JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 20 February 2008
Published in J Environ Qual 37:582-591 (2008)
DOI: 10.2134/jeq2006.0426
© 2008 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Griffith, D. W. T.
Right arrow Articles by Reisinger, A. R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Griffith, D. W. T.
Right arrow Articles by Reisinger, A. R.
Agricola
Right arrow Articles by Griffith, D. W. T.
Right arrow Articles by Reisinger, A. R.
Related Collections
Right arrow Global Change
Right arrow Landscape-Atmosphere Interactions
Right arrow Field evaluation techniques

Methane Emissions from Free-Ranging Cattle: Comparison of Tracer and Integrated Horizontal Flux Techniques

David W. T. Griffitha,*, Glenn R. Bryanta, David Hsua and Andy R. Reisingerb

a Dep. of Chemistry, Univ. of Wollongong, Wollongong, NSW, Australia
b National Inst. of Water and Atmospheric Research (NIWA), Wellington, New Zealand

* Corresponding author (griffith{at}uow.edu.au).

Received for publication October 2, 2006.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Summary and Conclusions
 REFERENCES
 
Accurate measurements of methane (CH4) emission rates from livestock in their undisturbed natural environments are required to assess their impacts on radiative forcing (i.e., enhanced greenhouse effect) and the environment. Here we compare results from two nonintrusive techniques for the measurement of CH4 emissions from cattle. The cows were kept in an outdoor feeding strip that allowed them to follow natural behavioral patterns but contained them within a well defined space. In the first technique, nitrous oxide (N2O) was released as a tracer at the upwind edge of the feeding strip, and the downwind concentrations of N2O and CH4 were measured simultaneously using Fourier transform infrared (FTIR) spectroscopy. Average CH4 emission per cow was calculated each half-hour on three separate days from the correlation between the two gases. The second technique was the integrated horizontal flux (IHF) or 1-D mass-balance method, in which we used the measured vertical profiles of CH4 concentration and windspeed downwind of the cows to determine the total CH4 emission. Comparing the IHF results to the known release rate of N2O allowed us to test the IHF technique independently. We found agreement within 10% for all comparisons on all days. The daily CH4 emission rate averaged over all tracer and IHF measurements was 342 g CH4 head–1 d–1. This is within the range of previous measurements for mature lactating dairy cattle (200–430 g CH4 head–1 d–1) but higher than expected for yearling cattle. The high CH4 emissions are accompanied by high CO2 emissions determined from the FTIR measurements. The bias is most likely due to the measurements being made during and after supplementary feeding of the cattle.

Abbreviations: IHF, integrated horizontal flux • FTIR, Fourier transform infrared


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results
 Discussion
 Summary and Conclusions
 REFERENCES
 
FOR COUNTRIES with a large agricultural industry, such as New Zealand and Australia, methane (CH4) emissions from cattle and sheep represent a significant fraction of the national greenhouse gas emissions inventory. Although standard methodologies exist to estimate average CH4 emissions from ruminant livestock (Houghton et al., 1996), they are generally based on emission measurements under controlled indoor feeding conditions (e.g., Gibbs and Johnson, 1994). In Australia and New Zealand, most cattle and sheep are kept under unrestrained outdoor conditions, so standard methodologies may not be readily applicable, and a need exists to develop and apply nonintrusive techniques for direct measurement of CH4 emissions from ruminant livestock under nationally representative conditions.

Several approaches have been successful. The SF6 tracer technique uses SF6 released from a permeation tube in the rumen and measures the CH4 emission directly at the mouth of individual animals by correlation of CH4 and SF6 concentrations (Johnson et al., 1994; Lassey et al., 1997). In the two-dimensional, mass-balance technique, a small number of stock are enclosed in a test plot, and emissions are determined from the balance of upwind and downwind CH4 concentrations measured at the plot boundaries over a range of heights (Denmead et al., 1998; Leuning et al., 1999). Judd et al. (1999) measured CH4 emissions from sheep using the flux-gradient micrometeorological technique. An alternative top-down approach lies in regional-scale boundary-layer budgeting, which derives area-wide emissions from comparing vertical CH4 profiles obtained via aircraft sampling with model calculations (Denmead et al., 2000; Lassey et al., 2000). Most recently, Laubach and Kelliher (2004; 2005a; 2005b) have presented a thorough assessment of open-path laser and in situ gas chromatographic measurements with flux gradient and 1-D mass-balance (or integrated horizontal flux [IHF]) methods and backward Lagrangian stochastic dispersion techniques (Flesch et al., 1995; Flesch et al., 2004) to determine CH4 emissions from a free-ranging dairy herd.

Tracer methods have much to offer in this area because they provide an alternative route to quantify atmospheric dispersion, which is not dependent on meteorological measurements. The SF6 technique is a tracer technique whereby the tracer is released in the rumen and the sample is collected at the cow's mouth. In this study, we used "external" tracers whereby the tracer release is external to the cow and is dispersed with cow emissions by the turbulent wind field. The main requirements are that (i) the tracer gas be released from the point or area colocated with the source gas whose unknown emission rate is to be measured, (ii) the tracer and source gases are then dispersed equally by atmospheric transport and turbulence, (iii) any natural emission or uptake of the tracer gas in the source region is negligible compared with the release rate, and (iv) tracer and target gas concentrations can be measured in the same air samples downwind of the source. If these conditions hold, the unknown source gas emission rate can be calculated from the correlation of measured downwind source and tracer gas concentrations. Examples of tracer techniques used to quantify atmospheric emissions have been described for CH4 emissions from landfills (Galle et al., 2001), effluent tanks (Kaharabata et al., 1998), and cattle housed in a barn (Marik and Levin, 1996) and for ammonia emissions from soils (Galle et al., 2000).

In this paper we report nonintrusive measurements of CH4 emissions from free-grazing yearling dairy cows. We use a tracer technique and a one-dimensional mass-balance or IHF approach. Measurements of all trace gas concentrations (i.e., CH4, N2O, and CO2) for both methods were made simultaneously by Fourier transform infrared (FTIR) spectroscopy with a single field instrument. We extend the tracer method to the measurement of CO2 emissions from the cattle, providing a quantitative indicator of the cows' level of metabolic activity that can be compared with the CH4 emissions.


   Materials and Methods
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
 Discussion
 Summary and Conclusions
 REFERENCES
 
Site Description
Measurements were made in the Wairarapa region of the North Island of New Zealand, a dairy and general farming district approximately 100 km northeast of Wellington. The measurement site was on a working dairy farm near Greytown (41°08' S, 175°25' E; 48 m above sea level); a typical layout is shown in Fig. 1 . Prevailing winds in the winter season are from the south to southwest, with typical daily temperatures ranging from 5 to 15°C.


Figure 1
View larger version (14K):
[in this window]
[in a new window]

  		Fig. 1. Schematic layout of the test plot site. Scale is approximate.

 
Measurements with the tracer-release method were made during a 2-wk campaign on 9, 11, and 16 Sept. 1999 when the available grazing area was upwind of the fixed measurement mast. Between 23 and 50 yearling Friesian dairy cows were kept in the fenced-off test plot of approximately 30 x 7 m in short pasture grass. They were fed silage in the enclosure at the start of each measurement period. A 6-m meteorological mast was located 30 to 45 m downwind of the cattle plot; this ensured that on average more than 50% of the cattle contributed to the measured emissions at the mast at any time (see later discussion of Eq. [3]). The mast was used to mount four cup anemometers spaced at approximately logarithmic heights: initially 0.5, 1, 2, and 4 m on 9 and 11 September and at 1, 2, 4, and 6 m on 16 September when it became clear that a significant fraction of the emissions passed above the 4-m level at the mast. The mast contained one sonic anemometer at a height of 2 m, two temperature probes at 1 and 2 m, an atmospheric pressure sensor, and a digital camera to monitor cow positions at 10-min intervals. Four gas sampling inlet lines were mounted with inlets at the same heights as the cup anemometers, and a fifth gas inlet was located 30 m upwind of the test plot at a height of 1 m. Meteorological data were recorded at 10-s intervals using a Campbell data logger (Campbell Scientific, Logan, UT) and stored by a PC along with the 10-Hz sonic anemometer data. For the present analysis, only the cup-anemometer wind speeds at four heights, wind direction from the sonic anemometer, temperature, and pressure data were required.

At the upwind edge of the test plot, pure N2O was released from a cylinder through an array of 11 capillaries connected to a common 1/4" (6.25-mm) nylon tube. The capillaries were adjusted in length so that each delivered the same flow, and the cylinder outlet pressure was regulated to give a total release rate of approximately 1 L N2O min–1. This flow ensured N2O enhancements several orders of magnitude larger than any natural emission or uptake rates in the fetch. Flow was measured twice per day with an electronic flowmeter calibrated in L min–1 (1 bar, 20°C). The capillaries were tied to stakes ~0.7 m above the ground at regular separations of about 3 m immediately behind the upwind electric fence.

Gas Analysis
Gas analysis was performed using FTIR spectroscopy. A detailed description of the technique can be found in Griffith (2002). In brief, air was drawn continuously at approximately 1.5 to 2 L min–1 from each inlet line through a 40-L buffer volume to smooth out short-term variations of gas concentrations. The trace gas concentrations in the five buffers were sequentially analyzed by drawing air from each buffer via a computer-switched manifold of solenoid-operated valves into an evacuated multipass absorption cell (22-m optical path length, 8.5-L volume) (model 22PA; Infrared Analysis, Anaheim, CA). Infrared transmission spectra were recorded at 1 cm–1 resolution using a Bomem MB100 Fourier transform spectrometer (ABB-Bomem, Quebec, Canada) with an InSb detector (liquid-nitrogen-cooled) with an integration time of approximately 3 min. The spectra were analyzed on-line for CH4, N2O, CO2, CO, and H2O using a nonlinear least-squares algorithm that iteratively calculates the spectrum from a database of spectral line parameters until best fit to the recorded spectrum is obtained (Griffith, 1996; Griffith et al., 2003). The total measurement time for each inlet line, including cell evacuation, flushing, spectrum measurement, and data analysis, was 5 min, so that a complete measurement cycle for four mast sample lines and the upwind reference line took 25 min, which is similar to the averaging time of the buffers. The measurements were fully automated from a controlling computer housed with the FTIR gas analysis system in a caravan about 25 m from the mast. Measured mixing ratios for N2O, CH4, and CO2 were corrected to dry air; the water vapor measurement was obtained from each relevant spectrum with approximately 5% accuracy. The relative precision (1{sigma}) of the measurements is 0.1 to 0.2% or better for N2O, CH4, and CO2. For the purposes of error estimation for flux calculations, we assume 1{sigma} errors in CH4, N2O, and CO2 measurements of 4 nmol –1, 0.5 nmol mol–1, and 1 µmol mol–1, respectively.

Tracer Method
For the tracer method, the CH4 emissions QCH4 from the cows contained in the test plot can be calculated from the ratio of measured CH4 and N2O mixing ratios ([CH4] and [N2O] in nmol mol–1) at any inlet (height), scaled by the measured N2O tracer emission rate QN2O:

Formula 1[1]
[CH4,bgnd] and [N2Obgnd] are the background mixing ratios, and {Delta} denotes the mixing ratio enhancement above the background level measured upwind of the cows and tracer release. The fluxes Q are in molar units (e.g., mol s–1, mol d–1) and can be converted to mass units by multiplying by the molecular weights of CH4 (16) and N2O (44). The enhancement of the downwind concentrations of N2O and CH4 (but not CO2) at all heights was substantially greater than any variability or gradients expected due to emission or uptake from the ground or other local sources over a measurement period; thus, the upwind reference from a single height could be used to calculate downwind enhancements at all heights. There was no significant dependence of calculated emission rate on inlet height, and calculated emission rates from each height using Eq. [1] were averaged over each 25-min measurement period for comparison with the IHF method.

Alternatively, Eq. [1] can be re-cast in the form

Formula 1A[1a]
The slope of a regression of downwind CH4 versus N2O concentration measurements over one 25-min measurement period using data from four heights provides the ratio of the CH4 and tracer fluxes. Using Eq. [1a] has the advantage that the CH4 background value is included in the regression via the fitted intercept and does not need to be determined explicitly.

Emissions of CO2 were determined by the same method. These measurements are approximate because CO2 is taken up and released by the surface soil and pasture through photosynthesis and respiration.

The One-Dimensional Mass-Balance (IHF) Method
In the 1-D mass-balance or IHF method, total gas emissions are calculated from the vertical profiles of gas concentration and wind speed downwind. Following Laubach and Kelliher (2004), the gas flux Q g s–1 per unit width L m of a linearly extended source through a plane perpendicular to the wind direction can be calculated as:

Formula 2[2]
where u(z) is the wind speed (m s–1) at height z m, and {Delta}c(z) is the enhancement of the gas concentration relative to the background level, expressed in concentration units (g m–3). The integrand is zero at the ground because the wind speed at ground level is zero. Above a certain height (zmax m), the concentration enhancement reduces to zero, corresponding to the plume dispersion height for a given source-sensor distance. Assuming mass-balance and no horizontal divergence perpendicular to the wind direction, the flux per unit width at the mast equals the emission per unit width upwind of the mast. The total source emission for all cows (Qtotal) may then be calculated by multiplying Eq. [2] by the width L of the array of sources (cows), assuming a uniform emission over the length of the source. The effects of uneven cow distribution are discussed below.

Because this approach represents an absolute measurement without requiring a tracer as reference, it may be applied not only to the CH4 emitted by the cows but also to the N2O released for measurements using the tracer method. This allows an independent and direct validation of the IHF approach by comparing the measured IHF N2O flux with the release rate measured directly at the tracer line capillaries.

Because the array of cows across the wind direction is not infinite in extent, we must account for variations in wind direction during an IHF measurement to allow for end-effects. We introduced an additional correction factor that describes the fraction of the linear array of cows that contributed to an air sample obtained at the mast. This (time-dependent) factor q for each measurement is defined as:

Formula 3[3]
where flag(i) = 1 if the wind comes from the direction of the source and flag(i) = 0 if the wind comes from outside. The sum is over the N 10-s wind measurements that apply to the accumulation period of each FTIR gas sample. Accounting for the effect of horizontal dispersion, flag(i) was also allowed to vary between 0 and 1 for wind directions within a given distance of either edge of the cattle/tracer array. The half-width of this horizontal dispersion was chosen to be equal to the approximate vertical dispersion height estimated from the vertical profile at the mast (8 m). Profile periods for which the wind direction was from the source area less than 20% of the time were excluded from the analysis.

Equation [4] describes the total emission strength of an upwind source (i.e., N2O tracer or CH4 from all cows in the test plot) as a combination of [2] and [3] with a maximum integration height zmax:

Formula 4[4]
where L is approximately 30 m. The integration is performed slab-wise between the measured heights up to the height where u x {Delta}c becomes zero. The total flux may then be compared with the known release rate of the tracer N2O or, assuming an even distribution of cows along the feeding strip, may be divided by the number of cows in the test plot to obtain an average CH4 emission rate per cow.


   Results
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
Discussion
 Summary and Conclusions
 REFERENCES
 
Data were collected on 9, 11, and 16 Sept. 1999 using 50, 30, and 23 cows, respectively, when the available grazing area was upwind of the (fixed) measurement mast. Measurement times and conditions are summarized in Table 1 . Figure 2 illustrates the raw data for 9 September when cows were brought into the enclosure at 1230 h and immediately fed silage distributed across the length of the enclosure to encourage uniform distribution. Cows were removed at 1800 h, allowing 13 individual profiles to be measured. The upper panel shows the CH4 and N2O mixing ratios, and the lower panel shows the horizontal wind speed plotted against time of day.


View this table:
[in this window]
[in a new window]

  		Table 1. Summary of measurement times and conditions.

 

Figure 2
View larger version (18K):
[in this window]
[in a new window]

  		Fig. 2. Time series of N2O (dashed lines) and CH4 (solid lines) mixing ratios and wind speed on 9 Sept. 1999. Data point labels are the measurement heights (m) for the corresponding point. The rightmost point in each series is the upwind mixing ratio.

 
Each group of points in Fig. 2 corresponds to one 25-min profile, with the data point labels corresponding to the measurement heights. The latest (unlabeled, lowest) point in each trace gas profile is the upwind mixing ratio. We can make a number of qualitative observations: (i) In each profile, the trace gas concentrations decrease and the windspeed increases with height, as expected for a source near ground level and normal wind profiles; (ii) N2O and CH4 mixing ratio enhancements are highly correlated; (iii) the upwind N2O mixing ratio remains steady at 310 to 312 nmol mol–1 throughout the day, reflecting the southern hemisphere clean air value (313 nmol mol–1 in 1999) within calibration uncertainty (http://www.cmdl.noaa.gov/ccgg/iadv/); (iv) upwind CH4 gradually increases from 1800 to 1900 nmol mol–1 during the day; this is higher than regional baseline values at the time (~1730 nmol mol–1; see http://www.cmdl.noaa.gov/ccgg/iadv/) and probably reflects the local and regional population of cattle in other herds leading to local buildup of CH4, whereas local sources of N2O are not large enough to cause a corresponding increase; and (v) the wind speed gradually decreased during the afternoon, with a corresponding general increase in downwind N2O and CH4 mixing ratios at the mast. In general, observations on the other two days are consistent with the observations from 9 September, but there are fewer measurements, and the patterns are less clear.

Tracer Method
Individual {Delta}CH4 and {Delta}N2O values at each height for the tracer (Eq. [1]) and IHF methods (Eq. [4]) were calculated as the differences between downwind values at the mast and the upwind value measured during the same 25-min profile period. The N2O release rate QN2O for this day was measured at 1230 h (1.0 L min–1) and at 1600 h (0.83 L min–1). This decline, which was also observed on other days, may be due to relaxation of the pressure regulator or temperature effects on cylinder pressure and flow as the cylinder cooled due to gas expansion.

For the tracer method, the total CH4 emission rate was calculated for each profile period following Eq. [1] from the weighted mean of the ratios of {Delta}[CH4]/{Delta}[N2O] at four heights and the mean value of the N2O release rate for the relevant period assuming a linear decrease between the measured values at 1230 h and 1600 h. The weights used in determining the weighted mean were the inverse-squared errors in the individual ratios. The results are shown in Fig. 3 (closed circles) and Table 2 .


Figure 3
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 3. Methane emission rates for 9 Sept. 1999 calculated using the tracer and integrated horizontal flux (IHF) methods. Closed circles, tracer method by ratio; open circles, tracer method by regression; open squares, IHF method. The individual points are offset on the x axis for clarity. The error bars are the relative precision (1{sigma}) errors on each point propagated through Eq. [1] assuming absolute errors of 0.5 nmol mol–1 for each N2O measurement, 4 nmol mol–1 for CH4, and 5% for the N2O release rate.

 

View this table:
[in this window]
[in a new window]

  		Table 2. Summary of measurements for all measured periods.

 
The alternative tracer analysis calculation (based on Eq. [1a] and regressions of five points per profile including the background point) provided similar results. A sample regression for the 1320 h period on 9 September is shown in Fig. 4 , and the resultant fluxes are shown for each period in Fig. 3 as open circles. The error bars in this case are the standard errors of the slopes and reflect the observed variability rather than that calculated from measurement precisions in the ratio method. The two measures of error are in good agreement. In the absence of error, both methods should yield identical results; the good agreement is due to the small amount of scatter in the measured data.


Figure 4
View larger version (9K):
[in this window]
[in a new window]

  		Fig. 4. Methane vs. nitrous oxide mixing ratios for the profile period beginning at 1320 h on 9 Sept. 1999. The labels are sample inlet heights and the error bars are the analytical errors, 4 nmol mol–1 for CH4, and 0.5 nmol mol–1 for N2O.

 
Results for the two other days were similar. Results for all three days are included in Table 2. All CH4 and N2O emission rates are normalized to per cow values for comparability.

Mass Balance or IHF Method
The dataset provides an excellent opportunity to test the mass balance or IHF method as an alternative analysis approach. The IHF approach can be validated by determining the N2O emission rate and comparing it with the known N2O tracer release rate, and it allows an alternative determination of the CH4 emission rate from the cows in the test plot to compare with the tracer method. The two CH4 determinations are quite independent except for the CH4 measurements themselves.

Because for N2O the true release rate is known, we first validate the IHF method for the measurement of N2O. Figure 5 shows examples of vertical profiles of {Delta}[N2O(z)] and u(z) x {Delta}[N2O(z)] measured on 9 and 16 Sept. 1999. On 9 September the inlets were at 0.5, 1, 2, and 4 m above ground, whereas on 16 September they were at 1, 2, 4, and 6 m. In both cases the N2O concentration at the top measurement level did not reduce entirely to the background reference concentration. However, in both cases the profiles allow an approximate extrapolation to a height of about 8 m where the integrand u(z) x {Delta}[N2O(z)] is forced to zero, and thus a height of 8 m was assumed as maximum integration height (zmax) in Eq. [4].


Figure 5
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 5. Sample profiles of {Delta}[N2O(z)] and u(z) x {Delta}[N2O(z)] on 9 and 16 Sept. 1999. The horizontal lines delineate the slabs whose areas are added to calculate the vertically integrated u x {Delta}C product.

 
To calculate the height integral of Eq. [4], the contributions of the individual slabs between measurements were added using a linear interpolation between the measured data points as illustrated in Fig. 5. The product of u x {Delta}[N2O] was assumed to be zero for heights of z = 0 m and z = zmax = 8 m. The wind speeds and directions, required to calculate the correction factors q(t), were averaged over 30 min preceding the time of measurement for each individual sample height. This period corresponds approximately to the time averaging in the 40-L buffers.

Uncertainties in the IHF-calculated fluxes were calculated by propagating errors through the slab-wise integration of Eq. [4]. The 1{sigma} errors in measured mixing ratios were taken as described previously, and the uncertainties in wind speed and direction were estimated from the standard deviations of measured wind speed and the calculated q(t) factor (Eq. [3]). For the uppermost slab, the error was taken to be 50% of its actual value as a conservative estimate of the uncertainty in the upper integration limit. Total error in the calculated fluxes is dominated by the uncertainty of wind speed and q(t) and, particularly for 9 and 11 September when the upper inlet was at 4 m and the upper slab makes a larger contribution, by the 50% uncertainty in the upper slab (see Fig. 5); the trace gas measurement error plays a minor role.

Figure 6 compares the IHF-derived N2O emission rate for 9 September with the release rate measured directly at the capillaries (assuming a linear decrease with time and a constant daily-average value). We find good agreement within uncertainty limits between the IHF-determined value and the release rate recorded directly at the capillaries (overall averages of 44.6 and 46.2 g head–1 d–1, respectively).


Figure 6
View larger version (9K):
[in this window]
[in a new window]

  		Fig. 6. Comparison of integrated horizontal flux (IHF)-derived N2O emission rate (points) and the release rate recorded at the capillaries assuming a constant emission at the daily average rate or a linearly decreasing rate between initial and final measured rates (lines).

 
Similarly good agreement was found for 11 and 16 September; the results are summarized in Table 2. Winds were significantly stronger and directionally more stable (i.e., q(t) from Eq. [3] was higher) on 9 September than on 11 and 16 September (see Table 1). As in the tracer analysis, profiles for which less than 20% of all wind data came from the source direction were excluded, which removed one profile from 11 September and two profiles from 16 September. The good agreement of the IHF N2O emission with the known release rate on all 3 d indicates that the wind speed correction factor q(t) satisfactorily corrects for end-effects.

Finally, we applied the same IHF method to the calculation of CH4 emissions from the cows. The results for 9 September are plotted alongside the tracer method results in Fig. 3 as open squares, and results for all days are shown in Table 2. The IHF and tracer results agree well within their uncertainties, as might be expected given the good agreement found for N2O.


   Discussion
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
Summary and Conclusions
 REFERENCES
 
Comparison of Methods
The results summarized in Table 2 show good agreement between the measured release rate of N2O and that determined by the IHF technique and for the CH4 emission rates between IHF and tracer methods. On 9 September the inlets were at 0.5, 1.2, and 4 m. In all cases, the agreement is well within the error limits of the respective measurements. Errors in the IHF technique are generally larger (i.e., up to 20% of the measured fluxes), whereas those for the tracer method are of the order of 5 to 7%. The IHF errors are dominated by the unavoidable variability of the wind speed and direction over each measurement period. The uncertainty in the upper slab contribution to the vertical integral in Eq. [4] is also significant but could be reduced in future studies with a selection of measurement heights extending to the top of the plume profile; in this case the IHF errors would reduce to 10 to 15%. In the case of the tracer measurements, only the trace gas measurements and the measured N2O release rate contribute to the errors; meteorological measurements with large variability such as wind speed and direction do not enter the quantitative calculations, and the consequent errors are smaller. Wind direction is used only to screen for periods when the measurement mast does not intercept the emission plume.

The tracer method has a clear advantage over the IHF approach in that it does not need quantitative wind speed or direction measurements and thus requires no correction for changes in wind speed or direction, such as q(t) in Eq. [3]. In effect, the tracer gas takes the place of meteorological measurements to quantify the dispersion of trace gases from the source and allows measurements in most micrometeorological conditions. The requirement that the tracer and target gases are equally dispersed by atmospheric wind and turbulence should be well met for N2O, CH4, and CO2. The tracer method may break down in very low winds or in a highly stable atmosphere with stratification due to lack of turbulent mixing; under these conditions, the assumption of constant background concentration may also fail. There is no need for a mast or vertical profiling; in principle, a single measurement in the plume suffices, although more measurements provide better statistics and averaging. The main drawbacks of the tracer technique are the requirement to measure at least two gases simultaneously in the same air mass (which represents little difficulty to FTIR measurements but requires additional equipment for other detection methods) and the requirement that tracer and emission source are sufficiently closely spaced that they are dispersed equally downwind. The tracer method achieved higher accuracy than the IHF method because only concentration measurements and the tracer release rate enter the calculations, and these quantities have less variability and measurement error than meteorological parameters such as wind speed and direction.

The advantage of the IHF technique is that it does not require a colocated tracer to determine absolute emission rates. Instead, it requires measurements of wind speed and mixing ratios at a minimum of four different heights to capture the entire plume. It can, therefore, be applied to situations where the boundary of the emission source is not as clearly defined as in the present case of a fenced-off test plot. The main requirements for an accurate measurement are (i) that the vertical profile measurements extend high enough to capture the vertical dispersion of the emitted trace gas, (ii) that the concentration enhancement be significantly larger than natural variations of the upwind reference concentration during a profile measurement, and (iii) that an adequate correction can be applied to the flux per unit width to account for the effects of changing wind direction on measured source emissions when the source is of limited cross-wind extent. The last condition requires a simple geometrical layout, where boundaries of the emission source area are perpendicular to the line of sight from the mast (e.g., a circular line source of a given length with its center at the measurement mast) or extend sufficiently beyond the range of wind directions encountered during a measurement so that the source spans the upwind area contributing to the measurements.

Nonuniform distribution of cows within the test plot is a further potential source of error for tracer and IHF techniques as implemented here. In the tracer technique, this may invalidate the assumption of colocation of CH4 and tracer release, and in the IHF technique it may invalidate the assumption of a uniform linear source. The mean differences between the IHF and tracer techniques for individual profiles were 7%, 10%, and 15% on 9, 11, and 16 September, respectively. However, these profile-to-profile differences average out considerably over a day of measurements; the corresponding differences between daily means were 2%, 2%, and 10%. We considered and rejected any attempt to introduce a quantitative correction factor to account for the nonuniform distribution of cows in the test plot. Although we have regular photographs of the cows' distribution, any such factor would entail a substantial degree of subjectivity arising from the need to not only average cow positions but also to relate this to the changing wind directions, which happen on different timescales. We feel this could not be calculated in a reliable and objective way.

A potential bias in the tracer measurements stems from the fact that the tracer release was on average further from the measurement mast than the cattle, by up to 5 m at a total distance of 40 m. Thus, the N2O tracer concentrations would be more diluted by atmospheric turbulence than the CH4 emissions, leading to an overestimate of the CH4 emissions. To estimate the magnitude of this effect, we have used a simple Gaussian plume dispersion model and simulated the experiment using the commercially available software package Windtrax (Thunder Beach Scientific, British Columbia, Canada), which uses backward Lagrangian trajectory analysis (Flesch et al., 1995; 2004) to calculate concentrations downwind of a source for given micrometeorological conditions. Both calculations suggest a possible bias of –5 to +15% for this effect under typical conditions, dependent principally on measurement height. Application of such a correction only to tracer measurements would decrease the systematic agreement between the tracer and IHF calculations. However, IHF calculations may also be biased to low CH4 emissions by the noninfinite extent of the test plot and nonuniform distribution of cows. In addition, the IHF technique is known to overestimate fluxes by 10 to 15% due to turbulent back-flux (Laubach and Kelliher, 2004; Leuning et al., 1999). We have made no attempt to make quantitative correction for these biases but recognize that both may overestimate emissions by up to 10 to 15%.

The tracer technique does not require individual handling of animals and careful laboratory analysis of permeation tubes like the SF6 tracer technique; rather, it delivers herd-average emissions and is unable to yield emission data for individual animals. On the other hand, IHF and tracer techniques are able to capture short-term variability in emission rates with a time resolution of the order of 30 min, compared with daily or longer averages provided by the SF6 tracer method (Fig. 3). As implemented here, the tracer technique is suited to short-term (a few days) investigations due to the consumption of tracer gas. It is applicable under a wide range of meteorological conditions, requiring mainly that the sampling point be downwind of the test animals and that the background concentrations are steady relative to the enhancements caused by the tracer release and test animal emissions.

Comparison with Literature Data
The average CH4 emission rate for this study, obtained by averaging all daily average CH4 emission data obtained by the two tracer and IHF methods (see Table 2), is 342 g CH4 head–1 d–1. Although this value lies within the range for mature lactating cows of 200 and 430 g CH4 head–1 d–1 reported by Lassey and Ulyatt (2000), it is perhaps surprisingly high for yearling heifers for which the UN Intergovernmental Panel on Climate Change (IPCC) estimate (Houghton et al., 1996) is 145 to 186 g CH4 head–1 d–1. However, from the known accuracy of the CH4 measurements (better than 0.2%) and the consistency of the results between the IHF and tracer methods (2–10% for the three days of measurements), we have no reason to believe the measurements to be in error. Further confirmation of their validity is available from analysis of the CO2 profile measurements taken concurrently. The CO2 emission rates provide a tracer for the general level of metabolic activity of the cows and provide a measure to which the CH4 emissions can be compared.

Tracer and IHF methods can be applied to the CO2 profile measurements to obtain an estimate of CO2 emissions from the cattle in the same way as for CH4. In the case of CO2, the interpretation is complicated by the uptake and emission of CO2 at the ground by photosynthesis and respiration of the growing pasture. Figure 7a shows the CO2 profiles together with those of N2O (as in Fig. 2). The CO2 mixing ratios are reduced due to photosynthetic uptake, particularly around the early afternoon and at the lower heights. In previous work (Griffith et al., 2002), we measured maximum CO2 uptake rates of up to 1 mg CO2 m–2 d–1 for fast-growing pasture in spring in southeast Australia. We can estimate the concentration gradient of CO2 due to such uptake near the ground from the flux gradient relationship:

Formula 4
For a gross uptake of Q = 1 mg CO2 m–2 d–1 and a realistic diffusion coefficient of K = 0.2 m2 s–1 for the wind conditions of 9 September, the resulting vertical gradient dC/dz is 2.7 µmol mol–1 m–1. The actual gradient is likely to be up to an order of magnitude lower for the cool, cloudy conditions on 9 Sept. 1999, but the drawdown of CO2 near the ground can be seen clearly in the CO2 profiles of Fig. 7a.


Figure 7
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 7. Times series of (a) vertical profiles of CO2 (solid lines) and N2O (dashed lines) and (b) calculated CO2 emission rates for 9 Sept. 1999.

 
Figure 7b shows the apparent CO2 fluxes calculated from the CO2 profiles of Fig. 7a using the tracer method. These represent a lower limit to the cow-emitted CO2 because some of the released CO2 is cancelled by the surface uptake, especially around midday and the early afternoon when photosynthetic uptake is strongest. We cannot make an accurate calculation of the net emission rate from the cows because the surface CO2 uptake is distributed over a broad upwind fetch and is not simply colocated with the tracer release or the cows. Nevertheless, we can make an estimate of the CO2 release rate from the cows from the later part of the record when we expect surface photosynthesis to be decreasing and roughly balanced by respiration. This estimate, approximately 3140 g C head–1 d–1, allows a comparison with the measured CH4 emission rates and previous measurements from the literature.

Table 3 summarizes measured CH4 and CO2 emissions from cattle from the present work and several studies from the literature. For consistency and easy comparison, in Table 3 all values are in g C head–1 d–1. The higher-than-expected CH4 emissions in the present study are accompanied by equivalently high CO2 emissions, such that the CH4/CO2 ratio is similar to those in other studies of dairy cattle. Both CO2 and CH4 emissions from the yearling heifers in this study are typical of those of mature lactating cows.


View this table:
[in this window]
[in a new window]

  		Table 3. CH4 and CO2 emission rates from the present and literature studies.

 
No feed intake data are available from the present study for comparisons, but the studies of Kinsman et al. (1995), Sauer et al. (1998), and Beauchemin and McGinn (2005) indicate that the respired CO2 represents approximately 40 to 50% of the C intake in the feed for dairy cattle and 30 to 40% for beef cattle, assuming that dry matter intake is approximately 45% C. The observed emission rates of CO2 and CH4 imply dry matter intake of the order of 15 kg head–1 d–1 for the period of our study, which is somewhat higher than the farmer's estimate of 10 kg d–1 (B. O'Neale, personal communication). However, our feeding and measurement period is confined to a few hours in the day, which began with high intake of silage feed; thus, we might expect that the measured average emissions are higher than a typical 24-h average, which includes resting times of lower emissions. Figure 3 illustrates that emissions were highest—above 500 g CH4 head–1 d–1—during the initial feeding but declined to below 300 g CH4 head–1 d–1 toward the end of the day as feeding activity slowed down. Thus, although results are quoted per day, the actual measurements may be biased to a period of relatively high metabolic activity and CH4 emissions.


   Summary and Conclusions
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Summary and Conclusions
REFERENCES
 
We have measured CH4 emissions from a small herd of free-grazing yearling dairy cows using two independent analysis techniques: a novel tracer technique and the more established IHF technique. The cattle were confined to a narrow feeding strip but free to graze normally, a tracer gas was released from the upwind boundary of the strip, and trace gas measurements were made in air sampled from a mast downwind of the strip. All trace gas measurements were made by FTIR spectroscopy. The results from both analysis methods agreed within 5 to 15% for individual profiles and 2 to 10% for daily means, with an average CH4 emission over three days of 342 g head–1 d–1. The level of emission was consistent with parallel measurements of CO2 emissions measured with the same setup but is higher than would be predicted using current IPCC Tier 1 guidelines. The discrepancy may be due in part to a high but unquantified bias of the order of 10% in both methods, in part to the restricted period during which measurements were taken (reflecting emissions after feeding rather than a true 24-h average), and in part be truly higher than IPCC guidelines infer for these cows.

The tracer method introduced here has several advantages, including higher precision than the IHF method and the fact that no quantitative meteorological measurements are required. On the other hand, it requires simultaneous measurements of at least two trace gas species. The CO2 emissions can be monitored in parallel with no extra experimental effort.

In a following paper, we will extend the technique such that the tracer gas release is made from a canister mounted to each individual cow and measure downwind concentrations using open-path FTIR spectroscopy to intercept the plume.


   ACKNOWLEDGMENTS
 
We thank the farmer, Brian O'Neale, who hosted this work and willingly provided the land, cattle, and infrastructure support. We also thank NIWA staff Tony Bromley, Ross Martin, and Mike Harvey for assistance and NIWA and the University of Wollongong for funding.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Summary and Conclusions
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Summary and Conclusions
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Griffith, D. W. T.
Right arrow Articles by Reisinger, A. R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Griffith, D. W. T.
Right arrow Articles by Reisinger, A. R.
Agricola
Right arrow Articles by Griffith, D. W. T.
Right arrow Articles by Reisinger, A. R.
Related Collections
Right arrow Global Change
Right arrow Landscape-Atmosphere Interactions
Right arrow Field evaluation techniques

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome