| Popis: |
Salt marshes are threatened by rising sea levels and human activities, and a major mechanism of marsh loss is edge retreat or erosion. To understand and predict loss in these valuable ecosystems, studies have related erosion to marsh hydrodynamics and wave characteristics such as wave power. Across global studies, erosion was found to be largely linearly related to wave power, with this relationship having implications on the resilience of marshes to extreme events such as storms. However, there is significant variability in this relationship across marshes because of marsh heterogeneity and the uniqueness of each physical setting. Here, we further investigate whether this linear relationship applies globally, and add a new dataset from the Great Marsh in Massachusetts (USA). We find that most marsh wave power and erosion data are not normally-distributed, and when statistically treating the data appropriately, the resulting relationships vary from previously published curves. We show the importance of maintaining statistical assumptions when performing regressions, as well as emphasize the site-specificity of these relationships. Without calibration using robust regressions at each marsh, erosion related to wave attack is not fully constrained, resulting in unreliable predictions of future marsh resilience and response to climate change. |
