针对TM/ETM^＋遥感数据的地表温度反演与验证

Inversion and verification of land surface temperature with remote sensing TM/ETM^+ data

利用2001、2004、2005年水体、裸地、农田、草地上红外辐射计的观测数据,结合大气下行辐射与比辐射率数据,通过多种方法分别获取了像元尺度的地表温度实测值,并对基于TM/ETM+数据的3种遥感反演算法得到的地表温度进行地面验证。结果表明:3种遥感反演算法得到北京地区地表温度的空间分布趋势一致。与地面实测数据相比,辐射传输方程算法的结果略高于地面实测值,单窗算法的结果与地面实测值一致性最好,而普适性单通道算法的结果明显低于地面实测值。从单窗算法反演的北京市地表温度分布图中可以看出,北京城市热岛效应显著,总体上城区地表温度高于郊区,水体温度最低,而且不同下垫面的地表温度差异明显。

Based on the observation data of infrared radiometer over water surface, bare soil, cropland and grassland in 2001, 2004 and 2005, as well as the downward longwave radiation and emissivity data, the authors calculated surface temperature at pixel scale using several schemes, and validated the surface temperature inversed by three algorithms with TM/ETM^＋ data. The results show that all the three algorithms have a consistent distribution of surface temperature in Beijing. By comparing the ground measurements, the temperature inversed by the radiative transfer equation algorithm is a little higher than that by ground measurements, the temperature inversed by mono-window algorithm is close to the observed temperature, and the result of the generalized single-channel algorithm is obviously lower. The distribution of surface temperature inversed from the mono-window algorithm indicates that the urban heat island effect in urban area is noticeable. The surface temperature in urban area is higher than the one in suburb, and the water surface temperature is the lowest. The surface temperature for different land cover types is quite different.