LEARNING FROM DAM REMOVAL MONITORING: CHALLENGES TO SELECTING EXPERIMENTAL DESIGN AND ESTABLISHING SIGNIFICANCE OF OUTCOMES

As the decommissioning of dams becomes a common restoration technique, decisions about dam removals must be based on sound predictions of expected outcomes. Results of past and ongoing dam removal monitoring are an important source of information that practitioners may utilize to evolve predictive and decision-making tools, emphasizing the need for thorough and defensible documentation of dam removal outcomes. However, as dam removals challenge many basic assumptions of conventional experimental designs and data analysis techniques, the quality of information available to aid decision-making may be questionable or misleading. Nevertheless, some study design principles and analysis procedures may be robust to the challenges presented by dam removal research. To assist managers in undertaking dam removal monitoring, this article discusses the assets and limitations of monitoring and analysis options available for dam removal studies, with emphasis on selecting a rigorous experimental design and determining significance of results. As the chosen monitoring design will influence the appropriateness of applying standard analytical methods, particularly statistical hypothesis testing, researchers should carefully consider constraints inherent to dam removal studies when designing a monitoring plan and assigning significance to observed changes. Ecological significance is often the most justifiable method for framing significance of dam removal outcomes, though it may be complicated by identification of environmentally significant thresholds. Another alternative is evaluation of the practical significance of results, when observed changes exceed measurement error and background variability. Establishing practical significance may be informative when statistical and ecological significance is inappropriate or impossible to determine. Copyright (C) 2010 John Wiley & Sons, Ltd.