Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River Basin

This study predicted daily-scale drought for the Fenhe River (FHR) Basin and applied the explainable artificial intelligence (XAI) method to the model’s prediction results. Daily-scale drought prediction can provide more timely and detailed drought information, while deep learning interpretable meth...

Full description

Saved in:
Bibliographic Details
Main Authors: Zixuan Chen, Xikun Wei, Guojie Wang, Yifan Hu, Haonan Liu, Jinman Zhang, Shuang Zhou, Zengbao Zhao, Yushan Liu
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-06-01
Series:Frontiers in Earth Science
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/feart.2025.1543497/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850207852985057280
author Zixuan Chen
Zixuan Chen
Zixuan Chen
Xikun Wei
Guojie Wang
Yifan Hu
Haonan Liu
Jinman Zhang
Jinman Zhang
Jinman Zhang
Shuang Zhou
Shuang Zhou
Shuang Zhou
Zengbao Zhao
Zengbao Zhao
Zengbao Zhao
Yushan Liu
Yushan Liu
Yushan Liu
author_facet Zixuan Chen
Zixuan Chen
Zixuan Chen
Xikun Wei
Guojie Wang
Yifan Hu
Haonan Liu
Jinman Zhang
Jinman Zhang
Jinman Zhang
Shuang Zhou
Shuang Zhou
Shuang Zhou
Zengbao Zhao
Zengbao Zhao
Zengbao Zhao
Yushan Liu
Yushan Liu
Yushan Liu
author_sort Zixuan Chen
collection DOAJ
description This study predicted daily-scale drought for the Fenhe River (FHR) Basin and applied the explainable artificial intelligence (XAI) method to the model’s prediction results. Daily-scale drought prediction can provide more timely and detailed drought information, while deep learning interpretable methods can help understand the impact of different predictors on droughts and improve the credibility of the model. The standardized antecedent precipitation evapotranspiration index (SAPEI) was selected as an index for evaluating drought conditions. Five classical deep learning prediction models, namely, long short-term memory (LSTM), gated recurrent unit (GRU), bidirectional long short-term memory (biLSTM) networks, transformer (TFR), and informer (IFR), were applied in the experiment, and the performance of each model was comprehensively evaluated. The results of the test set show that all models make effective predictions of droughts in the FHR Basin, with a Pearson correlation coefficient (R) higher than 0.75. BiLSTM performs better in short-term prediction, while TFR and IFR are better at long-term prediction. The results of the deep learning interpretable model show that, aside from the strong influence of the SAPEI itself in the prediction process, the mean temperature (TM) has the greatest influence among the auxiliary predictors, followed by precipitation (PRE) and relative humidity (RHU), with potential evapotranspiration (PET) being the weakest. Our work emphasizes the importance of timely warnings of drought and the role of XAI in the development of artificial intelligence.
format Article
id doaj-art-9b5da9e629c84e53a20be8c887aeefe7
institution OA Journals
issn 2296-6463
language English
publishDate 2025-06-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Earth Science
spelling doaj-art-9b5da9e629c84e53a20be8c887aeefe72025-08-20T02:10:23ZengFrontiers Media S.A.Frontiers in Earth Science2296-64632025-06-011310.3389/feart.2025.15434971543497Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River BasinZixuan Chen0Zixuan Chen1Zixuan Chen2Xikun Wei3Guojie Wang4Yifan Hu5Haonan Liu6Jinman Zhang7Jinman Zhang8Jinman Zhang9Shuang Zhou10Shuang Zhou11Shuang Zhou12Zengbao Zhao13Zengbao Zhao14Zengbao Zhao15Yushan Liu16Yushan Liu17Yushan Liu18China Meteorological Administration Xiong’an Atmospheric Boundary Layer Key Laboratory, Xiong’an, ChinaKey Laboratory of Meteorology and Ecological Environment of Hebei Province, Shijiazhuang, ChinaHebei Provincial Meteorology Service Center, Shijiazhuang, ChinaKey Laboratory for Climate Risk and Urban-Rural Smart Governance, School of Geography, Jiangsu Second Normal University, Nanjing, ChinaSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing, ChinaSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing, ChinaSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing, ChinaChina Meteorological Administration Xiong’an Atmospheric Boundary Layer Key Laboratory, Xiong’an, ChinaKey Laboratory of Meteorology and Ecological Environment of Hebei Province, Shijiazhuang, ChinaHebei Provincial Meteorology Service Center, Shijiazhuang, ChinaChina Meteorological Administration Xiong’an Atmospheric Boundary Layer Key Laboratory, Xiong’an, ChinaKey Laboratory of Meteorology and Ecological Environment of Hebei Province, Shijiazhuang, ChinaHebei Provincial Meteorology Service Center, Shijiazhuang, ChinaChina Meteorological Administration Xiong’an Atmospheric Boundary Layer Key Laboratory, Xiong’an, ChinaKey Laboratory of Meteorology and Ecological Environment of Hebei Province, Shijiazhuang, ChinaHebei Provincial Meteorology Service Center, Shijiazhuang, ChinaChina Meteorological Administration Xiong’an Atmospheric Boundary Layer Key Laboratory, Xiong’an, ChinaKey Laboratory of Meteorology and Ecological Environment of Hebei Province, Shijiazhuang, ChinaHebei Provincial Meteorology Service Center, Shijiazhuang, ChinaThis study predicted daily-scale drought for the Fenhe River (FHR) Basin and applied the explainable artificial intelligence (XAI) method to the model’s prediction results. Daily-scale drought prediction can provide more timely and detailed drought information, while deep learning interpretable methods can help understand the impact of different predictors on droughts and improve the credibility of the model. The standardized antecedent precipitation evapotranspiration index (SAPEI) was selected as an index for evaluating drought conditions. Five classical deep learning prediction models, namely, long short-term memory (LSTM), gated recurrent unit (GRU), bidirectional long short-term memory (biLSTM) networks, transformer (TFR), and informer (IFR), were applied in the experiment, and the performance of each model was comprehensively evaluated. The results of the test set show that all models make effective predictions of droughts in the FHR Basin, with a Pearson correlation coefficient (R) higher than 0.75. BiLSTM performs better in short-term prediction, while TFR and IFR are better at long-term prediction. The results of the deep learning interpretable model show that, aside from the strong influence of the SAPEI itself in the prediction process, the mean temperature (TM) has the greatest influence among the auxiliary predictors, followed by precipitation (PRE) and relative humidity (RHU), with potential evapotranspiration (PET) being the weakest. Our work emphasizes the importance of timely warnings of drought and the role of XAI in the development of artificial intelligence.https://www.frontiersin.org/articles/10.3389/feart.2025.1543497/fulldroughtpredictiondaily-scaledeep learningexplainable
spellingShingle Zixuan Chen
Zixuan Chen
Zixuan Chen
Xikun Wei
Guojie Wang
Yifan Hu
Haonan Liu
Jinman Zhang
Jinman Zhang
Jinman Zhang
Shuang Zhou
Shuang Zhou
Shuang Zhou
Zengbao Zhao
Zengbao Zhao
Zengbao Zhao
Yushan Liu
Yushan Liu
Yushan Liu
Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River Basin
Frontiers in Earth Science
drought
prediction
daily-scale
deep learning
explainable
title Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River Basin
title_full Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River Basin
title_fullStr Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River Basin
title_full_unstemmed Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River Basin
title_short Causes of watershed drought analyzed using explainable deep learning: a case study of the Fenhe River Basin
title_sort causes of watershed drought analyzed using explainable deep learning a case study of the fenhe river basin
topic drought
prediction
daily-scale
deep learning
explainable
url https://www.frontiersin.org/articles/10.3389/feart.2025.1543497/full
work_keys_str_mv AT zixuanchen causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT zixuanchen causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT zixuanchen causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT xikunwei causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT guojiewang causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT yifanhu causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT haonanliu causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT jinmanzhang causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT jinmanzhang causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT jinmanzhang causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT shuangzhou causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT shuangzhou causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT shuangzhou causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT zengbaozhao causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT zengbaozhao causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT zengbaozhao causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT yushanliu causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT yushanliu causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin
AT yushanliu causesofwatersheddroughtanalyzedusingexplainabledeeplearningacasestudyofthefenheriverbasin