Due to their effects on public health and to their potential increase linked to stratospheric ozone depletion mainly in the mid-latitude areas, solar UV radiations are intensively studied and monitored since more than ten years. Most studies show a good agreement between measurements and model-predicted values in clear sky or homogeneous cloud cover conditions. However in “broken clouds” situations the problem is largely more complex and the systematic disagreement between measurements and models appears.

We will show that it is possible to reproduce “real case measurements” by calculating, by means of a radiation transfer model, the total irradiance as the integral of elementary radiances on 2psr; each of the radiances being “corrected” for clouds effects by introducing neutral filters (assumed to represent clouds) positioned at various altitude. Sensitivity studies will also illustrate the impact of the “neutral filter” properties, optical density, positions, apparent surface… In addition to the other atmospheric properties, ozone, aerosols…, the importance of a realistic description of real cloud conditions is shown to be indispensable to obtain a correct agreement between experimental measurements and theoretical calculations.