*2.5.1. Planting recommendations*

Adhering to planting dates and planting plants at lower or optimal densities reduces mycotoxin accumulation during production [75–77]. Plants should be planted at recommended row widths and densities to specifically reduce water stress [78] and ensure optimal nutrient availability. Maize ears should be harvested from the field as soon as possible because favourable conditions for ear rot and/or mycotoxin accumulation may occur if harvest is delayed, thus leading to elevated mycotoxin levels [79, 80].

### *2.5.2. Crop rotation*

The primary objective of cultural control of mycotoxigenic fungi is to minimise factors that result in plant stress. Inoculum build-up on plant residues can be reduced by crop rotation practices, such as the rotation of maize with non-host crops [75, 81, 82]. Crop rotation with legumes, brassicas and potato could also significantly reduce *F. graminearum* contamination levels [83].

in managing mycotoxigenic fungi and their toxins. Furthermore, several studies report on a reduced fungal growth and mycotoxin contamination for *Aspergillus* and *Fusarium* using natural oils and phenolic compounds *in vitro*; however, the commercial value of such prod-

Preharvest Management Strategies and Their Impact on Mycotoxigenic Fungi and Associated Mycotoxins

http://dx.doi.org/10.5772/intechopen.76808

47

The use of biological control agents to manage mycotoxigenic fungi has been reported. Atoxigenic *F. verticillioides* strains competitively excluded FUM-producing strains and prevented them from producing FUM [103]. When these strains were applied by themselves through the silk channel, however, they resulted in high levels of FER. The effective control of toxigenic *F. verticillioides* and *F. proliferatum* by non-toxigenic *Fusarium* species in maize residues has also been observed [104]. Most success, however, has been achieved with the use of atoxigenic strains of *A. flavus* to control toxigenic *A. flavus* and *A. parasiticus*. When introduced into the soil, these atoxigenic strains reduced AF contamination of peanuts in the USA by 74.3–99.9% [105]. Atoxigenic *A. flavus* strains are now widely used to control AF in maize in several African countries (www.aflatoxinpartnership.org). Endophytic bacteria have been reported to control FUM-producing fungi by competitive exclusion [106], while *Trichoderma* strains controlled them through competition for nutrients and space, fungistasis, antibiosis, rhizosphere modification, mycoparasitism, biofertilisation and

An epidemic can be described as a 'change in disease intensity in a host population over time and space' [108]. Mathematical modelling of crop disease is a rapidly expanding discipline within plant pathology [109] with the first models developed by Van der Plank [110, 111]. In epidemiology, modelling aims to understand the main determinants of epidemic development in order to address disease management in a sustainable and efficient manner. It can, therefore, serve as an instrument to monitor and assess the risk of mycotoxin contamination in crops that would drive agronomic decisions during cultivation, in order to enhance management strategies [112].

Most research regarding disease forecasting of mycotoxigenic fungi has focussed on FHB of wheat. This disease is considered well suited for risk assessment modelling because of the severity of epidemics, compound losses resulting from mycotoxin contamination and relatively narrow time periods of pathogen sporulation, inoculum dispersal and host infection [113]. This can be seen from the online forecasting model FusaProg [114], which is a threshold-based tool to control *F. graminearum* with the optimised timing of fungicide applications and forecasts of DON content during flowering. DONCast is a prediction model from Canada that has been extensively validated and commercialised for wheat [112], while an adaption of this model has been proposed for maize. This model predicts the variation in mycotoxin levels associated with the year and agronomic effects from simple linear models using wheat samples from farmers. The DONCast model accounts for up to 80% of the variation in DON and is commercially employed for the past 10 years.

Field-based models to predict FUM B1 contamination in maize grain have been elusive, most probably due to the complexity of interactions between numerous abiotic and biotic disease

ucts has not been explored and may not be feasible [101, 102].

*2.5.6. Managing mycotoxigenic fungi with other microorganisms*

the stimulation of plant-defence mechanisms [107].

*2.5.7. Prediction systems*
