Agriculture Reference
In-Depth Information
Most forecast systems have been modified after their initial development. Several
modifications have been applied to Blitecast, perhaps because computers were just
coming into widespread use when Blitecast was first published. Blitecast itself is an
integration of two previous forecasts developed by Hyre (1954) and Wallin (1962).
The best known modification was to incorporate Blitecast into a comprehensive
potato management programme that was available on a diskette (Connell
et al
., 1991;
Stevenson
et al
., 1994). Attempts to incorporate weather forecasts (Raposo
et al
.,
1993) into practical use of Blitecast have not yet been successful.
For other purposes, such as estimating global severity of late blight based on
climate characteristics, analyses using Blitecast integrated with GIS technology was
informative (Hijmans
et al
., 2000). It seems likely that the next improvements in late
blight forecasting will be to include host susceptibility, pathogen occurrence and
virulence and aggressiveness as variables. To accomplish these tasks, simulation-based
forecast models need to become more flexible. Recently, SimCast, a simulation-based
forecast system, was modified to incorporate the effects of cultivars with high levels of
resistance and evaluated in Central Mexico (Grünwald
et al
., 2002a). Effective
management of late blight was obtained by combining host resistance and fungicide
sprays according to SimCast.
Local validation of any forecast system is necessary, because local environmental
conditions and the pathogen population may differ from those existing where the
forecast system was developed. A forecast may reflect an implicit assumption that is
perfectly appropriate for the location in which the forecast was developed, but is not
universally applicable. For example the rules of Van Everdingen would not be useful
in the Toluca Valley in Mexico, because the night- time temperatures are usually
below 10°C.
The assumption (typical of temperate zone potato production) that the pathogen is
at a very low population level at the beginning of the season may not be accurate for
all agro-ecosystems. In some areas of the tropics inoculum is available all year-round.
Using Burkard samplers, the concentration of sporangia in the troposphere (the
atmosphere layer that extends from ground level up to approximately 10 km height)
was monitored throughout the year (January 2004 to January 2005) in Viçosa, Brasil
(20º 45'S, 42° 52'W). Sporangia were collected in 41 of 51 (78%) sampled weeks
(Fig. 17.2) (M.A. Lima, L.A. Maffia and E.S.G. Mizubuti, unpublished data).
On the other hand, for countries in temperate climates, there have been periods in
which
P. infestans
was essentially absent. For example, some locations in the United
States were free of late blight for some years during the 1980s (Goodwin
et al
.,
1995b; Johnson
et al
., 1997). The incorporation of reliable estimates of incoming
inoculum would have contributed to more efficient disease management during
those years. In the mid 1990s in the USA, after the introduction of immigrant
strains, late blight was widespread and the assumption of inoculum availability was
again justified. However, in the future and in some locations we may again return to
the situation with rare occurrences of late blight and then information about the
probability of inoculum availability will again be useful in disease management.
Efforts to forecast the probable occurrence of the pathogen are justified, and when
successful, would be much welcomed. Disease forecasting is covered in detail in
Chapter 9.
