Meteorology has employed automatic computing machines since the early days of electronic computers. From the 1950s on, a large body of models used for “in silico” experiments (numerical simulation) has been built up, together with an international infrastructure of measuring, modeling, and testing. These outstanding developments— unique in science—led not only to an increasing standardization in developing and applying models but also to deepening the interlinking between modeling and generating evidence. The article explores needs and strategies for evaluating scientific results based on mass data output devices.
