Training & Education

An Introductory Guide to Adaptive Management for Project Leaders and Participants

Step 2. Design

The purpose of this step is to design a management plan and monitoring program that are informative and provide reliable feedback. The most informative plans are those that are deliberately designed as management experiments, to discriminate between the alternative hypotheses formulated in Step 1. Typically, this involves comparing a range of management actions. This approach is referred to as "active adaptive management". The alternative, referred to as "passive adaptive management", is to assume that the most plausible hypothesis is true, and then implement the action or set of actions that the model forecasts will have the best outcome.

Active adaptive management usually provides feedback that is more reliable and less ambiguous than passive adaptive management. However, passive adaptive management may be the best (or only) alternative where:

• it is impossible or impractical to design a powerful experiment;
• the ecological costs of testing a range of actions is unacceptably high;
• there is a high level of certainty and agreement about which hypothesis is true, and thus which action is best;
• past actions or natural disturbances provide reliable information about response over a range of conditions.

Often, it is worthwhile evaluating several designs, one of which may be a "passive" design. In some cases it may be valuable to test actions in a pilot project before testing them at a larger scale, in order to narrow the range of plausible actions, and refine methodologies. In situations or areas where the risk of damage is high and irreversible, Steps 1 and 2 may lead to the decision to postpone any management intervention until research and trials in less vulnerable areas provide more information.

At this stage it is also important to plan - at least in a preliminary way - how the data will be managed and analysed; how actions, objectives, and models will be adjusted; and how information will be communicated.

• Consider a number of management options, for example: a passive approach, where one action is implemented; an active approach, where several alternatives are compared; or testing a range of options at a pilot scale, before testing one or more at a larger scale.
• Ideally, a well-designed management experiment should include controls; replication of treatments in space and time; allocation of treatments to control for bias and environmental gradients, and to ensure statistical independence; and evaluation of confidence levels and power. Researchers and statisticians can provide valuable assistance in designing management experiments.
• If necessary, consider how and when to relax some of the design principles; note the consequences this will have for how the results are interpreted, and for the value of the resulting information.

• Evaluate the proposed plan or plans, on the basis of ability to meet long term objectives, ecological and economic costs, risk of negative outcomes, and ability to fill key gaps in understanding. Decide which proposed plan to implement.

2.3 Design monitoring protocol.

• the type and amount of baseline (pretreatment) data required;
• frequency, timing, and duration of monitoring;
• indicators to be monitored at each interval;
• appropriate spatial scales for monitoring different indicators;
• who is responsible for undertaking different aspects of monitoring.

• Specify method(s) that will be used to analyse data.
• Set up system for managing data over the long term (e.g., storage, analysis, access).
• Agree on who will interpret data and who will have access to it.

• Identify who needs what information when in order to make timely changes.
• Define the intensity and degree of response in an indicator that will trigger a change in management actions or objectives.
• Adjustments should reflect the trade-off between the costs of acting if preliminary results later prove to be incorrect, and the costs of not acting if they later prove to be correct.

2.6 Set up system to communicate results and information.

• quantitative decision analysis (see Keeney, 1982; McAllister and Peterman, 1992)
• project plan that documents uncertainties, design, and other parts of Steps 1 and 2