Near real-time retrieval of lake surface water temperature using Himawari-8 satellite imagery and machine learning techniques: a case study in the Yangtze River Basin

• DOI: 10.3389/fenvs.2023.1335725
• Publication Date: 2024-01-05T05:55:39Z
• Authors: Kaifang Shi; Jing-Cheng Han; Peng Wang
• Keywords: Yangtze river

Lake Surface Water Temperature (LSWT) is essential for understanding and regulating various processes in lake ecosystems. Remote sensing large-scale aquatic monitoring offers valuable insights, but its limitations call a dynamic LSWT model. This study developed multiple machine learning models retrieval of four representative freshwater lakes the Yangtze River Basin using Himawari-8 (H8) remote imagery in-situ data. Based on situ dataset Chaohu, were effectively configured validated to perform H8-based inversion . The test results showed that six provided satisfactory retrievals, with Back Propagation (BP) neural network model achieving highest accuracy an R -squared ( 2 ) value 0.907, Root Mean Square Error RMSE 2.52°C, Absolute MAE 1.68°C. Furthermore, this exhibited universality, performing well other within Basin, including Taihu, Datonghu Dongtinghu. ability derive robust estimates confirms feasibility real-time synchronous satellites, offering more efficient accurate approach Basin. Thus, proposed would serve as tool support implementation informed policies environmental conservation sustainable water resource management, addressing challenges such climate change, pollution, ecosystem restoration.