Environmental Engineering Reference
In-Depth Information
Measuring movement when individuals cannot be tracked is more problematic.
One possibility is to analyse the
genetic structure
of sub-populations and use a
genetic model to estimate recent rates of gene flow, and hence migration (Wilson
and Rannala 2003). Another possibility is to fit a
population model
incorporating
migration to catch and abundance data, and test whether this fits the data better
than a model without migration (Section 5.3.6). Finally, for propagules dispersed
by wind, current or animals, independent experiments might be carried out to
quantify the pattern of
settlement
. For example, for wind dispersed plants, an
array of seed traps (containers or sticky surfaces) can be placed at varying distances
from focal plants (e.g. Dauer
et al
. 2007; Gomez-Aparicio
et al
. 2007), while for
animal-dispersed seeds, behavioural studies of the key dispersing species can be
used to quantify movements (e.g. Weir and Corlett 2007).
Understanding the determinants of spatial variation in abundance can enhance
our ability to assess sustainability. Whenever an abundance survey is carried out by
sampling at a number of sites, it is therefore worth considering the possibility of
measuring some key environmental variables at each site. In designing a survey of
this kind, there are number of points to consider.
Survey aims
Studies of abundance in space fall broadly into two types: correlative and predic-
tive, and you should try to decide which approach you are aiming for when design-
ing a study. In a purely correlative study, you might simply attempt to detect an
effect of harvest, ideally quantifying this effect, and perhaps controlling for any
potentially confounding environmental variables. If the impact is expected to be
very strong and not complicated by habitat, relatively little sampling effort will be
required. Predictive studies use essentially the same approach, but take the analysis
a step further by using the results to predict abundance at times and places other than
those sampled (Box 2.19). As well as its potential value in guiding management,
predictive power can improve the precision of density estimates, and can allow the
estimation of total population size across heterogenous landscapes. However, to be
reliable, predictive studies generally need to ensure high sampling intensity across
several key environmental variables.
What to measure
Large-scale surveys are highly labour intensive, and may only be possible if the
effort required to estimate abundance at each site is minimised. In practice, this
often means a
single sample plot
at each site, such as a transect or point count,
yielding either presence-absence data, or an index of abundance in the form of raw
counts of individuals or signs seen. Analysis of indices of this kind can be revealing,
but it is essential to remember that these indices may reflect variation in detection
