Extreme rainfall events have been more frequent in recent decades, potentially as a climate change effect. This has been leading to a higher risk of the failure of existing hydraulic infrastructures, and to a higher awareness regarding the unreliability of design rainfall calculated with reference to historical data recorded in the last century. With this in mind, the present study questions the stationary assumption of the rainfall Depth–Duration–Frequency curves commonly used in Sicily, the biggest island of the Mediterranean Sea. Quantiles derived from the most up-to-date regional method, regarding Sicily, based on observations in the period 1928–2010, have been compared with those extracted from a high-resolution dataset related to the period 2002–2022, provided by the SIAS agency. The results showed a remarkable underestimation of the rainfall quantiles calculated with the regional approach, especially at the shortest durations and low return periods. This means that new hydraulic works should be designed with reference to longer return periods than in the recent past, and those that currently exist may experience a higher risk of failure. Future investigation of this aspect is crucial for enhancing the effectiveness of water management and detecting hydrological risks under a changing climate.

Sustainable use of water resources requires a modern approach for complex water systems management based on system’s modeling, as unsung operation alternatives may only be devised and tested through a model. The now available IT computing tools enable setting the traditional modelling methodologies for planning and operation of such systems in a new framework. The present work deals with the development of decision support models for optimizing the operation of complex water supply systems with multiple uses and focuses on the validation of the optimization model of the regional water supply systems managed by an Italian water utility. The software used is Aquator, a state-of-the-art commercial tool for generic water resources system simulation and operational optimization. The paper shows and comments the results of the validation of the model and uses them to draw some general principles for the validation of such kind of models: in fact, operational optimization models are conceived for the enhancement of systems management and performance. Consequently, while on the one side model output is expected to reflect the actual state of the system, on the other it can point out significant management enhancements.