Propagation of variability in climate projections within urban flood modelling: A multi-purpose impact analysis

Flooding is one of the most challenging weather-induced risks in urban areas. However, in a climate change perspective, significant gaps can still be observed in literature addressing the key role of rainfall input and related variability within urban flood impact models. The present research attempts to bridge this gap by investigating the effect of using a large ensemble of bias-corrected Euro-CORDEX climate projections on flood hazard estimations, with the goal of understanding the propagation of future climate variability in terms of flooding outputs for the urban environment. With this aim, a scenario analysis is performed for two return periods (10 and 200 years) building on nineteen climate projections for future horizon 2071-2100 under RCP 4.5 and 8.5 scenarios by means of CADDIES Caflood inundation model, using a test case in the City of Naples (Italy). Model outcomes are investigated and discussed in terms of relevant metrics and indicators available in literature targeting both general hazards, evaluated in terms of runoff volumes and flooded areas, and sectoral hazards, specified by a number of relevant literature hazard classifications. Modelling outcomes are analysed and normalised with respect to baseline values representing current climate conditions to emphasize potential impacts of climate change. Results show that flood features increase more slowly than rainfall, and the relationship between rainfall and flood increase is linear, with decreasing slope with increasing return period. Furthermore, the spread in rainfall input due to the use of different climate projections does not inflate through the impact modelling stages, albeit a dependence on the indicators used to model flood impacts can be observed. The outcomes of the research can be of aid to managers, designers and policy makers to understand the impacts of climate change in floodprone urban areas and, in perspective, to adapt urban areas exploring the feasibility and effectiveness of solutions.

Automated summary

This research investigates the impact of using a large ensemble of climate projections on flood hazard estimations in urban areas, aiming to understand the propagation of future climate variability in terms of flooding outputs. Results show that flood features increase more slowly than rainfall, and the relationship between rainfall and flood increase is linear, with decreasing slope with increasing return period. The spread in rainfall input due to the use of different climate projections does not inflate through the impact modelling stages, highlighting the potential impacts of climate change on urban flood risk.