Geoscience Reference
In-Depth Information
key factor analyses, yet older age classes may contribute little or even nothing
to the reproduction of the next generation. Brown et al. (1993) offered a tech-
nique they called structured demographic accounting that was designed to
solve this problem and others arising from key factor analysis. This method
was similar to that of Manly (1977) in that the total variation in
R
was decom-
posed into the variances and covariance of the component recruitment (births)
and survivals during particular age classes. However, this process was done
separately for each of several age classes so that variation in births as well as
deaths due to each class could be properly assessed. A modification of this
approach has been proposed by Silby and Smith (1998), who advocated cal-
culating the impact of each
k
-value on population growth rate (
r; e
r
=
R
) rather
than generational survival. Population growth rate is measured by standard life
table methods, which account for age-specific fecundity and survival.
The limitations of key factor analysis have been documented by various
authors. Kuno (1971) showed that sample error and compensatory density-
dependent mortality can lead to spurious conclusions in key factor analyses.
Manly (1977) acknowledges these as important limitations. Royama (1996)
identified several further problems. One of these was that the factors causing
population change may fluctuate on very different time scales and those
responsible most of the variation in
R
in a set of data may not be the factors
responsible for the onset or decline of high-density conditions for species
prone to outbreaks. Royama illustrated this with an example from spruce bud-
worm, a major defoliator of forest trees in Canada. The rise and eventual
decline of a population over a 10-year period were caused by a steady decline
in larval survival, whereas recruitment (egg laying) fluctuated more from year
to year and was identified by the techniques described earlier as the key factor.
This example illustrated that key factor analyses may give misleading results or
may not be easy to interpret. A single analytic process such as key factor analy-
sis may not suffice to unravel the causes of density change.
Experimental Manipulation
j
The techniques listed in this chapter have been typically applied to naturally
occurring, unmanipulated populations. Many ecologists have turned to exper-
imental manipulations to provide proofs of the effects of various factors on
population density or survival. For example, it is possible to establish experi-
mental populations that differ in density and then to measure the mortality
