Evaluation of High‐Resolution CORDEX‐CMIP6 Ocean Temperature Simulations for the Great Barrier Reef

Abstract

Longevity of the Great Barrier Reef (GBR) relies on specific oceanic and atmospheric conditions that are quickly deteriorating under climate change. While global climate models (GCMs) provide insights into large‐scale processes, they often lack the resolution to capture regional nuances affecting the GBR. Similarly, regional climate models (RCMs) are hindered by the complexity of atmosphere‐ocean interactions and potential biases in the representation of global processes in the regional domain. Here, we present a large ensemble of Conformal Cubic Atmospheric Model (CCAM) simulations that address challenges in GCMs and RCMs with various configurations in stretched grids, bias‐corrected SST forcings, spectral nudging and atmosphere‐ocean coupling and test the roles of model configurations in high‐resolution ocean temperature simulations. Near surface temperature (NST) and sea surface temperature (SST) were simulated at approximately 10 km resolutions and used to estimate heat stress with the Degree Heating Weeks (DHW), a commonly used thermal anomaly metric for coral reef ecosystems. Model performance was assessed using the Perkins skill score (PSS) and bias relative to observations. Model bias was significantly reduced throughout the year across the entire GBR when the models were both atmosphere‐ocean coupled and prescribed bias‐corrected SSTs. The same configuration also improved model performance in simulating extreme temperatures and DHW estimation across the GBR. Atmosphere‐ocean coupling without prescribed bias‐corrected SSTs showed considerable cold bias and reduced performance in the central and southern GBR. The improvements to high‐resolution ocean temperature simulations and heat stress estimations set an important benchmark to assess changing hazards across the GBR.