Agriculture Reference
In-Depth Information
The late blight model of Bruhn, developed for temperate North America and the
US-1 clonal lineage of
P. infestans
, has been modified to reflect: i) the more diverse
environmental conditions in the Andes; ii) the disease cycle more accurately; and
iii) the characteristics of a strain of
P. infestans
recently introduced to the Andes
(Andrade-Piedra
et al
., 2005a,b). Validation of the model was accomplished with
observed data not used in model construction. The observed data came from field
trials with three potato cultivars in three diverse locations in Peru for a total of
twelve epidemics. Observed and simulated epidemics were compared graphically
using disease progress curves and numerically using the area under the disease
progress curve in a confidence interval test, an equivalence test, and an envelope of
acceptance test. The level of agreement between observed and simulated epidemics
was high, and the model was found to be valid according to subjective and objective
performance criteria. The model was then evaluated against thirty-two epidemics in
diverse locations (Ecuador, Mexico, Israel, USA). If the susceptibility of the cultivar
was accurately known, the model provided realistic predictions in these diverse
locations.
The rather large effort devoted to the late blight disease has caused this disease
to become a model system for modelling activities. The wealth of information on the
disease, the significant modelling effort already devoted to it and its continued
economic importance, combine to provide continued motivation for improving
models of this disease.
17.4.2 Late blight forecasting
Potato late blight has been the subject of numerous forecasting efforts. Most efforts
have been directed at improving the efficiency and efficacy of fungicides, but a
forecast could be directed at any tactic. To define efficient use of fungicides is to
mean that fungicides are not used when they are
not
needed but are used when they
are
needed.
One of the first forecast systems with solid fundamental knowledge was
developed in The Netherlands by Van Everdingen (1926) and became known as the
'Dutch Rules'. According to these rules outbreaks of late blight are probable 15 days
after the occurrence of two days with two specific characteristics: the first day
requires more than 4 hours of dew during night time with a minimum temperature
above 10°C and the following day a mean cloudiness above 0.8 with measurable
rainfall. Many other late blight forecast systems have been developed over the years
(Hyre and Horsfall, 1951; Wallin, 1962; Krause
et al
., 1975; Johnson
et al
., 1998;
Singh
et
al
., 2000), with most utilizing empirical relations of weather variables with
disease development. Usually, the aim of these systems was the timing of the first
fungicide application. Despite the empiricism, the forecast systems helped, under
some environment conditions, to achieve greater efficiency in controlling late blight
compared with weekly fungicide spray applications. It seems that under non-
conducive environmental conditions, forecast systems tend to save sprays without
significant crop losses compared with weekly fungicide applications (Raposo
et al
.,
1993).
