Methodology For Retrieving Land Surface Temperature Under Cloudy Conditions From Passive Microwave Measurements

Land surface temperature(LST)is an important indicator of land surface energy balance and a key parameter for studying land surface physical process at regional and global scale.The emergence and development of remote sensing technology makes the acquisition of LST change from point measurement to regional measurement.Currently,the acquisition of LST using remote sensing technology is mainly on the basis of thermal infrared(TIR)and passive microwave(PMW)data.Compared with TIR signal,PMW signal is capable to penetrate cloud,rain,snow,vegetation and other arid surface features,which can cover the shortage of TIR data with its ability to acquire all-weather and all-time surface information.This paper aims to develop a practical and physically-based LST retrieval algorithm under cloudy condition.First,current status of researches on PMW-based LST retrieval is reviewed and fundamental principles and theories of PMW remote sensing are presented in this paper.Then detailed introduction of models and data are given.There are two main goals of this research:(1)to quantitatively analyze the atmospheric impact on satellite-measured brightness temperatures and propose appropriate atmospheric correction method at commonly-used PMW channels;(2)to develop a practical and physically-based LST retrieval algorithm under cloudy conditions based on the relationship of emissivities from different channels.There are three primary research contents as follows:(1)Quantitative analysis of atmospheric impact on satellite-measured brightness temperatures and proposal of atmospheric correction method.This study is performed under cloudy conditions,when the atmospheric impact is larger than clear sky.First,the atmospheric impact on satellite-measured brightness temperatures at commonly used microwave channels is analyzed using simulated data.The results show that atmospheric effect can be neglected when emissivity is higher than 0.7 at 1.4 GHz,emissivity is higher than 0.8 at 6.93 GHz and emissivity is higher than 0.9 at 10.65 GHz.Generally,atmospheric effect cannot be neglected at other channels,i.e.,18.7,23.8,36.5 and 89.0 GHz.Then,three atmospheric correction methods,i.e.,general atmospheric correction method,simple atmospheric correction method and emissivity-based atmospheric correction method,are proposed according to the requirement of accuracy and availability of input variables.This build a foundation for further researches.(2)Development of LST retrieval algorithm.In this paper,the relationship of emissivities among the channels at 18.7,36.5 and 89.0 GHz were built and then simplified.Assuming that the atmosphere and vegetation are negligible,a three-channel LST retrieval algorithm was proposed from the simplified RTM.Its fitting accuracy is 0.83 K when the emissivity of dry soil is given.The LST can be estimated only using the brightness temperatures at 18.7,36.5 and 89.0 GHz,with a fitting accuracy of 2.37 K when the emissivity of dry soil is approximately a constant.Additionally,the LST was estimated from the satellitemeasured brightness temperatures and the ground brightness temperatures,with the accuracy of 4.0-5.0 K and 3.0-4.0 K,respectively.(3)The retrieval and validation of LST under cloudy conditions.The LST was estimated using a two-step method across China.First,the satellite-measured brightness temperature from AMSR-E was converted to the ground brightness temperature using the general atmospheric correction method.Then,the LST was estimated from the ground brightness temperature using the proposed three-channel method.Compared to the air temperature under cloudy conditions,the accuracy of estimated LST is around 4.0 K.