The Inversion and Verification of Land Surface Temperature for Coal Fire Areas Based on TM Data

With the coal fire area of the Shuixigou mine in Xinjiang as an example and on the basis of observation data of infrared radiometer at the same time of passing aviation of Landsat 5 on June 31,2011,the authors calculated surface temperature at pixel scale using several schemes,comparatively studied the surface temperature of the coal fire area inversed by mono-window algorithm,and generalized single-channel algorithm and Weng algorithm with TM data.The results show that all the three algorithms show a consistent distribution of surface temperature of the Shuixigou underground coal fire area,and the mono-window algorithm and generalized single-channel algorithm have the smallest difference in the average surface temperature of the whole study area,which is about 1.60℃.Through a comparison with the ground measurements,a lower difference value is obtained by all the three algorithms,and the retrieved data by generalized single-channel algorithm are highly close to the data retrieved by mono-window algorithm,wih the regression coefficient and RMSE being 0.886 and 1.48℃ respectively.The retrieval results of generalized single-channel algorithm are in line with the spatial distribution law of the temperature of the underground coal fire area,and the high-temperature anomaly district is obvious.The result of the retrieved data of surface temperature is acceptable and the generalized single-channel algorithm is somewhat effective in acquisition of the LST of the underground coal fire area,thus providing a reference for the dynamic monitoring and evaluation of underground coal fire areas.