Identifying recommendation domains for targeting dual-purpose maize-based interventions in crop-livestock systems in East Africa

Livestock are recognized as essential assets for the livelihoods of poor farmers in the mixed crop-livestock systems throughout the tropics. In addition to that, the demand for livestock products is predicted to rise considerably in the next 20 years, presenting a unique opportunity to escape poverty and drive economic growth also in African countries. Increasing demand for feed, shortage of arable land and water together with shrinking and deteriorating common property resources are likely to put further pressure on the natural resources the small-holders are depending upon. There are real needs and opportunities for appropriate research to improve the livelihoods of poor crop-livestock farmers by addressing feed resource constraints, including through the development and use of better food-feed crops.

In Ethiopia, Kenya and Tanzania livestock productivity is closely linked to the quantity and quality of available fodder, of which a significant proportion may be sourced from maize stover. Two international agricultural research institutes, CIMMYT and ILRI, therefore joined forces in a project designed to improve the value of maize as livestock feed with an aim of enhancing the livelihoods and resilience of maize-livestock farmers in East-Africa.

The first issue addressed in the initiative and the topic of this paper was deciding where to work so as to maximize the chances of successful testing of the new varieties and at the same time grasp a better understanding of the context-specific incentives for adoption of these varieties.

For many years, breeders -and other agricultural researchers for that matter- have worked in relative isolation and focused on the bio-physical opportunities and constraints of varieties and other agricultural technologies. The recent focus on reaching the “millennium development goals” and the Paris declaration of 2 March 2005 mark an unprecedented level of consensus and resolve to reform aid to increase its effectiveness at combating global poverty. These initiatives have shifted the emphasis of donors and development agents from developing new technologies to enabling impact in the real world. In other words, the awareness has grown that the issue is much more complex. There is a need to include biophysical, economic, socio-cultural, institutional and environmental factors; all of these need to be considered in relation to feed resource innovations.

The selection of sites for research activities in feed resource innovations, therefore, equally requires a wide variety of detailed data on the climate and geography of the target area, and on the socioeconomic characteristics of the population that lives in this area.
In our study, GIS was used to stratify the project area according to criteria considered to be of strategic importance to the applicability and adoption of dual-purpose maize varieties. The use of standardized GIS layers ensured that criteria were applied consistently across countries. This provided a spatial framework for the structured exploration of opportunities to transfer knowledge and technologies.

As in any GIS application, the key to success is the availability of detailed spatial data. While remotely sensed information and GPS-based field surveys help plug some data gaps, much information is still difficult to obtain at a geographic scale that is relevant for operational impact. The need for improvement of data remains, but in these times of cutbacks and downsizing, we have to find ways to make more efficient use of existing resources and databases. A balanced choice had to be made between detailed data for only limited areas and data covering the different countries, enabling cross-country comparisons. In addition to that, decisions had to be made on whether to use existing secondary datasets or engaging in primary data collection. Due to cost and time constraints, we opted for the use the best available regional data, straddling across the three countries.

Relatively simple GIS functions and the use of regional datasets have proven to be helpful in this ex-ante selection of study sites. Considering the usual budget and time constraints of agricultural research projects, this methodology has proven to be relatively quick and inexpensive.