基于MODIS影像、中国湖泊数据集及气象数据,综合分析了2000—2019年赛里木湖湖冰物候特征变化及影响因素。结果表明：（1） 赛里木湖湖冰开始冻结和开始消融日期平均出现在11月2日和4月26日,湖冰完全封冻和完全消融日期平均出现在1月18日和5月17日,湖冰完全封冻期和湖冰冰期平均为99 d和196 d。（2） 近20 a赛里木湖湖冰开始消融和完全消融日期均呈提前趋势,而完全冻结日期也呈提前趋势,与对应月平均气温呈正相关;湖冰完全封冻期持续时间延长,而湖冰冰期呈缩短趋势。（3） 赛里木湖冻结和消融空间模式相同,即湖岸是湖冰形成最早也是消融最早的区域。（4） 赛里木湖湖冰物候变化是自身条件（湖泊形态因子、湖岸线复杂度等）及气候变化（气温、累积负积温等）共同作用的结果。

Lake ice phenology is a sensitive indicator of climate change. To clarify the changes in ice characteristics and influencing factors of lakes in Tianshan Mountains, which has alpine characteristics, Sayram Lake of Xinjiang in China is taken as the object of study. Sayram Lake is located in a closed alpine basin of brackish water. With its unique geographical location, the lake’s evolution and ice conditions are rarely influenced by human activities, and it can be an ideal research area for lake ice phenology in this region. This study conducted an analysis mainly based on MODIS, Chinese lake data, and meteorological data to reveal the phenological characteristics of lake ice on Sayram Lake during 2000—2019 by applying threshold, trend analysis, and other methods. We also identified the climatic factors that have influenced lake ice phenology over time and draw some conclusions. Results are summarized as follows: (1) Freeze-up start (FUS), freeze-up end (FUE), break-up start (BUS), and break-up end (BUE) on Sayram Lake usually occurred on November 2, January 18, April 26, and May 17, respectively. The average rates of FUE, BUS, and BUE were -0.25 d·a^-1, -0.03 d·a^-1, and -0.44 d·a^-1, respectively. The average complete ice duration (CID, between FUE and BUS) and the average ice duration (ID, between FUS and BUE) were 99 d and 196 d, respectively. (2) In the past 20 years, the BUS and BUE of Sayram Lake showed an earlier trend, and the FUE also showed an earlier trend, which was related to the decreasing trend of the average temperature in January. The BUS and BUE of Sayram Lake are related to the rising trend of average temperature in April to May and May to June. The ID was shortened, whereas the CID was prolonged. (3) Sayram Lake can be characterized by similar spatial patterns in both freeze-up (FU) and break-up (BU) processes, as parts of the surface that freeze earlier start to melt first. A more complex lake edge corresponds to an earlier start of the freezing time. Generally, freeze begins from the lakeshore and then gradually expands to the center of the lake until the lake is completely frozen. The ice begins to break in late April, and thawing begins from the northeastern lake shore. The lake ice from the shore moves gradually to the center of the slow melting and is completely melted by mid-May. Compared with the freezing process, melting on Sayram Lake is faster, lasting only 40 d, which is distinctly different from other lakes on the Tibetan Plateau. (4) The variations of phenology and evolution of the Sayram Lake ice are the results of local and climatic factors. The temperature, accumulated negative accumulated temperature, lake morphology factors, lake area, and shape of the shoreline are the main factors that affect the phenology of lake ice. However, the effect of lake salinity and geological structure on lake ice evolution cannot be ignored. This study on the phenological characteristics of lake ice on Sayram Lake can provide supplemental data for arid/semi-arid areas that lack data and provide more accurate parameter indicators and scientific basis for studying climate change in Sayram Lake and even Tianshan.