Assessment of the length of rainfall time series for rainwater harvesting in buildings

In this study, the possibility of using short-term instead of long-term rainfall time series in simulations of rainwater harvesting in buildings was evaluated. The main objective was to determine the required length of a short-term time series in order to produce results similar to those obtained with a long-term time series. The study was conducted using daily rainfall data on thirteen cities located in different countries. The 30-year time series were used as a reference and, from these, short-term time series were extracted, i.e., 30 series of 1-year length, 29 series of 2-year length, and so on. The computer programme Netuno was used to carry out the simulations with short-term and long-term time series and the results were compared. A simulation model was defined and some parameters were fixed, such as the catchment area and number of inhabitants, and others were varied, such as the time series length and the rainwater demand. The results obtained using the short-term time series were compared to those using the long-term time series considering three factors: potential for potable water savings obtained for the same rainwater tank size; the optimal rainwater tank size; and the ideal potential for potable water savings. For each city, the representative time series length was determined based on the frequency of similar results, i.e., when at least 90% of the results were similar to those obtained using a 30-year time series. A validation of the representative time series length was performed using a different period from that used in the initial simulation. Different representative time series lengths were found for each city, the shortest being 6 years and the longest 20 years. A time series of 15 years was sufficient to obtain results similar to the 30 year time series in seven out of the thirteen cities. It was concluded that the use of short-term time series instead of 30-year time series for the simulation of rainwater harvesting systems is valid, depending on the rainfall characteristics of the region.