Techniques for surveying exploited species - Conservation and Sustainable Use - page 63

a direct check on offtake to find the proportion of harvest that was not reported.
For example, this approach is used to monitor game species in North America in
situations where many of the animals are individually tagged (Rupp et al . 2000).
Registered hunters are contacted by post, requesting the numbers harvested as
well as any tag identifiers, while direct checks on offtake are also carried out to
detect unreported tags. Skalski et al . (2005b) and Pollock et al . (1995) provide
further details on this and other methods for estimating reporting probabilities.

absence of such pressures if respondents have little motivation to participate in the
survey, or if they simply have poor recall. Less often, offtake may be exaggerated if
harvesters see large catches as prestigious, or if harvesters that work in groups each
independently report their group's catch.
It is not always necessary to know the catch to estimate harvest mortality. For
example, if harvest takes place over a relatively short space of time, during which
there is no significant natural mortality, harvest mortality rate can be calculated
from population abundance estimates before and after harvest (Box 2.13). A
downside with this approach is that abundance estimates usually suffer from low
precision, and the combination of two imprecise estimates leads to extremely
imprecise estimates of mortality rate. Where it is impossible to get reliable abun-
dance estimates, this approach could in principle be taken with population
indices , such as basic encounter rates. However, this requires the strong assump-
tion that the index remains linearly related to abundance throughout. It is likely
that this assumption will often be violated, leading to biased results.
Where natural mortality during the period of harvest is an issue, we need to
disentangle this from the harvest mortality. The dead recovery form of mark-
recapture survival analysis described in Section 2.4.2.2 provides a means to do this
(Box 2.13). In this approach, if survival rate has been estimated from harvested indi-
viduals alone, the recovery rate is the probability that an individual is both harvested
and subsequently reported. If we can estimate independently the reporting probabil-
ity (see Box 2.12 for possible approaches to this problem), we can therefore derive the
harvest mortality rate. For example, Calvert and Gauthier (2005) used this approach
to estimate seasonal harvest mortality rates in greater snow geese Anser caerulescens .

Box 2.13 Indirect estimation of harvest mortality.

When natural mortality is negligible: using population estimates. Given population
estimates, N , (or population indices linearly scaled with abundance) before and
after harvest, assuming no natural mortality over the interval, harvest mortality is
given by:

N after
N before

M ˆ 1