A Nonlinear Iterative Method for Retrieving Land Surface Temperature by Using MODIS Data

Abstract

Land surface temperature(LST) is an important parameter in meteorology,hydrology,ecology and so on.This paper has proposed a nonlinear iterative method for retrieving land surface temperature(LST) from MODIS data which needs two parameters(land surface emissivity and atmosphere transmittance).At first,two radiance transfer equations for MODIS band 31 and 32 had been built,which involves two unknowns(land surface temperature and average atmosphere temperature).Because of complexity,a numerical algorithm has been developed to solve the equations using iterative approach.Then,the estimation of the two essential parameters of the nonlinear iterative method has been extensively discussed.The atmospheric transmittance can be estimated on the basis of the relationship between atmospheric transmittance and water content which can be retrieved from the visible and infrared bans of MODIS data.As MODIS image has the lower resolution,the pixels of MODIS for most natural surfaces can be viewed as composed of three land cover patterns: vegetation,bare soil and water.The composition ration of vegetation and bare soil can be estimated from NDVI,and the land surface emissivity of MODIS31 and MODIS32 can be estimated by using the land surface thermal radiation equations of composed pixel and the known emissivity of land cover patterns.Finally,the sensitivity of parameters is analyzed and the results indicate that the parameters are not sensitive to the LST retrieval method we have proposed.The retrieved LST in Jiangsu province indicated the space difference,urban heat island effect and different ground object types.