Modelling crop growth and crop water relations in South Africa : past achievements and lessons for the future

11 Aug 2010

Mathematical simulation of crop growth and water relations has become indispensable to agricultural science and practice. A critical assessment of how modelling has contributed to the development of crop science and to the management of crop production and natural resources in South Africa (SA) over the past 25 years could give new perspectives on the benefits derived from modelling, the appropriateness of approaches employed and the best way forward. The initial objectives of the major SA modelling initiatives (ACRU, BEWAB, CANEGRO, CERES, PUTU, SAPWAT, SWB) dictated the approaches that were followed and determined their impacts. Significant advances were made with regard to improved understanding of crop growth and water use and adapting models for local conditions such as dryland grain crop production under very low rainfall. Modelling provided invaluable support for strategic investigations into the impacts of climate change, land use and water use. Many of the models succeeded in providing much-needed information to improve tactical and operational management of irrigated and dryland agriculture. Some models have been (and are being) used operationally to forecast crop production (maize, wheat and sugar) and to monitor droughts in natural vegetation, adding value to the respective industries. Modelling has formed, in some cases, an integral part of tertiary education in crop science and hydrology. This should be strengthened to build more capacity to address the ever-increasing complexity of challenges in agriculture. The review identified factors that are crucial for modelling to maintain effective impacts on the science and practice of crop production and natural resource use. These were excellent scientific leadership, long term funding, effective collaboration between local and with international groups, expertise on local agronomy and high quality experimental data for model testing and adaptation. Future modelling efforts should explore opportunities to integrate information obtained from technologies such as remote sensing and genomics.