Barley modelling to improve the efficiency of field trials

Abstract

The application of crop growth models to the barley crop in Scotland was investigated. Three models, the ARCWHEAT winter wheat model (Weir et al.,1982), the CERES barley model (Ritchie et al., 1987, developed directly from the CERES wheat model, Ritchie et al., 1984) and a spring wheat model (van Keulen & Seligman, 1987) were evaluated using data from a range of seasons, cultivars and sites in South-East Scotland. The chosen models used different methods to predict crop development, based on the principle that development rate is a function of temperature and photoperiod. Specification of cultivars was based on differences in vernalisation requirement and photoperiod sensitivity which were found hard to quantify precisely. Crop performance was not modelled reliably enough to allow these models to identify Genotype x Environment interactions in variety testing. Thorough model validation required more data than those collected in the normal course of field trials.

Agronomic and physiological data from field experiments from a broad range of environments and cultivars were compiled into a database to examine the mechanisms controlling barley growth and development. Cultivars were classified according to genotypic characters which could be easily recognised such as winter/spring type, semi-dwarf/tall habit and ear row number. Rules and relationships derived from these data were used to build the modular, deterministic DAFS BARLEY MODEL which contained a range of options for simulating crop development. The pathway through the model has a large effect on the outcome and an optimisation procedure to select the most appropriate route and improve model accuracy is discussed. It is thus envisaged that the resulting barley model will be used as an adjunct to, rather than as alternative to the existing field trials program.

The model was constructed to: a) systematize historical knowledge gained from field trials b) enable data from existing field trials to be used as a guide to future research needs c) enable the results from trials to be reliably extrapolated to other sites and seasons d) aid in design of more cost-effective and efficient plans for field trial measurement, including the precise specification of data collection and recording methods for field trials.