| Rainfall intensity as an important indicator of precipitation information esp ecially on land, rainfall intensity, it is not only reflects the precipitation proble ms to some extent, but also effects of Land-surface Hydrological Processes,W ater Resources, Drought and Flood, Geological Hazard, Ecology and so on. But want to get rainfall intensity to any piece of information of land surface woul d be limited by traditional method: sparse and uneven distribution of sites, on the uncertainty of the ground radar technology and a hardware device, Low co verage and high cost is also affected the radar, making widespread rainfall inte nsity gets difficult. With development of the computer and remote sensing tech nology, fast and wide gets the intensity rainfall is possible at the land surface of any time, rainfall intensity information to obtain in mountainous, practical a pplication and the flood control and disaster reduction has a very important sig nificance.Taking Ganzhou as study regional, used TRMM satellite which carrying of TMI microwave imager instrument and the PR precipitation radar as remote s ensing data, and using remote sensing image software to process microwave da ta and radar data, after the basic pretreatment, using band tools calculate each effects factor of rainfall intensity, according to research area in the range of si ze established space location relationship among TMI microwave data, PR rada r data and the ground meteorological stations data, extraction bright temperature of corresponds to point, near-surface rainfall intensity and rainfall intensity dat a of meteorological stations, a method get the progressive regression model of fitting relationship between brightness temperature and near-surface rainfall inte nsity, which is the best model based on the correlation coefficient R2, comparis ons of calculated and measured to inverse mapping according to surface weath er station data and PR strong data. The research conclusions of this paper are as follows:(1) By establishing a single combination of factors regression model, found the multiple correlation coefficient increased with the accession number of brig htness temperature which contains various factors in the regression model; The model contains TPC is the best, the correlation coefficient R2 ranges from 0.543541~0.655674; The model contains ICM is return poor results, the best correlati on coefficient 0.080895~0.345175.(2) By establishing multiple combination of factors regression model, foun d the multiple correlation coefficient increased with the accession number of br ightness temperature, but it increased to a certain value does not change with t he number of brightness temperature, the range of multiple correlation coefficie nt increases and its estimated standard error reduction is not significant based on statistical analysis.(3) Through integrated compared single combination of factors regression model and multiple combination of factors model, found more combination of f actor model overall are better single combination of factors regression model; o n various combination factor model selected out relative better of model for an alysis found, R6 model in related coefficient and square errors for the best, its estimates precision reached to 86.86%, but compared with R4 model and R2 model and R3 model it does not significantly. Used 6 a model inverse of grou nd Rainfall intensity are less than PR rainfall radar near ground, 6 a model ea sily on rainfall strength of maximum value for anti-played; R6 model of algori thm about Rainfall intensity of size most close to PR rainfall radar near groun d, its R2 for 0.4356; R5 model for near ground PR rainfall intensity of inversi on effect is worst, R2 is only 0.1221, rainfall intensity of maximum value has not reaction; R4 model, and R3 model, and R2 model, and R1 model its quali ty decreased, but the size of the decline is not significant. |
PDF Full Text Request