作者: Alexandre Boilley , Lucien Wald
DOI: 10.1016/J.RENENE.2014.09.042
关键词: Irradiation 、 Retrospective analysis 、 Homogeneous 、 Mean squared error 、 Environmental science 、 Solar energy 、 North africa 、 Meteorology 、 Sky 、 Correlation coefficient 、 Climatology
摘要: This paper compares the daily solar irradiation available at surface estimated by MERRA (Modern-Era Retrospective Analysis for Research and Applications) re-analysis of NASA ERA-Interim reanalysis European Center Medium-range Weather Forecasts (ECMWF) against qualified ground measurements made in stations located Europe, Africa Atlantic Ocean. Using clearness index, also known as atmospheric transmissivity or transmittance, this study evidences that re-analyses often predict clear sky conditions while actual are cloudy. The opposite is true though less pronounced: predicted overestimation occurrence leads to an index MERRA. overall pronounced because observed counter-balanced underestimation cloudy conditions. squared correlation coefficient ranges between 0.38 0.53, showing a very large part variability not captured re-analyses. Within homogeneous area, bias, root mean square error surprisingly large. should only be used energy with proper understanding limitations uncertainties. In regions where clouds rare, e.g. North Africa, may provide gross estimate monthly yearly irradiation. Satellite-derived data sets offer uncertainty preferred. © 2014 Authors. Published Elsevier Ltd. open access article under CC BY license