FUSARIUM HEAD BLIGHT AND DON CONTAMINATION MANAGEMENT IN SOFT AND DURUM WHEAT CULTIVATION - Soybean and Wheat Crops: Growth, Fertilization, and Yield - page 147

in the model for its validation in the Pampas Region in 1993-1995, thus improving the fit
between the observed and predicted head blight incidence.
Both models developed in Pergamino satisfactorily predicted head blight incidence in
both Pergamino (Moschini and Fortugno, 1996) and in the Pampas region (Moschini et al.,
2001) showing that these models could be successfully used to predict FHB incidence in
those regions.
5.2. Argentina, Brazil, Uruguay
Fernandes et al. (2004) used a linked process based modelling approach to explain FHB
epidemics that developed at three sites in South America: Pergamino (Argentina), La
Estanzuela (Uruguay), and Passo Fundo (Brazil). In order to model FHB in wheat, they used a
wheat development model (Cropsim-Ceres 2002 included in DSSAT 4.0) and an FHB model
which calculate three parameters, i) proportion of susceptible tissue, ii) infection rate, iii)
spore cloud density. The final risk is calculated from the summation of the partial indices.
The rates and rules in the model are influenced by weather variables such as daily mean
temperature, daily mean relative humidity, daily solar radiation, and consecutive rainy days.
In Passo Fundo, Brazil, the model evaluation with disease data from 5 years of
epidemics, showed that the risk estimated by the model explained over 95% of variation in
disease field epidemics. Details on development of the model development still have to be
published.
5.3. Belgium
Detrixhe et al. (2003) developed an agrometeorological model that is able to estimate the
risk of FHB infection at a regional scale (1km x 1km grid) in winter wheat. Moisture on plant
surfaces has a considerable influence on fungal pathogen development. Therefore, estimation
of leaf wetness and duration is an important agrometeorological parameter that needs to be
taken into account in the development of FHB in wheat. This model is based on a leaf
wetness duration calculation module (Oger et al., 2002), which estimates the duration of
surface wetness from data provided by a weather station network (for standard meteorological
data) and a weather radar which cover the whole Belgian territory (for precipitation data).
Weather data (temperature, relative humidity, wind speed, short-wave and long-wave
radiation) are collected over a period that starts 8 days before wheat flowering and ends 7
days after and these data are interpolated to a grid size of 1 km x 1 km, which cover the entire
Belgian territory. A first evaluation of the model was carried out in 2002 with the collection
of 43 wheat samples and their analyses for Fusarium spp. incidence. The first evaluations
highlighted some instrumental limitations concerning the risk assessment in regions located at
more than 120 km from the weather radar and the necessity to include FHB parameters like
crop rotation, cultivar characteristics and field disease history in the model. Nevertheless, the
use of spatial interpolation of meteorological data and of the estimation of leaf wetness makes
this model very interesting and original. Further calibration/validation tests are in progress to
optimise the model. The institute (Agricultural Research Center, Gemloux, Belgium) is also
still working on the question of leaf wheat duration (Oger, 2008, personal communication ).