Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable Analysis
The Long Short-Term Memory (LSTM) neural network model is an effective deep learning approach for predicting streamflow, and the investigation of the interpretability of deep learning models in streamflow prediction is of great significance for model transfer and improvement. In this study, four key...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-09-01
|
| Series: | Water |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4441/15/18/3184 |
| _version_ | 1797576310472572928 |
|---|---|
| author | Qingqing Tian Hang Gao Yu Tian Yunzhong Jiang Zexuan Li Lei Guo |
| author_facet | Qingqing Tian Hang Gao Yu Tian Yunzhong Jiang Zexuan Li Lei Guo |
| author_sort | Qingqing Tian |
| collection | DOAJ |
| description | The Long Short-Term Memory (LSTM) neural network model is an effective deep learning approach for predicting streamflow, and the investigation of the interpretability of deep learning models in streamflow prediction is of great significance for model transfer and improvement. In this study, four key hydrological stations in the Xijiang River Basin (XJB) in South China are taken as examples, and the performance of the LSTM model and its variant models in runoff prediction were evaluated under the same foresight period, and the impacts of different foresight periods on the prediction results were investigated based on the SHapley Additive exPlanations (SHAP) method to explore the interpretability of the LSTM model in runoff prediction. The results showed that (1) LSTM was the optimal model among the four models in the XJB; (2) the predicted results of the LSTM model decreased with the increase in foresight period, with the Nash–Sutcliffe efficiency coefficient (NSE) decreasing by 4.7% when the foresight period increased from one month to two months, and decreasing by 3.9% when the foresight period increased from two months to three months; (3) historical runoff had the greatest impact on streamflow prediction, followed by precipitation, evaporation, and the North Pacific Index (NPI); except evaporation, all the others were positively correlated. The results can provide a reference for monthly runoff prediction in the XJB. |
| first_indexed | 2024-03-10T21:51:22Z |
| format | Article |
| id | doaj.art-436c4796d6264e66a366cd8f97b0a50f |
| institution | Directory Open Access Journal |
| issn | 2073-4441 |
| language | English |
| last_indexed | 2024-03-10T21:51:22Z |
| publishDate | 2023-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Water |
| spelling | doaj.art-436c4796d6264e66a366cd8f97b0a50f2023-11-19T13:25:00ZengMDPI AGWater2073-44412023-09-011518318410.3390/w15183184Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable AnalysisQingqing Tian0Hang Gao1Yu Tian2Yunzhong Jiang3Zexuan Li4Lei Guo5State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100038, ChinaSchool of Water Conservancy, North China University of Water Resources and Electric Power, Zhengzhou 450046, ChinaState Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100038, ChinaState Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100038, ChinaSchool of Water Conservancy, North China University of Water Resources and Electric Power, Zhengzhou 450046, ChinaHenan Water Conservancy Investment Group Co., Ltd., Zhengzhou 450002, ChinaThe Long Short-Term Memory (LSTM) neural network model is an effective deep learning approach for predicting streamflow, and the investigation of the interpretability of deep learning models in streamflow prediction is of great significance for model transfer and improvement. In this study, four key hydrological stations in the Xijiang River Basin (XJB) in South China are taken as examples, and the performance of the LSTM model and its variant models in runoff prediction were evaluated under the same foresight period, and the impacts of different foresight periods on the prediction results were investigated based on the SHapley Additive exPlanations (SHAP) method to explore the interpretability of the LSTM model in runoff prediction. The results showed that (1) LSTM was the optimal model among the four models in the XJB; (2) the predicted results of the LSTM model decreased with the increase in foresight period, with the Nash–Sutcliffe efficiency coefficient (NSE) decreasing by 4.7% when the foresight period increased from one month to two months, and decreasing by 3.9% when the foresight period increased from two months to three months; (3) historical runoff had the greatest impact on streamflow prediction, followed by precipitation, evaporation, and the North Pacific Index (NPI); except evaporation, all the others were positively correlated. The results can provide a reference for monthly runoff prediction in the XJB.https://www.mdpi.com/2073-4441/15/18/3184runoff predictionLSTM modelinterpretabilityXijiang River Basin |
| spellingShingle | Qingqing Tian Hang Gao Yu Tian Yunzhong Jiang Zexuan Li Lei Guo Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable Analysis Water runoff prediction LSTM model interpretability Xijiang River Basin |
| title | Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable Analysis |
| title_full | Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable Analysis |
| title_fullStr | Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable Analysis |
| title_full_unstemmed | Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable Analysis |
| title_short | Runoff Prediction in the Xijiang River Basin Based on Long Short-Term Memory with Variant Models and Its Interpretable Analysis |
| title_sort | runoff prediction in the xijiang river basin based on long short term memory with variant models and its interpretable analysis |
| topic | runoff prediction LSTM model interpretability Xijiang River Basin |
| url | https://www.mdpi.com/2073-4441/15/18/3184 |
| work_keys_str_mv | AT qingqingtian runoffpredictioninthexijiangriverbasinbasedonlongshorttermmemorywithvariantmodelsanditsinterpretableanalysis AT hanggao runoffpredictioninthexijiangriverbasinbasedonlongshorttermmemorywithvariantmodelsanditsinterpretableanalysis AT yutian runoffpredictioninthexijiangriverbasinbasedonlongshorttermmemorywithvariantmodelsanditsinterpretableanalysis AT yunzhongjiang runoffpredictioninthexijiangriverbasinbasedonlongshorttermmemorywithvariantmodelsanditsinterpretableanalysis AT zexuanli runoffpredictioninthexijiangriverbasinbasedonlongshorttermmemorywithvariantmodelsanditsinterpretableanalysis AT leiguo runoffpredictioninthexijiangriverbasinbasedonlongshorttermmemorywithvariantmodelsanditsinterpretableanalysis |