IPCC Tier 1 methodology overestimates the carbon footprint of smallholder cattle production systems in Kenya

doi:10.21203/rs.3.rs-3148140/v1

Download PDF

Research Article

IPCC Tier 1 methodology overestimates the carbon footprint of smallholder cattle production systems in Kenya

https://doi.org/10.21203/rs.3.rs-3148140/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Most Sub-Saharan countries still rely on the basic Intergovernmental Panel on Climate Change's (IPCC) Tier 1 methodology for reporting greenhouse gas (GHG) emissions from their livestock sector. This approach is, however, associated with high uncertainty and cannot detect changes in GHG emissions resulting from mitigation interventions. The present study aimed to quantify the effect of replacing the IPCC Tier 1 methodology with Tier 2 approaches in assessing total farmgate GHG emissions and products' emission intensities (EIs) in smallholder cattle production in Western Kenya. Additionally, the study assessed the effect of using locally measured methane yield (MY) instead of regional defaults in Tier 2 estimations. A partial life cycle assessment (LCA) was conducted using aggregated activity data from two contrasting study sites, one with moderately performing and another with low-performing cattle production systems. The total farmgate GHG emissions were allocated to the co-products, namely milk and live weight gain, using the novel energy expenditure allocation method, in comparison to the traditionally used protein mass allocation. Compared Tier 1 methodology, the use of Tier 2 approaches resulted in 18-44% lower total farmgate GHG emissions and milk EIs across the two study sites. When the IPCC regional MY default was substituted with locally measured MY, the total GHG emissions and milk EI increased by 19-25%. These findings suggest that the continued use of IPCC Tier 1 methodology or Tier 2 MY defaults by East African countries are resulting in considerable uncertainty in their livestock GHG inventory reporting. There is therefore a need for these developing nations to adopt higher-tier methodologies as well as measure and utilize locally relevant MY for more accurate inventory reporting and to be able to assess suitable GHG mitigation strategies for local cattle production systems. Additionally, the study found that the energy expenditure allocation method apportioned a lower proportion of GHG emissions to milk production compared to live weight gain, reflecting the greater nutrient use efficiency by cattle in producing milk than gaining weight. As developing countries are faced with the double challenge of increasing animal-sourced food to feed the growing population and a need to reduce GHG emissions associated with food production, it is sensible for these countries to report changes in product EIs rather than absolute emissions. It is, therefore, imperative to continually refine LCA methodology for more relevance in multifunctional smallholder systems for unbiased measurement of products' EIs from such systems.

Animal Science

Sub-Saharan Africa

emission factors

emission intensity

milk

live weight gain

life cycle assessment.

Use of Tier 2 resulted in 19-25% lower total GHG emissions compared to Tier 1.
Using locally measured methane yield instead of the defaults resulted in 19-25% lower GHG emission estimates.
Energy allocation method results in lower milk EI compared to using protein output.

Livestock production plays a significant role in supporting the livelihood and food security of the majority of the world's vulnerable populations, yet the sector has also associated with approximately 11.3% of the global anthropogenic greenhouse gas (GHG) emissions that are driving global warming (GLEAM, 2020). Cattle's milk and beef production account for about 61% of global anthropogenic methane (CH₄) emissions (Gerber et al., 2013), with the emissions varying from one region to another depending on animal numbers and level of productivity (GLEAM, 2017). Though livestock emissions from Sub-Saharan Africa (SSA) account for only about 5.5% of global livestock emissions, this is expected to increase significantly going into the future as the demand for animal products is expected to more than double by 2050 driven by the rapidly growing human population, increasing urbanisation and households' per capita income (Balehegn et al., 2021). Enteric CH₄ that is produced during microbes-mediated digestion of feed in the rumen is a highly potent GHG with a global warming potential of about 28 times that of carbon dioxide (CO₂) over 100 years (Myhre et al., 2014) and is the primary GHG emitted from livestock production activities. This is especially the case in the low-input ruminant production systems where animals are fed mainly on high-forage diets that are often of low quality and, consequently, associated with higher CH₄ production per unit feed intake because of low animal production (Poore and Nemecek, 2018).

In East Africa, cattle production is primarily characterized by two types of systems: smallholder mixed crop-livestock systems and extensive pastoral systems. These production systems are characterized by low capital investment, resulting in lower levels of animal production (Thornton, 2010). Consequently, animal products such as milk and beef in the region have been associated with higher emission intensities (EI), which refers to the amount of greenhouse gas (GHG) emissions produced per unit of animal product compared to similar products from the Global North (Opio et al., 2013). For example, the average carbon footprint of fat- and protein-corrected milk (FPCM) in Africa is estimated at 7.5–7.6 kg CO₂-eq./kg (Opio et al., 2013; Poore and Nemecek, 2018) compared to 1.6–1.7 kg CO₂-eq./kg for Europe (Opio et al., 2013), with the main differences attributed to differences in milk yield per animal. Recent LCA work quantifying milk and meat EIs from cattle production systems at the farm level in three counties in Western Kenya reported an average FPCM EI of 2.3 kg CO₂-eq./kg (Ndung'u et al., 2022), which was lower than what has been predicted by global models for SSA. The FPCM EIs were, however, highly variable from one county to another depending on the animal productivity, ranging between 2.1-5.0 kg CO₂-eq./kg.

Except for Ethiopia and Kenya, all SSA countries continue to rely on the most basic inventory reporting methodology, the Intergovernmental Panel on Climate Change's (IPCC) Tier 1 (Graham et al., 2022). This is despite the expected high uncertainty associated with this methodology. The continued use of Tier 1 is because of a lack of technical and structural capacity to generate and compile locally relevant data to allow them to transition to a higher tier methodology. The Tier 1 methodology employs default emission factors (EFs: total annual GHG emissions per animal) that have been extrapolated from studies conducted in the Global North, using temperate breeds of animals and feed resources. This raises concerns because of the known significant differences between temperate and tropical environments in animal diets (Archimede et al., 2018) and cattle breeds (Kurihara et al., 1999; Goopy et al., 2020). Also, the Tier 1 methodology cannot detect changes in GHG emissions as a result of changes in production practises and therefore deny these countries an opportunity to test and select locally appropriate GHG mitigation interventions for their livestock production systems. There is an urgent need, therefore, for these countries to move to a higher tier methodology to be able to meet their climate obligations that they signed into in the Paris Climate Change Agreement (UNFCC, 2015).

To date, SSA is lagging behind compared to the other regions of the world in terms of carrying out actual GHG emission measurements from its cattle production and the associated manure management systems (Graham et al., 2022). As a result, the default GHG emissions factors currently recommended in the IPCC Tier 2 methodology for the region are largely not informed by data that reflect the local cattle production situations. In the recent past, however, limited studies quantifying enteric CH₄ (Goopy et al., 2020; Korir et al., 2022; Korir et al., 2022b) and manure emissions (Pelster et al., 2016; Zhu et al., 2018; Zhu et al., 2020; Leitner et al., 2021; Zhu et al., 2021) have been published from East Africa (Supplementary Table 1). From these peer-reviewed studies, there is a consensus that the current default methane yield (MY; g CH₄ per kg dry matter intake) recommended for cattle production systems in SSA in the IPCC methodologies underestimate the true MYs by up to 25% (Goopy et al., 2020). On the other hand, the manure emissions factors (CH₄ and nitrous oxide (N₂O) output per unit of animal excreta) recommendations are thought to overestimate true emissions by at least two folds (Leitner et al., 2021). The available local measurementssubstantiating these two hypotheses, however, still limited and do not representative of all cattle production and manure management systems in SSA. There is a need, therefore, for more measurements from a broader range of production systems to validate the preliminary findings reported in the cited studies.

Life cycle assessment has been used extensively to assess GHG emissions and the carbon footprint of cattle products in commercial livestock production systems (Baldini et al., 2017), with some work also reported recently from smallholder systems in the Global South (Udo et al., 2016; Apdini et al., 2021; Ndung'u et al., 2022). The limited nature of LCA studies from smallholder systems in the Global South has been attributed to the scarcity of farm-level data from these production systems and also because of the complexities associated with such studies in these multifunctional production systems (Weiler et al., 2014; Salmon et al., 2018). The two common challenges that have been described in applying LCA in smallholder systems are, first, quantifying all system outputs, both tangible and intangible (Weiler et al., 2014) and, secondly, objectively allocating GHG emissions to all system outputs (co-products) in a manner that is not biased (Mackenzie et al., 2017). Because of these complexities, most LCAs conducted in these systems have led to the omission of some of the system outputs or skewed allocation of emissions to the products, and hence the net EIs reported are likely to have been overestimated (Weiler et al., 2014). Reduction in products' EIs is acknowledged by the United Nations Framework Convention on Climate Change (UNFCC) as one of the pathways that countries can explore in their Nationally Determined Contributions (NDCs) to meet their Paris Climate Change Agreement commitments. Therefore, applying LCA in smallholder cattle production systems to quantify Products' EIs needs to be optimised for more accurate measurements and reporting.

The present study used an LCA approach to assess the effect of using Tier 2 approaches instead of the IPCC Tier 1 in calculating the total farmgate GHG emissions and IEs in two contrasting smallholder cattle production systems in Western Kenya. The effect of using locally measured emission factors instead of regional defaults was also quantified. The effect of using the novel energy expenditure allocation method in LCA compared to protein allocation on total farmgate emissions and EIs was also tested. We hypothesised that: i) the use of Tier 2 approaches would result in lower EIs compared to using the IPCC Tier 1 approach, ii) the EI of animal products (milk and LWG) in the moderately producing site (highlands) would be lower compared to the low producing site (lowlands), iii) the substitution of IPCC Tier 2 MY with locally measured MY values would result in higher EIs irrespective of the study site and, iv) GHG allocation based on proportional energy expenditure on cattle production would result in similar EIs compared to using the protein mass allocation, irrespective of the level of production.

2.1. Study site and activity data collection

The present study was conducted using farm-level data collected from two study sites in Western Kenya: Nandi, characterised by moderately producing cattle production, subsequently called 'highlands', and Nyando, characterised by low-producing animals, hereafter referred to as 'lowlands'. The data was collected continuously for 12 months, with the surveys conducted in the lowlands between 2014 and 2015 (Goopy et al., 2018) and in the highlands between 2015 and 2016 (Ndung'u et al., 2018). The year-long data collected in the two study sites included animal productivity parameters (live weight changes and milk yield), animal numbers and their ages, production status, feed availability and quality over time and the average daily distance travelled by the animals in search of feed and pasture. A summary of climate, soil type, land ownership and herd compositions in the two study sites are shown in Table 1.

Table 1

Site description and herd characteristics of the two study sites, the lowlands (Nyando) and the highlands (Nandi) in western Kenya. The table values correspond to each study site's mean values.
Parameter	Highlands	Lowlands
Number of farms	126	56
Climatic factors
• Annual rainfall (mm)	1,200-2,000¹	1,100-1,725²
• Mean annual temperature (°C)¹	15–21	18–24
Annual seasons	Two rainy seasons and two dry seasons.	Two rainy seasons and two dry seasons.
Elevation (metres above sea level)¹	1,200-2,400	1,300-2,400
Agro-ecological zone (description)	Moderately cool and humid.	Mainly warm and subhumid and, in some regions, cool and semi-arid.
Soil type^{2, 3}	Nitisol, acrisol.	Nitisol, regosol, leptosol, vertisol, arenosol, adenosol, planosol.
Average land ownership¹ (hectares)	1.3	2.0
Cattle ownership no. per household¹	9	8
Herd structure¹
• Mature females (% of the total herd)	42	37
• Lactating animals (% of mature females)	40	16
• Mature males (% of the total herd)	4	23
• Heifers (% of the total herd)	14	3
• Growing males (% of the total herd)	50	3
• Calves (% of the total herd)	35	34
Animals sold (% of the total herd within a year)¹	24	16
Animals bought (% of the total herd annually)¹	11	9
Animal performance
• FPCM yield (L/d/animal)¹	4.1 ± 0.23	2.2 ± 0.04
• Average daily gain (g/day)⁴	170 ± 28	60 ± 15
Feed basket characteristics (± s.e)
• Predicted in vivo DMD⁵ (%)	63.4 ± 0.54	59.0 ± 1.07
• CP (% of DM)	8.7 ± 0.38	7.7 ± 0.81

Source: ¹Ndung'u et al. (2018); ²Ndung'u et al., 2022; ³Goopy et al. (2018); ⁴ This is an average of all animals in the herd; ⁵$DMD \left(g/100g DM\right)=83.58-0.824 \times ADF \left(g/100g DM\right)+(2.626 \times N \left(g/100g DM\right)$ (Oddy et al., 1983); DMD: dry matter digestibility; CP: crude protein; ADF: acid detergent fibre; N: nitrogen. The annual rainfall range represents measurements taken between 2001 and 2010; s.e: standard error.

Smallholder mixed crop-livestock farming was the predominant economic activity in both study sites. However, there were notable differences in terms of animal breeds, crops grown, and feed quality and availability across the seasons. These variations have been described by Ndung'u et al. (2018) for the highlands and Goopy et al. (2018) for the lowlands. The cattle kept in the lowlands were predominantly Bos taurus indicus (small East African Zebu), while in the highlands, up to 80% of the herd consisted of either pure-bred exotic dairy cattle (Friesian, Ayrshire, and Jersey) or indiscriminate crossbreds between dairy breeds and the native small East African Zebu. Most of the animals across the two sites were kept in grazing systems all year round, with the animals only kept in kraals at night for security reasons. Less than 20% of the animals were either under intensive (zero-grazing) or semi-intensive systems. The bulk of the animal feed available in both study sites consisted of natural pastures and planted fodder grasses grown on-farm with minimal fertiliser or manure amendments. To a small extent (less than 20% of the total ration), crop residues were fed to the animals, with the material varying from site to site and season depending on the crops grown. Maize stovers were the dominant crop residues in the highlands, and sugarcane tops in the lowlands. The seasonal feed basket (proportional contribution of the single feed item to the overall diet) dry matter digestibility (DMD) ranged between 62.6–64.9% across seasons in the highlands (Ndung'u et al., 2018), while in the lowlands, was between 56.9–60.9% (Goopy et al., 2018). Labour needs for animal production activities were provided almost exclusively by household members in both sites, and a negligible proportion of the farms surveyed used any form of mechanisation in animal-related activities.

The data collection was done in a total of five farm visits that were scheduled at an interval of about three months apart (at the beginning of the study, after 3, 6, 9, and 12 months). The visits were purposefully planned to capture seasonal changes in feed availability and quality over the four seasons (two rainy and two dry seasons) in each of the two study sites. Representative feed samples from all the feed types identified in the farms were taken for nutritional analysis. All cattle found on the farms during each visit, including calves born between the visits, were weighed using a digital platform scale (Model EKW, Endeavor Instrument Africa Limited, Nairobi). Their ages were determined using dentation or farmer's recall for young animals. Daily milk yield data for all the lactating animals in the surveyed households was obtained from the farmers' records. At the beginning of the study, all farmers were trained and supplied with graduated plastic containers for measuring the milk output and notebooks for recording the data. One composite milk sample per farm in the highlands was taken during the second farm visit for analysis to obtain an average milk quality in the farms. Milk energy content, a parameter required in estimating energy requirements for lactation and hence DMI by the animals according to IPCC methodology for calculation enteric CH₄ emissions, was calculated based on the equation by Tyrrell and Reid (1965) from milk's butter fat and total solid content. For the lowland, where milk samples were not collected because of limitations in analytical capacity in the region, an energy content of 3.054 MJ/kg derived from the Australian grazing livestock context (CSIRO, 2007) was used instead. Information on inputs imported into the farms and general management practices were captured through structured interviews with the household heads during each farm visit.

The feed samples collected during each farm visit were air-dried in the field to constant weight before being transported to the laboratory for further processing and analysis. At the laboratory, samples were dried at 50°C for 72 hours, ground using a hammer mill fitted with a 1 mm sieve, and pooled proportionately based on their contribution to the feed basket as reported by Ndung'u et al. (2018) and Goopy et al. (2018). The pooled samples were analysed for dry matter (DM), organic matter (OM), neutral detergent fibre (NDF) and acid detergent fibre (ADF) using standard Association of Official Analytical Chemists procedures (AOAC, 1990). The feed's apparent dry matter digestibility (DMD) was estimated using Eq. 1 below by (Oddy et al., 1983):

$$DMD (g∕100 g DM)=83.58-0.824 \times ADF (g∕100 g DM)+(2.626 \times N (g/100 g DM)$$

Where: DMD is apparent dry matter digestibility, ADF is acid detergent fibre, and N is nitrogen.

The CP content and the estimated apparent DMD in the annual feed basket in the two study sites are shown in Table 1. Further information on feed basket characteristics in the two study sites was reported in the accompanying studies by (Goopy et al., 2018) for the lowlands and Ndung'u et al. (2018) and the highlands.

The farm-level data from the sampled households in each study site were clustered and treated as two large data sets rather than individual farms. Thus, the calculations done in the present LCA study were at the study site level, with the data summarised in Table 1 used.

2.2. System boundary

The system boundary in the present study was set at "cradle to farm gate", thus, including CH₄ emissions associated with enteric fermentation and CH₄ and N₂O from dung, urine and manure management, but excluded emissions post-farm gate (Fig. 1). Upstream emissions, i.e., associated with farm inputs such as concentrates and fertiliser, were assumed to be negligible based on the findings by Ndung'u et al. (2022) while analysing data from the same households where these emission sources were estimated to account for less than 1% of total farm emissions cumulatively. Because there was low level of farm mechanisation in both study sites, GHG emissions associated with energy use were also assumed to be negligible and, therefore, not included in the assessment.

Cumulative milk output from all lactating cows and LWG by all the animal classes (male and female calves, growing young males, growing young females, mature males and mature females) in each herd were considered co-products from the two cattle production systems. Even though cattle are known to play additional economic and social roles in the two study sites, such as store of wealth (savings), providing draft power and source of manure, amongst others, these products and services were not considered in the present LCA because the necessary data to compute all these other outputs were not available.

2.3. Inventory analysis

2.3.1. Scenarios modelled and GHG emission calculations

Five scenarios were modelled for each study site to calculate on-farm GHG emissions associated with cattle production activities. The scenarios modelled estimated total GHG emissions at the farm level, including CH₄ enteric fermentation and CH₄ and N₂O from animal excreta. The other on-farm emission sources were considered negligible, as described earlier. Products' (milk and LWG) EIs were then computed by dividing the total annual GHG output (kg CO₂-eq.) by the total herds' level product output. In Scenario 1, "IPCC Tier 1", the IPCC Tier 1 methodology was used to compute enteric CH₄ emissions where the animals were first classified as either "lactating cows" or "others". The number of animals in each category was then multiplied by an appropriate IPCC Tier 1 EF (kg CH₄ per head per year), i.e., 76 kg for "lactating cows" and 52 kg for the "other animals" (Gavrilova et al., 2019). In Scenario 2 (IPCC Tier 2), the IPCC Tier 2 methodology was used as described by (Gavrilova et al., 2019). Briefly, the activity data collected at the farm level as described in section 2.1, including animal categories based on production systems, sex and age, animal live weights and live weight changes in each category, milk yield, distances travelled by the animals in search of feed and water, seasonal variation in feed quality and changes in animal numbers were used to estimate daily dry matter intake by the animals. The cumulative feed intake per season for each animal class was then multiplied by the default MY of 23.3 g CH₄/kg DMI as recommended in the IPCC methodology (Gavrilova et al., 2019) to obtain seasonal enteric CH₄ production. Seasonal CH₄ emissions were then summed to obtain annual emissions. Scenario 3, "IPCC Local MY", also used the IPCC Tier 2 methodology as described in Scenario 2, but the IPCC default MY used was substituted by 26.5 g CH₄/kg DMI derived from the animal respiration chamber measurements conducted with local cattle fed on locally available diets in Kenya (Korir et al., 2022; Korir et al., 2022b). For Scenario 4, "IPCC submaint." the approach and MY (26.5 g/ kg DMI) used in Scenario 3 were applied during the seasons when the animals were gaining weight (suggestive of adequate nutrition). For the season (one out of the four seasons) when the animals were losing weight (suggestive of underfeeding: sub-maintenance nutrition), an MY of 30.8 g/kg DMI derived from animal measurements conducted with cattle fed at restricted levels of feed intake – this representing underfeeding in the dry season that is known to increase enteric CH₄ emissions (Goopy et al., 2020). Scenario 5, "CSIRO Tier 2", employed a modified Tier 2 methodology as described by (Goopy et al., 2018) based on metabolisable energy (ME) equations derived from CSIRO (2007) in estimating feed intake by the animals, instead of net energy equations outlined in IPCC Tier 2 methodology. An MY of 20.7 g/kg DMI derived from the Australian cattle production systems (Charmley et al., 2016) was also used to calculate enteric CH₄ from the estimated DMI. Manure GHG emissions in the modelled scenarios were calculated as described in Section 2.3.2 below.

2.3.2. The excreta GHG emissions

The GHG emissions emitted from animal excreta (manure, dung and urine) were calculated by multiplying estimated total excreta by default emission factors (EF; CH₄ or N₂O emissions per unit excreta), taking into consideration the management practise: i.e., either managed manure or urine and dung deposited in pastures. The total dung excreted by the animals was estimated as the undigested fraction of the feed consumed and was calculated according to Eq. 2 below:

$\text{T}\text{o}\text{t}\text{a}\text{l} \text{d}\text{u}\text{n}\text{g} \text{o}\text{u}\text{t}\text{p}\text{u}\text{t} \left(\text{o}\text{r}\text{g}\text{a}\text{n}\text{i}\text{c} \text{m}\text{a}\text{t}\text{t}\text{e}\text{r} \text{o}\text{r} \text{N}\right)=Total feed intake\times (100\%-\%DMD)$ (2).

Where DMD is the apparent feed basket dry matter digestibility, calculated using the equation described earlier (Eq. 1).

Nitrous oxide was calculated based on the total N in the dung and urine output, while CH₄ was estimated based on the total carbon excreted. The manure N content was calculated using a mass balance approach described in the IPCC methodology (Gavrilova et al., 2019), where the excreted N was assumed to be the difference between the total N in the diet consumed by the animals and the total N spent on production, i.e., LWG and milk production. The manure carbon content was assumed to be 40% based on empirical data from smallholder cattle production conditions in Kenya, with cattle fed on comparable diets (Leitner et al., 2021). Based on farm survey data from the two study sites, as reported by (Ndung'u et al., 2022), the three excreta management practices were identified in both the highlands and the lowlands as follows: (i) 50% of the total urine and dung was excreted in the grazing fields during grazing time (~ 10 h); the remaining was deposited in the overnight kraals. Half of this excreta deposited in kraals was assumed to be (ii) piled (25% of total manure), and the other half was assumed to be (iii) unpiled (25% of the total manure). These three excreta management practices were considered in the five scenarios modelled.

When calculating GHG emissions from the excreta, Locally measured mean excreta N₂O EF of 0.46% (N₂O - N emissions/excreta N) and CH₄ EFs of 0.25% (CH₄/volatile solids in the excreta) were used for Scenarios 3, 4 and 5. A detailed summary of the available studies on cattle excreta emissions in Kenya that were included in generating the EFs used is provided in Supplementary Table 1. For Scenario 1 and 2, IPCC 2019 default factors: 1% for N₂O and 0.44% for CH₄ were used (Gavrilova et al., 2019). Indirect N₂O emissions associated with N leaching and volatilisation were assumed to be negligible because of the low CP content in the animal diets and hence N content in the animal excreta. The GHG emissions from manure utilised on-farm in fodder production were also assumed to be negligible because only a small proportion (> 10%) of the heaped manure was applied to the fields (Ndung'u et al., 2022).

2.4. Impact assessment and interpretation

The total farmgate GHG emissions from cattle production were considered as the sum of enteric CH₄ and manure CH₄ and N₂O emissions from the cattle in the farms surveyed. The total CH₄ and N₂O emissions were converted to carbon dioxide equivalent (CO₂-eq.) using the global warming potential (GWP) with an atmospheric lifespan of 100 years (GWP₁₀₀) for CH₄ and N₂O (28 and 265, respectively; (Myhre et al., 2014). Total GHG emissions were expressed per tropical livestock unit (TLU); kg CO₂ -eq. per TLU to correct for animal number differences across the two study sites and per unit product output (kg CO₂ -eq. per unit output; either FPCM, LWG or milk and meat CP).

2.5. Functional unit and allocation procedures

Emissions per unit CP output (CP in milk and LWG) (kg CO₂-eq./kg CP) was used as the primary functional unit, allowing GHG emissions to be attributed to milk and meat as co-products from the system. Emissions were also expressed per FPCM and LWG to allow a comparison of our findings with other studies reporting these parameters. Total milk CP was determined by multiplying cumulative annual milk yield from the herd by measured milk CP content (%), while meat CP was calculated from the total boneless meat plus CP in edible offals (liver, heart, intestines, tripe, lungs, kidney, tongue and pancreas). Meat output was estimated from total LWG by all the animals in the herd, assuming an average dressing percentage of 52.1% for all animal classes (calves, heifers, growing bulls, mature cows and mature bulls) (Muchenje et al., 2008), meat yield of 85.0% (Mummed and Webb, 2019) and meat CP content of 21.0% (on wet basis) (Muchenje et al., 2008). Edible offals were included in the meat output and were estimated at 5.3% of total LW with a CP content of 18.2% (on a wet basis) (Nollet and Toldrá, 2019).

Total GHG emissions were allocated to the co-products using three methods: 1) all emissions were allocated to either milk, milk CP LWG or meat CP (further on called 'no allocation'); 2) all emissions were allocated proportionally to milk and LWG (or milk and meat CP) based on CP mass in the two products produced, further on named 'protein allocation'), and 3) all emissions were allocated to the two products based on the proportional energy spent by the animals at herd level in producing milk to that spent in LWG (further on named 'energy allocation'). For the energy allocation approach, the energy (kilojoules) input required to produce a unit of milk (kg) and meat (kg) at the herd level was calculated, and the ratio between the two was used to apportion total GHG emissions to the two products. Energy spent by the animals on maintenance was partitioned proportionately between milk and meat based on the energy spent on producing the two products.

3.1. Effects of the study site, GHG emission accounting methodology and methane yield used on total farm greenhouse gas emissions and emission intensities.

3.1.1. Total emissions

Figure 2 below shows the proportional contribution of on-farm GHG emission sources (enteric CH₄ and manure emissions) to the total farmgate emissions for the five scenarios modelled across the two study sites. Irrespective of the GHG accounting methodology (scenario) and study site, enteric CH₄ formed the primary source of on-farm GHG emissions contributing between 94–97% to total farmgate GHG emissions. Using locally measured MYs (IPCC local and IPCC submaint.) to calculate enteric emissions resulted in the highest enteric CH₄ contributions in both the highlands (96%) and in the lowlands (97%). The lowest enteric CH₄ contribution (94%) was observed when the IPCC Tier 2 and CSIRO Tier 2 scenarios were modelled. However, the differences observed between the scenarios and across the study sites were minimal (Fig. 2).

When total farmgate GHG emissions were expressed per animal per year, IPCC Tier 1 methodology resulted in numerically higher EFs in both study sites (1.94 t CO₂-eq./TLU/year in the highlands and 2.13 t CO₂-eq./TLU/year in the lowlands). When either of the Tier 2 approaches modelled (IPCC Tier 2, IPCC local, IPCC submaint. and CSIRO Tier 2) were used in comparison to IPCC Tier 1, the GHG emissions were reduced by 22–41% in the highlands and by 19–44% in the lowlands, with the highest reductions reported when CSIRO Tier 2 was modelled in both study sites. There were, however, minimal variations between the study sites (Table 2).

Table 2

Comparison of total GHG emissions per head (t CO₂-eq./tropical livestock unit (TLU)/year) and carbon footprint (kg CO₂-eq./kg) of cattle products using three different allocation methods (no allocation, protein allocation and energy allocation) of fat and protein corrected milk (FPCM), live weight gain (LWG) and milk and meat CP output across five scenarios¹ modelled in the highland (Nandi) and lowland (Nyando) study site
	IPCC Tier 1		IPCC Tier 2		IPCC Local		IPCC submaint.		CSIRO Tier 2
Parameter	Highlands	Lowlands	Highlands	Lowlands	Highlands	Lowlands	Highlands	Lowlands	Highlands	Lowlands
Predicted Daily feed intake (kg/day/head)
Annual GHG emissions (t CO₂-eq./TLU¹)	1.9	2.13	1.20	1.36	1.47	1.67	1.52	1.73	1.15	1.29
No allocation
FPCM (kg CO₂-eq./kg)	5.1	24	3.2	15	3.8	18.2	4.0	19	3.0	13
Milk CP (kg CO₂-eq./kg)	153	712	97	466	116	552	120	572	92	397
LWG (kg CO₂-eq./kg)	36	90	23	59	28	70	29	72	22	50
Meat CP (kg CO₂-eq./kg)	358	876	226	573	272	679	281	704	216	489
Protein allocation (kg CO₂-eq./kg)
Emissions allocated to milk (%)	70	55	70	55	70	55	70	55	70	55
Emissions allocated to LWG (%)	30	45	30	45	30	45	30	45	30	45
FPCM (kg CO₂-eq./kg)	3.5	13	2.2	8.5	2.7	10	2.8	10	2.1	7.2
LWG (kg CO₂-eq./kg)	11	40	6.9	26	8.4	31	8.6	32	6.6	23
Milk + Meat CP (kg CO₂-eq./kg)	107	393	68	257	82	305	84	316	65	219
Energy allocation (kg CO₂-eq./kg)
Emissions allocated to milk (%)	60	13	60	13	60	13	60	13	60	13
Emissions allocated to LWG (%)	40	87	40	87	40	87	40	87	40	87
FPCM (kg CO₂-eq./kg)	3.0	3.0	1.9	2.0	2.3	2.4	2.4	2.4	1.8	1.7
Milk CP (kg CO₂-eq./kg)	92	92	58	60	70	71	72	74	55	51
LWG (kg CO₂-eq./kg)	14	78	9.3	51	11	61	12	63	8.8	44
Meat CP (kg CO₂-eq./kg)	143	763	90	499	109	591	112	613	86	426

¹One tropical livestock unit (TLU) = 250 kg LW (Jahnke and Jahnke, 1982).

3.1.2. Emission intensities

The use of IPCC Tier 1 methodology resulted in higher milk and LWG EIs compared to IPCC Tier 2 (34–37%), IPCC local (22–25%), IPCC submaint (20–22%) and CSIRO Tier 2 (40–44%), irrespective of the GHG allocation method used (Fig. 3). When the IPCC Tier 2 default MY (IPCC Tier 2) was substituted with locally measured values (IPCC local), milk and LWG EIs increased by 19–21% across the two study sites. When sub-maintenance nutrition during periods of feed scarcity was considered (IPCC submaint.), the EIs were increased by a further 3–5%. Modelling of the CSIRO Tier 2 approach resulted in 4–5% lower EIs in the highlands and 15% in the lowlands, compared to using IPCC Tier 2 methodology (Table 2).

Table 2 shows the EIs obtained from the five scenarios modelled using the three different GHG allocation methods tested. When no allocation or protein allocation methods were used, the lowlands (low production site) had 4–5 times higher FPCM and milk CP EIs (kg CO₂-eq./kg) and 2–4 times higher LWG and meat CP EIs compared to the highlands (moderately producing site), regardless of the GHG accounting methodology. The energy expenditure allocation method, however, resulted in almost similar EIs between the highlands and the lowlands, with less than 10% differences in the FPCM values obtained across the two study sites. The LWG and meat CP EIs were also about five times higher in the lowlands than in the highlands, irrespective of the GHG allocation methods tested (Table 2).

3.2. Effects allocation methods on emission intensity

Among the three allocation methods examined, attributing all greenhouse gas (GHG) emissions to a single production system output (either milk or weight gain; no allocation) resulted in the highest emissions intensities (EIs), regardless of the GHG accounting methodology used (Table 2). When the protein allocation method was applied in comparison to no allocation, FCPM and milk CP EI (kg CO₂-eq./kg) was reduced by 30% in the highlands and 45% in the lowlands, while LWG and meat CP EIs (kg CO₂-eq./kg) were reduced by 70% in the highlands and 55% in the lowlands, irrespective of the accounting methodology. In comparison to no allocation, using the energy allocation method reduced FPCM and milk CP EI (kg CO₂-eq./kg) by 40% in the highlands and by 87% in the lowlands across the five GHG accounting methods, while LWG and meat CP EI (kg CO₂-eq./kg) was reduced by 60% in the highlands and 13% in the lowlands. When comparing the energy allocation method to the protein allocation method, FPCM and milk CP EIs were reduced by 14% in the highlands and 77% in the lowlands. However, the LWG and meat CP EIs increased by 33% in the highlands and 94% in the lowlands, irrespective of the GHG accounting methods used (Table 2).

4.1. LCA application in multifunctional smallholder systems

The present study quantified total farmgate GHG emissions associated with cattle production activities and cattle products' (milk and live weight gain) EIs in two contrasting study sites in Western Kenya that predominantly practise smallholder mixed crop-livestock production, using an LCA approach. Cattle production is practised in the two study sites for multiple reasons, including as a source of food (milk) for both family consumption and for sale, as a store of wealth, for draft power used for land cultivation, cultural use such as paying dowry, for identity among other reasons. The multifunctional nature of these production units, compared to specialised commercial dairy or beef systems, poses unique challenges when conducting LCA. These challenges include quantifying the diverse non-homogeneous system outputs, expressing them in a standardized unit, and allocating environmental impact (GHG emissions). To date, there are no standardised methods for quantifying some of the services provided by cattle in the smallholder context, such as using livestock as a store of wealth and as an identity for communities, and consequently, some of these uses (system outputs) are often overlooked. The exemption of these outputs results in an underestimation of the actual cattle production systems output in such multifunctional systems, and as such, leads to an overestimation of their actual products' environmental footprint (Notarnicola et al., 2017). Regarding GHG allocation, the LCA ISO standards (ISO, 2006) only provide implicit guidelines on apportioning emissions between co-products, but their interpretation by researchers varies, leading to inconsistent application of allocation methods and conflicting outcomes. These complexities have contributed to the limited number of LCA studies conducted in SSA, exacerbated by the limited availability of farm-level data from such cattle production systems. Addressing these limitations is crucial to enhancing the accuracy of LCA application in smallholder systems and therefore enabling developing countries to track the effectiveness of GHG mitigation interventions in their livestock sector. Although the present study did not tackle the challenge of multi-functionality in smallholder systems LCA, GHG allocation, specifically to milk and live weight gain using the novel energy expenditure method, in comparison to protein mass allocation, was examined.

It is worth noting that both the products and services from cattle production systems, especially in the East African smallholder setup, play a significant role in farm-level management decision-making and, hence, GHG emissions. For example, where draft power is commonly used for ploughing land, the proportion of male animals kept in the herds is normally higher compared to where this is not the norm. These male animals are often also kept beyond their optimal meat harvesting age because of this service they provide on-farm. Failure to consider the draft power as a service these animals provide will create a wrong impression that these animals are non-productive, resulting in high EIs (Ndung'u et al., 2022). Weiler et al. (2014), while conducting a partial LCA on smallholder cattle production systems in Nandi, Western Kenya, observed the expansion of system outputs to include draft power, livestock as a source of finance, and manure, in addition to meat and milk, reduced FPCM EI by 20% compared to when emissions were only attributed to the latter two products. Omitting some of these animal products and services (outputs) based on methodological limitations is therefore prejudicial for such non-specialised production systems, hence the need to continually improve LCA application in smallholder systems. To this effect, the present manuscript considered milk and LWG from the two cattle production systems studied because of the lack of data to compute other products and services. The conclusions therefore drawn from comparative differences among the scenarios modelled rather than from the absolute EIs reported. It is our considered view, therefore, that the absolute EIs reported in the present study are likely overestimated for the studied systems and can only be interpreted under the assumptions made herein.

4.2. Effects of GHG accounting methodology and methane yield on total farm emissions and EIs.

Regardless of the study site (that influenced level of animal production), enteric CH₄ was the main on-farm source of GHG emissions, accounting for 94–97% of the total farmgate emissions. This observation agreed with the findings from other carbon footprint studies conducted in smallholder cattle production systems of the Global South (Kiggundu et al., 2019; Apdini et al., 2021; Ndung'u et al., 2022). The proportion of enteric CH₄ contribution, however, was higher than the proportions that have been documented for cattle production systems in the Global North (Beauchemin et al., 2010; O'Brien et al., 2014), where this source (enteric fermentation) accounted for 50–70% of the total farm level emissions. The higher enteric CH₄ contribution in the present study compared to the intensive cattle production systems in the Global North could be explained by the low GHG emissions from the other three common sources, including pre-farm emissions, animal excreta and feed production. First, pre-farm GHG emissions associated with farm inputs such as concentrate feeds, synthetic fertiliser, and energy and machinery use were low, accounting for less than 2% of the total emissions. This was because the cattle production systems in the modelled study sites were mainly low-capital systems implying that the use of the listed inputs was minimal (Goopy et al., 2018; Ndung'u et al., 2018). Secondly, GHG emissions from animal excreta (urine, dung and manure) accounted for only 3–6% of the total emissions (Fig. 2), which was lower than up to 20% reported for intensive cattle production systems (O'Brien et al., 2014). The lower manure emissions could be attributed to the primarily dry manure storage conditions in the two study sites that could have suppressed CH₄ and N₂O emissions. Also, the low CP content in the animals' diets (< 10% CP DM) that implied low manure N content could have contributed to this. Thirdly, because the main source of feeds was mainly natural pastures with limited amendment practises, this was considered carbon neutral and hence negligible GHG emissions. It can therefore be concluded from this finding that interventions towards reducing enteric CH₄ emissions at the farm level in smallholder cattle production, such as dietary changes, are likely to result in the highest farm-level reductions in East Africa compared to interventions from other sources, such as manure management. Currently promising interventions in reducing enteric CH₄ emissions, such as using feed additives including 3-NOP and seaweed, may not be attractive in these systems because of the additional input cost with minimal productivity gains. Without adequate incentives to encourage the uptake of such additives, e.g., carbon credits, this will not likely result in meaningful GHG reductions. In East Africa, carbon credit structures are still poorly developed, especially in the livestock sector. Such systems need to be developed not only to encourage farm-level enteric CH₄ mitigation but also to attract climate financing in the sector. Even though other dietary interventions that lead to improved animal production are likely to increase net GHG emissions at the farm level, this could be harnessed together with a reduction of less productive animals in the farms to achieve both an increase in farm-level outputs while reducing herd numbers. This strategy theoretically is likely to result in a reduction in total emissions without compromising animal production but would require a more holistic approach in its implementation to overcome the economic, social and ecosystem competing interests that are likely to be faced.

The higher total GHG emissions (t CO₂-eq./TLU/year) and EIs when IPCC Tier 1 methodology was used compared to IPCC Tier 2 approaches was likely a result of the overestimation of the EF used. For IPCC Tier 1 EF calculations, a diet DMD of 51% compared to 59–63% predicted in the present study is assumed. This could have substantially overestimated GE intake and hence enteric CH₄ emitted per animal (Gavrilova et al., 2019). Among the Tier 2 scenarios modelled, the main differences in total GHG emissions and EIs reported were mainly because of the MY used with the CSIRO Tier 2 scenario, which used the lowest MY (Charmley et al., 2016: 20.7 g/kg DMI) having the lowest emissions while the use of locally measured MYs (IPCC Local; 26.5 g CH₄ /kg DMI and IPCC Submaint.: 27.5 g CH₄ /kg DMI) resulting in the highest emissions. This finding highlights the need to conduct more local direct MY measurements on East African cattle to inform the refinement of the regional defaults recommended by IPCC for improved accuracy of inventory reporting by the countries in the region. More cost-effective direct animal measurement methods such as SF6, which also have higher throughput and are more suitable for extensive systems that are common in East Africa, may be desirable.

The locally measured MYs used in the present study were, unsurprisingly, higher than the revised IPCC recommendations (IPCC 2019). This is because, unlike in temperate countries where cattle are fed on more digestible diets at production levels, the animals used in the local studies were fed mainly on high-fibre low diets, which could have prolonged rumen retention time, increasing CH₄ output unit feed intake. Tropical C4 grasses are characterised by higher fibre and low CP contents compared with temperate C3 grasses, which compromise their digestibility and rumen passage rate. This consequently leads to longer rumen retention time and higher enteric CH₄ production per kg of DMI (Huhtanen et al., 2016; Goopy et al., 2020). In addition, when ruminants are underfed, this has been shown to increase MY compared to animals fed at the production level (above maintenance) (Goopy et al., 2020; Arndt et al., 2022). This underfeeding situation is common in smallholder contexts, especially during the dry season, and therefore, is likely to increase enteric CH₄ emissions. The IPCC recommendations assume that animals are consistently fed at the production level throughout the year. This is likely to underestimate the actual enteric CH₄ production during periods of feed scarcity when animals do not get sufficient feed, even just to maintain their body weights. When sub-maintenance nutrition was factored in (IPCC submaint. ), the total GHG emissions and EIs increased by only 3% in both study sites compared to when MY derived from animals fed at the production level (IPCC local) was used. The differences between IPCC local and IPCC submaint. were minimal because the sub-maintenance MY was only 16% higher than the above maintenance feeding MY (30.8 vs 26.5 g/kg DMI), and the higher sub-maintenance MY was only applicable in one of the four seasons in both study sites. The MY used in CSIRO Tier 2 approach derived from the Australian cattle production systems (Charmley et al., 2016: 20.7 kg/kg DMI) was unexpectedly lower than the current IPCC Tier 2 recommendation used despite this factor being partly derived from animals fed on tropical diets.

4.3. Effects of allocation method on emission intensities

The use of the three different allocation methods tested in the present study (no allocation, protein mass and energy expenditure allocation) to apportion GHG emissions between milk and LWG resulted in contrasting EIs, which agrees with findings from other studies applying different allocation methods on the same production system (Wilkes et al., 2020; Kyttä et al., 2022). When all GHG emissions were attributed to one product (no allocation) compared to the two other allocation methods, this resulted in more than 50% milk and LWG EIs because of overburdening of one product in this method. Compared to the protein mass allocation method, the energy expenditure method resulted in about 14% lower FPCM EI in the highlands and 76% in the lowlands. On the other hand, LWG EI decreased by about 33% in the highlands and 96% in the lowlands, irrespective of the scenario modelled. These two findings imply that when energy expenditure allocation is used, lower GHG emissions are attributed to milk production at the expense of beef production. This trend is more pronounced in the low-producing site than in the moderately producing site. This agrees with Kristensen et al. (2011), who observed the same trend while using the same allocation methods for two European dairy cattle production systems.

Inherently, cattle are known to be about three times more efficient in converting their dietary nutrients to milk CP than the CP deposited in body tissues (Flachowsky, 2002). When the energy allocation method was used, this approach tended to reflect the production efficiency of the products in partitioning emissions between them with more emissions apportions to the less efficient meat production. This was unlike the protein allocation methodology, where the basis of allocation is on the mass of CP in the products only without taking into consideration the relationship between inputs and the products. Because nutrient input, especially in cattle production, is a function of rumen digestion and hence CH₄ emissions, it is logical to allocate more GHG emissions to products produced less efficiently (meat CP) than those that are more efficient (milk CP). This was the case when the energy allocation method was used in the present study but not with the protein allocation that resulted in similar EI for a unit of CP in both milk and meat. Even though the protein allocation method is known to be insightful in some contexts, e.g., in informing consumer choices when selecting low-carbon footprint products based on their nutritional value, this has also been criticised extensively. Milk and meat are known to be rich in other nutrients beyond protein; therefore, a more appropriate metric in this context has been suggested to be expressing emissions per unit of total nutrient density in the product (Doran-Browne et al., 2015; Drewnowski et al., 2015). With different allocation methods yielding different outcomes that are informative depending on a specific question being answered in a given LCA study (Wilfart et al., 2021), using more than one allocation approach is more instructive to achieve a broader perspective of the production system under study. This will allow for more rational decision-making that takes into consideration a range of competing interests.

The proportion of GHG emissions allocated between milk and meat varied from one study site to another when energy and protein allocation methods were used, and this was likely related to the differences in the level of animal production, with the highlands having higher-performing animals than the lowlands. When the energy allocation method was used, more GHG emissions (60% of the total emissions) were allocated to milk in the highlands. This was in contrast to only 13% of the emissions attributed to lowland milk production. This higher GHG allocation in the highlands was because they were higher yielding (8.4 L milk/d, inclusive of the amount suckled by the calves) and, therefore, a higher proportion of energy intake spent on lactation, compared to the lowlands, which had a lower milk production (3.3 L/d) and hence a lower proportion of GHG emissions. This was in comparison to protein mass allocation, where a higher proportion of GHG emissions was consistently allocated to milk production across the two study sites, disproportionately increasing milk EIs in the lowlands, which is biased.

The continued use of IPCC Tie 1 methodology and the GHG emission defaults extrapolated from the Global North by East African countries for their livestock sector inventory reporting results in high uncertainties that were quantified in the present study. Our findings show that moving from using Tier 1 to Tier 2 methodology in calculating farm-level GHG emissions from smallholder cattle production systems in Western Kenya resulted in a 22–44% reduction in the total GHG emissions and EIs estimated. The findings of the present LCA study also suggest that substituting currently used MYs with locally measured values for the IPCC Tier 2 methodology will result in a 19–25% increase in EIs obtained. The accuracy of the national inventory reporting in these countries can therefore be improved by transitioning to using Tier 2 methodologies (from Tier 1) with the need to refine the MY defaults recommended in this higher tier methodology. It can also be concluded from the present findings that using energy expenditure to partition GHG emissions between milk and meat in multifunctional smallholder cattle production results presents an additional logical perspective in partitioning GHG emissions to co-products that seems to be ignored by the commonly used allocation methods. Future research in East Africa, therefore, should look into more cost-effective data collection methods to assist countries that are still using IPCC Tier 1 in the region to fully transition to using a higher Tier methodology for their livestock inventory to improve their accuracy of reporting and to be able to test suitable mitigation interventions for the local cattle production systems. There is also a need to continue conducting local MY measurements using local cattle breeds and feed resources to inform the revision of IPCC defaults to reflect the prevailing cattle production systems in East Africa. With the East African countries faced with an urgent need to increase their food production to feed the growing population, in addition to the need to meet their climate goal of reducing GHG emissions, the climate focus in these countries will inevitably be on reducing EIs rather than total emissions at least in the foreseeable future. The focus on EIs will allow them to increase food production while meeting their climate obligations. Tracking of reduction in EIs more accurately will, therefore, be crucial for their inventory reporting. Continuous refinements on the LCA application to improve the consistency and accuracy of quantifying EIs in smallholder cattle production are urgently needed to assist countries in SSA in identifying suitable mitigation practices for their local production systems and be able to track their mitigation contributions in their national inventories.

Declaration of conflict of interest.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Authors contribution

D. Korir: Conceptualization, Data curation and formal analysis, Writing - original draft; P. W. Ndung’u: Data curation, Writing – review and editing; A. Onyango: Data curation, Writing – review and editing; C. Arndt: Conceptualization, Formal data analysis, Writing - review and editing, Funding acquisition, Supervision; J. Goopy: Conceptualization, Writing – review and editing, Supervision; S. Marquardt: Conceptualization, Writing – review and editing, Supervision; and R. Eckard: Conceptualization, Writing – review and editing, Supervision, Resources.

Acknowledgements

The authors wish to thank Dr. Sonja Leinter and Dr. Michael Graham for their assistance in generating the local manure EFs used in the present study from the published studies from Kenya and proofreading our manure GHG emission calculations. We appreciate the help of Rodgers Rogito and Nelson Saya of ILRI’ Mazingira Centre in processing and analysing the feed samples collected during the field surveys. This work was made possible through financial support from the German Federal Ministry for Economic Cooperation and Development (BMZ) and the German Technical Cooperation (GIZ) under the project: Programme for Climate Smart Livestock (PCSL, grant no: 2017.0119.2) and Prince Albert II of Monaco Foundation through the IPCC scholarship programme. The authors also acknowledge the CGIAR Fund Council, Australia (ACIAR), Irish Aid, European Union, International Fund for Agricultural Development (IFAD), Netherlands, New Zealand, UK, USAID and Thailand for funding to the CGIAR Research Program on Livestock. This research was conducted as part of the CGIAR Initiative Livestock, Climate and System Resilience and Mitigate+: Research for Low Emissions Food Systems, which is supported by contributors to the CGIAR Trust Fund.

Apdini, T., Al Zahra, W., Oosting, S.J., de Boer, I.J.M., de Vries, M., Engel, B., van Middelaar, C.E., 2021. Understanding variability in greenhouse gas emission estimates of smallholder dairy farms in Indonesia. Int. J. Life Cycle Assess. 26, 1160-1176. doi:https://doi.org/10.1007/s11367-021-01923-z.
Archimede H., Rira M., Eugene M., Fleury J., Lastel M.L., Periacarpin F., Silou-Etienne T., Morgavi D.P., Doreau M., 2018. Intake, total-tract digestibility and methane emissions of Texel and Blackbelly sheep fed C4 and C3 grasses tested simultaneously in a temperate and a tropical area. J. Cleaner Production 185, 455-463. doi:https://doi.org/10.1016/j.jclepro.2018.03.059.
Arndt, C., Hristov, A.N., Price, W.J., McClelland, S.C., Pelaez, A.M., Cueva, S.F., Oh, J., Dijkstra, J., Bannink, A., Bayat, A.R., 2022. Full adoption of the most effective strategies to mitigate methane emissions by ruminants can help meet the 1.5° C target by 2030 but not 2050. Proc. Natl. Acad. Sci. U.S.A. 119, e2111294119. doi: https://doi.org/10.1073/pnas.2111294119.
Baldini, C., Gardoni, D., Guarino, M., 2017. A critical review of the recent evolution of Life Cycle Assessment applied to milk production. J. Clean. Prod. 140, 421-435. doi:https://doi.org/10.1016/j.jclepro.2016.06.078.
Balehegn, M., Kebreab, E., Tolera, A., Hunt, S., Erickson, P., Crane, T.A., Adesogan, A.T., 2021. Livestock sustainability research in Africa with a focus on the environment. Anim. Front. 11, 47-56. doi:https://doi.org/10.1093/af/vfab034.
Beauchemin, K.A., Janzen, H.H., Little, S.M., McAllister, T.A., McGinn, S.M., 2010. Life cycle assessment of greenhouse gas emissions from beef production in western Canada: A case study. Agric. Syst. 103, 371-379. doi:https://doi.org/10.1016/j.agsy.2010.03.008.
Charmley, E., Williams, S.R.O., Moate, P.J., Hegarty, R.S., Herd, R.M., Oddy, V.H., Reyenga, P., Staunton, K.M., Anderson, A., Hannah, M.C., 2016. A universal equation to predict methane production of forage-fed cattle in Australia. Anim. Prod. Sci. 56, 169-180. doi:https://doi.org/10.1071/An15365.
CSIRO, 2007. Nutrient requirements of domesticated ruminants. CSIRO publishing.
Doran-Browne, N.A., Eckard, R.J., Behrendt, R., Kingwell, R.S., 2015. Nutrient density as a metric for comparing greenhouse gas emissions from food production. Clim. Change 129, 73-87. doi:https://doi.org/10.1007/s10584-014-1316-8.
Drewnowski, A., Rehm, C.D., Martin, A., Verger, E.O., Voinnesson, M., Imbert, P., 2015. Energy and nutrient density of foods in relation to their carbon footprint. Am. J. Clin. Nutr. 101, 184-191. doi: https://doi.org/10.3945/ajcn.114.092486.
Gavrilova O, Leip A, Dong H, MacDonald JD, Gomez Bravo C, Amon B, Barahona Rosales R, Prado Ad, de Lima, MA, Oyhantcabal W (2019). Emmisions from livestock and manure management. In E. Calvo Buendia, K. Tanabe, A. Kranic, J. Baasansuren, M. Fukuda, S. Ngarize, A. Osako, Y. Pyrozhenko, P. Shermanau, & S. Federici (Eds), 2019 Refinement to the 2006 IPCC guidelines for national greenhouse gas inventories: Agriculture, forestry and other land use (Vol. 4, pp. 10.49–10.73). IPCC, Geneva, Switzerland.
Gerber, P.J., Steinfeld, H., Henderson, B., Mottet, A., Opio, C., Dijkman, J., Falcucci, A., Tempio, G., 2013. Tackling climate change through livestock: a global assessment of emissions and mitigation opportunities. Food and Agriculture Organization of the United Nations (FAO), Rome, Italy.
GLEAM, E.A., 2020. The global livestock environmental assessment model. Food and Agriculture Organization of the United Nations (FAO), pp. 22-26, Rome, Italy.
Goopy, J.P., Korir, D., Pelster, D., Ali, A.I.M., Wassie, S.E., Schlecht, E., Dickhoefer, U., Merbold, L., Butterbach-Bahl, K., 2020. Severe below-maintenance feed intake increases methane yield from enteric fermentation in cattle. Br. J. Nutr. 123, 1239-1246. doi:https://doi.org/10.1017/S0007114519003350.
Goopy, J.P., Onyango, A.A., Dickhoefer, U., Butterbach-Bahl, K., 2018. A new approach for improving emission factors for enteric methane emissions of cattle in smallholder systems of East Africa - Results for Nyando, Western Kenya. Agr. Syst. 161, 72-80. doi:https://doi.org/10.1016/j.agsy.2017.12.004.
Graham, M., Butterbach-Bahl, K., du Doit, C., Korir, D., Leitner, S., Merbold, L., Mwape, A., Ndung’u, P., Pelster, D., Rufino, M., 2022. Research Progress on Greenhouse Gas Emissions From Livestock in Sub-Saharan Africa Falls Short of National Inventory Ambitions. Front. Soil Sci. 2, 927452. doi: https://doi.org/10.3389/fsoil.2022.927452.
Huhtanen, P., Ramin, M., Cabezas-Garcia, E.H., 2016. Effects of ruminal digesta retention time on methane emissions: a modelling approach. Anim. Prod. Sci. 56, 501-506. doi:https://doi.org/10.1071/An15507.
ISO, 2006. Environmental Management – Life Cycle Assessment – Principles and Framework (Std. ISO 14044:2006[E]. International Standards Organization, Geneva, Switzerland.
Jahnke, H.E., Jahnke, H.E., 1982. Livestock production systems and livestock development in tropical Africa. Kieler Wissenschaftsverlag Vauk Kiel, Germany.
Kiggundu, N., Ddungu, S.P., Wanyama, J., Cherotich, S., Mpairwe, D., Zziwa, E., Mutebi, F., Falcucci, A., 2019. Greenhouse gas emissions from Uganda's cattle corridor farming systems. Agr. Syst. 176, 102649. doi:https://doi.org/10.1016/j.agsy.2019.102649.
Korir, D., Eckard, R., Goopy, J., Arndt, C., Merbold, L., Marquardt, S., 2022b. Effects of replacing Brachiaria hay with either Desmodium intortum or dairy concentrate on animal performance and enteric methane emissions of low-yielding dairy cows. Front. Anim. Sci. 96. doi:https://doi.org/10.3389/fanim.2022.963323.
Korir D., Marquardt S., Eckard R., Sainz-Sanchez A., Dickhoefer U., Merbold L., Butterbach-Bahl K., Jones C., Robertson-Dean M., Goopy J., 2022a. Weight gain and enteric methane production of cattle fed on tropical grasses. Anim. Prod. Sci. 63 120-132. doi: https://doi.org/10.1071/AN21327.
Kristensen, T., Mogensen, L., Knudsen, M.T., Hermansen, J.E., 2011. Effect of production system and farming strategy on greenhouse gas emissions from commercial dairy farms in a life cycle approach. Livest. Sci. 140, 136-148. doi:https://doi.org/10.1016/j.livsci.2011.03.002.
Kurihara, M., Magner, T., Hunter, R.A., McCrabb, G.J., 1999. Methane production and energy partition of cattle in the tropics. Br. J. Nutr. 81, 227-234. doi:https://doi.org/10.1017/s0007114599000422
Kyttä, V., Roitto, M., Astaptsev, A., Saarinen, M., Tuomisto, H.L., 2022. Review and expert survey of allocation methods used in life cycle assessment of milk and beef. Int. J. Life Cycle Assess. 27, 191-204. doi:https://doi.org/10.1007/s11367-021-02019-4.
Leitner, S., Ring, D., Wanyama, G.N., Korir, D., Pelster, D.E., Goopy, J.P., Butterbach-Bahl, K., Merbold, L., 2021. Effect of feeding practices and manure quality on CH₄ and N₂O emissions from uncovered cattle manure heaps in Kenya. Waste Manag. 126, 209-220. doi:https://doi.org/10.1016/j.wasman.2021.03.014.
Mackenzie S.G., Leinonen I., Kyriazakis I., (2017). The need for co-product allocation in the life cycle assessment of agricultural systems—is "biophysical" allocation progress? Int. J. Life Cycle Assess. 22, 128-137. doi:https://doi.org/10.1007/s11367-016-1161-2.
Muchenje, V., Dzama, K., Chimonyo, M., Raats, J.G., Strydom, P.E., 2008. Meat quality of Nguni, Bonsmara and Aberdeen Angus steers raised on natural pasture in the Eastern Cape, South Africa. Meat Sci. 79, 20-28. doi:https://doi.org/10.1016/j.meatsci.2007.07.026.
Mummed, Y., Webb, E.C., 2019. Carcass weight, meat yield and meat cuts from Arado, Boran, Barka, Raya cattle breeds in Ethiopia. J. Agric. Sci. 11, 45-51. doi: https://doi.org/10.5539/jas.v11n18p45.
Myhre, G., Shindell, D., Pongratz, J., 2014. Anthropogenic and natural radiative forcing. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Ndung'u, P.W., Takahashi, T., du Toit, C., Robertson-Dean, M., Butterbach-Bahl, K., McAuliffe, G., Merbold, L., Goopy, J.P., 2022. Farm-level emission intensities of smallholder cattle (Bos indicus; B. indicus–B. taurus crosses) production systems in highlands and semi-arid regions. Animal. 16, 100445. doi:https://doi.org/10.1016/j.animal.2021.100445.
Ndung’u, P.W., Bebe, B.O., Ondiek, J.O., Butterbach-Bahl, K., Merbold, L., Goopy, J.P., 2018. Improved region-specific emission factors for enteric methane emissions from cattle in smallholder mixed crop: livestock systems of Nandi County, Kenya. Anim. Prod. Sci. doi:https://doi.org/10.1071/AN17809.
Nollet, L.M., Toldrá, F., 2019. Handbook of analysis of edible animal by-products. CRC Press.
Notarnicola, B., Sala, S., Anton, A., McLaren, S.J., Saouter, E., Sonesson, U., 2017. The role of life cycle assessment in supporting sustainable agri-food systems: A review of the challenges. J. Clean. Prod. 140, 399-409. doi:https://doi.org/10.1016/j.jclepro.2016.06.071.
O'Brien, D., Capper, J., Garnsworthy, P., Grainger, C., Shalloo, L., 2014. A case study of the carbon footprint of milk from high-performing confinement and grass-based dairy farms. J. Dairy Sci. 97, 1835-1851. doi:https://doi.org/10.3168/jds.2013-7174.
Oddy, V., Robards, G., Low, S., 1983. Prediction of in vivo dry matter digestibility from the fibre and nitrogen content of a feed. Feed information and animal production: proceedings of the second symposium of the International Network of Feed Information Centres/edited by GE Robards and RG Packham. Farnham Royal, Slough [Buckingham]: Commonwealth Agricultural Bureaux, c1983.
Opio, C., Gerber, P., Mottet, A., Falcucci, A., Tempio, G., MacLeod, M., Vellinga, T., Henderson, B., Steinfeld, H., 2013. Greenhouse gas emissions from ruminant supply chains–A global life cycle assessment. Food and agriculture organization of the United Nations, Rome, Italy.
Pelster, D.E., Gisore, B., Goopy, J., Korir, D., Koske, J.K., Rufino, M.C., Butterbach-Bahl, K., 2016. Methane and Nitrous Oxide Emissions from Cattle Excreta on an East African Grassland. J. Environ. Qual. 45, 1531-1539. doi:https://doi.org/10.2134/jeq2016.02.0050.
Poore, J., Nemecek, T., 2018. Reducing food's environmental impacts through producers and consumers. Science 360, 987-992. doi:10.1126/science.aaq0216.
Salmon, G.R., Marshall, K., Tebug, S.F., Missohou, A., Robinson, T.P., MacLeod, M., 2018. The greenhouse gas abatement potential of productivity improving measures applied to cattle systems in a developing region. Animal 12, 844-852. doi:https://doi.org/10.1017/S1751731117002294.
Thornton, P.K., 2010. Livestock production: recent trends, future prospects. Philos. Trans. R. Soc. 365, 2853-2867. doi:https://doi.org/10.1098/rstb.2010.0134.
Tyrrell, H., Reid, J., 1965. Prediction of the energy value of cow's milk. Int. J. Dairy Sci. 48, 1215-1223. doi:https://doi.org/10.3168/jds.S0022-0302(65)88430-2.
Udo, H., Weiler, V., Modupeore, O., Viets, T., Oosting, S., 2016. Intensification to reduce the carbon footprint of smallholder milk production: fact or fiction? Outlook Agric. 45, 33-38. doi:https://doi.org/10.5367/oa.2016.0229.
Weiler, V., Udo, H.M.J., Viets, T., Crane, T.A., De Boer, I.J.M., 2014. Handling multi-functionality of livestock in a life cycle assessment: the case of smallholder dairying in Kenya. Curr. Opin. Env. Sust. 8, 29-38. doi:https://doi.org/10.1016/j.cosust.2014.07.009.
Wilfart, A., Gac, A., Salaün, Y., Aubin, J., Espagnol, S., 2021. Allocation in the LCA of meat products: is agreement possible? J. Clean. Prod. 2, 100028. doi:https://doi.org/10.1016/j.cesys.2021.100028.
Wilkes, A., Wassie, S., Odhong, C., Fraval, S., van Dijk, S., 2020. Variation in the carbon footprint of milk production on smallholder dairy farms in central Kenya. J. Clean. Prod. 265, 121780. doi:https://doi.org/10.1016/j.jclepro.2020.121780.
Zhu, Y., Merbold, L., Leitner, S., Wolf, B., Pelster, D., Goopy, J., Butterbach-Bahl, K., 2021. Interactive effects of dung deposited onto urine patches on greenhouse gas fluxes from tropical pastures in Kenya. Sci. Total Environ. 761, 143184. doi:https://doi.org/10.1016/j.scitotenv.2020.143184.
Zhu, Y.H., Merbold, L., Leitner, S., Xia, L.L., Pelster, D.E., Diaz-Pines, E., Abwanda, S., Mutuo, P.M., Butterbach-Bahl, K., 2020. Influence of soil properties on N₂O and CO₂ emissions from excreta deposited on tropical pastures in Kenya. Soil Biol. Biochem. 140, 107636. doi:https://doi.org/10.1016/j.soilbio.2019.107636.
Zhu, Y.H., Merbold, L., Pelster, D., Diaz-Pines, E., Wanyama, G.N., Butterbach-Bahl, K., 2018. Effect of Dung Quantity and Quality on Greenhouse Gas Fluxes From Tropical Pastures in Kenya. Global Biogeochem. Cy. 32, 1589-1604. doi:https://doi.org/10.1029/2018gb005949.

Supp.Tables.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

IPCC Tier 1 methodology overestimates the carbon footprint of smallholder cattle production systems in Kenya

Status:

Version 1

Abstract

Figures

Highlights

1. Introduction

2. Methodology

2.1. Study site and activity data collection

2.2. System boundary

2.3. Inventory analysis

2.3.1. Scenarios modelled and GHG emission calculations

2.3.2. The excreta GHG emissions

2.4. Impact assessment and interpretation

2.5. Functional unit and allocation procedures

3. Results

3.1.1. Total emissions

3.1.2. Emission intensities

3.2. Effects allocation methods on emission intensity

4. Discussion

4.1. LCA application in multifunctional smallholder systems

4.2. Effects of GHG accounting methodology and methane yield on total farm emissions and EIs.

4.3. Effects of allocation method on emission intensities

5. Conclusions

Declarations

References

Supplementary Files

Status:

Version 1