This WP will develop and apply techniques to evaluate and weight model projections based on aspects of the current climate that matter most for future projections. Ensembles of Earth System (ES) projections will be utilised to identify robust relationships evident across models, between key aspects of future projections and observable features of the current Earth System. In combination with observations, such relationships provide
Emergent Constraints on the key aspects of future projections, that we will use to weight projections and focus future model development. Emergent Constraints have been identified for physical climate feedbacks such as snow-albedo feedbacks (Hall and Qu, 2007), land carbon cycle feedbacks (Cox et al., 2013), and equilibrium climate sensitivity to doubling CO2 (Fasullo and Trenberth, 2012). Given the increasingly large archive of model outputs, there are risks associated with indiscriminate data-mining that can lead to the identification of spurious relationships between aspects of the current climate simulations and future projections (Caldwell et al., 2014). Similarly there is a risk of failing to identify emergent constraints if the selection of the long-term and short-term variables to be correlated is not based-on a thorough understanding of the processes that link them (Wang et al., 2014a). We will deal with these risks by building our search for Emergent Constraints on a sound physical/biogeochemical understanding and rigorous approaches used in statistical physics, such as the Fluctuation-Dissipation Theorem (FDT) and linear response theory (Lucarini and Sarno, 2011). These firm foundations will guide the search for emergent constraints on future projection uncertainties. The objective is to identify a suite of features of the current climate that are most relevant for the fidelity of future projections.