On-farm research leading to a dynamic model of a traditional chicken production system

A series of on-farm and on-station experiments on management and production of local chickens were conducted in Zimbabwe between 1998 and 2001. The overall aim of studies was to improve smallholder farmers’ livelihood and empower women through production of local scavenging chickens. Knowledge exists that poultry can be used as a tool for poverty alleviation and strengthening of women’s status. The combination of on-farm and on-station experiments complemented each other well and increased understanding of the complex factors influencing traditional chicken production. Farmers’ participation was a main focus throughout the on-farm studies and experiments were designed to rely heavily on farmers’ own data collection. The experiences were positive although missing data were a continuous problem throughout. Initially, the traditional chicken production in a study area in Zimbabwe was described and constraints and opportunities for production investigated. Thereafter an onstation experiment established that the production potential of the local chickens both for eggs and meat were higher than expressed under on-farm conditions. On the basis of these two trials, a third experiment was planned on-farm while using the obtained experiences. Possible improvement through crossbreeding was investigated in a fourth trial on station. The tested interventions did not lead to expected improvements of production. Merely results showed a tremendous variation in production between farms and this variation tended to overrule the effect of interventions. However, the findings suggest that there are room for improvements and that some farmers have extended knowledge about production. As a supplement to timeconsuming and labour intensive on-farm experiments and as a tool to systematically describe the production system, a dynamic model of a traditional African chicken production system was developed. The model: SimFlock can through simulations establish correlations between selected outputs (e.g. maximum number or kg of chickens produced, number of eggs produced and net return of the production) and production parameters (e.g. growth and survival). Thus it can provide answers to which parameters are of most importance in relation to outputs. SimFlock provides a holistic description of the production system and the outputs produced reflect the random variation within flocks as well as between flocks. This is important for representation of the risk of production. SimFlock might be used as a management and extension tool. The continuous interaction between researcher, farmer, onfarm experiments, on-station experiments and a systematic description led to an understanding of the system in a model, which is based on common values by the farmer and researcher. This is essential since common understanding aid improvements. Introduction Chicken production in most African countries is traditionally based on scavenging systems. Approximately 80% of chicken populations in Africa are reared in these systems (Guéye, 1998). This low input/output practice has been a component of smallholder farms for centuries and will probably continue to be so in the future. There are many advantages in this production system but also constraints. The advantages consist of a production based on free feed resources available in the surrounding environment and kitchen leftovers, using local chicken breeds, which has adapted to the conditions, and preserved their ability to incubate and brood naturally. All these factors maintain an inexpensive and low management level production. The constraints are high mortalities, low egg production and slow growth. Thus in recent years more focus has been on how the constraints can be minimised and thereby increase the overall production. A variety of factors act over time to influence productivity and thus dynamic management tools are useful. A dynamic stochastic model: “SimFlock” of the traditional African chicken production system has been developed as a supplement to onfarm and on-station trials. The objective of developing a dynamic and stochastic model of a traditional chicken production system is to test, theoretically, the effect of different interventions on chickens’ performance by simulating the transition of birds between various stages of the production cycle when taking several operations into account. This is done at flock level but the individual bird is an important feature. Given a number of available resources, constraints and interaction by the farmer, the model is able to generate an output in terms of e.g. sold and slaughtered chickens, egg production and net return of production. The model can also answer what the outputs are if changes occur in the system and as such be used as an extension tool. Further, SimFlock can establish correlations between outputs and given production parameters. The aim of the experiments was to improve traditional chicken production by increasing survival and growth. Experiences from Bangladesh showed that farmers’ livelihood and women empowerment were increased through improved chicken production (Alam, 1997; Danida, 1998). Further, it is a well known fact that most of the income, controlled by women in developing countries, is spent directly on sustaining and improving the livelihood of the families (Miller, 2001). Moreover, due to the low investment costs in chickens, it is an enterprise that even the poorest farmers are capable of venturing into. Methods and materials After farmers had identified a need for research on chicken production (Muchenje and Sibanda, 1997), a series of on-farm and on-station experiments were conducted in Zimbabwe between November 1998 and May 2001. On-farm experiments were conducted in Sanyati communal farming area 250 km Northwest of the capital Harare. On-station experiments were conducted at Henderson Research Station in Mazoe. Initially an on-farm study was conducted to describe the existing chicken production system and characterise the chickens. Constraints and opportunities for improvements in the production were identified and the survey was used to plan further research. Since very little was known about the production potential of the local chickens in Zimbabwe it was decided to carry out an on-station experiment to establish the growth potential as well as the egg production potential of local chickens under improved conditions. The on-station experiment showed that chickens’ production potential for both meat and eggs was much higher than what is usually found on-farm. Another on-farm trial was thus planned with the purpose of improving the traditional chicken production in such a way that the number of chickens ready for slaughter and sale was increased. Housing of chicks (0-3 weeks), feed supplementation and anthelmintica were all thought to have a positive effect on growth rate and survival of chickens. No emphasis was put on egg production since farmers’ priority was meat production. Farmers were responsible for record keeping and Field Technicians visited farms once a week to help farmers fill in missing records and weigh chickens. In the same time as the second on-farm trial was running another on-station experiment was conducted to establish the growth potential of crossbred chickens (Cobb #500 parents x local), since crossbred chickens in Bangladesh had proven to produce better than exotic breeds and local breeds under scavenging conditions (Barua et al., 1998; Haque et al., 1999). Based on knowledge obtained and data collected in these four experiments an object-oriented dynamic stochastic model SimFlock was developed. Such a model can serve as a supplement to time-consuming and labour intensive on-farm experiments. The objective of the model was to predict the effect of different interventions on selected outputs and to establish correlations between outputs and production parameters. Figure 1 illustrates the series of on-farm and onstation experiments leading to a dynamic model.