Hydraulic simulation of an urban river affected by treated effluent based on signal processing theory and physically based models

The Tonghui River–a treated effluent-affected urban river located in Beijing, China. Inspired by the signal processing theory, this study presented a simulation scheme for the treated effluent-affected river based on hydrologic monitoring, pattern recognization, pattern extraction, and hydrologic/hydraulic modelling. It aimed to precisely depict the river flow patterns when detailed wastewater treatment plant effluent data was absent and to fill in the gap of the application of signal-based hydrological time series processing methods in physically based hydraulic simulation. Diurnal and semidiurnal patterns caused by the wastewater treatment plant (WWTP) effluent were recognized from the water level series using the continuous wavelet transform. Due to their small amplitudes, they were masked during flood events but dominated the flow regime in dry seasons. Based on the discrete wavelet decomposition and Fourier series fitting, these periodical patterns were extracted and fitted. With a preliminarily calibrated hydraulic model and a linear signal amplifier, a simulated WWTP effluent was retrieved. Dry seasons simulation utilizing the simulated effluent obtained significantly better performance than using the average effluent data from the aspects of conventional evaluation metrics, cross-wavelet transform, and wavelet coherence. [Display omitted] •The effluent-affected urban river flow series was interpreted as a complex signal.•Diurnal effluent patterns were extracted based on signal-processing approaches.•An improved dry-season flow regime simulation performance was obtained.•The diurnal patterns were correctly simulated though the effluent data was absent.