Our understanding of future climate extremes relies extensively on climate projections by Earth system models. Such model projections are nevertheless subject to substantial uncertainties, impeding their reliability in assisting climate adaptation planning. Different emergent remedies have been proposed to constrain the projections with observations. They are end-to-end constraints, if not all, that is, observation constraints are applied directly and separately to individual extreme climate variables. I will illustrate the need of process-oriented constraints for advancing our physical understanding of future extreme climate changes and developing storylines to better inform adaptation practitioners and stakeholders. Then, I will introduce some initial attempts on such process-oriented constraints. 

Climate extremes,Observational constraints,Climate projections