Abstract
Experience with the two types of radiometers in the experiment VIRGO on the ESA/NASA mission SOHO launched in December 1995 has revealed important information about the long-term behaviour of such radiometers in space. Exposure-dependent changes can be determined from comparison of radiometers with different periods of exposure to the sun, as was done for, for example, the ACRIM series. The physical mechanisms for the long-term changes are not yet well understood. However, with the available data a model is developed which describes the temporal behaviour quite well and allows some inference of the mechanisms involved. The direct comparison of the two types of radiometers, after they have been individually corrected for the exposure-dependent part, reveals changes that are independent of the exposure to solar radiation. From this comparison an upper limit for the uncertainty of the long-term changes can be deduced. The uncertainty of the composite TSI depends mainly on the uncertainty of the tracing of ACRIM II to I, which amounts, together with the uncertainty of the HF correction during the gap between the two ACRIM experiments, to ±62 ppm (i.e. 62×10−6). Together with the result from a comparison with the independent record of ERBE the long-term uncertainty for the whole record from 1978 to present is estimated to ±85 ppm.