Productivity assessments in small ruminant improvement programmes : a case study of the West African dwarf goat

Livestock production in the tropics is characterised by a high degree of variability in terms of composition, setting and aims. A good understanding of these characteristics is a prerequisite for the planning of a successful improvement programme. A frequently used criterion to assess the suitability of an innovation or intervention is productivity which indicates the efficiency of output generation within the production setting. This setting can be at different levels varying from the individual animal to the national level. In the general introduction (Chapter 1) the difficulties involved in defining productivity were pointed out. The major problems concern valuing and aggregating the output and formulating the scaling factor, reflecting the scare resources. A distinction was made between scare resources that are manipulable at the level of assessment and those that are not. This distinction was made because of the implication for the value of efficiency assessments. Generally if one moves from the level of the individual animal to the national level the number of resources that are variable (as well as the degree of variability) increases. Thus even if the value of the output does not vary between these levels, productivity parameters may change.This study examined how productivity assessments can be used to characterise livestock production systems and to explore ways for improvements, using a case study of the West African Dwarf goat in South-Western Nigeria. The focus of the study was limited to small ruminants meat breeds and the households level.The main questions investigated were:1. What methods are suitable for estimating productivity given the aim of the assessment and the requirements and availability of data? How is productivity affected by changes of the production environment (management, nutrition) and how can these effects be investigated?2. What role can productivity assessments play in the identification of constraints and the testing of innovations at farm level?In a biological sense, production in meat breeds (P) consists basically of the accumulation of liveweight through the growth of the individual animal and the increase in the number of animals through births. The net production (NP) of a flock is the balance between P and losses.Parameters used to assess the productivity of small ruminants meat breeds can be roughly grouped into two categories of indices: 1) those that focus on the multiplication aspect; 2) those that focus on the accumulation aspect. Indices in the first category principally concern reproductive parameters, whereas those in the second category chiefly consider NP. The review, modification and testing of existing indices (Section 2.1) resulted in the following recommendations: the ReProduction Index (RPI) expressing the litter liveweight at a predefined age or weight aggregated over the reproductive life cycle of the doe/ewe or the observation period and standardised to year-basis. For the assessment of the productivity at flock level it was recommended to use NP, calculated as the difference between the outflow and inflow of the flock with a correction for the change in stock at the beginning and the end of the observation period, and standardised to year-basis.The choice of the scaling factor (apart from the time aspect already included) depends on the purpose of the productivity assessment, the production environment and the type of records available. Normally the scarce resources manipulable at the level of assessment should be included or be represented.In the case of reproduction, the default is the individual eweldoe and in the case of productivity at flock level, the flock. If the scaling factor is to reflect the value of the animal and if this value is reflected in the liveweight, the average liveweight of the doe/ewe in the case of reproduction and the average flock weight in the case flock productivity can constitute the scaling factor. If the scaling factor is to reflect feed requirements, it can consist of the metabolic liveweight of the doelewe in the case of reproduction, or the average metabolic flock liveweight in the case of flock productivity. A minimum data collection of two years is recommended for both assessments. It is concluded that it is only appropriate to use productivity indices to compare animal sub-systems, if the differences between the total production systems, which may possibly interact with the animal component are represented in the index.In Section 2.2 it was concluded that the effect of changes of the production environment on these productivity parameters can best be investigated by the use of a dynamic stochastic simulation model that simulates the career of a small ruminant within a production system. The final output of such a model should include productivity parameters. Furthermore, the data required as inputs into the model should be easily obtainable, also in the tropics. Because none of the existing models met these requirements a new model, PC-Flock, was developed specifically for small ruminant meat breeds. The major inputs into the energy-driven model are breed characteristics, flock dynamics (inflow, outflow, mortality) and feed (quality, amount). The output includes the above recommended productivity indices. A sensitivity analysis singled out adult weight, weight and age at sexual maturity as the most important breed characteristics. Validation of the model with data from Wageningen Agricultural University and South-Western Nigeria showed that the model functions satisfactorily. It is illustrated how PC-Flock can be used to test nutrition and management interventions, even if, as is often the case in the tropics, not all input parameters are fully known.In Section 3.1 the results of various nutritional experiments carried on-station and surveys conducted on-farm were presented. An on-station study showed the selective ability of the West African Dwarf goat (Section 3.1.1). These studies also indicated clearly that pen-fed West African Dwarf goats, may not be able to maintain themselves on poor quality tropical grasses only, not even if offer levels are high. The browse species <em>Gliricidia sepium</em> and <em>Leucaena leucocephala</em> can be valuable supplements to these grasses. When these browse species constitute the sole diet, weight gains of 35-40 g d <sup><font size="-2">-1</font></SUP>are feasible. This can be further improved by the inclusion of cassava, a highly digestible energy-rich product.Two intake trials with these browse species demonstrated that a wide range of offer levels is required to obtain a reliable estimate of the relationship between feed offer and intake (Section 3.1.2). Feeding at a fixed refusal rate to determine the quality of feeds may lead to misjudgment as feeds do not attain maximum intake at the same refusal rate. If a feed is heterogenous and thus offers opportunity for selection, high offer levels and accompanying high refusal rates may have to be accepted if the objective is to maximise animal production per unit of available feed. When feeding leaves and racemes of <em>Gliricidia sepium</em> and <em>Leucaena leucocephala</em> maximum production per unit of feed offered was estimated to be obtained at refusal rates above 30%.The on-farm research reported in this thesis was carried out in six village in South- Western Nigeria. In one of the six villages confinement was mandatory. Three major goat management categories were distinguished: a) free roaming (FR);b) confinement by tethering (CT); and, c) confinement in enclosures (CE).Two surveys were conducted to investigate goat feeding practices. The results showed that in the traditional free roaming system considerable quantities of supplements are regularly fed, consisting often of cassava products (tuber and peel) and maize offal, a by- product from local maize processing with a commercial value (Section 3.1.3). These products are highly digestible but the protein content of the total diet fed was, however, generally low. The animals may be able to balance their diet through scavenging and browsing. Under permanent confinement more feed was given, but the protein content remained low. This may have limited weight gain.Some feeds frequently reported to be fed, such as the residues from bean processing, did not contribute much quantitatively. This shows the care that must be taken when quantitatively interpreting the results of a qualitative survey.It is concluded that <em>Gliricidia sepium</em> and <em>Leucaena leucocephala</em> could replace maize offal or increase the total amount of feed offered per animal. The suitability appeared to depend on the labour needed to harvest and feed the browse, the cost of maize offal and the value of the extra production. Within the socio-economic setting at the time of the survey, feeding small amounts of browse to balance proteindeficient diets seemed the most attractive option. Feeding of larger amounts of browse or using browse to replace maize offal seems to be feasible only if harvesting and feeding can be combined with other on-farm activities.The results of production studies carried out on-station (Section 3.2.1) showed that the West African Dwarf goat is potentially a good reproducer, an important characteristic for a meat breed. Goats fed on good quality diets, could be mated successfully during the lactation period, but the effective mating did not occur during the first 11 to 12 weeks of lactation. Under the same weaning practice, improved feeding increased litter size at birth. Although this may result in somewhat higher preweaning mortalities, the resulting litter size at weaning was affected positively. The kidding intervals were not affected.Analysis of the health records routinely kept on-station (Section 3.2.2) showed that when goats are yearly vaccinated against PPR, regularly washed or sprayed to control ecto- parasites, housed permanently on slatted floors and properly fed, mortality can be kept low, especially after weaning. Based on this information a management package was developed, consisting of a health, nutrition and housing component. Testing of this package in pilot units on-station showed that overall mortality figures were lower than recorded at village level. The difference, however, depends on the basis of comparison. It is larger when mortality is expressed as percentage of the average stock size or as percentage of the total number of exits but much smaller when comparisons are based on kg liveweight, because other factors such as the age structure of the deaths and the other exits come into play. Thus, if one wants to compare production systems which differ as widely as the pilot units on-station and free roaming system, one has to take proper account of the impact of the factors underlying the parameters concerned.During the testing of the management package in six villages in South-Western Nigeria, the package was presented as a set of possibilities from which the farmer could choose and adapt any (Section 4. 1). Thus no treatments groups were formed. The results showed a varied response, in line with the constraints reported by the goat keepers. The health innovations were adopted widely with some minor adaptations but problems in the area of nutrition and management were less uniform, so the response to (these) innovations was more diverse. The project was too short to evaluate these adaptations properly.If innovations offer many possibilities for adaptation, farmers should be given sufficient time (at least three years in the case of the browse innovation and likely even more in the case of the housing innovation) to determine the optimum for their conditions. Only then can possible benefits be properly assessed.In Section 4.2 an analysis is presented of goat keeping in six villages in SouthWestern Nigeria. A paradigm was developed to capture all benefits a farmer may derive from goat keeping in order to assess productivity at the household level. It is postulated that if formal finance and insurance markets are functioning poorly or are absent, livestock constitute a source of easily convertible capital for financing (B _{<font size="-2">f</font>} ) and insurance (B _{<font size="-2">i</font>} ). The total benefits (B _{<font size="-2">t</font>} ) derived from livestock is estimated as: value added + B _{<font size="-2">f</font>} + B _{<font size="-2">i</font>} . The estimate of B _{<font size="-2">f</font>} is based on the flock outflow and the estimate of B _{<font size="-2">i</font>} on the stock. It is estimated that in the most prevalent production system, i.e. free roaming, B _{<font size="-2">i</font>} and B _{<font size="-2">f</font>} constitute the major part of B _{<font size="-2">t</font>} (≈80 %). The total benefits scaled to labour at zero capital costs give returns which are comparable to labour returns from crop production and cassava processing, which is the most common income-generating activity of those who provided the major share of the labour spent on goat keeping.Confinement obligation resulted in smaller flocks (about half) and a shift in the labour division. The wives provided most of the labour in the free-roaming system, whereas 'other members' and to a lesser extent the husbands provided most of the labour in the two confinement systems. The involvement of the wives in goat keeping in terms of actual time spent remained approximately the same in all systems and husbands and 'other members' took charge of the additional tasks in CT and CE. Expenditure on purchased feed appeared to be substantial in all systems.In all three categories flock size remained fairly constant over the two-year study period. The proportion of adult female stock varied around 40% and did not differ between categories. However, in both CE and CT relatively more adult males and less young stock were kept.In FR and CE deaths per household approximately equalled total outflow, whereas in CT total outflow exceeded the number of deaths. The major proportion of the stock movements concerned the transfer of goats through caretaking arrangements, followed by sales and slaughters.Flock productivity in these systems was much lower than on-station results, partly because of poorer reproductive performance, higher mortality after weaning (in spite of the widely adopted health interventions) and lower weight gains and also because the flock outflow pattern, which is directly related to B,, is different.The individual performance of does, as expressed in the reproduction index (RPI), was about half of the performance recorded on station on high feeding levels. This difference was mainly due to the combination of lower weight gains and smaller litters. The latter was especially prevalent in confined flocks and might have been related to the level of nutrition of the does. Parturition intervals under village conditions did not differ from on-station results.Assessment of interventions within the wider perspective on goat keeping shows that innovations which may increase biological productivity (or total physical production) may not necessarily increase the total benefits (Chapter 5). It is shown how PC-Flock can be used to explore effects of innovations provided these effects can be translated into input into the model. For the free-roaming system such simulations seem to suggest that an increase of the feed resource results in higher biological productivity and larger total benefits. However, also these extra benefits do not seem to compensate the extra costs, if browse harvesting cannot be combined with other on-farm activities.It is suggested that the success of improvement programmes to address problems at farm level, could likely be increased by a heavier involvement of the farmers themselves, not only in the constraint analysis but also in the development of possible innovations. In such an approach the 'constraint' should be the point of departure. However, it is recognised that the objectives of farming systems research programmes may be wider than to solve the immediate constraints at farm level. A multi-track approach seems to be a good strategy to pursue these wider objectives. In such a set-up several productivity assessments may be required, each tuned to the aim and the level of appraisal of the programme components.It is concluded that the role of productivity assessments in small ruminant improvement programmes is limited because of the difficulties involved in its definition and because productivity parameters are relative measures.Although it may be possible to assess the suitability of innovations at a particular site based on the adoption rate, estimation of the economic attractiveness may be required to explore the suitability for other areas or to evaluate government involvement. Because the economic value of innovations in livestock can only be judged if both physical and non- physical benefits are taken into account, it is recommended that such assessments are included in small ruminant improvements programmes.