Reliability of Initial-Value Predictions and Climate Attributions: The Importance of Seamlessness (Invited Presentation)

Climate attribution is a rapidly growing field. But are our models good enough to reliably attribute specific weather events to anthropogenic forcing? Contrary to some conclusions in the literature, I argue that diagnostics of forecast reliability from initial-value predictions (e.g. on medium-range, intraseasonal and interannual timescales) contain vital information for assessing the reliability of attribution estimates. I will discuss an idealised example which demonstrates the crucial importance of developing seamless weather and climate prediction systems.