Component Temperature Retrieval By Integrating MODIS Data Onboard Terra And Aqua Satellites

Land surface temperature is a key parameter to the physical process of earth surface. Generally land surface is mixed-pixel, and the result of retrieval is so-called the average temperature of mixed-pixel. If the temperatures of the vegetation canopy and soil surface in the vegetation-soil mixed-pixel are separated, it could provide much precise parameters for estimation of land surface energy balance and vapor transpiration. So it is important to study the temperature of the earth's surface component for the global energy balance.Atmospheric correction is an important part to retrieve land surface temperature. Based on the knowledge of atmospheric radiative transfer, the modeling scheme is ensured. And the MODTRAN is selected to model the atmosphere. By the different simulation parameter, atmosphere is modeled on different condition. So the influence of atmosphere is reduced in the process of atmospheric radiative transfer. The sensor zenith angle is also needed in the process of atmospheric correction. Considering the repetition visit cycle of the MODIS data is 16 days, A cycle of the MODIS angles data is gained. And a database has been designed for managing it by Visual Basic and Oracle, a commercial software.Based on analyzing MODIS data, the 32 and 33 bands of MODIS arechosen to carry out the component temperature retrieval. The morning data of Terra and the afternoon data of Aqua satellites are selected to retrieve synchronously seven parameters, including surface component temperature, vegetation coverage degree and the earth's surface component radiance respectively.Because the inversion parameters are more than the number of equation, It belongs to underdetermined system of equations. Genetic algorithm is employed to retrieve component temperature. Thus component temperature inversion precision is guaranteed.To validate the retrieval results by MODIS data onboard Terra and Aqua satellites, the retrieval process of the genetic algorithm is investigated during this study. Terra-MODIS data at 11:12 on 16 June 2005, Aqua-MODIS data at 12:50 on 15 June 2005 are selected to retrieve component temperature. At the same time, the field data of Huailai in the Hebei province are measured synchronously. The retrieved Vegetation component temperature of Terra-MODIS data has a difference of 1.69°C to the field data, and the retrieved soil component temperature of Terra-MODIS has a difference of 1.4°C to the field data. On the other hand, the retrieved Vegetation component temperature of Terra-MODIS and Aqua-MODIS data has a difference of 1.49°C to the field data, and the retrieved soil component temperature of Terra-MODIS and Aqua-MODIS has a difference of 0.69°C to the field data.