应用Landsat-8数据分析山地地表温度格局及影响要素

Patterns of Land Surface Temperature over Mountainous Area and Its Influencing Factors Using Landsat-8 Data

地表温度是研究陆地表面与大气之间相互作用的重要参数,分析山地地表温度格局及作用机理,对了解山地生态系统的时空分布特征,准确刻画山地地表水热环境具有重要意义。以云南滇中地区为研究区,应用Landsat-8数据、多窗口区域匹配算法(IMW),结合近地表常规气象站及微气象台站同步观测数据进行地表温度反演,分析地表温度空间格局及其与地理环境和人居环境因子的定量关系,构建了山地地表温度多因子作用模型。结果表明:遥感影像与地面同步观测数据相结合反演的地表温度场平均绝对误差为2.01℃、平均相对误差为2.35%、均方根误差为5.09℃,反演结果优于MODIS地表温度产品。地表的温度场空间格局与地理环境及人居环境密切相关,归一化植被指数、地形起伏度、海拔、坡度、水域影响与地表温度场成负相关,而居民区与地表温度成正相关关系;从相关性大小看,山地温度受归一化植被指数影响最大,其次是地形起伏度、海拔、坡度、水体、居民区,坡向影响最小,说明提高植被覆盖具有地表降温的重要作用,城市人居环境具有地表增温效应。因此,构建山地温度时空模拟模型,应充分考虑地理环境微格局与人居环境影响。

Land surface temperature(LST) is an important parameter to study the interaction between land surface and atmosphere. It is of great significance to analyze the pattern and mechanism of LST in mountain areas for understanding the temporal and spatial distribution characteristics of mountain ecosystems and accurately depicting the mountain surface water and thermal environment. Taking the central Yunnan region as the research area, using Landsat-8 data, multi-window region matching algorithm(IMW), combined with the synchronous observation data of near-surface conventional meteorological stations and micrometeorological stations, the land surface temperature spatial pattern and its quantitative relationship with the geographical environment and human settlement environment factors were analyzed, and the multi-factor action model of mountain land surface temperature was constructed. The average absolute error, the average relative error, and the root mean square error of land surface temperature field are 2.01 ℃, 2.35%, and 5.09 ℃, respectively. The spatial pattern of the surface temperature field is closely related to the geographical environment and human settlement environment. The influence of NDVI, topographic relief, dem, slope, and water area is negatively related to surface temperature field, while the influence of residential area is positively related to surface temperature. In terms of correlation, mountain temperature is most affected by NDVI, followed by topographic relief, dem, slope, and waterbody, residential area, and slope aspect have the least impact, indicating that improving vegetation coverage has an important role in surface cooling, and urban human settlements have the effect of surface warming. Therefore, the microgrid of the geographical environment and the impact of human settlements should be fully considered in the construction of a spatiotemporal simulation model of mountain temperature.