Tidal marshes are coastal systems that provide valuable ecosystem services, highlighting coastal protection and carbon burial. For centuries, these dynamic ecosystems have kept pace with sea level rise through organic and mineral matter accumulation. In the current situation of accelerated sea-level rise and changes in suspended sediment concentrations, the evolution of these systems has gained special attention across scientific fields. Several methodologies like process-based models and machine learning algorithms have been applied to assess the evolution of tidal marshes in different sea level rise and suspended sediment concentration scenarios. However, up to now, these methodologies have not been integrated to assess and model the evolution of marshes. In this study, we have successfully combined a machine learning algorithm with a dynamic process-based eco-geomorphic model to assess and evaluate potential distributions of three Spanish marshes, under a selection of potential sea-level rise and suspended sediment concentrations which may unfold during this century. Results obtained from this study have proven that through the integration of existing methodological approaches and their adaptation to test contexts, we can better simulate the potential evolution of marsh systems on a local scale considering potential sea level rise and suspended sediment concentration changes. Under current sediment supply and public land availability, marshes in Oka estuary, Bay of Santander, and Cadiz Bay could lose 6.7-28.9 %, 33.1-87.5 %, and 41-86.4 % of their area, respectively. The integrated marsh evolution model presented in this work can be extrapolated and/or customised to other coastal and marine systems, fostering its reusability.
Keywords: Machine learning; Sea level rise; Suspended sediment concentration; Tidal marshes.
Copyright © 2024 Elsevier B.V. All rights reserved.