An eco-geomorphological modelling framework for vulnerability assessment of coastal wetlands

Coastal wetlands refer to biomes localised within the tidal prism, such as mangrove forests and salt marshes, where exposition to salt or brackish water play a role in the organisms’ functions. These wetlands are very important due to multiple factors such as coastal protection to storm surges, habitat and foraging for fish, crustaceous, water birds and other animals, water filtering and carbon sink. However, this ecosystem is under threat due to sea-level rise (SLR). Although the coastal wetlands have mechanisms to tackle natural variations in the sea level, we rely on numerical models to predict their fate, given the expected high rate of SLR for the 21st century. Over the last two decades, new models have been proposed, and their complexity has increased. In this thesis, we developed and tested an eco-geomorphological modelling framework that includes the essential mechanisms to estimate vegetation cover and soil elevation changes but still maintain reasonable computational and input data demands. We used this model to assess how different types of artificial structures can change the evolution of a coastal wetland. We observed that drainage ditches improve water conveyance in and out from the floodplains, favouring mangrove forests' expansion. Conversely, embankments with culverts to restrict the tidal flow helps the maintenance of salt marsh areas. However, under the high rates of SLR expected by the end of the current century, we can expect a significant suppression of coastal wetlands. One common simplification used to permit the simulation for a large period (e.g. the entire century) is to use sinusoidal waves to represent the average properties of the tidal cycles. We confronted such a condition with a Monte Carlo experiment using observed data, therefore estimating the uncertainty in the model outputs due to the variability in water levels and sediment concentration. The results showed that vegetation cover could vary significantly, especially in transition areas between different vegetation types, while elevation change is less affected. The use of simple inputs could track the primary trend but underestimated vegetation cover on high areas. Lastly, we collected a database from 20 sites spread worldwide to implement the framework under a simplified domain, aiming to obtain vegetation change and soil elevation trends on each site. The results showed more resilience in areas covered by mangrove than in areas covered by salt marsh, although sites where marshes are found on a high plateau showed strong resilience as well.