Land Surface Temperature And Net Radiation Retrieval In Heihe River From HJ Satellite Data

Photosynthesis,transpiration quantity,land surface temperature and atmospheric temperature and the net radiation are very closely related,the net radiation play a key role in the system,accurate surface net radiation data in human ecological environment also is very important.There is no doubt that the land surface radiation observation sites can provide the most accurate measurement value of the net radiation flux,but the site distribution is always limited and expensive maintenance,which can’t meet in the field of climate research,weather and other related research needs,so the research on the land surface net radiation retrieval method still has great practical significance.The surface energy balance and global scale physics is an important factor in the process of land surface temperature,so the LST is the core of the earth’s surface temperature energy exchange factor.So it becomes one of important tasks in quantitative remote sensing research to make use of satellite remote sensing to retrieve land surface temperature at present.Through remote sensing image LST is the most convenient and most effective method of retrieving temperature from thermal infrared remote sensing,getting the image information is the precondition of application of thermal infrared remote sensing technology.So accurately to retrieve the land surface temperature,will promote the development of drought prediction and crop water research,substantial increase in crop yield estimation,the forecast accuracy and the precision of the global climate change.In this paper,by adopting Heihe River satellite images in the study area,using relevant satellite data and ground observation data,in remote sensing retrieval we carry out some basic study in the LST model retrieval,surface albedo retrieval,near surface temperature retrieval and net radiation retrieval.（Ⅰ）The surface albedo is a key variable for retrieving surface net radiation.In this paper,HJ-1B satellite due to a single point of view,so every direction usually assume that the surface is gay lambert surface,using the method of retrieval of surface reflectivity,to as an approximation of the surface albedo,therefore there are errors.Firstly,the bi-directional reflectance and surface albedo in MODIS spatial resolution at HJ-1B viewing and illumination geometries were derived using kernel-driven BRDF model and the MODIS BRDF model parameters product.Assumed the ratios of albedo and reflectance are similar at HJ-1B and MODIS spatial resolutions,and then the HJ-1B satellite narrowband albedos were calculated by applying these ratios to the HJ-1B surface reflectance.Finally using 6 s radiation model,from the Angle of energy,completed the narrow band albedo is converted into a wide band.The accuracy of this approach had also been evaluated using ground measurements and albedos from MODIS.The results showed that the HJ-1B total shortwave albedos were close to the ground measurements with MAE=0.025,RMSE=0.032.By compared with the albedos from MODIS,they have a consistent distribution of surface albedo and the broadband albedos inversed from HJ-1B data provided more detailed landscape texture;the MAE and RMSE for MODIS albedos and HJ-1 albedos were less than 0.013 and 0.016,respectively.（Ⅱ）The LST is a key parameter for retrieving surface net radiation.This paper adopts three models to retrieve the LST of the Heihe River suitable to HJ Satellite.From HJ satellite average surface temperature of the three kinds of model retrieval,the MODIS LST difference histogram and the measured ground station value,analyzing and evaluating three kinds of HJ precision results,the results show that in this paper the land surface temperature and the underlying surface condition has a good corresponding relation in the study area,the spatial distribution trend is roughly same,has certain practical value.Three kinds of retrieval model,mono-window algorithm accuracy is higher,for HJ satellite using mono-window algorithm retrieve the LST to get the temperature measured values to compare with the ground,it can be seen that the overall measured value is slightly lower than the ground,mean absolute error（MAE）1.47 K,root mean square error（RMSE）1.55K.The quantitative application of HJ satellite thermal infrared data demonstration,the purpose of the monitoring the thermal environment has a very good application,the LST in this paper is also important in the land surface net radiation retrieval of the study area.（Ⅲ）So from the perspective of sensitivity analysis,adopting the changing rate of the principle,the partial differential equation for sensitivity analysis is used to analyze five factors.This paper adopt mono-window algorithm to analysis parameter sensitivity.The established partial differential equation showed that the emissivity error of 0.01 resulted in LST error of around 0.6 K.The LST error was negative correlation to the atmospheric transmittance and positive correlation to the error of atmospheric transmittance;the error of0.1 resulted in LST error of around 1 K.Meanwhile,the atmospheric water vapor error and the LST error exhibited a linear relation;the atmospheric water vapor error of 0.1g/cm²resulted in LST error of around 0.2 K.The LST retrieval error were positive correlation to both the near-surface air temperature error and the average atmospheric error;the near-surface air temperature error of 1 K led to the LST retrieval error of around 1 K.Overall,the LST retrieval error and interval ratio are related to the average atmospheric temperature error as well as the near-surface air temperature error.If the LST error is less than 1K,the error of emissivity,the error of atmospheric transmittance,the error of average atmospheric temperature,the error of water vapor content and the error of near-surface temperature should be less than 0.01,0.50,0.02,0.50 and 0.10 respectively.By deriving the partial differential equation of the revised mono-window algorithm,a more detailed and precise analysis is performed on the LST retrieval in HJ-1B/IRS.This study also provides a reference for other similar LST retrieval algorithms based on environmental satellite thermal infrared band,as well as a scientific basis for future improvement of LST retrieval accuracy.（Ⅳ）According to the characteristics of HJ-1B data,combing several parameters of retrieval about surface albedo,land surface temperature,the near surface temperature and emissivity based on the methods,adopting Iqbal model and Steffan-Boltzmann model to establishing a set algorithm scheme applicable to HJ-1B data retrieval of instantaneous net radiation,After statistics,the average absolute error（MAE）31.25 W·m^-2,the root mean square error（RMSE）33.93 W·m^-2,and on the basis of instantaneous net radiation,using sinusoidal interpolation method to estimate the total surface net radiation in the day,and then calculate the average net radiation in the day.Through the statistics that the average absolute error is 27.35 W·m^-2,root mean square error is 30.27 W·m^-2.Verification results show that the method is feasible,for quickly accessing to large scale of surface net radiation time and space information provides a new way.