Decision models for the integrated use of fungicides on wheat

Downward pressure on the unit cost of wheat production requires that fungicides should be applied only where an economic return will result. Several decision models have been devised, in the UK and Europe, to allow growers to assess the risk of disease development and hence, tailor fungicide inputs to disease pressure. Field experiments, at six sites in three seasons, tested decision models on varieties of contrasting resistance, under varying pressure from Septoria tritici and powdery mildew. The models tested were: the Bayer Cereal Diagnostic System, the Long Ashton Splashmeter, ADAS Managed Disease Control and a developmental Integrated Disease Risk (IDR) system (the last, designed to the support the use of appropriate fungicide doses). Untreated controls and prophylactic one- and three-spray programmes were also included. Measures of disease control, yield, grain quality and fungicide input costs allowed the performance of systems to be assessed. The main findings of the project are summarised below.

Disease control and fungicide input costs achieved by the decision models were compared against those achieved in commercial practice, using disease survey data. The results suggest substantial potential for improvement in unit cost, if decisions could be based on better intelligence of disease risk.

Prophylactic programmes can be devised, which produce a net benefit across a range of sites and seasons. However, such programmes have to over-apply on average, in order to prevent occasional severe losses under high disease pressure.

Combining assessments of inoculum, weather, host resistance and the sensitivity of the crop to disease, provided a robust quantification of disease risk which could be used to guide fungicide dose decisions. However, the complexity of combining risk factors in order to improve precision, conflicts with the need for simplicity of operation on-farm.

Decision models for individual diseases are inappropriate for commercial use and need to be combined, so that decisions are optimised simultaneously for all the major foliar diseases. Interactions with stem-based and ear diseases also need to be taken into account. Furthermore, disease control decisions form only part of the crop management complex.

It was concluded that disease control decision models would be implemented most effectively as a component of a computer-based modular system, covering a range of cropping decisions. Such a system should be implemented as a support to, rather than a replacement for, current methods of technology transfer and decision making, and could act as a direct route for transfer of future improvements in disease risk assessment to the industry.