EGU24-12335, updated on 09 Mar 2024
https://doi.org/10.5194/egusphere-egu24-12335
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Application of artificial neural networks for modeling the climate change impacts on Mediterranean coastal lagoons:  the Venice Lagoon example.

Fabio Bozzeda1, Marco Sigovini2, and Piero Lionello1
Fabio Bozzeda et al.
  • 1Department of Biological and Environmental Sciences and Technologies, University of Salento, DiSTeBA, S.P. Lecce-Monteroni, 73100 Lecce, Italy
  • 2Consiglio Nazionale delle Ricerche, Istituto di Scienze marine (CNR-ISMAR), Venice, Italy

Mediterranean coastal lagoons play a pivotal role in the environmental, social, and economic facets of the coastal areas.  In general, coastal lagoons serve as biodiversity hotspots, supporting diverse ecosystems, including wetlands, marshes, seagrass beds, and unique fauna. They function as carbon sinks, sequestering substantial amounts of atmospheric carbon dioxide. Coastal lagoons also act as hydrological regulators, serving as natural buffer zones during extreme weather events, regulating hydrological cycles, and minimizing the impacts of flooding.  Mediterranean coastal lagoons hold significant socio-economic value, providing essential fishing grounds and supporting local fishing communities, and as tourist. Finally, coastal lagoons have been integral to Mediterranean cultures for centuries, holding historical and cultural significance. Understanding the impacts of climate change on coastal lagoon is crucial for coastal planning and adaptation strategies. In this study artificial neural networks (ANNs)are applied to estimate the impacts of anthropogenic climate change on water masses characteristics of coastal lagoon. Specifically, ANNs are used to model the associations between climate variables and water mass properties (namely temperature and salinity), which can be used for future projections. The developed ANNs approach can be applied to generic coastal lagoons, if sufficient in situ data of temperature, salinity and sea level are available for developing the model. The driving meteorological variable can be extracted from meteorological reanalysis and model climate projections if their resolution is sufficient to describe the relevant mesoscale features. The method is applied to the Venice Lagoon, the largest Mediterranean lagoon. The Venice lagoon is an ecologically and socio-economically relevant environment with notable susceptibility and a comprehensive description of climate change impacts is essential for its conservation and sustainable management. An advantage of the Venice lagoon is the relative richness of in situ observations, because of extensive past field campaigns. Here we provide and estimate of the expected changes of its temperature and salinity that are produced by low and high climate change scenario, with corresponding uncertainties. The ANN was parameterized using a combination of field data of temperature, salinity, and sea level, along with reanalysis data for ground temperature, wind speed (v and u components), temperature at 2 m, precipitation, evaporation, and humidity. Field data were obtained from a 10-year monitoring campaign, during which 30 stations within the lagoon were sampled. Reanalysis data were downloaded from the Copernicus ERA5 database. The climate scenarios used for projections were obtained from the Med-CORDEX network.

How to cite: Bozzeda, F., Sigovini, M., and Lionello, P.: Application of artificial neural networks for modeling the climate change impacts on Mediterranean coastal lagoons:  the Venice Lagoon example., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12335, https://doi.org/10.5194/egusphere-egu24-12335, 2024.