Modelling barley disease epidemics for use with decision support systems
In a field trial during 1995/96, epidemics of Pyrenophora teres and Rhynchosporium secalis were studied in winter barley with concurrent records of weather data to identify key environmental parameters that affect epidemics. Temperature was identified as a key influence in the onset of P. teres epidemics. Disease symptoms were observed to progress when daytime temperatures consistently reached 10°C and minimum nightime temperatures for the same period remained above 5°C. Short leaf wetness periods and longer photoperiods also correlated with increased disease levels during the P. teres epidemic. In R. secalis, relationships between disease onset and individual environmental parameters were not consistent, however, high rainfall events and prolonged leaf wetness periods were recorded prior to greatest disease increase. Hypotheses based on individual and combined weather criteria, based on the results of the 1995/96 field trials, were tested in controlled conditions. The effect of temperature on P. teres was confirmed, with small differences between ascospores and conidiospores. Latent period of both P. teres and R. secalis was influenced by cultivar resistance, inoculum concentration and plant growth stage. In a second field trial in 1996/97 reduced dose fungicide programmes, using hypotheses of epidemic development based on environmental criteria, were tested and compared favourably to a standard programme with greater fungicide doses. Environmental criteria were combined within a decision model for timed reduced-dose fungicide programmes for each pathogen, where risk scores were allotted for each set of criteria and fungicide treatment decision was based on the cumulative risk score. Both the P. teres and R. secalis decision models were tested in a final field trial in 1997/98. Lower disease levels and greater yield response, with lower fungicide input, was achieved from both models compared to a standard fungicide programme. The potential for using the P. teres and R. secalis decision models in a decision support system for cereals is discussed.