黄河源永曲河流域地表温度与土地覆盖的时空变化特征分析

Analysis on the Characteristics of Temporal and Spatial Changes of Surface Temperature and Land Cover in Yongqu River Basin at the Source of the Yellow River

为明确黄河源地区地表温度变化特征及其影响因素,以永曲河流域为例,利用2008年、2020年Landsat数据,通过辐射传输法反演地表温度,分析地表温度的时空变化特征及其与土地覆盖类型、地形因子和归一化植被指数之间的相互关系。结果表明:①永曲河流域地表平均温度由2008年的24.330℃上升至2020年的24.592℃。②地表温度具有显著的地形分异特征,与海拔高度呈现负相关关系,高海拔地区尤为显著。地表温度随坡度增加而降低,按平缓坡、阳坡、半阳坡,半阴坡、阴坡的顺序,地表温度逐渐降低。③不同土地覆盖类型的地表温度差异较明显,草地的地表温度较高,小灌木丛较低,平均地表温度相差4.688℃。④地表温度与归一化植被指数的地理加权回归结果呈正相关关系,且自西向东相关性逐渐增强,地理加权回归法由于引入了空间权重,能定量研究空间要素的非平稳关系,并解决其异质性问题,有效避免了传统回归模型的局限性,具有一定的优势。

In order to clarify the variation characteristics and influencing factors of surface temperature in the source area of the Yellow River,taking the Yongqu River Basin as an example,the landat data in 2008 and 2020 were used to inverse the surface temperature by radiative transfer method.The temporal and spatial variation characteristics of surface temperature and its relationship with land cover type,topographic factor and normalized difference vegetation index were analyzed.The results show these as follows.The average surface temperature of Yongqu River Basin increase from 24.330℃in 2008 to 24.592℃in 2020.Surface temperature has significant topographic differentiation characteristics,and negatively correlated with altitude,especially in high altitude areas.The surface temperature decreases with the increase of slope.According to the order of flat slope,sunny slope,semi-sunny slope,semi-shady slope and shady slope,the surface temperature decreases gradually.The surface temperature difference of different land cover types is obvious.The surface temperature of grassland is higher,and that of small shrub is lower,and the average surface temperature difference is 4.688℃.The geographically weighted regression results of surface temperature and normalized differental vegetation index were positively correlated,and the correlation gradually increased from west to east.Due to the introduction of spatial weight,the geographically weighted regression method can quantitatively study the non-stationary relationship of spatial factors and solve the heterogeneity problem,which effectively avoids the limitations of traditional regression models and has certain advantages.