Work

A proposal for bidirectional long-short-term memory rainfall-runoff predictive modeling (BiLSTM-RPM) with a sliding window generator (SWG) technique for an urban watershed

Year: 2025

Type: article

Abstract: Neural networks (NNs) have recently gained attention for establishing Rainfall-runoff Predictive Models (RPMs) to receive accurate upcoming hydrographs. This study established Bidirectional Long-Short-Term Memory Model-based RPMs (BiLSTM-RPMs) for a small-to-medium-scale urban watershed, the Upper Kanda basin, occupying nearly 11 km2. The average rainfall of six stations and streamflow of a water level gauge were considered to gather a hundred events with minute-to-minute intervals for 12 years from 1999. The influence of batch-wise shuffling was investigated by developing BiLSTM-RPMs with a sliding window generator (SWG) technique for target lengths (TLs) such as 10 minutes (TL10), TL20, TL30, and TL60. Batch-wise shuffling was supported to predict seamless hydrographs, where all TLs achieved Nash–Sutcliffe model efficiency (NSE) above 0.8, Root Mean Square Error (RMSE) below 0.01 mm/min, and coefficient of determination (R2) for peak alignment above 0.95. Long-Short-Term Memory Model-based RPMs (LSTM-RPMs) derived appreciable forecasted hydrographs for TL10 and TL20; however, they performed poorly for TL30 and TL60, where the R2 was 0.88 and 0.62, respectively.

Source: Hydrological Research Letters

DOI: 10.3178/hrl.24-00003

Authors: Cabila Subramaniyam, Yoshiyuki Imamura, Hideo Amaguchi

Institution: Tokyo Metropolitan University

Language: English

Topic: Hydrological Forecasting Using AI

Subfield: Environmental Engineering

Field: Environmental Science

Domain: Physical Sciences

SDGs: Sustainable cities and communities, Clean water and sanitation, Climate action

Open Access status: gold

Estimated APC paid: $414