South Africa experiences a high interannual variability of rainfall which, in a region with abundant solar radiation, is the main determinant of year-to-year variations in crop yields. The coefficient of variation of annual rainfall ranges from less than 20% to about 40% across the country's arable area (Schulze 1997). As a result maize, which is the country's staple food, exhibits a coefficient of variation in annual yields ranging from less than 15% to over 60% (Schulze 2003). The variability in crop production has implications for food security in the country, particularly at household level amongst resource-poor farmers, whose livelihoods are heavily dependent on agriculture.
Seasonal climate forecasts are available for South Africa, the main source of operational forecasts being the South African Weather Service (SAWS). Although seasonal climate forecasts are being applied ever-increasingly in agriculture to aid in climate sensitive decision-making, efforts to do so are confined mainly to commercial agriculture. These efforts are to be encouraged and supported, since commercial agriculture is an important economic activity in the country and is the key to national and regional food security (du Toit et al. 1999). However, relatively little research is conducted to support the application of climate forecasts in decision-making in the small-scale/subsistence agriculture sector, where farmers are particularly vulnerable to the vagaries of climate.
The usefulness of climate forecasts for applications in agriculture can be enhanced if the forecasts are translated into agricultural outlooks, where the information is targeted for decision-making. Translation of climate forecasts into agricultural outlooks can be facilitated through the generation of crop yield forecasts using crop simulation models. This approach has the benefit of accounting for factors which affect crop growth that would not be represented in climate forecasts alone, such as antecedent soil moisture conditions and crop management practices.
Given the context described above, a desktop research project (Lumsden and Schulze 2004) was undertaken with the following objectives: (1) to research methodologies required to produce crop yield forecasts for small-scale/subsistence agriculture in South Africa, (2) to evaluate the quality (accuracy) of crop yield forecasts produced using the above methodologies, (3) to assess the potential to apply the crop yield forecasts to improve crop management decisions, (4) to make recommendations for further development of the products of the research and, using insights gained in the project, to make broader recommendations on future research and operational needs.
In this chapter the methodology developed to produce crop yield forecasts will be presented. The results shown will focus on the assessment of the potential to improve crop management decisions, given a crop yield forecast. Key recommendations emanating from the project will be discussed.
Was this article helpful?