Environmental Engineering Reference
In-Depth Information
will clearly be difficult to satisfy in most cases, and is often a huge source of bias in
this method.
More complex
statistical catch-at-age
methods have been developed that
ameliorate some of these problems (Hilborn and Walters 1992), and these are a
mainstay of commercial fishery stock assessments. However, the approach requires
a lot of accurately sampled data on the ages or sizes of harvested individuals
over a long period of time, which is likely to be prohibitive where the harvest is
smaller scale and less intensively monitored. We do not therefore provide details
here, but a software implementation and further information are available in
MULTIFAN-CL and its associated publications (Section 2.7.1).
2.3.6.1 Indirect sign
For animals that are hard to observe, it is often easier to survey their signs, such as
dung, prints, burrows, temporary nests or calls. Physical signs are usually relatively
easy to survey using plot or distance sampling. However, in order to estimate
absolute abundance from these signs, we also need to know their
rates of produc-
tion and decay
(Box 2.7). This can be achieved by monitoring a sample of live ani-
mals to estimate the number of signs produced per unit time, and monitoring a
sample of the signs in order to estimate the proportion that disappear per unit
time. However, once all of the sampling and observation error is taken into
account for sign density as well as production and decay rates, the net result is often
very low precision
for the abundance estimates (Plumptre 2000).
A further problem with this method is that production and decay rates are
difficult
to estimate
appropriately. By its nature, the method is applied to animals that are hard
to see, making it difficult to quantify sign production rates. For example, while tame
elephants can be used to obtain useful defecation rates if they are foraging naturally,
Box 2.7
Calculation of density from indirect signs.
Given sign density,
Y
ˆ
, and assuming that rates of production (
p
) and decay (
d
) are
constant over the study period, animal density is estimated by:
Y
ˆ
p
D
ˆ
Each of the three variables that determine the abundance estimate have associated
errors that contribute to the uncertainty of the final estimate. The standard error
of any estimate derived from a multiplicative combination of several others can be
approximated by a process known as the delta method (Seber 1982):
SE(
D
ˆ
)
ˆ
CV(
Y
)
2
CV(
d
)
2
CV(
p
)
2
