Agriculture Reference
In-Depth Information
5.4. Canada
DONcast is a prediction model that has been used successfully in Ontario, Canada, as a
forecasting tool mainly to help with fungicide spray decisions at heading and for grain
marketing decisions (Hooker et al., 2002). Growers and crop advisors in Ontario have used it
since 2000 and it continues to evolve with the inclusion of environmental and agronomic
diversities. It was developed from data collected from over 750 farms across Ontario, since
1996, using multiple regression techniques which allowed the most important climatic
variables associated with the increase or decreases of DON to be indentified. The model is
based on three equations, related to three different critical periods around wheat heading: i)
from 4 to 7 days before heading, ii) from 3 to 6 days after heading, and iii) from 7 to 10 days
after heading. The first critical period (i) corresponds to inoculum production: in this period,
rainfall amounts of>5 mm per day trigger and increase the DON potential while daily
minimum air temperatures of less than 10°C limit the DON potential. Similarly, weather
variables in critical periods ii and iii correspond to infection during flowering and fungal
growth. Here, the number of rainy days and days with relative humidity over 75% at 11.00 h
increase the DON potential; daily maximum temperatures over 32°C and average
temperatures of less than 12°C instead limit the DON potential. Daily weather data are
converted to binary values using a set of criteria for each variable within each critical period.
The binary values are summed within each weather variable and critical period and the
complex summations are plugged into empirical equations to forecast the concentration of
DON in wheat grain at harvest. Validation analysis has shown a greater accuracy than 80% in
determining whether a fungicide application is able to reduce DON. In 2004, a web-based
interactive model was developed for industry in Ontario. This model allows input of field-
specific weather and agronomic variables to be used as input for more accurate predictions.
The model is available on the website www.ontarioweathernetwork.com/DONcast.cfm.
Since 2000, DONcast has been validated and calibrated not only in Canada, but also in
other regions including the United States, Uruguay and France (Hooker and Schaafsma, 2004;
Schaafsma and Hooker, 2006; Schaafsma et al., 2006; Schaafsma and Hooker, 2007).
5.5. The Czech Republic
Data about DON contamination in wheat grain, weather conditions during the growing
season and cultivation practices from two field experiments conducted in 2002-2003 were
used for the development of a neural network model designed for DON contamination
prediction in The Czech Republic (Klem et al., 2007). Using the data from field experiments,
Klem et al. trained neural networks to predict DON content, on the basis of weather data, as a
continuous input variable and preceding crop as a categorical input variable. The neural
network works with five input variables: the categorial variable (preceding crop) and
continuous variables (average April temperature, sum of April precipitation, average
temperature five days prior to anthesis, sum of precipitation five days prior to anthesis).
Search WWH ::




Custom Search