The relationship between weather data and agronomical factors and deoxynivalenol (DON) levels in oats was examined with the aim of developing a predictive model. Data were collected from a total of 674 fields during periods of up to 10 years in Finland, Norway and Sweden, and included DON levels in the harvested oats crop, agronomical factors and weather data. The results show that there was a large regional variation in DON levels, with higher levels in one region in Norway compared with other regions in Norway, Finland and Sweden. In this region the median DON level was 1000 ng g⁻¹ and the regulatory limit for human consumption (1750 ng g⁻¹) was exceeded in 28% of the samples. In other regions the median DON levels ranged from 75 to 270 ng g⁻¹, and DON levels exceeded 1750 ng g⁻¹ in 3-8% of the samples. Including more variables than region in a multiple regression model only increased the adjusted coefficient of determination from 0.17 to 0.24, indicating that very little of the variation in DON levels could be explained by weather data or agronomical factors. Thus, it was not possible to predict DON levels based on the variables included in this study. Further studies are needed to solve this problem. Apparently the infection and/or growth of DON producing Fusarium species are promoted in certain regions. One possibility may be to study the species distribution of fungal communities and their changes during the oats cultivation period in more detail.