LONG TERM ECOLOGICAL RESEARCH AND REPRODUCIBILITY: LESSONS FROM WHOLE LAKE EXPERIMENTS
Science has recently been criticized for producing studies that cannot be replicated. Replication in this context is repeating an experiment or study and obtaining the same result. This criterion is problematic for long-term ecological research which cannot be replicated as typically uncontrolled single or multiple ecosystems are studied over time. Here, we argue that repeated observations of dynamics as well as repeated manipulations provide a means for evaluating reproducibility given that replication is not possible. We use a specific example of Peter Lake which we observed over more than 30 years and fertilized with inorganic nitrogen and phosphorus nine different times. Phytoplankton responded to nutrient increases but dynamics varied reflecting different loading rates, food webs and environmental conditions. Nonetheless, the mean responses to nutrient loading were similar when differences among years in zooplankton size and dissolved organic matter concentrations were accounted for. Overall, responses to the manipulations varied in detail but were reproducible. Long term research can powerfully address reproducibility through weight of evidence approaches where repeated observations and manipulations are evaluated against models.
Pace, M. L., University of Virginia, USA, email@example.com
Wilkinson, G. M., Iowa State University, USA, firstname.lastname@example.org
Carpenter, S. R., University of Wisconsin, USA, email@example.com
Location: 308 A/B
Presentation is given by student: No