Agriculture Reference
In-Depth Information
Fig. 2.12. (a) Area under the disease progress curve (AUDPC) for a hypothetical epidemic.
(b) Illustration of the midpoint rule (or trapezoidal integration method) for calculating
AUDPC. (Campbell and Madden, 1990b).
which are late in the crop growth cycle and where yield is accumulated over a short
period of time. However, integral models that use AUDPC cannot distinguish
between early and late-occurring epidemics without applying weighting factors to
assessments at different growth stages (Hills et al., 1980).
Generalized or non-linear models are sometimes more appropriate where the shape
of the loss-disease curve dictates that this approach should be used; many such
models can have variability in the shape of the curve relating yield to
the disease descriptor. Synoptic or multivariate statistical models are used where
multiple diseases and other constraints may be determining the yield-loss relationship,
a situation often encountered in actual crop production systems; data for such models
often derive from surveys, in which no manipulation has been carried out to obtain
specific disease levels. Complex multivariate techniques for analysis of the data,
such as principal components and correspondence analyses, may be required.
Models can be expanded to account for control costs and resulting economic yield,
both in quantity and quality. Expert systems and geographic information systems
(GIS) can also be used to provide regional estimates of losses in agricultural
production.
2.7 CONCLUSIONS AND FUTURE DEVELOPMENTS
The future of agricultural research will depend on understanding higher levels of
organization, and these are the levels addressed by the science of applied ecology;
thus, the crop is a plant population and pathogens are populations of organisms with
which the plant population interacts (Weiner, 2003). This concept will demand that
 
Search WWH ::




Custom Search