Probabilistic Models for Optimal Rainwater Harvesting Tank Sizing: a Comparison with Traditional Approaches

Abstract With growing water scarcity and increasing variability in rainfall patterns due to climate change, along with the pressures of urbanization and population growth, the adoption of non-conventional water resources such as rainwater harvesting and water reuse has become a crucial sustainable strategy. Traditional approaches to the design of rainwater tanks, based on the demand volume in a predefined dry period, don’t take into account properly the risk of failure. Two probabilistic methods are then proposed, to address this issue. The first method is a modification of the conventional “demand-side” approach, by the use of a probabilistic estimation of the inter-event time. The second one, more reliable and more complex, takes into account in a parametric way the full stochastic rainfall process, allowing to consider also the pre-filling possibility due to consecutive storm events. Complexity in this second method is managed in order to develop direct relationships for practical applications. Although enough simple, these methods improve the design of Rainwater Harvesting Systems (RWHs), allowing to apply cost-benefit analysis procedures. Methods are compared and evaluated through the application to a case study in Milan, Italy.