ABSTRACTThe aim of this study was to model fusarium mycotoxins against agronomic factors in order to identify those that have the greatest impact on mycotoxin levels in harvested wheat. To achieve this, fusarium mycotoxins levels were monitored, and associated agronomic data collected, in approximately 150 English wheat fields/year between 2006 and 2013. Results showed large seasonal variation in fusarium mycotoxin levels, with high levels in 2008 (13% and 29% exceeding legal limit for unprocessed soft wheat intended for human consumption for deoxynivalenol (DON) and zearalenone (ZON), respectively) and 2012 (10% and 15% exceeding legal limit for unprocessed soft wheat intended for human consumption for DON and ZON, respectively) and low levels in 2006 and 2011 (no samples exceeding legal limits for unprocessed soft wheat intended for human consumption for DON or ZON). Analysis of agronomic factors identified previous crop, cultivation and variety as the greatest risk factors. The greatest risk of mycotoxin development in grain was following maize as a previous crop and minimum tillage. The combined effect of these factors gave respective average DON and ZON levels 20 and 14 times higher than other previous crop and cultivation combinations. A newly quantified risk factor was harvest date. A 1-month delay in harvest resulted in a 10 and 25 times greater mean DON and ZON concentration, respectively, when compared to crops harvested around the long-term regional average harvest date. These results highlight the highly seasonal variation in fusarium mycotoxins in wheat and the agronomic factors that should be avoided to minimise fusarium mycotoxin levels in harvested wheat.