Pekka Huhtanen, SLU: Dietary and animal factors influencing methane production in dairy cows and models to predict methane formation
1. Dietary and animal factors
influencing methane production in
dairy cows and models to predict
methane formation
Pekka Huhtanen, Edward Cabezas-Garcia
and Mohamamd Ramin
Swedish University of Agricultural Sciences, Sweden
Valio Ltd, Finland
4. Greenhouse gas and agriculture
~ 30% GHG emissions relate to agriculture and land use
18% livestock alone (more than the transportation sector)
Globally livestock sector contributes
9% of CO2 emissions
37% of CH4 emissions
65% of N2O emissions
64% of NH3 emission
(Steinfeld et al. 2006)
6. Fermentation stoichiometry
Production and utilisation of hydrogen (H2) determines
how much methane is produced
Carbohydrate polymers (fibre (NDF), starch, etc.) are
degraded to glucose
Glucose (hexose = C6) is degraded to 2 pyruvate
Pyruvate is fermented to VFA, CO2 and CH4
C6 2 Acetate (C2) + CO2 + CH4
C6 2 Propionate (C3) + 0.5 CO2 – 0.5 CH4
C6 1 Butyrate (C4) + 1.5 CO2 + 0.5 CH4
8. Simulation from raw data by Kittelmann et al. (2014)
y = 0.053x - 4.76
R² = 0.16
y = 0.037x + 1.14
8
10
12
14
16
18
20
22
330 350 370 390 410
CH4,g/kgDMI
CH4VFA (mmol/mol)
Lin. (Observed) Lin. (Predicted)
CH4VFA = 0.5 × C2 - 0.25 × C3 + 0.5 × C4 (Wolin, 1960)
Cabezas-Garcia (2017)
9. Wolin equation / conclusions
Good basis for understanding factors influencing CH4
production
Does not take into account other H2 sinks (microbial
cells, nitrate)
Is consistent with observed changes in VFA and CH4
when diet composition manipulated
Feed lot diets >90% concentrates, >40% propionate of
rumen VFA and <3% CH4-Energy of GE intake
Forage based dairy cow diets <50% concentrate DM, 17-
22% propionate and 6-7% CH4-E of GE
Is also related to between animal variation, but
explains rather small proportion of variability
10. How to express methane production?
Total production (g/d, L/d, MJ/d)
Mainly influenced by intake
Positively correlated to production
Methane yield (g/kg DMI, % of GEI)
Describes how much one unit of intake produce CH4
Decrease as feeding level / intake increase
Biologically more correct than total CH4
CH4 per NDF intake biologically meaningless, since other
dietary components also produce CH4
Methane intensity (g per unit of product; e.g. kg ECM,
kg LWG, kg carcass gain)
Mainly influenced by production level
Practical target to reduce CH4 per unit of product
Especially in growing cattle Body size influence
Life time intensity – takes into account longevity
11. Effects of dietary factors
on CH4 - Intake
Feed intake is the most important factor
affecting total CH4, the effects of other
factors small but significant
Explained 85 – 95% of variation in total CH4
production (treatment means data)
Small R2 for individual animal data (between
animal variation, greater random error)
Relationship between intake and CH4 not
linear; power function biologically most
relevant: CH4 = a * DMIb
12. Intake vs. methane yield
Total CH4 increase with increased DMI, but
CH4 yield decrease, because of
1. Digestibility decreases due to faster passage rate
– less truly fermented carbon per unit of intake
2. Efficiency of microbial cell synthesis increases –
more carbon to cells and less to VFA and gasses
3. Microbial cells are more reduced than fermented
CHO – microbes are H2 sink
4. Small changes in rumen fermentation pattern
towards propionate
The decrease about 10% per multiple of
maintenance
13. Factors influencing methane yield (1)
Digestibility
Methane yield increase per kg DMI or GEI with
improved digestibility, but reduce when expressed
per digestible energy
Methane can only be produced from fermentable
material
The effects of forage (grass silage) digestibility
have been conflicting
Can be confounded by intake level, microbial efficiency and
fermentation pattern
Nitrates in early harvested silages
Methane intensity most likely decrease due to
increased DMI and production
14. Factors influencing methane yield (2)
Carbohydrates (starch, fibre, NFC)
Within the range of dairy cow diets (<70-75%
concentrate DM) the effects are relatively small
With high concentrate feed-lot diets much less CH4 (2-
3 vs 6-7% of GEI)
With grass silage-based diets increased concentrate
proportion does not markedly influence rumen
fermentation: mainly increased butyrate at the expense
of reduced acetate – minor changes in propionate
The effects of concentrate supplementation variable;
sometimes even increases in CH4 with increased
concentrates (barley): e.g. Beever et al. (1988); Moss et
al. 1994)
15. Factors influencing methane yield (2)
Carbohydrates (starch, fibre, NFC)
Rather small effects within typical dairy cow diets
Grass silage grain: the increase in GHG during feed
production greater than decrease in CH4 (in a Danish
study greater GHG of barley compared with grass silage
was 14 g CH4 per kg DM)
More fermentable material to faeces – increased CH4
production from manure?
CH4 intensity can improve due to increased yield
Sustainability? Human edible output/input decreases.
Pigs and poultry utilise grain more efficiently for
protein production with much less CH4
16. Factors influencing methane yield (1)
Fat
Reduce CH4 yield
Replace fermentable CHO, often starch, in the diet
(main factor)
Biohydrogenation of unsaturated fatty acids; with
practical dietary fat levels minor contributor (1 mol
C18:2 (280 g) takes 4 h-atoms = 1 mol CH4 = 16 g (about
3.5-4.0% of daily CH4 from dairy cows)
Changes in rumen fermentation pattern towards
propionate
Grass silage-based diets rather resistant to changes
in fermentation pattern
17. Mitigation strategies (3)
Additives
Hypothesis: Influence rumen microbes
Plant extracts: not economically feasible: effective doses
used in vitro studies impossible for dairy cows
Monensin and other ionophores not allowed in EU
Nitrates; effective and could be used to replace urea as N
supplement
More expensive than urea
If NPN not required N losses increase
Alternative H2 acceptors (e.g. funarate)
3-nitroxypropanol
Has been effective (20-30% reduction) even in long term studies,
but not yet accepted
21. Between-cow variation in CH4
Highly variable values for between cow
coefficient of variation (CV) in total CH4 or
CH4 yield presented
High values usually when emissions determined
by the “Sniffer” methods
Realistic values <12-15% CV – even less
Our understanding of mechanisms of CH4
production do not support high between cow-
CV
27. 0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12 14
Days
HeadIn(proportion)
Cow1
Cow2
Cow3
Cow4
CH4 (g/d)= 230 + 276 * HP
CO2 (kg/d) = 6.3 + 6.7 * HP
28. CH4
CO2
A B
C
Gas fluxes, g/d
”Flux vs Sniffer” methods
The GreenFeed System Sniffer refers to a method which
mimics gas analysers (Garnsworthy et
al. 2012)
Sniffer, No air
flow
Sniffer, No air
flow
29. Exp. Design Diet N
Mean
(g/d)
CV
Res.
SD
Rep1
Latin square Forage 20 405 0.09 21.8 0.72
Latin square Forage 30 421 0.08 29.3 0.53
Latin square Straw 16 419 0.12 37.3 0.60
Switch-back Concentrate 16 451 0.12 22.1 0.84
Cyclic-change over For. x Conc. 16 443 0.11 37.3 0.60
Latin square Oats 16 454 0.12 21.7 0.86
Cyclic change-over Protein 29 455 0.10 24.3 0.76
Cyclic change-over Protein 24 395 0.09 31.1 0.51
Cyclic change-over Protein 25 453 0.11 30.2 0.65
Switch-back Glycerol 22 452 0.13 27.6 0.81
“GreenFeed” studies at RBD (CH4, g/d)
1Repeatability = δ2 Cow / (δ2 Cow + δ2 Residual)
31. Animal factors
Rumen microbiome
Intensive subject of research to relate variation in
methane production to rumen microbial population
Many statistical relationships found, but are they really
causal relationships
Large number of microbes and variables many
significant relationships by random
Large between animal variation can be associated with
very similar end-products
“Super-bacs” may not survive in the rumen – need a
specific niche
For example, acetogenic bacteria (use H2 to produce
acetate from CO2) exist in the rumen in small numbers
32. Animal factors
Rumen microbiome
Sometimes abundance of different methanogens have
been associated with differences in CH4, sometimes not
causal relationship?
It could be expected that if rumen microbiome is
involved in between animal variation in CH4, it should be
seen in rumen VFA pattern
Analysis of Kittelmann et al. data indicated that the
relationship between Wolin CH4/VFA and CH4 yield was
close to expected, but it explained only 16% of variation
in observed CH4
These data included selected low and high emitters, i.e. variability
likely to be greater than in “unselected” population
34. Cabezas-Garcia et al. (2017) J. Dairy Sci.
Overall (including diet, animal and period) CV 4.3% - with diets based on grass
silage differences in rumen fermentation pattern play only minor role in
Variation of CH4
36. Variance components (II)
Item
NDF pool iNDF-kp
(g/kg BW) (1/h )
Mean 19.4 0.023
CV
Total 0.255 0.242
Diet 0.088 0.086
Cow 0.130 0.090
Repeatability1
0.72 0.39
1Repeatability = δ2 Cow / (δ2 Cow + δ2 Residual)
37. Animal factors
Much more variability in variables related to
animal physiology than rumen microbiome
Repeatability higher for physiological variables
E.g. higher CV and rep for rumen VFA concentrations
than molar proportions (passage rate, rumen volume,
absorption rate)
Variation in physiological variables analogous to
intake effects
Passage rate increases
Digestibility decreases with increased passage rate
Microbial efficiency increases with passage rate
Fermentation pattern change (minor) with passage
rate
38. Methane and Digestibility
Digestibility (D) is a function of digestion rate (kd)
and passage rate (kp)
D = kd / (kd + kp)
Digestion rate is a feed (NOT ANIMAL) character
it is impossible to reduce CH4 and maintain (or improve
digestibility) by selecting low emitters (not the same
as e.g. production and fertility
Reduced D more fermentable substrate to manure
more CH4 from manure
39. Methane vs. digestibility
High Low High Low
In vivo
Goopy et al 66.4 64.8 23.5 20.8 0.6
Pinares‐Patino et al. Pasture 62.9 59.5 24.6 21.7 1.2
Pinares Pellet 63.4 58.1 10.8 6.4 1.2
Modelling
Huhtanen Cow 75.8 71.5 24.4 20.4 1.1
Sheep 77.3 73.1 29.2 25.0 1.0
Methane, g/kg DMDigestibility, %
Cellulose digestibility vs. CH4 yield (r = 0.66)
NDF digestibility vs CH4 yield (r = 0.74)
Passage rate vs. Digestibility (r = - 0.78)
Energy loss per 1 g/kd DMI in CH4 (0.185 MJ) is
greater than gain from reduced CH4 (0.055 MJ)
40. Adapted individual cow data by Shiemman et al. (1971)
y = 25.8x - 11.1
R² = 0.81
y = 37.5x - 19.0
R² = 0.85
5,0
6,0
7,0
8,0
9,0
10,0
11,0
12,0
0,68 0,70 0,72 0,74 0,76 0,78 0,80 0,82
CH4/GEI,%
GE Digestibility, g/kg
Lin. (Lactation)
Lin. (Maintenance)
41. Huhtanen et al. 2016; Anim.
Prod Sci
10 g/kg decrease in OMD decreased
CH4 yield by 0.9-1.0 g/kg DM
1 h increase in MRT 0.33 – 0.37 g/kg DM
increase in CH4 yield
42. Importance of CH4 energy
Johnson & Johnson (1995) reported a range from 2 to
12% in CH4-energy of GEI
This represents the whole range between intake levels
(maintenance vs. ad libitum) and diet composition (highly digestible
diet at low intake (12) vs. feedlot diet ad libitum (2)
Within practical dairy cow diets between-cow CV about 10%
Mean CH4 = 6.5% of GEI
GEI = 400 MJ/d (21-22 kg DM)
CH4 = 0.065 * 400 = 26 MJ/d
20% reduction = 0.2 * 24 MJ = 5.2 MJ
ME requirement of 1 kg ECM = 5.2 MJ potential for 1.0 kg ECM
At zero ME balance partitioning of incremental ME about 50:50
between body tissues and mammary gland 0.5 * 1.0 = 0.5 kg ECM
Rather small potential to increase ME intake
43. Conclusions
1. Potential to reduce CH4 in our condition rather
limited
1. High forage quality + moderate amounts of
concentrates + optimal protein
2. Economically optimal level of fat (e.g. replacement
of barley/wheat with oats)
2. Increased longevity – less CH4 from replacement
3. Selection for improved feed efficiency less CH4
per unit of product (no need to measure CH4)
4. We need to be careful in selection programs to avoid
reducing diet, especially cell wall digestibility –
ruminants’ special character (and niche) in human
food production
5. The whole system (feed production – animal –
manure) should be taken into accounr