On limiting the influence of serial correlation in metocean data for prediction of extreme return levels and environmental contours

Abstract

Metocean conditions change slowly, over the course of hours, sometimes even days, as storms develop and swells travel across the globe. Thus, measurements of these conditions are often serially correlated. However, many commonly employed methods for predicting metocean conditions for engineering design analyses are built upon an assumption of statistical independence of the data (e.g., hourly significant wave heights). In this brief study, we present an assessment of the serial (temporal) dependence in a selected metocean dataset. A method for processing a dataset that identifies and groups data sequences as “storm” events, and thus reduces serial dependence, is proposed and tested for estimating extreme metocean return levels. The results of this study show that the proposed procedure does indeed limit dependence and that environmental contours produced using this storm grouping procedure are reasonable based on the original dataset and when compared with associated alternative contours that ignore temporal dependence.