Marenco, R.E. and Canals, M.F., 2025. Numerical modeling of wave transformation in San Juan, Puerto Rico: Comparison between SWAN and SWASH. Journal of Coastal Research, 41(2), 347–365. Charlotte (North Carolina), ISSN 0749-0208.

This study compares SWAN and SWASH results in a realistic two-dimensional scenario to describe wave transformation over a reef environment in San Juan, Puerto Rico. Both models were evaluated with U.S. Geological Survey observations between November 2018 and April 2019. SWAN resolves wave transformation using spectra, including the wind field, white-capping, wave breaking, and bottom friction. SWASH predicts wave propagation with high resolution and different vertical layer configurations, accounting for wave breaking through shock-capturing mechanisms. SWAN was run with an 8 m spatial resolution and forced with buoy wind and wave data, showing very good results across a range of sea states. The SWASH model was run for 20 representative sea states and compared with SWAN. SWASH was also run at 8 m horizontal resolution and was forced with wave buoy data. It was run using three vertical configurations: one layer, two layers, and three layers. A notable improvement in performance was observed when transitioning from a one-layer to a two-layer configuration; however, additional increases in layer count to three layers did not deliver significant benefits. In general, SWAN demonstrated superior performance in predicting wave heights across the examined area for most sea states, except for longer wave periods (>11 seconds), where SWASH showed marginally better accuracy. This pattern shows the robustness of SWAN's phase-averaged methodology under a diverse array of wave conditions. The results also shed light on the significant computational trade-offs between the computationally efficient, phase-averaged SWAN model and the more resource-intensive, phase-resolving SWASH model. Despite SWASH's sophisticated physics, its heightened computational demands did not uniformly result in superior performance, especially for shorter wave periods. SWASH's performance was comparable to SWAN for lower frequency events, but its poor performance in shorter wave periods may have been affected by the limitation of the 8 m spatial resolution.