Comparison of different prediction systems for blossom blight risk assessment on pear and apple in Aegean Region of Turkey

The Maryblyt version 4.3 model, BRS (Billing Revised System) and BIS 95 (Billing Integrated System) were evaluated for the prediction of blossom blight for the growing season in apple and pear orchards in the Aegean Region. Temperature and rainfall data were collected by automated data loggers and phenological disease observations were recorded personally. In Bornova in 1996 the Maryblyt model indicated one infection period in the first part of the bloom and first blossom infection was observed at the middle of April but model predicted it at the beginning of May on pear. In 1997 and 1998 the computer program was evaluated in two regions at a distance of 150 km. In the first region in both years the model predicted high risk of infection and three infection periods during entire bloom period. In both years first symptoms on pear were observed earlier than predicted. In the second region the model predicted one infection period at the last part of bloom but actual infection was observed earlier than infection risk day. In 1998, the Maryblyt model was evaluated in two apple growing regions. The model predicted one infection period but early symptoms occurred earlier than predicted in both orchards. BIS 95 developed by Eve Filling has been field tested and compared with the Maryblyt model. Generally, both models seemed to provide accurate prediction of infection periods but actual symptoms observed earlier than the model's prediction. The BRS risk assessment system apparently indicated more days with an infection risk than infections actually took place. However in no case did the system failed to indicate a real infection date.