Predicting impactful Earth system change on climate timescales is a research challenge with clear societal relevance that has seen rapid progress in recent years. Initialized ensemble climate simulations, such as the decadal prediction experiments coordinated by the Decadal Climate Prediction Project (DCPP) of the World Climate Research Programme, have considerably expanded the temporal and phenomenological horizons of what we consider to be potentially predictable environmental change. At the same time, new questions have emerged relating to model fidelity, sources of predictability, and optimal prediction system design. This session aims to bring together researchers focused on seasonal, multiyear, and decadal prediction to promote cross-fertilization across communities that share the common goal of advancing our understanding of predictable climate signals and our ability to harness them for societal benefit. We invite prediction-focused abstracts from diverse disciplines (ocean, atmosphere, land, sea ice, biogeochemistry) and particularly welcome contributions that bridge disciplines and/or timescales.