Research On Sea Surface Temperature Retrieval And Application Of GF-5 Satellite

The GF-5 satellite full-spectrum imager has four temperature retrieval channels(B09: 8.01 ~8.39?m,B10: 8.42 ~8.83?m,B11: 10.3 ~11.3?m,B12: 11.412 ~ 12.5?m)in the 8-13?m thermal infrared spectrum,and the spatial resolution is less than 40 m.It is the first satellite with four thermal infrared channels in China,which provides a good basis for the research of temperature retrieval and other fields.Based on the global atmospheric profile data of MODTRAN(Moderate resolution atmospheric Transmittance and Radiance code)and TIGR(Thermodynamic Initial Guess Retrievl),this paper adopts Jiménez-Mu?oz & Sobrino's single channel algorithm,QIN single window algorithm and the multi-channel algorithms are used to simulate the sea surface temperature(Sea Surface Temperature,SST)retrieval of GF-5 satellite thermal infrared data,and the GF-5 thermal infrared data is used to verify and analyze the SST retrieval.In the SST simulation retrieval,the data with GF-5 channel response function are simulated by MODTRAN in three atmospheric models: mid-latitude summer,mid-latitude winter and mid-latitude,and the radiance of different observed zenith angles and emissivity are obtained under the condition of no year-on-year emissivity.The parameters such as transmittance and water vapor content are used to obtain the radiative brightness temperature of GF-5 by Planck function.The retrieval temperature value is obtained by using the three algorithms mentioned above.The error between the retrieval temperature value and the true value of SST is taken as the evaluation standard for the accuracy of temperature retrieval.By comparing the errors under the conditions of different channels,different observation zenith angles and no year-on-year emissivity of GF-5 satellite,it is shown that in the single channel temperature retrieval,B11 channel has the highest retrieval accuracy and the best effect,and the B09 channel has the lowest retrieval accuracy,B12,B10 was the second;Among the multi-channel temperature retrieval algorithms,B09-B10-B11-B12 channel has the highest accuracy,the next are B10-B11-B12,B09-B11-B12,B11-B12,B09-B10-B11,B09-B11,B09-B10-B12,B09-B10,B10-B11,B10-B12,the retrieval accuracy of B09-B12 channel is the lowest(When the zenith angle is 0 °),the SST retrieval accuracy is the highest,and the specific emissivity has great influence on the SST retrieval accuracy,and the higher the specific emissivity is,the higher the retrieval accuracy is.Different atmospheric models have a great influence on the accuracy of SST retrieval,among which,the mid-latitude winter atmospheric model has the highest retrieval accuracy,followed by the mid-latitude atmospheric model retrieval accuracy is higher,the mid-latitude summer atmospheric model retrieval accuracy is lower.Overall,the retrieval accuracy of multi-channel algorithm is better than that of single-channel algorithm.Because the retrieval error of the single channel algorithm is large and does not meet the requirement of the standard SST retrieval precision,the second revision is carried out,and the effect of the revision is obviously improved,and the effect of the second revision of the universal single channel algorithm is not as good as that of the QIN single window,and after the second revision,The results of B09 channel of GF-5 are improved but still do not meet the standard SST retrieval accuracy.The two-channel split window algorithm has several channel combinations that do not meet the precision requirements of the standard SST retrieval,so the second revision of the temperature retrieval algorithm is also carried out,and the split window algorithm itself has a smaller error(the retrieval temperature is evenly distributed on both sides of the real temperature).After revision,the error changes little,and the effect of revision is poor.Finally,the satellite-ground synchronization experiment of GF-5 thermal infrared data is verified.The results show that the retrieval accuracy of five combinations of GF-5 thermal infrared data is better than 2K,and the root mean square error of B09-B10 channel combination is the least.0.57 K.This paper studies the temperature retrieval and temperature rise extraction of the temperature drainage experiment using the satellite retrieval data of this channel.It is found that the temperature rise extraction has 5 levels.It can be seen that the GF-5 thermal infrared data has higher resolution and higher temperature retrieval accuracy.