Delivering probabilistic climate hazards assessments

Abstract Decision-making regarding adaptation to climate hazards frequently requires a quantitative assessment of the likelihood of future hazards. This is typically done using climate model projections conditional on future greenhouse gas (GHG) and aerosol concentrations, as well as land-use changes. This leaves decision makers with the responsibility to select the concentration pathway corresponding to their risk tolerance, with little guidance from the scientific community. Here we propose a climate service estimating the occurrence of future climate hazards based on a weighted mixture distribution calibrated on an ensemble of hundreds of simulations from different global climate models and GHG concentration pathways. Each simulation is weighted according to its performance over the historical period, the equilibrium climate sensitivity of climate models, and the likelihood of future GHG concentrations. A prototype implementation for Canada was evaluated by groups of professionals from engineering firms and governments to collect criticism and suggestions. This feedback underlines the interest for climate products that integrate seamlessly with established risk-based frameworks familiar to decision makers.