3S Integration Model Of Sunshine Duration

Based on previous studies, the 3S integration model of sunshine duration was developed in this thesis, which gave a comprehensive consideration on terrain and sky factors. Taking China as study area, using 1 km×1km DEM data, MODIS cloud product from 2001 to 2003 and meteorological data from 2001 to 2003, including cloud fraction, percentage of sunshine and sunshine duration, the sunshine duration of China was simulated, and the annual mean and monthly mean sunshine duration (daily amount) in lkmxlkm resolution from 2001 to 2003 were mapped.The major works of this thesis were as follows:(1) Based on total cloud fraction observed by ground meteorological stations, five correction methods were put forward in this thesis. These are difference correction, ratio correction, difference-mixed correction, ratio-mixed correction and normalization-mixed correction. MODIS cloud products (MOD06) from 2001 to 2003 were corrected by using them. Accordingly, five models for calculating low cloud fraction were established, by which the grid images of low cloud fraction from 2001 to 2003 were obtained. The performance evaluation shows that all of the five correction methods are effective. Comparatively, the ratio correction method has the highest precision and is easy to implement. After correction, the cloud fraction of MODIS is identical with that of ground observations, both in spatial distribution and in quantitative value. Furthermore, intensive observations tests show that monthly Mean Absolute Bias Error (MABE) of cloud fraction obtained from MODIS after ratio correction is less than 5%. Likewise, ratio method is the best model for calculating low cloud fraction by the tests, which has the stability and the reliability in spatial dimension and time dimension.(2) According to the good linear relationship between cloud fraction and percentage of sunshine, and considered the two factors, total and low cloud fraction, the remote sensing model for calculating percentage of sunshine was created by introducing remote sensing data. Using the model, the percentage of sunshine in China from 2001 to 2003 was simulated. The annual MABE of simulation is about 5%, and annual RABE is around 10%. In the general trend, the spatial distribution of the simulation is consistent with that of observations. The local distribution characteristics show that the simulation of percentage of sunshine is clearer and more painstaking than that observed by ground meteorological stations, embodying the advantage of satellite observation.(3) Using DEM data and remote sensing data as the general characterization of terrain and sky, the 3S integration model was established, the simulation of sunshine duration in 1km×1km resolution of China from 2001 to 2003 were achieved. The monthly MABE of simulation results is 0.15 to 0.16 hours in one day, monthly RABE is 2.34% to 4.64%. The intensive observations tests show that the 3S integration model has good simulation ability in spatial dimension and time dimension. The simulation of sunshine duration in China reflects the macroscopic and local distribution characteristics.The thesis is characterized by:(1) The correction methods of MODIS cloud product based on ground observation data give full play to the advantage of satellite observation, which makes up the deficiencies of ground observation in large range and time-space observation. This study gives a reference for relative researches of using ground observation data to correct the related satellite digital product, which is of great significance to the related investigation of plateaus, mountains and desert with sparse stations.(2) By means of GIS/RS technology, the 3S integration model of sunshine duration is coupling the geographic data, remote sensing data and meteorological data, in which the inhomogeneity of terrain and sky is solved, which could be provided to other related investigation as more objective data and has popularization value.