Techniques for surveying exploited species - Conservation and Sustainable Use - page 44

African forest elephants Loxodonta cyclotis have never be tamed, and are too danger-
ous to follow in the wild (Hedges and Lawson 2006). Decay rates are easier to esti-
mate, but are generally extremely variable in time and space, violating the assumption
of constant decay rate. While this problem can be avoided by relaxing the assumption
of constant decay rates (Plumptre and Harris 1995), this requires the variation in the
decay rate to be tracked over time and space, and the intense effort required for this is
unlikely to be feasible in most cases. As a result, there is often a strong temptation to
borrow production and decay rates estimated at times and places other than those at
which the sign density estimates are made. This introduces huge scope for bias, which,
in combination with low expected precision, makes this a technique to use with more
than usual care. Returning to the elephant example, the CITES MIKE (Monitoring
the Illegal Killing of Elephants) programme suggests that typical defecation rates can
be used to estimate Asian elephant abundance in weakly seasonal moist forests because
rates have been shown to be fairly constant and consistent across sites, but this
approach is not appropriate in strongly seasonal environments, where we only know
that the defecation rate is likely to be highly variable (Hedges and Lawson 2006).
In principle, call rates might be used as a crude index of abundance for vocal ani-
mals, but they will rarely be linearly related to abundance because of a variety of
confounding factors, particularly time of day, time of year and weather conditions.
While it may be possible to control for these factors by using regression models to
examine their relationship with call rate (residual variation then hopefully being
directly related to abundance), social facilitation of calls cannot be dealt with in
this way. For example, many territorial species, such as the green peafowl discussed
in Box 2.19, tend to call more when local densities are higher. When the calls of
individuals can be distinguished (e.g. some birds, Gilbert et al . 1994), the territory
mapping approach (Section 2.3.1) might be used.
2.3.6.2 Presence-absence survey
The primary aim of a presence-absence survey is to estimate proportional occu-
pancy across a number of sites, as a proxy for abundance. For example, if presence
were detected in 10% of sites visited, it would suggest a much lower population
size than if 90% of sites were apparently occupied. The definition of a site here is
case-specific. For example, it may be a discrete habitat island, or, in continuous
habitat, a systematically defined grid square. Surveys of this kind can use any com-
bination of cues to confirm presence, including direct sightings, indirect signs,
calls or automatic monitoring systems such as camera traps. This flexibility means
that for some species, particularly those that are rare or hard to see, occupancy
surveys may provide an efficient means of assessing changes in abundance
(Thompson 2004; Joseph et al . 2006).
This approach has two main drawbacks. First, it is a relatively crude index of
abundance . For example, population size at a site is likely to decline substantially
well before the population is recorded as absent. Second, the failure to detect a
species does not necessarily mean that it is absent; it may simply be hard to detect.