受全球气候变暖的影响，青藏高原湖泊、河流的温度显著升高，水温变化与水生生物的繁育息息相关，但目前对于青海湖裸鲤洄游产卵的重要场所——河流入湖口的水温研究较少，水温变化对于裸鲤产卵的影响也尚不明确。因此本研究基于刚察气象站观测数据、中国区域地面气象要素驱动数据集（CMFD）、欧洲中期天气预报中心第五代陆面再分析数据集（ERA5-Land）、第六次国际耦合模式比较计划（CMIP6）的多模式数据，尝试利用Fresh water Lake Model（简称FLake模式）模拟青海湖最大支流布哈河入湖口处水温并评估其适用性，探讨了再分析数据和CMIP6多模式数据驱动模拟历史（1981—2014年）水温的集合的优越性，预估了未来时期（2024—2100年）三种情景下的水温演变及成因，结果表明：（1） CMFD和ERA5-Land模拟水温的集合优于单个模式的模拟水温且在短期和长期模拟结果精度都较好，CMIP6多模式长期的模拟水温集合优于单个模式的模拟水温，可以较好的再现再分析数据模拟水温的集合。（2） 未来布哈河入湖口的水温随着排放强度的增加显著升高，与水温呈正相关的气象因子从大到小依次为气温、比湿、向下长波辐射和向下短波辐射，呈负相关的气象因子是风速；除远期高排放情景（SSP585）的向下短波辐射，其他情景的气象因子与水温的关系均通过了95%的显著性检验。（3） 近期（2024—2040年），三种情景的水温较历史时期均有少许升高，水温间的差异不大，青海湖裸鲤产卵的窗口期略微缩短，水温升高会对裸鲤的产卵造成轻微影响；中期（2041—2060年），三种情景的水温进一步升高，水温间的差异逐渐显现，裸鲤的产卵窗口期明显缩短，水温升高对裸鲤的产卵有一定的危害；远期（2081—2100年），三种情景的水温差异更显著，SSP126和SSP245情景的升高速率减慢；SSP585情景的水温升高速率依旧加快，水温持续升高导致裸鲤的产卵窗口期显著缩短，一定程度限制了青海湖裸鲤的产卵活动。

Under the influence of global warming, the temperature of lakes and rivers in the Qinghai-Xizang Plateau has increased significantly, and the change is closely related to the breeding of aquatic organisms. However, there are few studies on the water temperature at the estuary of rivers, which is an important place for the migration and spawning of Gymnocypris przewalskii in Qinghai Lake, and the influence of water temperature change on the spawning of G. przewalskii is still unclear. Therefore, this study is based on the observation data of Gangcha Meteorological Station, China Meteorological Forcing Dataset (CMFD), European Centre for Medium-Range Weather Forecasts Reanalysis v5-Land (ERA5-Land) and Coupled Model Intercomparison Projects 6 (CMIP6). The Fresh water Lake Model (FLake Model) is used to simulate the water temperature at the estuary of Buha River, the largest tributary of Qinghai Lake, and evaluate its applicability. It also discusses the superiority of reanalysis and CMIP6 multi-model data to drive the collection of simulated historical water temperatures (1981-2014). The evolution and causes of water temperature under three scenarios in the future period (2024-2100) are estimated. The results show that: (1) Average water temperature simulated by CMFD and ERA5-Land was better than that of the single model, and the accuracy of the simulation results was better in both short and long terms, of the average simulated water temperature of CMIP6 multi-model was better than that of a single model, which reproduced the simulated water temperature of reanalysis data set effectively. (2) In the future, the water temperature at the estuary of the Buha River may increase significantly with the increase of emission intensity. The meteorological factors that were positively correlated with the water temperature are air temperature, specific humidity, downward long wave radiation and downward short wave radiation from the largest to the smallest, while the wind speed was negatively correlated is; except for the downward short-wave radiation of the long-term high emission scenario (SSP585), the meteorological factors of the others all passed the 95% significance test. (3) In the recent period (2024-2040), the water temperature under the three scenarios may elevate slightly, increased compared with the historical period, and there is little difference between the water temperatures. The spawning window period of G. przewalskii is slightly shortened, and the rising water temperature will have a slight impact on the spawning of G. przewalskii. In the middle period (2041-2060), the water temperature of the three scenarios increased further, and the difference between the water temperatures gradually appeared. The spawning window period of G. przewalskii was obviously shortened, and the rising water temperature had certain harm to the spawning of G. przewalskii. In the long term (2081-2100), the water temperature difference between the three scenarios is more significant, and the rise rate of SSP126 and SSP245 scenarios slows down. In the SSP585 scenario, the rate of water temperature rise was still accelerated, and the continuous rise of water temperature resulted in a significant shortening of the spawning window period of G. przewalskii, which limited the spawning activities of G. przewalskii to a certain extent.