Environmental Engineering Reference
In-Depth Information
Dead recovery analysis is particularly suited to exploited species because har-
vest can be an excellent source of recoveries. On the face of it, it may seem that
only harvest mortality is accounted for in this case, but this is not so. Regardless
of the source of recoveries, the method estimates overall survival rates, including
both natural and harvest mortalities (see Section 2.4.2.4 for ways to separate
these out). The
assumptions
of the method are essentially the same as for the live
recapture approach, and the above discussion of these assumptions also applies
here. The main difference is that, although the marking occasions should be as
brief as possible, and take place at a time when there is minimal mortality or
movement, the recoveries need not be restricted in time. Also, if
age-specific
variation
in survival rate is to be estimated, it is essential to mark a combination
of both adults and young—marking young alone will not provide the necessary
information.
The
sample size
required to obtain an estimate with a given degree of precision
also depends on the expected survival and recovery rates. As an approximate rule of
thumb, if the average survival rate is to be estimated for a study period with six
marking occasions, and the probability of a dead individual being reported is
expected to be around 0.1, to achieve a coefficient of variation of around 10%, a
few hundred individuals would have to be marked on each occasion if the survival
rate were high (around 0.8), rising to a few thousand individuals per occasion if the
survival rate were low (around 0.2). Increasing the number of parameters to be esti-
mated (for example, by estimating time- or age-specific survival rates) increases the
required sample size by a factor of at least two, whereas improving the reporting
rate can give an equally dramatic reduction in the sample size required. Software
BAND2 (Section 2.7.1) provides a tool for estimating sample sizes required to
obtain a given level of precision and a given set of conditions.
Known fates
When individuals are marked and released, usually only a small proportion of
them are found when they die. The dead recovery analyses described above enable
us to deal with this situation. However, sometimes it is possible to follow fates per-
fectly, such that the time of death is known, at least approximately, for all marked
individuals. For example, this may be the case for highly visible or sessile species
that can easily be observed directly, or for animals that can be located at will using
radio-tags (e.g. Nybakk
et al
. 2002). Ideally, all individuals are of known age at the
time of marking, as this allows age-specific survival rates to be estimated. One way
to achieve this might be to follow individuals from birth.
The simplest way to analyse data of this kind is to treat them as a life table. The
method in Box 2.9 can be used to calculate the survival rate for each period, using
the numbers of marked individuals known to be alive after each time interval since
marking instead of sample counts. This approach is known as a
cohort life table
.
Unlike static life tables, no assumptions are made about the population rate of
change.
