barn was scraped from the floor daily and stored in tanks below floor level. These holding tanks were periodically pumped to an open manure-holding facility located 200 m south of the buildings. Of the 321 cattle at this dairy farm, 152 were lactating cows, 48 were dry cows or heifers over 2 yr, 90 were heifers or steers between 1 and 2 yr, and 31 were heifers or steers less than a year.

View larger version (45K): [in this window] [in a new window]   Fig. 1. Arial view of the study site showing the position of the barn and open pens housing cows, the release points for CH4 and SF6, and the laser and sample lines. The laser in the northwest corner was turned to a north or east reflector depending on wind direction.

Measurements
Three open-path lasers (GasFinder2; Boreal Laser, Spruce Grove, AB, Canada), tuned for detecting average CH₄ concentration along a path (1 m above the ground), were situated 75 to 100 m away from the nearest farm building (Fig. 1). Lasers were aligned to enclose the north, east, and south perimeters of the farm; an array of retroreflectors terminated the laser path. When wind direction was westerly, the laser along the north perimeter was turned to a second array of retroreflectors located 200 m to the north of the laser (Fig. 1). In this manner, with minimum effort, there was always one laser recording concentration of the target gas upwind of the farm (background concentration) and one recording downwind (plume) concentration. The average concentration along the laser path was scanned approximately every 5 s and averaged every 15 min. Laser correction factors were determined repeatedly throughout the experiment. When possible we adjusted each laser to match our gas chromatograph (GC) measured CH₄ concentrations (see below), if the wind direction indicated both the laser and GC samples measured the same concentration (i.e., background). These corrections varied from 0.83 to 0.97. These corrections were needed to remove bias instrument error that was evident when the three lasers were set side-by-side (2-m separation) in the field with no obvious upwind source. This side-by-side comparison resulted in average CH₄ concentrations of 1488 (±93), 1560 (±99), and 1593 (±112) µg m^–3, which were significantly different from each other (P < 0.05). Accuracy analysis (Allen and Raktoe, 1981) was used to decompose the differences. On average, this analysis attributed 80% to bias (offset), 13% to random, and 7% to regression error. Assuming nothing changed for the on-farm study, the corrections to the lasers using the GC data removed the largest (bias) error. According to the laser specifications, the accuracy of the laser is 661.2 µg m^–3 per m (at 20° and 101 KPa), or 3.3 and 1.8 µg m^–3 over path distances of 200 to 360 m used in our study, respectively. However, our experience suggests that when working with lasers with different histories, and at the low concentrations associated with on-farm CH₄ plumes, it is necessary to revisit the between-laser precision for each new study. We estimate in our case that maximum differences between lasers from the side-by-side comparison were 79 µg m^–3, but during our on-farm study these differences were most likely closer to 20 µg m^–3 after correcting for bias using the GC data as a common reference.

The average alongwind horizontal velocity (U); average wind direction (dir); the standard deviation of alongwind, acrosswind, and vertical velocity fluctuations (_u, _v, and _w, respectively); friction velocity (u*); surface roughness (z_o); and the Monin–Obukhov stability length (L) were calculated from data collected at 10-Hz sampling interval (CR23X datalogger; Campbell Scientific, Logan, UT) using a three-dimensional sonic anemometer (CSAT3; Campbell Scientific). These statistics were calculated following Flesch et al. (2004), using the same 15-min averaging period as for gas concentrations. The anemometer was located at the laser locations in the southeast corner of the farm (Fig. 1). At this location, the influence of the farm buildings was deemed to be far enough away to have minimal impact on the sonic data.

Additional line-averaged concentrations of CH₄ and sulfur hexafluoride (SF₆) were measured downwind of the farm by drawing air through regulated intakes (using flow-holder valves) every 20 m along a long polyethylene tube (0.6-cm i.d.) at a 1-m height. On 18 November the sample line was lengthened from 100 m (5 intakes) to 200 m (10 intakes) long. A 12-vdc pump (TD3LS7; Brailsford and Company, Rye, NY) was connected to one end of the sample line. The flow rate through each intake was adjusted to be equal and approximately 2 L min^–1. A 2-L polyethylene bottle was connected to the exhaust of the pump to provide a 1-min time filter. A 20-mL air sample was taken every 5 min using a syringe and septum port. The content of the syringe was immediately discharged into an evacuated vial (6.8 mL volume) through a double septum cap (follows from Rochette and Bertrand, 2003), and the time and date recorded on the vial. Every 30 min, the background concentration upwind of the farm was sampled using three air samples drawn at 1.5-m height with a 20-mL syringe (after purging the syringe three times between sampling). At the end of each day, the vials were taken to the laboratory for analysis of CH₄ and SF₆ concentrations on a GC the following day (within 24 h of sampling). The over pressured vial provided a sufficiently large air volume to pass through the electron capture detector (ECD) column in the GC.

Before the start of the experiment, two lines of high-density polyethylene tubing (6-mm i.d.) were laid out to allow independent release of SF₆ and CH₄ from 10 release points throughout the barn and pen areas (Fig. 1). The release lines where connected to mass-flow controllers and cylinder tanks of ultrapure SF₆ and CH₄. The gas flow at each release point was adjusted using flow hold valves to be one-tenth of the gas flow set by mass-flow controller, where the mass-flow controller was 0.2 L min^–1 for SF₆ (GFCS-011806; Aalborg Instruments and Controls, Orangeburg, NY) and 54.2 L min^–1 for CH₄ (GFCS-011246; Aalborg Instruments and Controls). For each gas, three release points (20 m from each end and at the middle) were located 1 m above the peak of the roof. There were also three release points on the east side of the barn along the ventilation openings (aligned with roof locations). Four release points (2-m height) were spaced along the south perimeter of the open pen that housed the pregnant, nonlactating cows. The mass-flow controller units were checked in the laboratory by releasing gas over a long period of time and measuring the mass loss of the cylinder.

Emission Calculations
We used the inverse-dispersion technique outlined by Flesch et al. (2004). A Lagrangian stochastic (LS) model predicts the theoretical ratio of the targeted gas concentration to the emission rate (C/F)_sim. An emission rate (F_LS; g s^–1) is then determined after measuring the target gas concentration in the plume (C_obs; g m^–3):

[1]

The commercially available software product WindTrax (Thunderbeach Scientific, 2005) was used in this calculation. A global positioning system was used to mark the farm structures, our gas release points (10 points), and the laser paths around the farm. WindTrax determined (C/F)_sim in one of two modes, depending on how we modeled the emission source. Treated as 10 point sources (the actual configuration), WindTrax calculated 5000 trajectories downwind of each emission point (forward mode). Alternatively, when we treated the farm as two surface area sources, WindTrax calculated 10 000 trajectories upwind of the laser path (backward mode). These sets of trajectories define (C/F)_sim.

The WindTrax model assumes idealized wind conditions appropriate for simple undisturbed terrain (using Monin–Obukhov wind relationships). However, in our farm setting the buildings and trees created wind complexity. Another difficulty in calculating (C/F)_sim was that we had to assume an emission source configuration (i.e., location and relative proportion of emissions). While we knew the "synthetic" SF₆ and CH₄ release configuration, we did not know the details of the "natural" CH₄ emissions; this is needed to separate the synthetic from the total emissions. Considering these complications, Flesch et al. (2005) concluded, "...with careful selection of a measurement location, an idealized inverse-dispersion technique can give the emissions from a farm complex within an error of ±10%...." They list three suggestions for dealing with wind and source complexity, and making accurate measurements:

• The farm should be isolated on the landscape, with winds measured in the ambient regime.
  
• Concentration should be measured at many obstacle heights h (i.e., barn or tree heights) downwind of the farm (they conservatively propose 20 h).
  
• Concentration should be measured at multiple "source-separation" distances x_s downwind of the farm (distance between separate emission sources, in our case the distance between the barns housing the lactating and the nonlactating cows was 90 m). It is recommended that distances beyond two x_s be employed.
  

Our site meets the first condition, but not the last two. At times our concentration measurements were less than 10 h and 1 x_s from the farm. Although the measurement location was closer than recommended, it was chosen to ensure a detectable concentration rise for the CH₄ lasers.

Sampling Periods
Our analysis scheme was based on 15-min sampling periods ending at 15, 30, 45, and 60 min past the hour (with corresponding wind measurements). The CH₄ gas was turned on for 35 min, starting 5 min before a 15-min period, to allow the downwind concentrations to reach equilibrium before the measurement began. At the end of the 35 min the CH₄ gas was shut off for 25 min. This cycle was repeated each hour (Fig. 2 ). This on–off release pattern for CH₄ was adopted to separate the farm CH₄ emission from the known synthetic emission. After gas release stopped at the end of a 15-min period, we waited for concentrations to fall back to an equilibrium level (typically 5 min) before calculating concentration. From this pattern we created two "gas-on" and two "gas-off" observations each hour. Assuming the natural CH₄ emissions were constant during the hour, the difference between the gas-on and gas-off emissions was taken as our estimate of the synthetic gas release rate (and should equal the actual 54.2 L min^–1 release rate). Because there were no natural SF₆ emissions from the farm, it was not necessary to cycle between gas-on and gas-off conditions. Therefore, the SF₆ gas was left on for the complete measurement period.

View larger version (12K): [in this window] [in a new window]   Fig. 2. Illustration of the on–off timing of CH4 (solid) and SF6 (dashed) for a sampling period.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Concentration Signal
An example of CH₄ concentrations during a release cycle is shown in Fig. 3 . There were a number of difficulties encountered using this approach. First, the downwind CH₄ signal results from a combination of natural and synthetic sources. During gas-off periods the concentration was 5 to 10% above the upwind value because of emissions from the cows. During gas-on periods the downwind concentration generally rose by another 5 to 10% because of our synthetic gas additions. Separating the natural and synthetic sources was problematic because of the likelihood of differences in the two source configurations. Here we assumed the configuration of natural CH₄ emissions was identical to the synthetic. To get the synthetic emission rate we subtracted the gas-off emissions from the gas-on values, and assumed the natural emissions were the same for both groups. This may not be true if the natural emissions changed with time. This possibility was reduced, however, by our use of short, alternating gas-on and gas-off periods (30 min). This difficulty does not exist for SF₆.

View larger version (20K): [in this window] [in a new window]   Fig. 3. Example of the time series of line-average methane concentration (CL) taken upwind and downwind of the farm on 9 November.

Gas Recovery Using the Actual Source Configuration
In the following discussion we judge the success of our inverse-dispersion calculations in terms of recovery rate, which expresses the inferred emission rate as a percentage of the known release rate. When we treat our farm emissions as coming from 10 point sources, using the actual outlet locations (including the height of release) in WindTrax, the average recovery rate for CH₄ was 86% (Table 1). More impressive was the 91, 105, and 89% recovery on three of the four days. The cause of the larger error of 59% recovery rate on 1 December was not identified. In terms of wind conditions, there was nothing dramatically different on this day. On average the SF₆ results were better, with a perfect recovery of 100% (±29%; Table 1). The better recovery for SF₆ compared with CH₄ was expected, since we did not have to deal with confounding natural emissions.

The perfect average recovery of SF₆ hides a large day-to-day variability, from 148% on 9 November to 74% on 1 December. This large variability (compared with CH₄) may be partially due to the short sampling line. While the CH₄ concentration path length was between 250 and 360 m, and positioned so that the downwind laser usually captured the full width of the farm emission plume, the SF₆ line was only 100 to 200 m long, which was not long enough to capture the full plume. This increases the uncertainty in our calculations, as the SF₆ concentration along the sampling line would be sensitive to lateral dispersion from the farm (whereas if the sample line spans the plume it is insensitive to lateral dispersion). Likewise, our calculated F_LS would be similarly sensitive to the modeled dispersion. This is an important consideration, as modeling horizontal dispersion is more error prone than modeling vertical dispersion (Flesch and Wilson, 2005).

Inadequate sampling may also cause some of the SF₆ uncertainty. From Fig. 3 we can see the strong temporal variability in CH₄ concentration, which also exists for SF₆. Given this variability, our SF₆ sampling frequency of once every 5 min may be insufficient to give an accurate 15-min average (to correspond with the 15-min wind statistics).

Gas Recovery Using a Generic Source Configuration
In theory, accurate inverse-dispersion calculations require that we model the actual emission source configuration. In many real-world situations we do not have the privilege of knowing these details. For example, we do not know the proportion of natural CH₄ emissions coming from the barn and the dry stock pens. Flesch et al. (2005) discussed a solution to this problem: locate the concentration measurements where the modeled (C/F)_sim will be insensitive to the source configuration details. Generally, this sensitivity is reduced as we move downwind. Our dataset allows us to consider the problem of unknown source configuration, and how it influences inverse-dispersion calculations.

As a hypothetical exercise, we applied the inverse-dispersion technique to our farm, but now assume a generic emission configuration. From a distance we might broadly identify two emission sources: the milking barn and the dry stock pens (i.e., where the cattle are). In WindTrax we thus treat the farm as two surface sources, corresponding to the barn and pens, and assume each has the same real emission rate. Although this does not match either the natural or synthetic source configuration, it has obvious similarities in terms of the relative spread of release, and is a rational choice given limited information. The question is how much this assumption affects the emission calculations.

We recalculated the CH₄ and SF₆ recoveries using WindTrax with the generic configuration (Table 2). Not surprisingly, there was a decline in accuracy. We found the average CH₄ recovery was reduced from 86 to 66%, and SF₆ recovery from 100 to 77%. From Flesch et al. (2005) we conclude that this undesired sensitivity is the result of concentration being measured too close to the farm.

View larger version (8K): [in this window] [in a new window]   Fig. 4. Hypothetical ratio of emissions calculated using a source configuration of 10 point sources with that using two area sources (Farea/Fpoint), plotted versus distance from the farm boundary to the line-average concentration measurement (fetch).

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
We found that inverse-dispersion calculations gave gas emissions from a dairy farm with reasonable accuracy. On average we "recovered" our synthetic gas emissions of CH₄ and SF₆ to within 15%. As a further indication of success, "natural" CH₄ emissions were calculated to be 286 g d^–1 animal^–1, almost identical to that derived from previous studies. Our results are strong evidence that an idealized inverse-dispersion technique that ignores wind complexity can be used for some real-world settings. In applying this technique, it is advisable to first show this is a valid assumption, using a release-recovery study as in our study.

Several factors were important to our successful application. First, the farm studied was relatively isolated on the landscape. This meant a clearly defined ambient wind regime, and locating our anemometer in this regime was straightforward. It also meant that we did not have other nearby emission sources (for CH₄) to confound our concentration signals. A second factor was that we used relatively long line-average concentration measurements. The advantage of this should not be ignored. As argued by Flesch and Wilson (2005), utilizing crosswind average concentration reduces one of the major sources of uncertainty in the inverse-dispersion technique, specifically inaccurate modeling of lateral dispersion. Another factor in our success was ignoring periods known to be problematic for inverse-dispersion calculations (i.e., low winds, extreme stabilities, and measurements of downwind concentration that do not sample enough of the farm plume). One difficulty is the tracking of trajectories when wind directions are not consistent over the averaging time, common under low winds and very stable periods. Shortening the averaging time may avoid this particular problem in the inverse-dispersion calculation.

A fourth factor in our success was having a concentration measurement location that allowed us to ignore wind complexity. Even though we were as close as 75 m to the farm, this was evidently far enough (and much closer than recommended by Flesch et al., 2005). But to achieve accurate results at this location, it was necessary to use the actual source configuration in the dispersion calculations (i.e., 10 outlets). In many real-world situations these details will not be known, requiring assumptions about the source configuration. This will mean locating the measurement location further from the farm, keeping in mind that the location is a compromise between being far enough from the source to ensure the plume is uniform, yet close enough to ensure the target gas concentration is detectable above background concentration.

The selection of a proper concentration measurement location is probably the foremost issue in applying the inverse-dispersion technique. Concentration must be sufficiently high (relative to background) for accurate detection with the measuring devices. This creates a tension between wanting to be close to better detect concentration rise, and wanting to be far to minimize dispersion modeling errors caused by wind and source complexity. Our results highlight how the accuracy of the modeled source configuration can influence this distance decision. We conclude that the more one knows the source geometry and their relative strengths (e.g., the relative location, size, and type of animals, barn construction), and incorporating these to give an accurate source configuration in the dispersion model, the greater the likelihood of a successful application.

	ACKNOWLEDGMENTS

 
This study was funded by National Carbon and Greenhouse Gas Emission Accounting System (NCGAVS) and the Canadian Foundation for Climate and Atmospheric Sciences (CFCAS). Technical support for the study was provided by Trevor Coates, Nathan Dyck, Linda Kremenik, Darrell Vedres, and Jason Seabrook at Agriculture and Agri-Food Canada's Lethbridge Research Centre.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 

• Allen, O.B., and B.L. Raktoe. 1981. Accuracy analysis with special reference to the prediction of grassland yield. Biom. J. 23:371–388.
• Flesch, T.K., and J.D. Wilson. 2005. Estimating tracer emissions with a backward Lagrangian stochastic technique. In J. Hatfield (ed.) Micrometeorology in agricultural systems. ASA, Madison, WI.
• Flesch, T.K., J.D. Wilson, L.A. Harper, and B.P. Crenna. 2005. Estimating gas emission from a farm using an inverse-dispersion technique. Atmos. Environ. (in press).
• Flesch, T.K., J.D. Wilson, L.A. Harper, B.P. Crenna, and R.R. Sharpe. 2004. Deducing ground-to-air emissions from observed trace gas concentrations: A field trial. J. Appl. Meteorol. 43:487–502.[CrossRef]
• Kaharabata, S.K., P.H. Schuepp, and R.L. Desjardins. 2000. Source strength determination of a tracer gas using an approximate solution to the advection-diffusion equation for microplots. Atmos. Environ. 34:2343–2350.[CrossRef]
• Kinsman, R., F.D. Sauer, H.A. Jackson, and M.S. Wolynetz. 1995. Methane and carbon dioxide emissions from dairy cows in full lactation monitored over a six-month period. J. Dairy Sci. 78: 2760–2766.[Abstract]
• Marik, T., and I. Levin. 1996. A new tracer experiment to estimate the methane emissions from a dairy cow shed using sulphur hexafluoride (SF₆). Global Biogeochem. Cycles 10:413–418.
• Moss, A.R., J.-P. Jouany, and J. Newbold. 2000. Methane production by ruminants: Its contribution to global warming. Ann. Zootech. 49:231–253.[CrossRef]
• Rochette, P., and N. Bertrand. 2003. Soil air sample storage and handling using polypropylene syringes and glass vials. Can. J. Soil Sci. 83:631–637.
• Thunderbeach Scientific. 2005. WindTrax software. Available at www.thunderbeachscientific.com (verified 13 Sept. 2005). Thunderbeach Scientific, Nanaimo, BC, Canada.
• Wilson, J.D., G.W. Thurtell, G.E. Kidd, and E.G. Beauchamp. 1982. Estimation of the rate of gaseous mass transfer from a surface source plot to the atmosphere. Atmos. Environ. 16:1861–1867.[CrossRef]

This article has been cited by other articles:

				S. M. McGinn, K. A. Beauchemin, T. K. Flesch, and T. Coates Performance of a Dispersion Model to Estimate Methane Loss from Cattle in Pens J. Environ. Qual., July 23, 2009; 38(5): 1796 - 1802. [Abstract] [Full Text] [PDF]

				L. A. Harper, T. K. Flesch, J. M. Powell, W. K. Coblentz, W. E. Jokela, and N. P. Martin Ammonia emissions from dairy production in Wisconsin J Dairy Sci, May 1, 2009; 92(5): 2326 - 2337. [Abstract] [Full Text] [PDF]

				S. M. McGinn, T. K. Flesch, B. P. Crenna, K. A. Beauchemin, and T. Coates Quantifying Ammonia Emissions from a Cattle Feedlot using a Dispersion Model J. Environ. Qual., October 16, 2007; 36(6): 1585 - 1590. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (11) Citing Articles via Google Scholar Google Scholar Articles by McGinn, S. M. Articles by Beauchemin, K. A. Search for Related Content PubMed PubMed Citation Articles by McGinn, S. M. Articles by Beauchemin, K. A. Agricola Articles by McGinn, S. M. Articles by Beauchemin, K. A. Related Collections Dispersion