Environmental Engineering Reference
In-Depth Information
Type of
data
Known
individuals
Counts
Type of
count
Mark-recapture
design
Combined closed
and open population
Open population with
random re-encounters
Sample of
population
Population totals
or estimates
Individuals and
associated offspring
Jolly-Seber
mark-recapture
Robust
design
Individual
age ratios
Sampled
age ratios
Population
age ratios
Fig. 2.13
A decision tree for identifying the most appropriate productivity
estimation method for a given type of data.
2.4.3.3 Which method is best?
When estimating productivity, the main choice is a relatively simple one between
count-based or individual-based methods (Figure 2.13). Appropriate survey tim-
ing and careful representative sampling can make count-based methods very
effective, and these are the obvious choice for relatively visible species. In species
that are more difficult to count or follow up reliably, extensions of mark-recapture
methods can be used, and the choice among these depends primarily on the
resources that can be devoted to sampling. While more complex and intensive
sampling is required for the robust design method, it can greatly reduce the risk of
bias and poor precision inherent in Jolly-Seber and related methods. Robust
design is therefore preferable if resources allow.
In Chapter 1, we introduced density dependence as a central process for exploited
species. In principle, the strength and form of density dependence have a crucial
influence on the impacts of harvest, and defining the process can therefore be very
helpful in understanding and predicting outcomes. One approach, covered earlier
in this chapter, is to assume that the
logistic model
applies. Because this model is
defined by a linear decline in growth rate with increasing population, we can fully
parameterise it with only two parameters,
r
max
and
K
, and these can be estimated
either independently (Sections 2.4.1 and 2.3 respectively), or jointly by fitting the
model to a series of catch and population index data (Section 2.4.1). This approach
can be seen as a parsimonious default option, fitting the simplest possible density-
dependent model to the available data. The benefit of this is that it allows us to
work with cases where information on the precise form of density dependence is
lacking, which is a common situation.
