Agriculture Reference
In-Depth Information
3.7.3 Early warning systems
Occasionally, the results of survey indicate the need for action by farmers. The
appearance of new pathotypes of an important fungus may suggest that they should
choose alternative varieties for the next cropping year or else budget for additional
fungicide applications. In the UK, such information of immediate value is
announced in the farming press. The fact that this is done rarely, however, is seen as
a mark of success for the UKCPVS, the aims of which are to:
•
protect cereal growers from unexpected epidemics
•
support plant breeding for improved and more durable resistance
•
underpin cultivar evaluation and recommendation schemes
(see http://www.hgca.com/content.output/56/56/Crop-Research/Crop-Research/ UK
Cereal-Pathogen-Virulence-Survey.mspx). Frequent press releases announcing, for
example, the unexpected appearance of new races of fungi would indicate failure to
achieve those aims.
In Denmark, use of the internet for this purpose has gone further than in any
other country, just as it has with variety recommendations. Early warnings of
important problems are disseminated to farmers through the national crop advisory
system, PlanteInfo (http://www.planteinfo.dk), which is now used by the majority of
Danish farmers.
3.7.4 Decision support
Computerised decision support systems are becoming important in agricultural
advisory services. In particular, they can help farmers to manage variable costs such
as that of pesticide applications. Systems such as PlanteInfo in Denmark offer the
potential to provide advice relevant to different varieties, rather than to a crop
species as a whole. One of the criteria on which the decision support should be
based is disease resistance, so accurate, up-to-date knowledge about the
susceptibility of varieties to current pathogen populations and new pathotypes is
essential.
An important influence on the susceptibility of a variety to a disease is the
frequency of matching virulence. A simple model has been devised to predict the
frequencies of virulences in the barley mildew pathogen population in a six-month
period, from information about the host and pathogen populations in the previous six
months. Hovmøller
et al.
(1993) applied this model to data from an area of 1 km
radius. The model was largely successful in predicting virulence frequencies but
there was one striking exception, in that a much higher frequency of the virulence
Va9, corresponding to
Mla9
resistance, was observed in one sample than the model
predicted. This was because there was a particularly large number of volunteer
seedlings of Ida, which has the resistances
Mla9
+
Mlk1
+
Mlg
.
If such a model is to be incorporated into a decision support system, predicted
virulence frequencies must be converted into predictions of disease severity. This
will need two additional kinds of information: meteorological data, to predict mean
