Grid Simulation Of Sunshine Duration

The dissertation used DEM data of 1km × 1km spatial resolution,the MODIS cloud cover products from 2016 to 2018,meteorological data from 2016 to 2018,including cloud cover,sunshine percentage and sunshine duration combined with the distribution of possible sunshine duration under undulating terrain.The model realized the grid simulation of the sunshine hours in the study area,and draws the spatial distribution map of the annual average of the sunshine duration and the average daily total of the four seasons with the spatial resolution of 1km × 1km in 2016-2018.The work included the following aspects:(1)Firstly,the comparative analysis of the cloud cover of MODIS in the study area in 2018 and the total cloud cover of the meteorological station was conducted,and it was found that the two had a good correlation,but there was a difference in the numerical value.Then,taking the station data of the total cloud cover of the meteorological station as the true value,the difference correction,ratio correction and normalized hybrid correction methods were proposed to revise the cloud cover of MODIS,and the difference correction method with the best effect was used to revise the cloud cover product of MODIS from 2016 to 2018,so as to obtain the cloud cover raster image data of 5km×5km from 2016 to 2018.The total cloud cover data of the meteorological stations were highly consistent with the revised MODIS cloud cover in spatial distribution.The average monthly absolute error of the revised MODIS cloud cover was generally less than 10%,which indicated that the revised MODIS cloud cover had good reliability and stability in space and time.(2)Before and after the reduction of the MODIS cloud cover had sunshine percentage and weather station data were analyzed,found a significant negative correlation between cloud cover and sunshine percentage,in order to both linear fitting sunshine percentage of grid simulation model was established,implemented in 2016-2018 the sunshine percentage grid simulation in the study area.Through the verification of randomly selected sites,the annual average absolute error of the grid simulation results of percentage sunshine from 2016 to 2018 was less than 4.5%,and the annual average relative error was less than 9%.According to the grid simulation results of sunshine percentage and the spatial distribution of sunshine percentage of meteorological stations,it can be found that the distribution characteristics of the two were highly similar to the trend.Compared with the sunshine percentage of the meteorological station,the grid simulation results were clearer and clearer in local details,which reflects the advantages of introducing satellite remote sensing data.(3)According to the simulated raster image of sunshine percentage in the study area from 2016 to 2018,combined with the distributed model of possible sunshine duration under undulating terrain,the grid simulation of actual sunshine hours in the study area was realized,and the spatial mapping of sunshine duration in the study area under 1km×1km resolution from 2016 to 2018 was completed.Grid simulation results and the site data of monthly average absolute error was 0.33-0.34 hours,monthly average relative error of 5.82-6.17%,and a random sample of data validation site monthly average absolute error was 0.33-0.45 hours,on the average relative error was 5.99 7.57%,the results showed that sunshine time of grid simulation result had better ability in space and time simulation,can reflect a certain distribution trend and characteristics of local distribution as a whole.