Improving LSTM hydrological modeling with spatiotemporal deep learning and multi-task learning: a case study of three mountainous areas on the Tibetan Plateau
Improving LSTM hydrological modeling with spatiotemporal deep learning and multi-task learning: a case study of three mountainous areas on the Tibetan Plateau
复制标题
DOI：
10.1016/j.jhydrol.2023.129401
复制
发表时间：
2023-03
影响因子：
6.4
通讯作者：
Bu Li;Ruidong Li;Ting Sun;Aofan Gong;F. Tian;Mohd Yawar Ali Khan;G. Ni
中科院分区：
文献类型：
--
作者：
Bu Li;Ruidong Li;Ting Sun;Aofan Gong;F. Tian;Mohd Yawar Ali Khan;G. Ni
长短期记忆（LSTM）网络在处理长时间时间动态方面表现出优异的能力，并已被证明在降水径流模拟中是有效的。然而，目前的LSTM水文模型缺乏多任务学习和空间信息的结合，限制了其充分利用气象水文数据的能力。为了解决这一问题，本研究提出了一种基于时空深度学习（DL）的水文模型，该模型将二维卷积神经网络（CNN）和LSTM相结合，并引入实际蒸发（E a）作为额外的训练目标。此外，采用探针方法对所提出的深度学习模型的内部嵌入层进行了解码。结果表明，LSTM和CNN-LSTM水文模型均能较好地模拟径流Q和E a， Nash-Sutcliffe效率系数（nse）分别大于0.82和0.95。较高的nse表明，在仅lstm模型中引入空间信息可以提高模型的整体性能和峰值性能。此外，仅使用lstm模型的多任务仿真在估计Q体积和性能方面显示出更好的准确性，nse增加了大约0.02。探测方法还表明，CNN可以在CNN-LSTM模型中捕获盆地平均气象值，而LSTM Q （ea）模型包含已知ea (Q)过程的信息。总的来说，本研究证明了空间信息和多任务学习在LSTM水文建模中的价值，并为解释DL模型的内部嵌入层提供了一个视角。
Long short-term memory (LSTM) networks have demonstrated their excellent capability in processing long-length temporal dynamics and have proven to be effective in precipitation-runoff modeling. However, the current LSTM hydrological models lack the incorporation of multi-task learning and spatial information, which limits their ability to make full use of meteorological and hydrological data. To address this issue, this study proposes a spatiotemporal deep-learning (DL)-based hydrological model that couples the 2-Dimension convolutional neural network (CNN) and LSTM and introduces actual evaporation (E a) as an additional training target. The proposed CNN-LSTM model is tested on three large mountainous basins on the Tibetan Plateau, and the results are compared to those obtained from the LSTM-only model. Additionally, a probe method is used to decipher the internal embedding layers of the proposed DL models. The results indicate that both LSTM and CNN-LSTM hydrological models perform well in simulating runoff (Q) and E a, with Nash-Sutcliffe efficiency coefficients (NSEs) higher than 0.82 and 0.95, respectively. The higher NSEs suggest that introducing spatial information into LSTM-only models can improve the overall and peak model performance. Moreover, multi-task simulation with LSTM-only models shows better accuracy in the estimation of Q volume and performance, with NSEs increasing by approximately 0.02. The probe method also reveals that CNN can capture the basin-averaged meteorological values in CNN-LSTM models, while LSTM Q (E a) models contain the information about the known E a (Q) process. Overall, this study demonstrates the value of spatial information and multi-task learning in LSTM hydrological modeling and provides a perspective for interpreting the internal embedding layers of DL models.