MODELING THE HYDROLOGICAL BEHAVIOUR OF GREEN ROOFS WITH NASH MODEL

Green roof is a performing tool able to effectively contribute to the stormwater management in urban areas increasingly affected by severe flood events. Knowledge of hydrological behavior of vegetated roofs could support the strategic planning of green systems in urban areas so as to prevent the occurrences of damaging events. In the context of assessing the potential of eco-covers to mitigate the urban flooding, it is necessary to restrict the horizon of the investigation from the long-term scale to the event scale. Many models exist able to reproduce the hydraulic and hydrological performances of green roofs at the event scale (HYDRUS, SWMM). They appear very accurate but significantly computational efforts and technical expertise are required for their implementation. In addition, detailed input parameters (soil properties and boundary conditions) are needed to run the models which can only be extracted from time-consuming field campaigns or laboratory tests. Parsimonious conceptual models could help overcome this issue. The Nash model is one of the simplest but very practical conceptual hydrological models whose ability to reproduce the hydrological behavior of green roofs is still almost unexplored, although the few studies on this topic available in the scientific literature confirm its high prediction performances. Therefore, practical and scientific interest in this direction is encouraged. The present work aims to test the accuracy of Nash model in predicting the performance of green systems. The model has been calibrated and validated using thirteen rainfall/runoff events, collected at two green roof experimental benches differing for the composition of the drainage layer (modular trays vs expanded clay) and located in the campus of University of Salerno under Mediterranean climate. The calibrated hydrological parameter was the storage delay coefficient “k”. The events have been characterized in term of duration, peak intensity, return period and volume and the relationship between the events properties and the reservoir delay has been investigated. The results show that the model, despite its simplicity, has proved to satisfactorily reproduce the hydrological behavior of both the experimental roofs. The k-parameter appears consistently similar for the two test beds, with a very moderately larger delay in the case of the system, equipped with plastic trays which stores, retains and delays water until the maximum capacity of the trays is reached.