利用2012年1-12月在新安江水库(千岛湖)6个点位的每月一次的水温及其他环境因子的周年观测资料,分析了水库水温逐月变化、季节变化、垂直分布及温跃层的形成与变化,探讨了温跃层特征量(温跃层深度、厚度、强度)与表层水温、水体透明度的关系.新安江水库表层和中层水温与气温存在显著的线性相关,又以表层水温线性关系最好,而下层水温与气温没有显著相关性,说明下层水温受气温的影响很小,全年处于相对恒温状态.水库表层和中层水温逐月变化明显,呈现夏季最高、春秋季次之、冬季最低的变化趋势,其中中层水温最高值出现的季节较表层水温明显后延,下层水温没有明显的逐月变化和季节变化.水温垂直分布显示,4个季节均存在不同程度的温跃层和温度分层现象,其中水深最深的大坝前水温分层最明显.小金山、三潭岛和大坝前3个典型点位从春季的4月份到冬季的2月份温跃层深度由1.61±0.47 m逐渐增加至39.37±5.35 m,而温跃层厚度和强度则在夏季最高、冬季最低,温跃层随着季节的变化呈现增强稳定减弱消失的周期变化.温跃层深度与水体透明度存在显著正相关,与表层水温存在显著负相关,并基于透明度和表层水温建立温跃层深度的多元线性回归模型.;Based on the in situ investigation data obtained during the period from January to December in 2012 in Xin'anjiang Reservoir, the monthly and seasonal variations, and vertical distributions of water temperature are analyzed. In addition, the thermocline depth, thickness and strength were calculated and the correlations between thermocline depth and water temperature at the surface and secchi disc depth were developed. Significant and positive linear relationships were found between air temperature and water temperature at the surface and middle layers with the marked high correlation coefficient for water temperature at the surface layer. However, no significantly positive relationship was found between air temperature and water temperature at the bottom layer where temperature was maintained almost as an approximate constant indicating that water temperature at the bottom layer was not affected by air temperature. The monthly and seasonal variations of water temperature at the surface and middle layers were as follows:summer > spring, autumn > winter, with a maximum in July or August and a minimum in February or March. The maximal water temperature at the middle layer was lagged to September or October compared to that of the surface layer. In addition, no marked monthly and seasonal variations were found for water temperature at the bottom layer. There is no obvious temperature differences and thermal stratification between the surface and the deep layer water in all four seasons. From April to February, the mean thermocline depth gradually increased from 1.61 ± 0.47 m to 39.37 ± 5.35 m for the three typical sites (Xiaojinshan, Santandao and Dabaqian). In contrast, the thermocline thickness and strength had the highest values in summer and the lowest values in winter. Seasonally, a periodic variation of thermocline from amplification to stabilization, to weakness and further to disappearance was recorded from April to March. Significant and positive linear relationships between the thermocline depth and secchi disc depth, but negative linear relationships between the thermocline depth and water temperature at the surface layer were found. A multiple linear regression model between the thermocline depth and secchi disc depth, water temperature at the surface layer was developed to model the thermocline depth in Xin'anjiang Reservoir.