Genetic Monitoring for Managers


Elements of GEM Project Design

Genetic Monitoring Program Design

Most regular monitoring efforts will yield useful information, even though the most valuable aspects might not be apparent until decades later. To maximize prospects of obtaining useful results, numerous factors should be considered in monitoring program design:

Identify objectives: Long-term monitoring programs might begin with rather general objectives that are likely to evolve over time. Shorter-term efforts generally have more focused objectives that should be as specific as possible. For example: (i) estimate abundance or Ne of a target population with a coefficient of variation = 0.5; or (ii) have a 95% probability of detecting invasive species (or hybrids or immigrants) in the study area. Alternatively, the primary objective might be to draw general conclusions about ecological impacts that will be applied to other populations (thus increasing the scope of inference); for example, the evaluation of whether dams alter migration patterns in an aquatic species over time.

Evaluate potential sampling and analytical methods: Potential sampling and analytical methods should be evaluated with respect to: (i) relevance to objectives. Can the method provide the right type of information? In general, long-term projects should begin with a relatively simple sampling design to provide flexibility for future changes; (ii) logistics. Is it feasible to collect the necessary data?; (iii) power. Given a feasible sampling regime, can the analytical method deliver the necessary statistical power to accomplish the objective?; and (iv) robustness. Unexpected events will occur; how sensitive are the methods to unforeseen developments and model assumptions?

Include an experimental design: Appropriate controls should be established whenever possible; if this is not feasible, changes in key population parameters can still be monitored, but it will be more difficult to establish cause-and effect relationships. If the objectives involve broad inferences about ecological processes, a more complex experimental design is called for. For example, the traditional BACI (before-after, control-impact) design can be modified to enable more temporal and spatial replication of monitoring efforts (Underwood 1994).

Manage adaptively: The program should allow for maximum flexibility to respond to new information and unanticipated factors, including changes in the natural system (e.g. flood, drought or cyclic climate regimes); anthropogenic changes (uncontrolled modifications to system); and new analytical methods. If the program extends for any period of time, it is likely that important new methods will emerge. Can the data and sampling regime accommodate new methods? Can voucher specimens be archived for subsequent verification or to provide material for new (as yet unforeseen) analyses?