IMPACT ASSESSMENT FOR SILVICULTURAL AND RESTORATION PROJECTS

摘要

Three basic principles characterize the design of true experiments: replication, randomization, and local control. These principles are best ensured if investigators use appropriate sampling designs that allow monitoring studies to detect whether there was an effect after treatment/restoration and also whether that effect was due to treatment/restoration. Thus, in an optimal project there would be adequately replicated control and experimental units at appropriate spatial and temporal scales, using asound experimental design that determines the statistical analysis of response variables that represent salient patterns and processes of interest. So designed, long-term monitoring allows adaptive management to proceed that relies on feedback obtainedthrough monitoring results that would guide managers ever closer toward their management or restoration goals. It is a tragedy when you get none of this. Your study lacks a control, you are unsure about a benchmark to measure success or failure of your actions, there are too many variables to measure, and adequate replication is either impossible or your blocks differ with respect to management histories, vegetation, and topographic and edaphic features so that they cannot be considered valid replicates. Recent advances in research design and statistical analysis will be reviewed that may allow you to still salvage your study and may help you to develop a better design for the next one.