This thesis is about the production of knowledge in meteorological research. Meteorology is an interesting case because of its crucial role in defining the climate change problem, but also because of its reliance on simulation modeling, a comparatively little studied scientific practice. The thesis provides an analysis of the central practices of simulation modeling and field experimentation. It draws upon concepts from social world theory, where practice is closely related to work. The thesis also employs the notion of translation, as developed by the actor-network approach, in order to analyse how different practices and entities become associated with each other. Empirical data was collected with interviews and participant observation and most of the fieldwork was conducted at the Department of Meteorology, Stockholm University.