Modeling Dissipative Effects of Large Scale Coastal and Oceanic Regions on Tides in the Indian and Western Pacific Ocean

High-resolution circulation models are important tools for understanding complex physical processes along coastlines. Large-scale models can aid our assessment of the combined effects of tides and storm surges and the risks of communities in vulnerable regions. In this study we use the unstructured grid circulation model ADCIRC run in horizontal two-dimensional and hydrostatic (2DHH) mode to validate the eight main harmonic constituents of tides in two large basins: the Indian Ocean and Western Pacific Ocean. We have developed a large-scale high-resolution regional model for each of these two domains that provide localized resolution of gradients in geometry, bathymetry/topography, and flow processes nearshore. Mesh model resolution varies within the same mesh from deep water (≈ 25 km), to the continental shelf (≈ 2 km), to coastlines (≈ 1 km), and small island features, channels and bays (≈ 150 m). Each mesh contains roughly 3-4 million nodes and 6-8 million triangular elements. Utilizing message passing interface (MPI) parallel computing constructs, model simulations can be decomposed between 1000 or more computational cores which allows 45 days of simulation time to be computed in a few hours of real time.  

The validation of tidal constituents is an important step to show that the models can approximate the necessary long wave physics in a region before attempting storm surge analysis. The eight main tidal constituents consisting of four semidiurnal (M2, S2, K2, N2) and four diurnal (K1, O1, P1, Q1)  are used to drive the models through lateral free surface boundary conditions provided by the TPX08 data assimilated global model and astronomical potential functions. The free surface time series is then deconstructed into those constituents and compared with data assimilated global tide models and tide gauges in the region. An important consideration in large-scale basin simulation is that ADCIRC in 2DHH does not generate three-dimensional dissipative physics, such as internal tide generation and eddies formed from island chains, continental shelf breaks, and oceanic ridges and trenches. This cannot be corrected for by data assimilation methods and must be accounted for through terms such as bottom friction and horizontal eddy viscosity factors. Thus, the development of such region circulation models strongly focuses on the selection of these parameters. Through this parameter selection process we gained extensive insight into the physics of the flow in certain smaller basins and marginal seas. Different strategies affect diurnal and semidiurnal tides differently; bands of high friction along deep ocean ridges and island chains were found to impact diurnal tides strongly, while basin-wide friction on continental shelves and in gulfs effect semidiurnal tides more. This implies a combination of friction patterns based on local geometry and bathymetry is necessary to approximate the dissipative effects on the overall tidal signal in these complex regions.

Finally, a major issue that plagues regional models is that solutions generated by lateral boundaries will undoubtedly become incompatible with the solutions from global models that use different governing equations, discretizations, and grid resolution. Lateral boundaries also induce reflections and poor solutions near them. Moreover, the advective terms in the governing equations are typically ignored at or near the boundary in order to achieve model stability. We aim to combine different regional models throughout the globe in a seamless fashion in order to eliminate such boundary problems while maintaining high resolution nearshore. As a proof of concept we show that by joining the Indian and Western Pacific models at a common boundary the solutions within the interior of each region improve. It also allows a more consistent formulation by restoring the advective terms to all parts of the domain. In the future, we hope to extend this model to include other regional models in the Eastern Pacific and Atlantic Ocean basins.