Matthias Heymann Centre of Science Studies, Department of Physics and Astronomy, Aarhus University Constructing evidence and trust: How did climate scientists' confidence in their models and simulations emerge? Climate simulation started as an offspring of numerical weather forecasting in the mid-1950s and has gained increased attention since the early 1960s. In the late 1970s climate simulation was largely accepted in the climate-modelling community as a means of providing long-term climate projections. Based on climate simulation, climate scientists had reached agreement that climate warming was likely to occur in future decades, if carbon dioxide concentrations continued to increase. A consensus was reached, even though global temperatures had been stagnating since the 1950s and empirical evidence of climate warming was lacking, and even though climate scientists were aware of significant uncertainties involved in climate simulation. This paper aims at investigating the emergence and construction of confidence and trust in early climate simulation from the mid-1950s until about 1980. The investigation focuses on an analysis of the writings of climate scientists William Kellogg, Stephen Schneider, and James Hansen. From the presentation, formulation and justification of findings based on climate simulation, typical sources of confidence in climate models can be inferred. The importance of social, political and biographical contexts for the emergence of confidence and trust in climate modelling and its prediction of future climate change is granted, but beyond the scope of this paper. NBA Home