Environmental Engineering Reference
In-Depth Information
2.3.5.2 Change in ratios
In this method, information on the proportion of the population observed comes
from sampled changes in the
population structure
following a change in that
structure. An example of this might be a harvest that targets males so that the
population sex ratio is significantly altered, alongside independent visual surveys
of sex ratio made before and after harvest (Box 2.6). In order to yield reliable,
precise estimates, a
substantial change in ratios
is required between surveys, and
this requires that offtake is both substantial and strongly targeted at one of the
categories, a constraint that means this method is rarely used.
The key
assumptions
of the basic change in ratios method are:
●
The population is closed (apart from offtake);
●
The catch is measured accurately;
●
Survey detection probabilities are equal across categories.
The first two assumptions also apply to catch-effort methods, and the discussion
of these issues in Section 2.3.5.1 is also relevant here. The assumption of
equal sur-
vey detection probabilities
across categories is required to ensure that estimated
ratios are an accurate reflection of the actual ratios, although methods exist to allow
this assumption to be relaxed (Udevitz and Pollock 1991, 1995). The method can
also be extended to situations with more than two categories (Otis 1980; Udevitz
and Pollock 1991), for example, if several size classes are recognised, or if juveniles
Box 2.6
Change in ratios abundance estimation in action.
Given categories
x
and
y
(e.g. male and female), numbers removed,
C
, and num-
bers seen,
n
, during surveys 1 and 2, population sizes before and after harvest can
be estimated by:
n
1
(
C
y
n
x
,2
C
x
n
y
,2
)
N
ˆ
1
n
1
n
x
,2
n
2
n
x
,1
N
ˆ
2
N
ˆ
1
C
Below is a simulated dataset, based on a population of 1000 before harvest, with
40% males. A reasonably heavy harvest of 20% was applied, with 80% of the off-
take males. Ratio surveys before and after harvest observed 25% of the population,
encountering each sex randomly. The counts obtained were as follows:
Male
Female
Total
Harvest,
C
16 0
4 0
2 0 0
Pre-harvest survey,
n
1
105
145
250
Post-harvest survey,
n
2
61
139
20 0
