Medium-Term Urban Water Demand Forecasting with Limited Data Using an Ensemble Wavelet–Bootstrap Machine-Learning Approach

Abstract

Accurate and reliable weekly and monthly water demand forecasting is important for effective and sustainable planning and use of urban water supply infrastructure. This study explored a hybrid wavelet–bootstrap–artificial neural network (WBANN) modeling approach for weekly (one-week) and monthly (one- and two-month) urban water demand forecasting in situations with limited data availability. The performance of WBANN models was also compared with that of standard artificial neural networks (ANN), bootstrap-based ANN (BANN), and wavelet-based ANN (WANN) models. The proposed WBANN method is aimed at improving the accuracy and reliability of water demand forecasting by incorporating the capability of wavelet transformation and bootstrap analysis using artificial neural networks. Daily and monthly maximum temperature, total precipitation, and water demand data for almost three years obtained from the city of Calgary, Alberta, Canada were used in this study. For weekly and monthly lead-time forecasting, the hybrid WBANN and WANN models were determined to be more accurate compared with the ANN and BANN methods. The WANN and WBANN models simulated peak water demand very effectively. The better performance of the WANN and WBANN models for weekly and monthly water demand forecasts indicated that wavelet analysis significantly improved the model’s performance, whereas the bootstrap technique improved the reliability of water demand forecasts by producing ensemble forecasts. WBANN models were also found to be effective in assessing the uncertainty associated with water demand forecasts in terms of confidence bands, which is helpful in operational water demand forecasting. This study was conducted with a very short length of available data, indicating the effectiveness of WANN and WBANN modeling approaches in situations with limited data availability.