Data variability and uncertainty limits the capacity to identify and predict critical changes in coastal systems - A review of key concepts

How do inherent variations and uncertainties in empirical data constrain approaches to predictions and possibilities to identify critical thresholds and points of no return? This work addresses this question in discussing and reviewing key concepts and methods for coastal ecology and management. The main focus is not on the mechanisms regulating the concentration of a given variable but on patterns in variations in concentrations for many standard variables in entire lagoons, bays, estuaries or Fjords (i.e., on variations at the ecosystem scale). We address and review problems related to(1)The balance between the changes in predictive power and the accumulated uncertainty as models grow in size and include an increasing number of x-variables.(2)An approach to reduce uncertainties in empirical data.(3)Methods to maximize the predictive power of regression models by transformations of model variables and by creating time and area compatible model variables.(4)Patterns in variations within and among coastal systems of standard water variables.(5)Based on the results of the review, we also discuss the concept "Optimal Model Scale" (OMS) and an algorithm to calculate OMS, which accounts for key factors related to the predictive power at different time scales (daily to yearly prediction) and to uncertainties in predictions in relation to access to empirical data and the work (sampling effort) needed to achieve predictive power at different time scales.