Determination of surface global radiation using METEOSAT images and ground based visibility measurements

Our purpose is the forecast of the global radiation. As a first step we try to determine the global radiation as function of other predictable parameters. First the daily average values of the relative global radiation was considered as parabolic function of the cloud coverage obtained from METEOSAT images, an empirical formula was determined for calculation of the relative global radiation from the cloud amount. The correct this formula the cloud coverage and ground based visibility data were used. The value of the multiple correlation coefficient presenting the accuracy of the new formula was 0.96 for Budapest, and 0.93 for region of Hungary. This fact indicates that we have a sufficiently correct formula for calculation of the global radiation.     

Particle Acceleration in a Three-Dimensional Model of Reconnecting Coronal Magnetic Fields

Particle acceleration in large-scale turbulent coronal magnetic fields is considered. Using test particle calculations, it is shown that both cellular automata and three dimensional MHD models lead to the production of relativistic particles on sub-second timescales with power law distribution functions. In distinction with the monolithic current sheet models for solar flares, particles gain energy by multiple interactions with many current sheets. Difficulties that need to be addressed, such as feedback between particle acceleration and MHD, are discussed.     

Determination of surface global radiation from meteosat images using relative brightness as new parameter to characterize the cloudiness

For the investigation of radiative properties of clouds using satellite images an objective parameter is required which is independent of conditions of illumination and of recorders. We found a new parameter - the relative brightness - which seems to serve this purpose. Using the relative brightness the daily amount of surface global radiation can be determined easily by linear regression with sufficient accuracy from METEOSAT images. Period between November 1981 and July 1983 was studied and our results are not worse than the ones using more complicated methods.