Although East Africa is home to one of the most advanced dairy industries in Sub-Saharan Africa, regional annual milk production is insufficient to meet the demand. The challenge of increasing milk yields (MYs) among smallholder dairy cattle farmers (SDCFs) has received considerable attention and resulted in the introduction of various dairy management strategies (DMSs). Despite adoption of these DMSs, MYs remain low on-farm and there is a large discrepancy in the efficacy of DMSs across different farms. Therefore, the present study sought to: (1) identify on-farm DMSs employed by East African SDCFs to increase MYs and (2) summarize existing literature to quantify the expected MY changes associated with these identified DMSs. Data were collected through a comprehensive literature review and in-depth semi-structured interviews with 10 experts from the East African dairy sector. Meta-analysis of the literature review data was performed by deriving four multivariate regression models (i.e. models 1 to 4) that related DMSs to expected MYs. Each model differed in the weighting strategy used (e.g. number of observations and inverse of the standard errors) and the preferred model was selected based on the root estimated error variance and concordance correlation coefficient. Nine DMSs were identified, of which only adoption of improved cattle breeds and improved feeding (i.e. increasing diet quality and quantity) consistently and significantly (P < 0.05) increased daily MYs across the available studies. Improved breeds alongside adequate feeding explained ≤50% of the daily MYs observed in the metadata while improved feeding explained ≤30% of the daily MYs observed across the different models. Conversely, calf suckling significantly (P < 0.05) reduced MYs according to model 2. Other variables including days in milk, trial length and maximum ambient temperature (used as a proxy for heat stress) contributed significantly to decreasing MYs. These variables may explain some of the heterogeneity in MY responses to DMSs reported in the literature. Our results suggest that using improved cattle breeds alongside improved feeding is the most reliable strategy to increase MYs on-farm in East Africa. Nevertheless, these DMSs should not be considered as standalone solutions but as a pool of options that should be combined depending on the resources available to the farmer to achieve a balance between using dairy cattle genetics, proper husbandry and feeding to secure higher MYs.

Cattle emit over 65% of enteric methane (CH₄) in sub-Saharan Africa (SSA), making them the focus of many mitigation strategies targeting livestock emissions. Since measured feed intake data are sparse, emission factors for enteric CH₄ (EF_CH4) are mainly estimated indirectly from gross energy intake (GEI) using the net energy (NE) requirements for different metabolic processes in cattle. However, all NE requirement systems commonly used for cattle in SSA were developed for cattle in temperate regions. Therefore, we assessed the suitability of different enteric CH₄ models for estimating GEI of cattle in SSA. The Intergovernmental Panel on Climate Change (IPCC) and South African models were identified as the main Tier 2 based methods used to estimate enteric CH₄ emissions from cattle in SSA. In the IPCC model, EF_CH4 was estimated as {GEI * (Ym/100)}/55.65, where Ym is the conversion factor (%) of gross energy in feed to CH₄ and 55.65 the energy content of CH₄ (MJ/kg). The GEI was estimated based on NE requirements for different metabolic processes in cattle as per the American National Research Council. In the South African model, EF_CH4 was estimated as {Y/100 * GEI/55.22}, where Y is the CH₄ yield and 55.22 the energy content of CH₄; Y was calculated from the dry matter (DM) digestibility while GEI was calculated by predicting DM intake and multiplying it by 18.4 MJ (gross energy per kg DM). Also, the suitability of the British and German NE requirement systems was assessed as alternatives used for cattle nutrition in SSA. These NE systems were implemented in the IPCC model to yield the “AFRC” and “GfE” models, respectively. The four models were then evaluated using an evaluation dataset summarizing feed quality and DM intake results from 21 studies conducted in SSA, with 125 dietary treatments, and 822 cattle observations. The relative prediction error (RPE) and concordance correlation coefficient (CCC) were used to evaluate the models’ accuracy. Only the South African model estimated the GEI in dairy cattle with an acceptable RPE (18.9%) and highest CCC (0.87), while the other three models yielded estimates with RPE > 20%. None of the four models we assessed estimated GEI for other cattle (i.e., non-dairy) with an RPE < 20% or CCC > 0.30. The inaccuracy in GEI estimates suggests an error of the same magnitude in EF_CH4 estimates. Therefore, a concerted effort is needed to improve the accuracy of enteric CH₄ estimation models for cattle in SSA.

India is home to more than 525 million ruminants, which are major contributors to global warming via enteric methane (Ent_CH4) emissions. Various mitigation strategies exist to reduce Ent_CH4 emissions but accurate emission estimates are needed to establish the true potential of these strategies. Measuring Ent_CH4 emissions is expensive and unrealistic at such a large scale, so an urgent need exists for accurate Ent_CH4 prediction models. The present study evaluated the accuracy of various existing models and developed a new model to predict Ent_CH4 emissions from cattle in India. Six Ent_CH4 prediction models based on either DMI or gross energy intake (GEI) were identified as applicable to and suitable for the Indian context. Models based on DMI and GEI were derived from various works, including those of the Intergovernmental Panel on Climate Change and others (designated as IPCC_DMI, IPCC_GEI, Ribeiro_DMI, Ribeiro_GEI, Patra_DMI, and Patra_GEI). These were evaluated using 2 independent datasets characterizing 528 lactating (dairy) and 122 nonlactating (nondairy) cattle from 15 and 13 studies, respectively, under different management practices across 13 Indian states. Furthermore, the same datasets were combined to develop an empirical Ent_CH4 prediction model using a linear mixed-effects framework. The relative prediction error (RPE) and mean bias error (MBE) were used to evaluate model accuracy. A model's prediction was considered acceptable when RPE was <20%. None of the 6 models predicted Ent_CH4 for nondairy cattle with an RPE <20%. None of the 6 models predicted Ent_CH4 for both dairy and nondairy cattle with an RPE <20%. For dairy cattle, only the Ribeiro_DMI and Patra_DMI models approached this threshold, producing RPE values of 22.6% and 22.9%, respectively. The linear mixed-effects model (Alam's model, described herein: Ent_CH4 [g/d per cow] = 15.45 + 1.91 × DMI [kg/d], conditional R² = 0.94), developed for both dairy and nondairy cattle, achieved a substantially lower RPE (9.48%) than any of the 6 previously tested models. Therefore, the Ribeiro_DMI and Patra_DMI models could acceptably predict Ent_CH4 emissions from dairy cattle in India, but none of the models evaluated in our study was suitable for nondairy cattle. Our linear mixed-effects model provides more accuracy than the latter 2 in estimating emissions for dairy cattle and also offers a suitable option for nondairy cattle in India.

Introduction: Climate-smart livestock (CSL) practices have the potential to boost food production while improving the resilience and environmental sustainability of agricultural systems in sub-Saharan Africa. However, their adoption rates in the region remain low. Focusing on Uganda as a case study, multivariate Probit and Tobit regression models were applied to determine the factors influencing the adoption of CSL practices by dairy farmers. Methods: Data were collected from 626 cattle-keeping households located in southwestern, central and eastern Uganda. The Adoption and Diffusion Outcome Prediction Tool (ADOPT) was used to predict the peak adoption levels and the number of years to reach these for CSL practices. The CSL practices identified and considered were forage cultivation, artificial insemination, routine vaccination and planting trees. Results and Discussion: The results showed that 37%, 14%, 45% and 19% of cattle-keeping households grew forage, used artificial insemination, carried out routine vaccination and planted trees respectively. Forage cultivation, artificial insemination use and planting trees were positively and significantly influenced by access to credit and keeping farm records. Use of artificial insemination was associated with higher milk revenues. The household head’s education level significantly influenced both the area allocated to forage cultivation and the number of trees planted. Households that allocated a large portion of their land to grazing cattle were less likely to grow forage and use artificial insemination. The number of trees planted was positively and significantly influenced by the age of the household head, household size, keeping farm records, access to credit and total adult equivalent (total number of working hours among adults per household). Forage cultivation and use of artificial insemination were practiced interdependently, pointing to the existence of complementarities and synergies between these two practices. The complementarities and synergies between both practices should be explored for increased adoption. Conclusion: CSL practices should be introduced as a package of practices rather than as stand-alone practices, as their complementarities can enhance adoption. Development programs and policies should focus on improving access to credit, while investing in capacity building for record keeping, to increase the likelihood for adoption and intensification of CSL in the region.

Large-scale land use change (LUC) of African Guinea savannas to crop fields is expected to cause negative impacts on ecosystem functions (ESF) and long term land productivity. The complex interactions of key processes in savannas evoked by LUC calls for a process-based modelling approach. We employed the dynamically coupled Land Use Change Impact Assessment (LUCIA) model and the Livestock Simulator (LIVSIM) which represent LUC impacts on soil processes, landscape-scale matter fluxes, seasonal grass and crop growth, and livestock nutrition, production and reproduction, depending on seasonal feed availability and quality on accessible pastures. For a rangeland in Borana, Ethiopia, two different LUC scenarios were evaluated in comparison to the baseline of traditional pasture-based land use. In the intensive LUC scenario 52% of grassland was converted into unfertilized maize fields, inaccessible for livestock. The integrated LUC scenario of the same grassland conversion rate allowed feeding maize straw and provided high-quality feed reserves from seasonally managed pastures. LUC in the intensive LUC scenario led to declining yields in the second year after conversion. Feed production on the remaining rangeland patches was insufficient for livestock nutrition, causing drops of herd body weight and herd size particularly in drought years. Resilience of herd performance to LUC was enhanced in the integrated LUC scenario when feeding maize straw and high-quality feed reserves. In both LUC scenarios, topsoil organic carbon storage decreased after ploughing shrub grassland for cultivation, and so did soil water storage capacity due to soil pore destruction. Soil erosion of less than one cm after 10 years occurred under cultivation. The simulation results indicated that the well validated model framework could predict impacts of LUC and simple crop-livestock integration on savanna ESFs, grass growth dynamics and livestock production during seasonal and inter-annual rainfall variation. This study lays the foundation for further land use scenario simulations to improve the understanding of benefits and risks caused by savanna grassland conversion.