Complexity to Forecast Flood: Problem Definition and Spatiotemporal Attention LSTM Solution

A gated recurrent unit (GRU) network, which is a kind of artificial neural network (ANN), has been increasingly applied to runoff forecasting. However, knowledge about the impact of different input data filtering strategies and the implications of different architectures on the GRU runoff forecasting model’s performance is still insufficient. This study has selected the daily rainfall and runoff data from 2007 to 2014 in the Wei River basin in Shaanxi, China, and assessed six different scenarios to explore the patterns of that impact. In the scenarios, four manually-selected rainfall or runoff data combinations and principal component analysis (PCA) denoised input have been considered along with single directional and bi-directional GRU network architectures. The performance has been evaluated from the aspect of robustness to 48 various hypermeter combinations, also, optimized accuracy in one-day-ahead (T + 1) and two-day-ahead (T + 2) forecasting for the overall forecasting process and the flood peak forecasts. The results suggest that the rainfall data can enhance the robustness of the model, especially in T + 2 forecasting. Additionally, it slightly introduces noise and affects the optimized prediction accuracy in T + 1 forecasting, but significantly improves the accuracy in T + 2 forecasting. Though with relevance (R = 0.409~0.763, Grey correlation grade >0.99), the runoff data at the adjacent tributary has an adverse effect on the robustness, but can enhance the accuracy of the flood peak forecasts with a short lead time. The models with PCA denoised input has an equivalent, even better performance on the robustness and accuracy compared with the models with the well manually filtered data; though slightly reduces the time-step robustness, the bi-directional architecture can enhance the prediction accuracy. All the scenarios provide acceptable forecasting results (NSE of 0.927~0.951 for T + 1 forecasting and 0.745~0.836 for T + 2 forecasting) when the hyperparameters have already been optimized. Based on the results, recommendations have been provided for the construction of the GRU runoff forecasting model.

Section: Introductionmentioning

confidence: 99%

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Wang

Liu

Yue

et al. 2020

“…Unlike physical models, the latter model analyzes the relationship between the input and output variables from observed data. However, these models are data intensive and require a large number of in situ observations to make predictions reliable [38,39]. The outcomes of both probabilistic and deterministic models are, however, influenced by the choice of flood influencing factors, such as topography.…”

Section: Introductionmentioning

confidence: 99%

Water Level Forecasting Using Spatiotemporal Attention-Based Long Short-Term Memory Network

Noor

Haq

Rakib

et al. 2022

Bangladesh is in the floodplains of the Ganges, Brahmaputra, and Meghna River delta, crisscrossed by an intricate web of rivers. Although the country is highly prone to flooding, the use of state-of-the-art deep learning models in predicting river water levels to aid flood forecasting is underexplored. Deep learning and attention-based models have shown high potential for accurately forecasting floods over space and time. The present study aims to develop a long short-term memory (LSTM) network and its attention-based architectures to predict flood water levels in the rivers of Bangladesh. The models developed in this study incorporated gauge-based water level data over 7 days for flood prediction at Dhaka and Sylhet stations. This study developed five models: artificial neural network (ANN), LSTM, spatial attention LSTM (SALSTM), temporal attention LSTM (TALSTM), and spatiotemporal attention LSTM (STALSTM). The multiple imputation by chained equations (MICE) method was applied to address missing data in the time series analysis. The results showed that the use of both spatial and temporal attention together increases the predictive performance of the LSTM model, which outperforms other attention-based LSTM models. The STALSTM-based flood forecasting system, developed in this study, could inform flood management plans to accurately predict floods in Bangladesh and elsewhere.

“…Hu et al [29] described the development of a framework using long short-term memory (LSTM) and the reduced order model (ROM) to represent the spatiotemporal distribution of floods, considering the uncertainty in flood-induced conditions. Wu et al [30], in an attempt to decrease the uncertainty arising from long-time forecasting, integrated LSTM and the spatiotemporal attention module for flood forecasting. Reliable flood mapping could also benefit from remote-sensing data and satellite imageries.…”

Section: Introductionmentioning

confidence: 99%

Understanding Uncertainty in Probabilistic Floodplain Mapping in the Time of Climate Change

Zahmatkesh

Han

Coulibaly

2021

An integrated framework was employed to develop probabilistic floodplain maps, taking into account hydrologic and hydraulic uncertainties under climate change impacts. To develop the maps, several scenarios representing the individual and compounding effects of the models’ input and parameters uncertainty were defined. Hydrologic model calibration and validation were performed using a Dynamically Dimensioned Search algorithm. A generalized likelihood uncertainty estimation method was used for quantifying uncertainty. To draw on the potential benefits of the proposed methodology, a flash-flood-prone urban watershed in the Greater Toronto Area, Canada, was selected. The developed floodplain maps were updated considering climate change impacts on the input uncertainty with rainfall Intensity–Duration–Frequency (IDF) projections of RCP8.5. The results indicated that the hydrologic model input poses the most uncertainty to floodplain delineation. Incorporating climate change impacts resulted in the expansion of the potential flood area and an increase in water depth. Comparison between stationary and non-stationary IDFs showed that the flood probability is higher when a non-stationary approach is used. The large inevitable uncertainty associated with floodplain mapping and increased future flood risk under climate change imply a great need for enhanced flood modeling techniques and tools. The probabilistic floodplain maps are beneficial for implementing risk management strategies and land-use planning.