Forecasting of tobacco downy mildew based on information of the last 30 years in north of Iran

Downy mildew, caused by Peronospora tabacina, is the most important disease of tobacco in the world which came to Iran in 1963. The northern part of Iran has the highest levels of tobacco cultivation and more than 70 to 80% of the extracted tobacco is obtained from these areas. During the last 30 years, the disease has appeared in 18 years with varying severity in Tirtash region. This research was conducted with the aim of examining the relationship between the amount of disease and climate variables. For this purpose, the amount of disease was recorded as severity percent (EP3) and converted to outbreak occurrence (EP1) and severity levels (EP2) during the 30 years (1984-2014). The growth period of tobacco was considered as a model base and four time windows were defined in this time period and predictor variables were determined among 21 meteorological variables. Correlation of the variables with the disease indicate that the best time window for disease warning was complete length of Farvardin (March 20 to April 20) and average of mean temperature (ATA), total precipitation (SP) and average of wind speed (AWS) had the highest correlation with dependent variables. The highest efficiency of prediction of developed models were 80.7%, 64.5% and 33.8% based on EP1, EP2 and EP3, respectively. The best combination of two predictor variables for disease prediction was obtained by ATA and SP.