Optimising cattle genotypes and systems in Sub-Saharan Africa to increase productivity and reduce GHG emissions
Bunning, Harriet Jean
Cattle are important in supporting development in Sub Saharan Africa (SSA) as they are a source of both food and income. However studies show there is a substantial gap between the potential and realised production levels, leading to potential profit and efficiency levels not being reached. In particular, herds that have poor efficiency also tend to produce more greenhouse gases (GHGs) per unit of food product. The lack of consistent recording systems for both pedigree and performance data makes traditional selective breeding to improve profitability and efficiency, difficult. Crossbreeding strategies require comparatively little data, are generally easy strategies to follow and results can be achieved more quickly, so are a good fit in SSA systems. Crossbreeding also allows farmers to take advantage of the complementary fitness traits from local breeds and production traits from exotics, as well as providing favourable heterosis. In order to determine an optimal crossbreeding strategy, models which predict the longterm outcomes of varying strategies are needed. Therefore, this thesis describes number of models which can be combined for this purpose. Initially, a metaanalysis of heterosis in cattle in the tropics was carried out in order to characterise the expression of heterosis, which is an important factor in determining the performance of crossbreds and crossing strategies. In particular, the effect of trait, breed pair and climate were examined. In total, 62.5% of estimates were found to be significantly different from zero, the majority of which (89.8%) were beneficial for the studied trait. Milk, longevity and health traits were found to show the greatest heterosis, which showed great potential of crossbreds to increase performance for these traits which are strongly linked to profitability and efficiency. Crosses between more distantly related breeds showed moderate to high heterosis, whereas crosses between breeds of a similar type did not express heterosis that was significantly different from zero. These results show that heterosis has significant and favourable impact on productivity of cattle farming in tropical production systems. In order to model how herd composition changes over time, fertility parameters are needed for different crossbred individuals. In particular, agespecific calving rates, the probability a cow will calve at a given age, were needed. The results of the metaanalysis showed that the fertility traits more commonly recorded were age at first calving and calving interval. Therefore three variations of a model that used these as input parameters to predict age-specific calving rates were developed. These were tested using both input parameters for Ethiopia, but also from UK dairy cattle, where the predicted values could be compared to observed values. All three models performed well under both scenarios (R2 from 0.981.00), with the model in which estimation errors were reduced by reducing the size of age class considered, performing the best. Next, a deterministic herd model, which predicted the effect of crossing strategy on herd composition, using input parameters from the fertility model, was combined with a genetic model which used breed additive, heterosis and recombination effects from studies in the metaanalysis, to predict the performance of varying types of crossbred individuals for a given trait. This allowed for the prediction of herd performance for a given trait under varying crossbreeding strategies. These models were tested using a case study of Boran-Holstein crossbreeding in Ethiopia. Herd performance for annual milk yield and yearling weight was predicted under a range of crossing strategies. For milk yield, strategies which increased the proportion of Holstein genetics, whilst maximising heterosis and reducing recombination, tended to perform best. For yearling weight, all strategies increased the herd performance of this trait compared to the initial herd of purebred Borans, which is undesirable as heavier yearlings have greater feed costs. Strategies that minimized the proportion of Holstein genetics, such as using a crossbred sire, tended to perform best. In order to predict an overall optimal strategy for a given system, rather than considering traits individually, models which combined the results for multiple traits were needed to predict annual herd profit and GHG emissions. These models were developed using the International Panel on Climate Change (IPCC) guidelines for a tier II approach and input parameters specific to the Ethiopian case study were used where possible. For annual profit, crossbreeding strategies that maximised heterosis tended to perform best. In particular, a true rotation strategy, where sire breed was alternated every generation, allowed for heterosis to be maximised and so led to the greatest increase in annual profits. For GHG emissions, crossbreeding strategies that minimized the proportion of Holstein genetics tended to perform best, producing the fewest kilograms of carbon dioxide (CO2) equivalents. In particular, a strategy where the herd were graded up to a maximum of 50% Holstein, using an first cross (F1) bull, consistently produced the least GHGs, compared to all but a herd of purebred Boran. However, in order to reduce GHGs whilst maintaining food production levels, emission intensities (kg CO2 equivalents per unit of food produced), rather than gross emissions can be considered. Results showed that the true rotation strategy, which produce high levels of milk but not the highest gross GHG emissions perform best when the aim is to minimize GHGs per unit of milk produced.