SEAS Colloquium in Climate Science

Speaker: John Scinocca, Canadian Centre for Climate Modelling and Analysis, University of Victoria

Title: Empirical Runtime Bias Correction of Climate Models 

Abstract:  Most climate studies that rely on models use a validation process to evaluate how well the models match historical climate observations. The idea is that the smaller the model's errors in representing past climate conditions (climatological biases), or the better its overall accuracy (fidelity), the more reliable its predictions and projections of future climate will be. However, it can be challenging to distinguish the impact of model fidelity from the impact of model formulation on these predictions and projections, mainly because improvements in model fidelity often involve changes in model formulation.  Empirical Runtime Bias Corrections (RBCs) offer a unique approach to address this challenge. They provide a way to enhance a model's fidelity while keeping its underlying formulation constant. This presentation will provide two methods for deriving RBCs and explore their potential for improving future climate predictions and projections.