The Lancang-Mekong River Basin (LMRB) is an important transboundary river basin containing large-scale cascade hydropower dams. There has been increasing interest in the cross-border impacts of cascade reservoirs and climate change on the water temperature of the LMRB. However, there remain limited related basin-scale quantitative studies. This study applied a deep learning method (Long short-term memory neural network, LSTM) to reconstruct the natural water temperature of the Lower Mekong River (LMR), which was compared with observation data influenced by dam operation. The aim of this study was to assess the impacts of the dams and climate change on the dynamics of water temperature in the basin. The results showed that: (1) Climate change had a relatively minor effect on the water temperature of the LMRB, with the warming effect of climate change gradually increasing from upstream to downstream. (2) The operation of the cascade reservoirs has resulted in seasonal warming and cooling of water temperature, although the average annual water temperature of the LMRB has increased by 0.27–0.59 °C. (3) The slight increase in annual average water temperature due to reservoir operation gradually weakened from upstream to downstream. (4) The increase in annual river temperature became more pronounced from 2008 to 2016 with the completion of Xiaowan, Nuozadu, and other power stations. This study characterized the responses of streamflow temperature to climate change and cascade reservoirs and can act as a reference for water resources management and hydropower development in the LMRB.

中文翻译：

---

梯级水库跨境对澜沧江-湄公河水温的影响

澜沧江-湄公河流域（LMRB）是一个重要的跨界河流流域，拥有大型梯级水电大坝。人们越来越关注梯级水库和气候变化对 LMRB 水温的跨境影响。然而，相关的流域尺度定量研究仍然有限。本研究应用深度学习方法（长短期记忆神经网络，LSTM）重建湄公河下游（LMR）的自然水温，并将其与受大坝运营影响的观测数据进行比较。本研究的目的是评估大坝和气候变化对流域水温动态的影响。结果表明：（1）气候变化对LMRB水温的影响相对较小，气候变化的增温效应从上游到下游逐渐增强。 (2)梯级水库的运行导致水温季节性变暖和变冷，尽管LMRB年平均水温升高了0.27~0.59℃。 （3）水库调度导致的年平均水温小幅上升从上游到下游逐渐减弱。 （4）2008—2016年，随着小湾、糯杂渡等电站的建成，年河流水温上升趋势更加明显。该研究表征了径流温度对气候变化和梯级水库的响应，可为LMRB水资源管理和水电开发提供参考。