## Feed intake and enteric methane emissions

The feed intake of individual animals grazing in a flock or herd cannot be directly measured. To circumvent this dilemma, methods have been developed to estimate the ME requirement of grazing ruminants, including CSIRO (2007). The ME requirement for maintenance (MEm) has been determined by the amount of feed ME needed to maintain an animal's weight (Blaxter, 1989), called the live weight (LW), on a daily basis. Living can be expensive. For the ruminant, MEm sets a lower limit based on the ME required to maintain its 'machinery of life'. Based on energy dissipation via carbohydrate oxidation, measurement has focused on the respiration rate of healthy, awake, inactive, non-reproductive, fasting adults in a moderate temperature environment. Energy loss from a body depends on its surface area. For the homeotherm, this leads to a dimensional argument for the dependence of MEm on animal size via the square of a linear size variable (Brody, 1945). Extending the argument, MEm would depend on animal volume raised to a power of two thirds. Consider density and MEm would depend on the animal's live weight (kg LW), also raised to a power of 0.67. The curvilinear relation between MEm and kg LW may be portrayed graphically as a straight line if both axes are transformed to logarithmic scales. The classic synthesis of Kleiber (1932) showed the line's slope, the power coefficient, was equal to 0.75 for values of kg LW across four orders of magnitude based on measurements done for animals from rats to steers. This was verified across 27 orders of magnitude (10-18 to 1010kg LW) by West and Brown (2005). Alternatively, if the number of cells in an animal was proportional to w, one might have expected a power coefficient = 1.0, so MEm per unit of kg LW was constant. Recently, compiling the largest database to date (including 3006 species) and averaging data on the basis of life forms, Makarieva et al (2008) calculated MEm was 0.3-9W kg-1 LW (W = J s-1), a 'strikingly' narrow range they suggested was evidence for 'life's metabolic optimum'. This was indeed remarkable, but recalling Equation 8 and FCH4 ^ R, the FCH4 inventory compiler must accurately account for a heavier animal having a lower MEm per unit of w than a lighter animal.

To estimate a ruminant's ME requirement according to CSIRO (2007), feed ME content and digestibility are required. As an example, we have utilised indicative values of 10MJ ME kg-1 DM and 70 per cent, respectively, for beef cattle grazing temperate forage. The MEm was 15 per cent larger for intact males than castrated males and females, decreasing about 2 per cent per year with age and the net efficiency of use of ME for maintenance was 70 per cent, calculated from the feed ME content. As an example, for 150-700kg LW beef cattle, MEm was estimated to be 30-96MJ ME d-1. The ruminant's ME requirement may have exceeded MEm owing to the needs of food gathering and growth and for breeding females, pregnancy and lactation. These requirements have been expressed in the form of MEm multipliers. For grazing beef cattle in the same LW range as before, the multiplier was proportional to the LW, adding 11-32 per cent to the MEm. For weight gain of 0.2kg LW d-1, the ME multiplier added a further 20 per cent to the MEm. For breeding females, the pregnancy and lactation ME multiplier added another 13 and 20 per cent, respectively, to the MEm. As an example, for grazing, breeding females that weighed 430kg LW, the total ME requirement was 104MJ ME d-1, 66 per cent more than MEm. As an example for males, a 700kg LW grazing bull had a total ME requirement of 134MJ ME d-1.

While FCH4 implicitly depends on feed intake according to variable R in Equation 8 and the animal's ME requirement may be estimated, the relation between feed DMI and FCH4 may also be determined by measurement. This was applied earlier to the determination of variable m in Equation 8. Here, we propose a different approach based on the analysis of different data from an

Figure 9.1 The relationship between daily feed DMI and enteric CH4 emissions for 23 weaned lambs fed cut and carried grass in calorimeter chambers

Note: Each data record was a two-day mean during 3-13 June 2008. Linear regression yielded a slope of 17.4 ± 1.2g CH4 kg-1 DMI, offset of 3.4 ± 0.8g CH4 and 91 per cent of variability in the CH4 emissions was associated with feed DMI.

experiment purposely designed as a verification test. For weaned lambs in calorimeter chambers fed cut and carried grass, so DMI and FCH4 could be directly measured, the relation that best fitted the data was linear (Figure 9.1). The 23 data records portrayed in Figure 9.1 came from an experiment denoted FLL for feed level lamb, comprising 23 sheep <1 year old that weighed 35-41kg. The linear relation between DMI and FCH4 was interpreted to have suggested the CH4 yield, variable m in Equation 8, was constant with feed intake, indicated by the relation's slope (17.4 ± 1.2g CH4 kg-1 DMI). Although this regression accounted for 91 per cent of the variability, the relation included an offset of 3.4 ± 0.8g CH4 d-1. For the 23 records, the mean CH4 yield was 23.6 ± 0.5g CH4 kg-1 DMI.

We also explored the relation between CH4 yield and feed intake expressed as a proportion of MEm, independently calculated according to the animal's metabolic weight following CSIRO (2007). While this expression of the independent variable is different to the feed intake, the feed intake remained part of it. Thus, the relation between CH4 yield and intake as a proportion of MEm could only be explored, recognizing the limitation of having both the independent and dependent variables determined using the measured feed intake.

Determination of MEm for weaned lambs in the calorimeter chambers included resolution of an issue about an appropriate value of live weight required for the calculation. This was instructive so, candidly, we shall explain our approach. In the calorimeter chambers, the lambs were fed twice daily. For a grazing lamb, the maximum daily DMI was estimated to have been 3 per cent of the live weight. For a lamb confined to a calorimeter chamber, as an approximation, this proportion was reduced to 2 per cent. Thus, after each completed meal in the chamber, we estimated the DMI could have been up to 1 per cent of the live weight. This quantity of (ingested but undigested) feed was called 'gut fill'. We emphasize the feed intake was actually measured and this estimate has been formed only to explain the gut fill issue. Over time after the meal, digestion reduced the gut fill depending on the feed passage rate. For calculation, we considered a lamb of LW 40kg. For grass, we assumed the DM content was 15 per cent. Thus, the lamb's meal could have included up to 0.4kg DM. Including water contained in the fresh cut and carried grass, the maximum gut fill was estimated to have been 2.7kg. Consequently, the estimated maximum gut fill was equivalent to 7 per cent of the lamb's LW. For determination of the lamb's MEm based on LW, unknown gut fill represented a potentially significant bias error. For each lamb, a fasting LW was determined as the weight measured 24 hours after feeding and before placement in a calorimeter chamber. Fasting weight was used to determine MEm, providing a rational basis for the calculation.

As the feed intake, expressed as a proportion of MEm increased, the CH4 yield decreased (Figure 9.2). Increasing the proportional expression of feed intake from 1 to 2 corresponded with an 18 per cent reduction in the CH4 yield

Feed intake as a proportion of the maintenance of ME required

Figure 9.2 The relationship between feed intake as a proportion of the maintenance energy (ME) requirement and CH4 yield for 23 weaned lambs fed cut and carried grass in calorimeters chambers

Note: Linear regression yielded a slope of -4.5 ± 0.8g CH4 kg-1 DMI, offset of 29.5 ± 1.0g CH4 and 62 per cent of variability in the CH4 yield was associated with the proportional expression of feed intake.

according to the regression. Though preliminary, this analysis suggested a limitation of the common approach to estimating FCH4 according to feed intake and a mean value of the CH4 yield. To visualize this limitation, consider that increased feed intake must have physically (due to a fixed rumen volume) corresponded with increased passage rate through the lamb, so decreased time available for microbial metabolism of dietary carbohydrate in the rumen. The alternative approach to estimating CH4 yield suggested by the relation in Figure 9.2 could readily be incorporated calculations including enteric CH4 emissions inventories. Though not shown here, preliminary analyses have also suggested separating animals into types based on their age as well as physiological state may be necessary. Further research is warranted to verify these suggestions and the merit of an alternative approach for enteric CH4 emissions inventories in particular. The alternative approach would be more complex than the common, current method but should give a more accurate estimate of the 'true' CH4 emissions.