Demonstration of the impacts of anti-sedimentation techniques on Japanese reservoir siltation via mass data ANN analysis

Abstract

Reservoirs have been installed as long-term assets to guarantee water and energy security for decades, if not centuries. However, the effect of siltation undermines reservoirs' sustainability because it significantly reduces the reservoirs' original capacity. Extreme events such as typhoons, floods and droughts are posited to have extreme impacts on sediment inflow and deposition in reservoirs. The same holds true for ISMTs (implemented sediment management technologies), such as dredging, spilling and bypassing. However, the large-scale analysis of their effects on reservoir sedimentation progression, recovery and development was not feasible due to data scarcity and technological restrictions. The present paper closes this information gap by conducting a GRU (gated recurrent unit) neural network analysis of 1,224 Japanese reservoirs, for which the sedimentation, local precipitation, extreme events and ISMTs were monitored between 2000 and 2017. The network reveals the beneficial impacts of dredging, spilling and bypassing. The results also demonstrate the potential of smart management and improved monitoring for sedimentation threat abatement. Thus, foresighted engineering and dedicated governance action in flood and drought scenarios can significantly strengthen the sustainable behavior of key infrastructure elements such as reservoirs. HIGHLIGHTS Unique data set: 1,225 dams with 18 years of sediment record each.; Anti-sediment management notations.; Artificial neural networks with gated recurrent units as methodology.; Unique conclusions regarding the efficiency of sediment management methodology.; Generalized results represent multifaceted types of reservoirs.;

Keywords

• japan
• mass data
• neural networks
• reservoirs
• rnn
• sediment management