卫星遥感反演地表温度的时间和角度归一化研究

Aiming at the inconsistent problem of land surface temperature (LST) derived from different satellite data, this project will conduct the study on the temporal and angular normalization of satellite-derived LST. First, based on the diurnal temperature cycle (DTC) model for clear sky, a local LST variation model will be developed for partly cloudy day, and it will be used for temporal normalization of LST derived from both polar-orbiting and geostationary satellite data. Second, a radiative transfer modelling experiment will be carried out, in which the directional land surface emissivity is simulated using the Scattering by Arbitrarily Inclined Leaves Hotspot (SAILH). An empirical angular normalization model will be proposed according to the relationship between the simulated directional LST and observation angles for non-isothermal pixel. Meanwhile, inspired by the development of the bi-directional reflectance distribution function (BRDF) model, we try to develop a kernel-driven and semi-empirical angular normalization model of LST. Finally, based on the temporal normalization model、intercalibration and atmospheric profile retrieval, a new algorithm will be proposed to retrieve directional LST from multi-source and multi-angle satellite data. The new algorithm will be applied to directional LST retrieval from combined polar-orbiting and geostationary satellite data, and temporally and angularly normalized LST products will be generated. The temporally and angularly LST products will be validated with the in-situ LST measured at the Heihe field experimental site.

针对从卫星数据反演的地表温度在时间和角度上不一致问题，本项目将开展地表温度时间和角度归一化研究。首先，在晴空日温度循环模型的基础上，发展出长时间段（大于1小时）有云条件下的局域地表温度变化模型，实现极轨和静止卫星地表温度的时间归一化；其次，利用SAILH模拟红外波段方向比辐射率，开展辐射传输数值模拟实验，分析非同温混合像元地表温度随观测角度变化的规律，发展出地表温度的角度归一化经验模型；同时，受二向反射率分布函数核驱动模型的启发，分析地表温度方向性与二向反射率分布在物理机理上的异同，建立几何和物理模型，尝试发展出核驱动的地表温度角度归一化半经验模型。最后，在时间归一化、交叉辐射定标和大气温湿廓线反演的基础上，提出基于多源多角度卫星数据的方向地表温度反演新算法，并开展反演应用，用角度归一化模型对结果进行角度归一化，形成时间和角度归一化的地表温度产品，用黑河流域试验场实测数据对结果进行验证。

针对从卫星数据反演的地表温度（Land Surface Temperature, LST）在时间和角度上不一致问题，本项目开展了LST角度和时间归一化研究。首先，发展出昼夜温度循环（Diurnal Temperature Cycle, DTC）模型，实现了对地表温度的时间归一化。其次，受核驱动BRDF（Bi-directional Reflectance Distribution Function）模型的启发，发展出地表温度角度归一化模型。再次，发展出大气温湿度廓线反演算法，从FY-3D微波温度计(Microwave Temperature Sounder, MWTS)和微波湿度计（Microwave Humidity Sounder, MWHS）数据反演得到大气温湿度廓线。然后，提出从多源卫星数据反演方向地表发射率新算法，在辐射定标再处理的基础上，联合FY-3D/E先进中分辨率光谱成像仪(MERSI-2)数据和MODIS数据反演得到方向地表发射率。最后，发展出极轨和静轨卫星LST反演算法，从先进葵花成像仪（Advanced Himawari Imager, AHI）数据、FY-3D/E先进中分辨光谱成像仪(Advanced MEdium Resolution Spectral Imager, MERSI-2)数据反演得到LST，完成精度验证，生成时间和角度归一化结果。