Evaluating CMIP5 Models' Representation of Oceanic Drivers of North African Climate

Abstract:

North Africa is highly vulnerable to hydrologic variability and extremes, including impacts of climate change. Prior North African studies largely disagree regarding the dominant oceanic drivers of the region’s hydrologic variability and the likely impacts of climate change. Here, we aim to apply a multivariate statistical method, the Generalized Equilibrium Feedback Assessment (GEFA), to evaluate the representation of the oceanic drivers of North African rainfall in the Coupled Model Intercomparison Project Phase Five (CMIP5) in comparison to a GEFA-based observational benchmark. The reliability of the statistical GEFA method is first evaluated against dynamical experiments within the Community Earth System Model (CESM), with initial focus on the atmospheric response to sea-surface temperature anomalies in the tropical Pacific and Indian Oceans. In order to estimate the minimum number of years of data needed to obtain stable GEFA response estimates to individual SST forcings, given large atmospheric internal variability, we apply GEFA to data records of varying durations from the CESM Large Ensemble historical simulations. After validating the GEFA method, we apply the statistical approach to both CESM and a range of CMIP5 models in order to elucidate the models’ representation of oceanic drivers of North African rainfall. The observed oceanic forcings of North African rainfall, as investigated in a parallel study, serves as the benchmark for evaluating each model’s performance and credibility in terms of future climate projections.