Effect of Wetness on Blossom Infections by Erwinia amylovora - Impact on Forecasting Models

Infection of flowers by fire blight requires wetting in order to facilitate the movement of the bacteria from the stigma surface to the nectaries on the hypanthium. Current forecasting models (MARYBLYT (TM), Billing's Integrated System, "Cougarblight") call for rain or heavy dew for indicating infection risk days. In South West Germany, heavy dew is measured by leaf wetness sensors (Lufft) installed in the canopy of orchards. Leaf wetness lasting for more than two hours is taken as heavy dew. However, infections without any detectable leaf wetness were observed in the past. Therefore, the effect of low amounts of precipitation on infection of flowers was investigated using the detached blossom assay (Pusey, 1997). The flowers were inoculated on the stigma surface. After the bacteria had multiplied to the capacity of the stigma surface (about 10(7) cfu/flower) the flowers were sprayed with sterile, deionized water. If the flowers were sprayed with 0,1 mm of water, 52% were infected at 24 and 18 degrees C and 6% at 16 degrees C, whereas without spraying 8% of the flowers were infected at 24 degrees C, 3% at 18 degrees C and 0,5% at 16 degrees C. To investigate the reliability of measuring wetness in orchards four sensors for leaf wetness, relative humidity and temperature were installed at different positions in an orchard tree. The data were collected in April and May 2008 and 2009. In addition, leaf wetness was estimated by eye on ten days. After dewfall at night the leaf wetness sensors did not detect any leaf wetness although the leaves were obviously wet. Further, leaf wetness values deviated to a great extent from each other independent of their position in the tree. It seems that the amount of 0,1 mm dew which is required for an infection by Erwinia amylovora cannot be measured reliably by leaf wetness sensors. The data of the humidity sensors were more consistent. Therefore, leaf wetness was calculated from relative humidity data (70-80%, depending on the temperature). Infections which occurred without any measured wetness in the past could be predicted by this method. Relative humidity data will be used for estimating dew for the forecast of fire blight in southwest Germany in the future.