Performance assessment of rainwater harvesting systems: Influence of operating algorithm, length and temporal scale of rainfall time series

Abstract Rainwater harvesting (RH) provides an important alternative for alleviating urban water scarcity while also providing stormwater control benefits. In this study, a hydrologic model was developed and applied to systematically investigated the influence of operating algorithms (e.g., “yield before spillage”, YBS and “yield after spillage”, YAS), length and temporal scale of rainfall time series on the assessment of water saving and stormwater control performance of RH for a case study application in Seattle, USA. Three performance indices (water saving efficiency, reliability, and stormwater capture efficiency) were analyzed considering four water demand scenarios. The results show that the YBS algorithm results in overestimation of the three indices while the YAS tends to underestimate them. The influence of operating algorithm selection can be diminished by using smaller simulation time steps for RH systems with larger tanks. The indices obtained using daily time series are close to those generated with the hourly ones, while the use of monthly time series generates inaccurate results. As the time series length of simulation increases, the difference between the indicator values obtained using short-term simulation periods and the reference 49-year period reduces. Representative rainfall time series length of 13, 15, 17, and 15 years for toilet flushing, laundry, lawn irrigation, and mixture of the demands, respectively, is determined and verified to be sufficient to generate equivalent indicator values to those obtained using a 30-year long time series. These results provided reference for proper selection of operating algorithms, length and temporal scale of rainfall time series for continuous modeling and performance assessment of RH systems.