PhD student: Elise VISSENAEKENS (2019-2021)
Supervisor (ASR): Katell Guizien
To provide good management advice, connectivity in coastal areas must be fully understood. The common thread of this thesis is the analysis of the uncertainty of the SYMPHONIE2015 model and its effect on larval dispersal simulations. In the first chapter, the robustness of the model to the violation of the assumptions was tested. This was done by calculating six relative and absolute statistical indicators during and outside of wind, wave and stratification events. The results showed that the model performance is not affected by these events. In the second chapter, the instantaneous error was calculated. Then, the cumulative error distributions were compared to each other in space and time. Over time, intraseasonal differences in error distributions were smaller than interseasonal differences. In space, eight groups of error distributions could be formed. No relationship was found between model performance and stratification, water depth, resolution and slope. However, a strong correlation between current velocity and error distributions was found. In chapter three, instantaneous error was added as noise to the Lagrangian dispersion simulations and compared to the initial run to assess the effect of model error on connectivity. The median difference in transfer rate between analyzes with and without noise is zero for most areas. However, the relative transfer rate difference can vary from -100% to 100%. Knowing the uncertainties in the dispersion simulations can help to use them for management advice.