An explainable deep learning model based on hydrological principles for flood simulation and forecasting

Abstract

Deep learning (DL) models always perform well in hydrological simulation but lack physical-based principles. To address this gap, we integrate the relatively complex runoff generation and flow routing principals of Xinanjiang (XAJ) model into the architecture of recurrent neural network (RNN) units and establish a physical-based XAJRNN layer. Subsequently, this layer is fused with LSTM layers to construct an explainable deep learning (EDL) model, which underwent testing at the Lushui River and Qingjiang River basins in China. Compared to benchmark models, the proposed EDL model performs very well, the average Nash-Sutcliffe efficiency (NSE) values for these two basins are 0.98 and 0.94, respectively. The flood peak relative errors (PRE) and peak timing difference (ΔT) are close to zero, which demonstrate that the EDL model can accurately simulate flood events. Notably, the EDL model incorporated physical principles not only can improve flow simulation accuracy, but also enhance interpretability, which offer fresh insights for the fusion of DL and hydrological models for flood simulation and forecasting.