| The land surface temperature is an important factor of the earth's surface energy balance, area and the whole world dimension earth's surface physics process. It's also the leading factor of the various circulation process of water, heat and carbon of the earth'surface energy exchange and the earth system. Besides, the LST (land surface temperature) is also the core information of the land surface energy exchange, and to carry out an surface temperature back calculation will bring great advance to the study in many fields, such as the drought forecasting, the crop lack of water study, crop yield driving condition forecasting, figure weather forecasting,the whole world change and the whole world carbon balance progressing.In Xinjiang, the area is big and the topography and physiognomy are complex. So we cannot monitor the ground temperature in all areas. Therefore, the tradition observation method has already had no way to satisfy the need in large area real time observation in reality studying. The appearance of satellite remote sense has made it possible to solve this problem, so the LST satellite remote sense back-calculation has got many researchers as a result to take it seriously.This thesis, basing on the ecology, geography and the 3S technology, have carried out back-calculation on Xinjiang entire the earth's surface of boundary temperature with the broad disruption window algorithm by making use of the MODIS L1B data of 2001 and 2005. It back-calculated the LST of April, July and October in 2001 and 2005. Making use of the MODIS data, the land utilization has been classified at the same time. Under Arcgis9.1 and ENVI4.3, the distribution characteristics of the 6 phases of the LST in Xinjiang entire boundary time and space have been analyzed. And we make the conclusion as follows:(1) According to MODIS - the remote sensing image characteristics, after completing the geometry emendation, data pretreatment work of radiation calibration in foundation, with the data of research area meteorological and the measured result, this paper develops the surface temperature inversion model using the present inversion precision higher fission window the research. Its inversion precision is higher, the inversion method is suitable for the arid area surface temperature research, the result may be more precise, the results can be a real reflection of the research area surface quantity and the difference of heat space condition.(2) The topographical condition can seriously influence the entire boundary spatial distribution of the surface temperature in Xinjiang. The temperature overall distribution is decided by the sierra and the basin, and the temperature appears the opposite pattern with the actual land form, that is the temperature of mountainous region is low and the temperature of basin is high. Under this surface temperature overall pattern, the land surface temperature of southern regions in Xinjiang is higher than the northern regions in Xinjiang, regardless of in the desert or in the water body. This displays the nonuniformity of the distribution of the thermal resources in the space. In the time criterion, the change of the land surface temperature in season displays the same to the air temperature, it changes with the season change; but in 2001 and 2005, the change of the surface temperature is not very obvious.(3) Xinjiang's land utilization type in 2001 and 2005 has been classified by the BP algorithm, and through the classification accuracy test, the precision of various land utilization type user are all above 0.8, and the total precision is higher than 0.9, and the Kappa coefficient is also above 0.9, which is higher than the request of the lowest permission distinction precision 0.7. This indicates that the algorithm is better to the MODIS data class.(4) With the correlation analysis of the altitude and the surface temperature, we discovered the vertical lapse rate of Kunlun Mountain, Tianshan and the A'le Taishan is 6.56℃/Km, 5.73℃/Km and 4.52℃/Km respectively, among which the value of A'le Taishan is the smallest, but Kunlun Mountain is the biggest. The land surface temperature vertical lapse rate, which is obtained from this the method, is consistent to the data that many years temperature average vertical lapse rate obtained by surface weather observation in the low elevation area.(5) The relation between the vegetation coverage degree and the surface temperature assume negatives correlation. In April, June, October in 2001 and 2005, the correlation coefficient of low vegetation coverage degree is bigger than the middle and the high coverage degree; In April, 2001, the correlation coefficient of the low vegetation coverage degree and the LST is only -0.34, but the correlation coefficient of the middle and high vegetation coverage degree and the LST is -0.57 and -0.71 respectively; The situation in July, 2001 also is similar. The correlation coefficient of the low vegetation coverage degree and the LST is -0.31,the middle is -0.77 and the high is–0.81; In October, 2001, the low is -0.42, the middle is -0.64, and the high is-0.89. The situation in 2005 is similar to 2001, but we discover that, by comparing,in July,2001 the correlation coefficient between the high vegetation coverage degree and the LST is bigger than that in July,2005.(6) Through the ratio index of cool island, we discover that, the ratio index of cool island is the largest in July, 2001 of the northern regions in Xinjiang, which is 0.0074; the next is in April, is 0.007; October is the lowest, is 0.001. In the season, the cold island ratio index displays the tendency that increase first and then reduces. Comparing with 2005, the scope and the intensity of the cold island ratio index appears drop.
|
PDF Full Text Request