Julie Jebeile (Université Catholique de Louvain), Anouk Barberousse (Université Paris-Sorbonne)
Computer power and our understanding of climate processes continue to improve, and are expected to reduce the uncertainty affecting model projections. One way to assess such a reduction is to estimate the model spread, which measures the range of estimates from multiple models. A decrease in uncertainty is expected to result in a decrease in model spread. However the models considered in successive reports by the IPCC do not exhibit any reduction in model spread. Does this mean that there has been no model improvement resulting in a decrease in uncertainty? Our aim is to explain this seemingly paradoxical situation.
