How Fine is Fine Enough? Effect of Mesh Resolution on Hydrodynamic Simulations in Coral Reef Environments

Coral reef environments are biodiversity hotspots that provide many services to coastal communities. They are currently facing an increasing anthropogenic pressure that jeopardises their survival. Numerical ocean models are an important tool to understand the functioning of coral ecosystems and the mechanisms ensuring their resilience. However, simulating the water circulation through reef systems is challenging because of their naturally complex topography and bathymetry. Many ocean models used for such applications have a spatial resolution coarser than the scale of individual reefs, which puts into question their suitability for the task. Here, we assess the sensitivity of a coastal ocean model's outputs to its spatial resolution when simulating the water circulation in the entire Great Barrier Reef (Australia), the largest coral reef system in the world. We consider the same model with five different resolutions near reefs, ranging from 250 m to 4 km, and compare the model outputs at different locations around reefs and in the open ocean. We also simulate the transport of passive particles released from those different locations. Our results show that the simulated tidal signal is similar for all five resolutions. However, strong discrepancies (> 10 cm/s) in the current velocity are observed near the reefs and along the rugged coastline. When using a coarse-resolution model, the amplitude of the currents is overestimated over reefs, and underestimated between them. We find that validating the model at deep water mooring sites is not sufficient to ensure it performs well close to reefs. Discrepancies in currents lead to more directional and uniform tracer dispersal patterns on coarse-resolution meshes that contrast with the more dispersive patterns observed on fine-resolution meshes. Those differences at the reef level have a large cumulative effect when simulating transport processes over several weeks. Our results suggest that ocean circulation and transport simulations in coral reef environments should be based on model resolutions finer than the reef scale, which generally means a maximum resolution of about 250-500 m.