Assessing current sustainability of use - Conservation and Sustainable Use - page 137

long-term data on catch and effort or population size data (Section 4.3.3),
although the analysis is then focused on trends rather than a one-off comparison.
Finally, non-linear patterns of density dependence are thought to be wide-
spread in nature (Sibly et al . 2005), and the simple linear density dependence of the
logistic model is therefore often likely to be inappropriate. Density dependence is
difficult to detect and measure in practice (Brook and Bradshaw 2006; Freckleton
et al . 2006, Section 2.4.4), and for many species it will be necessary to proceed
without a clear understanding of the process. In this case, the logistic model
remains useful as a default option; however, it is important to use it in the full
knowledge of the potential bias that might result.
In summary, biological reference points are very widely used, despite their prob-
lems. The reason for this is that they are simple, they can be calculated when there
are only limited data available and they don't need complicated statistics or mod-
elling. This makes them attractive as a 'quick and dirty' way of assessing sustain-
ability. It requires care and understanding to apply them in a way that is still
meaningful, rather than being downright misleading.
4.3 Trends over time
The dynamic nature of sustainability means that monitoring trends over time is the
most direct way of assessing sustainability. Putting it most simply, if there is a nega-
tive trend in a variable that is associated with system sustainability, this is an indica-
tor of concern. Trends are analysed using regression, in which the trend over time in
the variable of interest (population size, for example), is related to trends in other
variables (for example, number of hunters in the area). A number of issues arise here:
● Is the trend in the variable of interest actually a reflection of system sustain-
ability ? Trends in prices of wildlife products, for example, usually have mul-
tiple causes and are not always directly traceable back to reductions in
population size.
● Is any association between the trend in this variable and in other variables
actually reflecting causation ? Might there instead be other factors, not
included in the regression, which are impacting on both variables separately
or together and so causing a spurious association? For example, perhaps bad
weather leads to fishing fleets being unable to leave port, and so a reduction in
catch, and at the same time reduces spawning success in that year. This would
lead to an association between poor catches and low fish recruitment which
has nothing to do with the biological sustainability of fishing, and would only
be properly explained if the weather was included in the regression model.
● Is the trend real , or is it masked or exacerbated by sampling error or monitor-
ing biases? We discussed this issue in Chapter 2 with respect to ensuring good
experimental design and come back to it in Chapter 7 with respect to long-
term monitoring.
Sometimes spatial variation is used as a proxy for temporal variation. For example,
differences in animal abundance between locations that have been hunted for a