Long-Term Reconstruction and Variation Characteristics of Water Temperature in Mainstream Yangtze River
Year: 2025
Type: article
Abstract: [Objective] Due to climate change and human activities, water temperature in the Yangtze River basin is gradually increasing. Although the Hydrological Yearbooks accurately record water temperature data, the lack of long-term continuous observations and limited observation years make it difficult to precisely quantify and characterize long-term trends, posing potential challenges to the stability and health of its ecosystem. [Methods] This study innovatively combined measured water-temperature data from the Hydrological Yearbooks with ERA5-Land climate reanalysis data, employed the XGBoost machine-learning algorithm, and developed a water-temperature estimation model based on meteorological and hydrological variables to reconstruct daily water-temperature data for seven hydrological stations along the Yangtze River mainstream from 1980 to 2022. [Results] (1) At Zhutuo, Hankou, and Datong stations, the model RMSE and R2 values were 0.831-1.021 ℃ and 0.951-0.987, respectively. (2) From Zhutuo to Datong, the warming rate was 0.20-0.32 ℃ per decade. (3) At Yichang Station, the correlation between water temperature and water level reached R2=0.666. (4) Compared with 1980-1996, water temperature increased by 0.38-0.75 ℃ during 1997-2012 and continued to rise by 0.17-0.38 ℃ from 2013 to 2022. [Conclusions] (1) The XGBoost machine-learning model performs excellently and is robust in capturing river thermal dynamics. (2) Among the seven mainstream stations, those connected to lakes show the greatest warming. (3) Monthly-scale analysis indicates that water-temperature rise is synchronous with air temperature and surface downward long-wave radiation, while changes in water level lag behind. (4) Along-channel water-temperature changes in the last decade are more pronounced than in earlier periods.
Author: WANG Sheng-hui
Language: English
Subfield: Water Science and Technology
Field: Environmental Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7091262908
FWCI:
Cites:
Cited by:
Related to: