Multi-model chain for climate change scenario analysis to support coastal erosion and water quality risk management for the Metropolitan city of Venice

Under the influence of anthropogenic climate change, hazardous climate and weather events are increasing in frequency and severity, with wide-ranging impacts across ecosystems and landscapes, especially fragile and dynamic coastal zones. The presented multi-model chain approach combines ocean hydrodynamics, wave fields, and shoreline extraction models to build a Bayesian Network-based coastal risk assessment model for the future analysis of shoreline evolution and seawater quality (i.e., suspended particulate matter, diffuse attenuation of light). In particular, the model was designed around a baseline scenario exploiting historical shoreline and oceanographic data within the 2015–2017 timeframe. Shoreline erosion and water quality changes along the coastal area of the Metropolitan city of Venice were evaluated for 2021–2050, under the RCP8.5 future scenario. The results showed a destabilizing trend in both shoreline evolution and seawater quality under the selected climate change scenario. Specifically, after a stable period (2021–2030), the shoreline will be affected by periods of erosion (2031–2040) and then accretion (2041–2050), with a simultaneous decrease in seawater quality in terms of higher turbidity. The decadal analysis and sensitivity evaluation of the input variables demonstrates a strong influence of oceanographic variables on the assessed endpoints, highlighting how the factors are strongly connected. The integration of regional and global climate models with Machine Learning and satellite imagery within the proposed multi-model chain represents an innovative update on state-of-the-art techniques. The validated outputs represent a good promise for better understanding the varying impacts due to future climate change conditions (e.g., wind, wave, tide, and sea-level). Moreover, the flexibility of the approach allows for the quick integration of climate and multi-risk data as it becomes available, and would represent a useful tool for forward-looking coastal risk management for decision-makers.