Development and validation of a high-resolution ROMS model for the Central Great Barrier Reef using LP-optimised bathymetry and realistic river forcing

The Great Barrier Reef (GBR) shelf is a dynamically complex coastal system where tidal currents, wind-driven circulation, river plumes, and the East Australian Current (EAC) interact across a wide range of spatial and temporal scales. Accurate numerical models of the GBR shelf require realistic representation of the steep and intricate bathymetry, yet terrain-following coordinate models are constrained by the need to smooth bathymetry in order to suppress pressure gradient errors. Here we present a new high-resolution ( ∼ 1 km horizontal, 25 vertical sigma layers) Regional Ocean Modeling System (ROMS) configuration for the central GBR shelf, nested within the Copernicus Marine (CMEMS) global ocean analysis and forecast system. A key innovation is the application of a modernised linear programming (LP) approach to bathymetry smoothing that minimises the total departure from observed 30 m resolution bathymetry while strictly satisfying both r x 0 and r x 1 roughness constraints. The LP formulation incorporates depth-dependent r x 0 targets and employs a heuristic parallel decomposition strategy using connected components and scipy’s HiGHS solver, making it computationally feasible for large operational grids. For an equivalent r x 0 constraint, the LP method requires 25% less bathymetric modification (RMS change 14.7 m vs. 19.5 m) and modifies 30% fewer grid cells than conventional Laplacian smoothing, preserving reef passages, shelf-break morphology, and nearshore channels critical for cross-shelf exchange. The model is forced with real-time discharge from 18 scaled gauged rivers distributed across 104 coastal grid cells, 9 tidal constituents (Q 1 , O 1 , P 1 , K 1 , N 2 , M 2 , S 2 , K 2 , M 4 ) at the open boundaries, and GFS 25 km atmospheric forcing. Tidal validation against five tide gauge stations yields amplitude RMSE of 1.69 cm and phase RMSE of 3.5° across seven major constituents. Validation against four IMOS/ANFOG Slocum glider missions spanning wet and dry seasons (January 2025 to February 2026) demonstrates the model’s ability to reproduce observed thermohaline structure across contrasting forcing conditions, including extreme wet-season freshwater plumes with salinities below 23. Systematic comparison with IMOS Himawari satellite nighttime SST over 355 days yields an RMSE of 0.38 °C and correlation of r = 0.98 . This configuration provides a realistic, validated tool for studying physical processes on the GBR shelf and their implications for reef ecosystem health.