Early warning model for cucumber downy mildew in unheated greenhouses

An early warning model for occurrence of cucumber downy mildew in non-heated greenhouses was developed based on disease records and microclimatic parameter analysis. It also integrated empirical and fundamental forecasting models, early warning theory and plant protection product risk analysis. The thresholds for infection were a daily temperature range 5'C, daily mean RH > 80% and daily mean temperature at 15-25'C in autumn. The temperature contribution rate for symptom appearance was measured by the reciprocal of the days of the inoculation period and the relationship was fitted by a polynomial of degree 3. Data recorded in two greenhouses were used to construct the model and its predictions were evaluated with an independent dataset. The warning source search, warning obviation and plant protection production risk analysis were based on China's Good Agricultural Practices (CHINAGAP). The implications of the variables and modules of the model in the safety production are discussed.