Interpreting Conv-LSTM for Spatio-Temporal Soil Moisture Prediction in China
Feini Huang,
Yongkun Zhang,
Ye Zhang,
Wei Shangguan,
Qingliang Li,
Lu Li,
Shijie Jiang
Affiliations
Feini Huang
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies, School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai 519082, China
Yongkun Zhang
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies, School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai 519082, China
Ye Zhang
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies, School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai 519082, China
Wei Shangguan
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies, School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai 519082, China
Qingliang Li
College of Computer Science and Technology, Changchun Normal University, Changchun 130032, China
Lu Li
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies, School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai 519082, China
Shijie Jiang
Department of Computational Hydrosystems, Helmholtz Centre for Environmental Research, 04318 Leipzig, Germany
Soil moisture (SM) is a key variable in Earth system science that affects various hydrological and agricultural processes. Convolutional long short-term memory (Conv-LSTM) networks are widely used deep learning models for spatio-temporal SM prediction, but they are often regarded as black boxes that lack interpretability and transparency. This study aims to interpret Conv-LSTM for spatio-temporal SM prediction in China, using the permutation importance and smooth gradient methods for global and local interpretation, respectively. The trained Conv-LSTM model achieved a high R2 of 0.92. The global interpretation revealed that precipitation and soil properties are the most important factors affecting SM prediction. Furthermore, the local interpretation showed that the seasonality of variables was more evident in the high-latitude regions, but their effects were stronger in low-latitude regions. Overall, this study provides a novel approach to enhance the trust-building for Conv-LSTM models and to demonstrate the potential of artificial intelligence-assisted Earth system modeling and understanding element prediction in the future.
