We have proposed a deep learning-based prediction model using Long-short Term Memory (LSTM) and the seq2seq structure to estimate hourly rainfall‐runoff for the next 24 hours. Focusing on two Midwestern watersheds, namely, Clear Creek and Upper Wapsipinicon River in Iowa, these models were used to predict hourly runoff for a 24‐hr period using rainfall observation, rainfall forecast, runoff observation, and empirical monthly evapotranspiration data from all stations in these two watersheds. The models were evaluated using the Nash‐Sutcliffe efficiency coefficient, the correlation coefficient, statistical bias, and the normalized root‐mean‐square error.
The results show that the LSTM‐seq2seq model outperforms linear regression, Lasso regression, Ridge regression, support vector regression, Gaussian processes regression, and LSTM in all stations from these two watersheds. The LSTM‐seq2seq model shows sufficient predictive power and could be used to improve forecast accuracy in short‐term flood forecast applications. In addition, the seq2seq method was demonstrated to be an effective method for time series predictions in hydrology.
Related Articles
- Xiang, Z., Yan, J., & Demir, I. (2020). A rainfall‐runoff model with LSTM‐based sequence‐to‐sequence learning. Water Resources Research, 56, e2019WR025326. (DOI: https://doi.org/10.1029/2019WR025326)
Two watersheds and their locations in the State of Iowa. (bottom left) The detailed streamflow map of the Clear Creek Watershed and (right) the Upper Wapsipinicon River Watershed with sub-watershed boundary (red line) and USGS stream gauges (green markers).
Observations and model predictions of the strongest rainfall events in WY2016. The left figure is the 24h-ahead predictions in a 3-day window. The right figure is a single prediction for the next 24 hours on November 26, 2015, 10 a.m. Proposed LSTM-seq2seq model predicts the best.
