Quality Evaluations And Feature Analysis Of Radiation Data At Background Stations In China

An international BSRN method of radiation data quality assessment and quality control (QA-QC) is applied to investigate data quality of radiation measured from the outset to2011at Lin’an (LN), Longfengshan (LFS), Shangdianzi (SDZ) and Waliguan (WLG), four background Stations in the mainland of China. The results show that more than95%of radiation data both at LN and LFS meeting the requirements of QA-QC, but the qualitative percentage are only71.6%and42.3%respectively at SDZ and WLG. The error of solar tracker in tracking sun-disk is the main reason for those non-qualitative direct and diffuse radiation data. The thermal offset of PSP pyrometer is from-24W-m-2to-5W-m-2, about five times of CM21pyrometer. With the compensation of thermal offset, the qualitative rate of radiation data has been obviously elevated.The qualitative rates of downward long-wave radiation (DLR) at all stations are more than98%. However, failure of temperature sensor caused the data missing rate of DLR at SDZ up to15%.Analysis on the qualified radiation data which han been carried out. Under the same solar zenith angle (SZA) with cloud free sky, it is identified that the highest values of WLG global and direct radiation irradiances are sequentially followed by LFS, SDZ and LN, while the order of diffuse radiation values at these sites are at an opposite direction. The magnitudes of the solar radiation data are attributed to the atmospheric transparency and turbidity. Annual mean solar radiation and long-wave radiation at the sites are also calculated. An obvious decreasing trend of DLR at LN is detected. DLR at LN is also higher than that at SDZ and LFS. A positive trend of cloud free global solar radiation at LFS is due to the decreasing trend of aerosol loading. Cloud cover status is probably the main cause for the changes of surface solar radiation. The annual global solar radiation is biggest at WLG and least at LN. The surface albedo and surface net radiation at SDZ show a distinctly diurnal and seasonal variation.According to the empirical formula Y=αμ0b proposed by Long (2000), cloud-free detection method for October is tentatively developed by using the qualitative1-min of global and diffuse shortwave irradiance at LFS. Referring to meteorological observations from2005to2010, the linear tendency between measured and fitted cloud free global solar radiation data is1.007and the correlation coefficient is0.9989. The annual variations of monthly regressed coefficient b shows a "U" type from January to December with the smallest value in July. This phenomenon is partly related to the annual variation of Earth-Sun distance and atmospheric transparency. Mean while the corresponding coefficient, a varies irregularly with little changes. Therefore, monthly rather than annually coefficients a and b are recommended for the future cloud-free detection method.