Introduction
Roughages constitute the main part of ruminant diets. Despite this importance, relatively few systematic studies have been conducted to evaluate the effect of roughage characteristics on enteric methane (CH4) emissions. There is substantial evidence for an indirect effect of the roughage feeding value on CH4 emission. Improving the feeding value of roughages leads to increased animal productivity, and reduced CH4 intensity (i.e. g CH4/kg ruminant product). However, the direct effects roughage characteristics have on enteric CH4 emissions are less well studied and, due to large variations in experimental conditions, difficult to interpret. The present work gives an outline of recent experimental work aimed specifically at evaluating the direct effect of variation in roughage characteristics on enteric CH4 emissions, and discusses the possibility of quantifying this with a process-based model.
Material and methods
The results of recently conducted trials studying the effects of roughage characteristics on enteric CH4 emissions were evaluated. Four trails with a randomised block design were conducted in climate-controlled respiration chambers at Wageningen UR. Enteric CH4 was measured as affected by N fertilisation rate and number of days regrowth of grass fed as herbage (Warner et al., Reference Warner, Hatew, Podesta, Klop, Van Gastelen, Van Laar, Dijkstra and Bannink2015; 28 cows); by N fertilisation rate and number of days regrowth of grass cut for ensiling (Warner et al., Reference Warner, Podesta, Hatew, Klop, van Laar, Bannink and Dijkstra2016; 54 cows); by number of days of regrowth of grass and stage of lactation (i.e. feed intake; 56 cows; Warner et al., unpublished results); and by date of cutting of whole-plant maize for ensiling (Hatew et al., Reference Hatew, Bannink, Van Laar, De Jonge and Dijkstra2016; 28 cows).
An existing dynamic model was used to predict enteric CH4 from feed intake and dietary characteristics as described by Bannink et al. (Reference Bannink, Van Schijndel and Dijkstra2011). The model has contributed a Tier 3 approach to measure CH4 emissions for the Emission Inventory in the Netherlands since 2005, and it represents the dynamics of enteric fermentation and digestion. It predicts CH4 emissions from enteric fermentation processes in dairy cows based on feed intake, feed substrate degradation characteristics and feed chemical composition as inputs (Heeren et al., Reference Heeren, Podesta, Hatew, Klop, Van Laar, Bannink, Warner, De Jonge and Dijkstra2014; Bannink et al., Reference Bannink, Warner, Hatew, Ellis and Dijkstra2016). The formation of volatile fatty acids as fermentation end-products is predicted, including the net production of hydrogen to be converted into CH4. As a result, this Tier 3 approach can predict how variation in roughage characteristics affects variation in CH4 yield, in contrast to the adoption of CH4 emissions as a constant percentage of gross energy intake (e.g. 6.5% in the IPCC Tier 2 approach; Intergovernmental Panel on Climate Change (IPCC), 1997).
Results and discussion
It is well established that dry matter (DM) intake is the most important driver of enteric CH4 emission rate, as established by Hammond et al. (Reference Hammond, Humphries, Crompton, Kirton and Reynolds2009), however digestibility and chemical composition did not explain observed variation in CH4. Nevertheless, model simulations by Bannink et al. (Reference Bannink, Smits, Kebreab, Mills, Ellis, Klop, France and Dijkstra2010) using the Tier 3 approach indicated potentially large effects on CH4 emission. Relatively few studies to evaluate the effects of grass characteristics on CH4 emission in ruminants have been conducted under highly controlled conditions however (Warner et al., Reference Warner, Hatew, Podesta, Klop, Van Gastelen, Van Laar, Dijkstra and Bannink2015). The recent trials evaluated in the present study were specifically designed to study the effect of variation in roughage characteristics (digestibility, rumen degradation characteristics, chemical composition) on enteric CH4 emission. Within limits, the confounding effect of feed intake and variation in feed intake between individual cows was restricted (up to 85% of voluntary intake of the cow in a single block with the lowest DM intake). The results for grass herbage (Warner et al., Reference Warner, Hatew, Podesta, Klop, Van Gastelen, Van Laar, Dijkstra and Bannink2015) and grass silage (Warner et al., Reference Warner, Podesta, Hatew, Klop, van Laar, Bannink and Dijkstra2016; Warner et al., unpublished) clearly indicated effects of grassland management and stage of lactation (i.e. DM intake level) on enteric CH4 emissions, with a profound increase of CH4 emission with increased maturity of the grass sward, and a relatively small increase with lower DM intake and lower N fertilisation rate. Methane emission differed substantially irrespective of the unit it was expressed in; by 23% for % of gross energy intake (from 5.6% to 7.3%), by 22% for g/kg DM intake (from 19.3 to 24.6 g/kg), by 28% for g/kg digested organic matter (from 27.4 to 38.1 g/kg) and by 41% for g/kg fat- and protein-corrected milk (from 10.3 to 17.4 g/kg). Grass digestibility and fibre content was also strongly affected by maturity of the grass sward, and experimental results indicated an increase in CH4 emission of 1.0 g/kg DM (about 5%) per 50 g/kg DM increase in grass NDF content, and an increase of 1.3 g/kg DM per 5% decrease in organic matter digestibility. Using the Tier 3 approach to predict these effects on CH4 (Bannink et al., Reference Bannink, Warner, Hatew, Ellis and Dijkstra2016), predicted values reproduced observed CH4 emissions with a prediction error of 14%, although a good correlation with between predicted and observed values (r=0.73). This size of prediction error falls within the estimate made for the Dutch Tier 3 approach (Bannink et al., Reference Bannink, Van Schijndel and Dijkstra2011). More careful evaluation of the model predictions indicated that predicted rumen volatile fatty acids as end-products of microbial fermentation was a main cause of CH4 prediction error for grass silages, but not for grass herbage. Prediction of feed digestibility was not a main cause.
There is also clear experimental evidence of less CH4 emitted from maize silage compared with grass silage (Brask et al., Reference Brask, Lund, Hellwing, Poulsen and Weisbjerg2013; Hammond et al., Reference Hammond, Muetzel, Waghorn, Pinares-Patino, Burke and Hoskin2015; Van Gastelen et al., Reference Van Gastelen, Antunes-Fernandes, Hettinga, Klop, Alferink, Hendriks and Dijkstra2015). However, the size of difference between these two types of roughage varies (10% to 15%), and probably must be explained from both variation in maize and grass characteristics. Testing the effect of a change in maize silage characteristics, with increased maturity from 25% to 40% DM in whole-plant maize at cutting, indicated a decrease in CH4 emission by 7%, 10% and 8% when expressed as percentage of gross energy intake, g/kg DM and g/kg fat- and protein-corrected milk, respectively (Hatew et al., Reference Hatew, Bannink, Van Laar, De Jonge and Dijkstra2016). Hence, when comparing CH4 emission from maize silage and grass silage, variation in maize silage characteristics needs to be taken into account besides that in grass silage. No model predictions are yet available for the effect of maize cutting on CH4 emissions, whereas preliminary model predictions confirmed the 15% lower CH4 emission from maize silage compared with grass silage.
The recent experimental work and the model predictions by a Tier 3 approach discussed here, indicate that a simultaneous evaluation is required of the effects of variation in grass and maize characteristics with changes in farm management and farming conditions, in order to realistically estimate consequences on enteric CH4 emission.
Conclusions
With changes in grassland management and maize harvest management there is variation in the chemical composition, rumen degradation characteristics and digestibility of grass and maize. This variation exerts a profound impact on CH4 yield with differences up to 30% and 10% for grass (silage) and maize silage, respectively. The size of this variation warrants the inclusion of the processes underlying rumen methanogenesis in modelling studies that aim to quantify the effects of roughage production and harvest, or grazing management, on on-farm greenhouse gas emissions.
Acknowledgements
This study was funded by the Ministry of Economic Affairs (The Hague, the Netherlands), the Product Board Animal Feed (Zoetermeer, the Netherlands) and the Dutch Dairy Board (Zoetermeer, the Netherlands) through the project Low Emission Animal Feed, and by the Ministry of Economic Affairs (The Hague, the Netherlands) through the international research project FACCE MACSUR – Modelling European Agriculture with Climate Change for Food Security – a FACCE JPI knowledge hub.
